key: cord-0010296-x8ysndmi
authors: nan
title: Scientific Section
date: 2004-08-26
journal: Transfusion
DOI: 10.1111/j.0041-1132.2004.449_5.x
sha: 44061e9f19ee656be068431a9f67caa62087ee84
doc_id: 10296
cord_uid: x8ysndmi

nan

We demonstrate an efficient method for stable galactosylation of PC prior to refrigeration. Simple refrigeration itself maintains platelet function in vitro for at least 14 days. At the same time, galactosylation prevents in vitro phagocytosis of refrigerated platelets. These findings are an important step towards development of method for routine refrigeration of platelets for transfusion.

P4-030A Phage Display Tools for Automated Blood Typing D L Siegel, University of Pennsylvania, Philadelphia, PA Background: Our long-term objective is to use a set of novel molecular technologies to develop a class of renewable, inexpensive, high-quality blood bank testing reagents that will function in a rapid, high-throughput, automatable assay system. At the core of the system are bacteriophage particles (BPs) which display RBC mAbs on their surface, physically linking the phenotype of an antibody (the antigen-binding receptor) with its genotype (the DNA sequence within the particle that encodes that particular antigen receptor). Such BPs are capable of self-replication in bacteria and can be used to phenotype cells using agglutination-based methods in tube, plate, and gel. Here, our goal is to extend this technology further by exploiting the naturally-occurring presence of unique DNA sequences within the BPs to enable the phenotype of an RBC to be determined by assaying the genotype of the detecting reagent. Such an approach could then take advantage of nucleic acid detection schemes which offer exquisite sensitivity/specificity, require minute amounts of sample, and are easily adapted to automation. Furthermore, development of a "phenotyping-by-reagent-genotyping" methodology would enable multiplexed assays in which extended RBC antigen profiles could be obtained in a single tube at one time, something not possible by the use of conventional agglutination. Methods: To test the feasibility of this approach, a model system was designed in which anti-B and anti-D BPs were created, combined together, and incubated with RBCs of each of the four possible B/D phenotypes. After removing unbound BPs, PCR was used to amplify a unique DNA tag within each bound BP, each tag having a characteristic length. Gel electrophoresis was then used to detect the presence of amplified tags. Results: By detecting the size of the PCR product, the presence or absence of template anti-B and/or anti-D BP DNA was assessed and the results were in perfect concordance with the expected binding pattern of the BPs based on the known phenotypes of the B-, B+, O-, and O+ RBCs. Quantitative analysis revealed that the typing assay required a total of only ~1500 RBCs (135 pL) and the equivalent of only 20 attograms or ~100 anti-B or anti-D IgG molecules. Conclusions: Using bacteriophage-displayed mAbs, the simultaneous binding of anti-RBC antibodies of differing specificity can be assessed by the amplification and detection of antibody-associated DNA. The phenotyping of cells by this type of approach requires minuscule amounts of test sample and reagents. To facilitate high-throughput automation superior to currently-available agglutination-based platforms, methods are being adapted which use high-speed PCR and a rapid, fluorescence DNA hybridization scheme. Background: West Nile virus (WNV) NAT screening was implemented in minipools (MPs) to prevent transfusion-associated transmission. Individual donation testing (IDT) was necessary for areas of the US that experienced epidemic levels of WNV. A trigger was developed to manage conversion from MP to IDT. Methods: WNV NAT was performed in MPs of 16 or by IDT (Gen-Probe TMA); WNV reactives were confirmed positive (pos) by repeat TMA, PCR (NGI) and IgM (Abbott) in index and then again in follow-up (f/u) samples. About 4.125 million donations were screened from 6/23/03 to the end of 2003 with two epidemic areas (NE and KS) having IDT performed from 8/20-10/4 for 30,501 donations. 18,037 donations collected in NE from mid-July to the initiation of IDT (MP neg) are being retested IDT; IDT confirmed pos donations will trigger lookback. Results: Overall WNV prevalence was 1 : 5700 (415 confirmed positive (pos) donors of 839 NAT reactives); prevalence for the two epidemic regions was 1 : 175 (N = 206; 0.57%) for NE and 1 : 243 (N = 99; 0.41%) for KS. Of the 415 confirmed pos donors, 291 (70%) were repeat TMA and/or PCR reactive at index without IgM, with the remaining 124 IgM reactive at index or f/u, with or without repeatable RNA (75 RNA pos at index/IgM pos at index, 47 RNA negative (neg) at index/IgM pos at index and 2 RNA neg at index/IgM pos only at f/u). Of the 415, NAT reactivity was repeatable by TMA in 334 (80%); the remaining 81 required PCR and/or IgM for confirmation. IgM frequency in confirmed pos donors at index increased linearly as the season progressed. The overall peak week of confirmed pos donors was 8/25-8/31; with cases identified from 6/29-12/1. Based on 4 NAT reactives and a frequency of 1 : 500-1 : 1000 (trigger evaluated from 2002 retrospective studies indicating one IDT confirmed pos/MP neg for every 4 MP reactives), IDT for 30,501 donations in NE and KS resulted in 349 initial reactives (1.14% vs a MP initial reactive rate of 0.73%) of which 181 (51%) confirmed pos. Of the 181, 85 were MP reactive; 96 were not consistently MP reactive (0 or 1 pos of 2 tested). Of the 96, at index 50 were PCR neg/IgM pos, 38 were PCR pos/IgM pos and 8 were PCR pos/IgM neg. This is a yield of 1 : 3800 low level viremic donors versus 1 : 359 yield of donors detected by MPs and 1 : 318 total pos donors requiring IDT. The 8 PCR pos/IgM neg samples had viral loads of 100-950 copies/mL; 950 was the highest viral load for an IDT pos/MP neg donation. Conclusions: WNV NAT prevented the release and transfusion of 835 components from 415 confirmed pos donors. MP NAT serves as a surveillance tool to monitor the need to convert to IDT during WNV epidemic periods. IDT prevented the release of an additional 1 : 318 pos donations of which 1 : 3800 likely would have been infectious.

Background. Screening donors for West Nile virus (WNV) RNA using nucleic acid amplification technology (NAT) in 2003, led for the first time to detection of humans during the viremic, pre-seroconversion phase of infection. An understanding of the time course and dynamics of WNV RNA and serological markers following acute infection has important implications for donor screening and deferral policies, as well as for diagnosing WNV infection in clinical settings. Methods. Donations were screened for WNV by either Mini-Pool NAT (MP) with 16 units/pool, or individual donation NAT (ID-NAT) using the Gen-Probe/Chiron TMA assay. Index unit viral loads were determined using a target-capture/real-time PCR assay (Chiron Corporation) and IgM and IgG status using EIAs (Focus Technologies). NAT-reactive donors were offered enrollment into a follow-up study that included a symptom questionnaire at enrollment and sampling at weekly intervals, with follow-up ending when donors tested negative by ID-TMA and developed WNV IgM.

Follow-up samples were tested for RNA by TMA and RT-PCR, as well as for IgM and IgG antibodies. Results. From July 1 through October 31, 2003, NAT screening of 681,567 donations yielded 220 confirmed viremic donations with 39 (18%) identified by ID-NAT and 181 (82%) identified by MP-NAT. Median viral load of confirmed positive donations was 2370 gEq/mL (range: <50-690,159); 43 (20%) confirmed viremic index donations tested IgM-reactive and 39 (18%) IgG-reactive. 182 (83%) of the confirmed positive donors enrolled into the follow-up protocol. The first follow-up specimen was obtained a median of 9 days (mean 15 days) following the index viremic donation. Enrolled donors gave an average of 2.5 follow-up specimens (range: 1-8). Of the 140 IgM negative index donations, seroconversion occurred on the first f/u bleed in 113 (81%) cases and the second f/u bleed in 27 (19%) cases. The time to IgG SC following IgM SC averaged 3.7 (95% CI 2.6, 4.8) days. The time to loss of detectable RNA by ID-TMA averaged 7.9 (95%CI 6.2, 9.6) days after IgM SC and 4.1 (95% CI 2.3, 5.9) days after IgG SC. Conclusion. IgM and IgG seroconversion occur sequentially and consistently over the 2 weeks following viremic donations, supporting use of SC for donor confirmation and reinstatement, and for clinical diagnostic testing. However, low-level WNV RNA continued to be detectable by TMA for ~8 days following IgM and 4 days following IgG SC. The infectivity of transfusions with such low-level viremia post-SC is under investigation. We are also conducting further viral load analyses to estimate the duration and kinetic parameters of the RNA+ pre-SC phase.

Background: Transfusion decisions depend on many factors including physiologic alterations, comorbidities and recent test results. Computerized decision support (DS) has been shown to improve medication ordering and compliance with some guidelines such as thromboembolism prophylaxis. We sought to evaluate the effect of computerized physician order entry (CPOE) with DS and realtime feedback on adherence to transfusion guidelines. Study Methods: This prospective, randomized study was conducted in a 700-bed tertiary care teaching hospital between April 2003 and April 2004. New transfusion guidelines were disseminated to staff and followed by educational lectures. All transfusions of RBCs, platelets and FFP are ordered with CPOE. Transfusion order appropriateness was assessed by DS and based on institutional guidelines, test results and patient-specific clinical data entered during CPOE. Orders were excluded from DS evaluation if they were emergent or originated from non-study locations. Chart review was done for all orders rated by DS as inappropriate (DS-INAPP) and a 15% sample of orders rated as appropriate (DS-APP). Three periods were studied: baseline, post-education and DS intervention. The unit of randomization for DS was the physician. Junior housestaff physicians were stratified by specialty and post-graduate year, and then randomized into either the control group receiving no DS feedback (DS-OFF) or the intervention group receiving DS feedback (DS-ON). Orders rated by DS as appropriate (APP) were processed. Orders rated inappropriate (INAPP) resulted in the appearance of a new DS recommended order. Physicians could then either accept or decline the DS recommendation, or cancel the order. Final orders changed to satisfy DS recommendations were re-classified and no longer considered INAPP. Results: Conclusion: Education followed by DS resulted in a decrease in the proportion of orders that did not comply with institutional guidelines from 68% to 50% (p < 0.001). Computerized physician order entry with DS feedback recommendations is effective in improving transfusion practices. Future studies are needed to determine the impact of such interventions on specific inpatient populations as well as patient outcomes and costs.

Background. The TEG is used in situations were point of care testing of hemostasis is desired, although its value is still controversially because of insufficient test validation. The main parameters of the TEG are (a) the reaction time ®, the time until the initial fibrin formation and comparable with the coagulation times PT and APTT; (b) clotting time (K), the time until a fixed level of clot firmness is reached; (c) the angle (a) is closely related to K and measures the rapidity of fibrin build up and gives information about the clot strength; R, K and a are prolonged by anticoagulants and factor deficiencies; (d) maximum amplitude (MA) is a measurement of maximum strength or stiffness of the developed clot; it is especially influenced by platelets and fibrin. We studied the effects of thrombocytopenia on all the above described TEG parameters. Methods. In 17 patients treated with chemotherapy for hematologic malignancies, TEG's were performed after the first chemotherapy course, when blood count became normal. During the subsequent decrease of platelet count and the aplastic period TEG parameters were measured in native whole blood samples. The findings were compared with those obtained in 60 normal persons. Results. MA the parameter considered to be influenced mostly by platelet count decreased with platelet counts below 50 ¥ 10 9 /l: platelets 20-50 (n = 86), MA 41 ± 12 mm (controls 46 ± 7, p = 0.004), platelets <20 (n = 25), MA 34 ± 8 mm. However, the parameters considered to be influenced mostly by coagulation factors were significantly prolonged (R and K) or dereased (a) with platelet counts below 100. Platelet count 50 -100 (n = 37), R = 49 ± 31 min (controls 21 ± 5, p < 0.0001), platelets 20-50, R = 66 ± 35, and in platelets <20 R = 57 ± 26. In patients with a platelet count 50-100, K = 21 ± 13 min (control 9 ± 3, p < 0.0001), with platelets 20-50, K = 39 ± 21; with platelet < 20, K = 49 ± 26. The angle is in patients with platelets 50-100, 15 ± 9 deg (controls 26 ± 8, p < 0.0001); in platelets 20-50, a = 9 ± 7 deg, and in platelets <20, 7 ± 8. In all patients PT, APTT and fibrinogen levels were normal during the study period. Conclusion. Platelet count not only influence MA but even more the coagulation parameters R, K and a, which is in agreement with the new conceptual cell based model of hemostasis; secondly the TEG should be considered as additive to platelet count and plasmatic coagulation tests and not as a replacement.

N Blumberg, L Fine, K F Gettings, J M Heal, University of Rochester, Rochester, NY Background: Randomized trials and animal model studies have demonstrated that in some instances leukoreduction of blood transfusions reduces the risk of post-operative infections. This benefit has been attributed to reduction of harmful immunomodulatory effects of blood transfusions. Linerelated infections are among the leading infectious complications occurring in hospitalized patients. We performed a study of line related infections in patients before and after the July 2000 implementation of universal leukoreduction of transfusions in our hospital. Methods: Line related infection data were collected from hospital infection control records for the period 18 months prior to and 18 months after July 2000. Transfusion histories were obtained from transfusion service records. Results: 334 line related infections occurred during the study period of 1999-2001. The number of line related infections decreased from 150 to 98 (-35%) in transfused patients after implementation of universal leukoreduction, whereas line related infections increased from 41 to 45 (+10%) in non-transfused patients (p = 0.04). This corresponded to a 38% decrease, from 5.3 to 3.3 infections in transfused patients per 10,000 patient days (p = 0.002). The infection rate remained stable in non-transfused patients at 1.5 infections/10,000 patient days both pre-and post implementation of universal leukoreduction. Virtually identical decreases were seen on surgical, medical and pediatric services. The patterns of infecting organisms, types of lines infected and transfusion practices remained unchanged during this three year period.In the pre-implementation cohort, we discovered two patients with line related infections who should have received only leukoreduced transfusions by protocol, but received non-leukoreduced blood through error or oversight. In patients with hematologic malignancies, who receive leukoreduced transfusions by protocol since 1989, the number of line related infections increased from 24 in the 18 months prior to July 2000 to 29 in the subsequent 18 months (not significantly different). Thus the decrease in line related infections after July 2000 is due primarily to decreases in patients who previously did not receive leukoreduced transfusions by protocol. Conclusions: A substantial and statistically significant decrease in line related infections occurred in our hospital coincident with implementation of universal leukoreduction. This result was observed in transfused but not untransfused patients and throughout all clinical services. This association is thus most likely cause and effect, but further studies are needed for confirmation.

E C Wong, C S Smith, V R Criss, N L C Luban, Children's National Medical Center, Washington, DC Background: With the advent of FDA approved methodology for bacterial testing of plts and the demonstration of adequate apheresis plt function at 7 days of storage, 7 day old plts should be suitable for release. Recently the FDA has cleared the Extended Life Plt (ELP) collection/storage bag that is part of the COBE Spectra Apheresis System (Gambro BCT, Lackwood CO) for 7 day plt storage when the extended storage is coupled with 100% screening for bacterial contamination using a method cleared by the FDA. Given our use of an FDA approved methodology for bacterial testing of apheresis plts (BacT Alert, Bio-Merieux, Durham, North Carolina), we determined the in-vivo efficacy of transfused 7 day old apheresis plts. Study Design and Methods: Prospective study of 7 day old apheresis plt txns to pts at a tertiary care pediatric hospital. Apheresis plts were collected on the COBE Spectra using the single needle LRS procedure and stored using the ELP collection/storage bag at 20-24C under standard blood bank conditions. Plt products were cultured at both 24 hours and 7 days post collection using the BacT Alert system. Prior to release, pH and glucose were measured on the iStat (Abbott Laboratories, Chicago, IL). Pre-and post-txn (between 1 and 4 hours post-txn) plt counts and corrected count increments (CCIs) were determined on recipients. Txn recipients were monitored for adverse reactions. Results: Overall, a total of nineteen 7 day old apheresis plts were transfused to 14 pts (12 males and 2 females) ranging from <1 month to 17 years in age. All pts were being treated on the hematology/oncology service for hematologic malignancies/congenital immunodeficiency syndrome (n = 10), in the neonatal intensive care unit for necrotizing enterocolitis with perforation/sepsis/respiratory failure (n = 3) or in the pediatric intensive care unit post cardiac surgery (n = 1). Apheresis plts released on day 7 had median pH of 7.0 (range 6.6 to 7.3) and median glucose of 235 mg/dL (range 154 to 317). None of the samples from 24 hours or 7 day old apheresis plts demonstrated bacterial growth after 6 days. Median CCI was 8, 635 (range 1, 359 to 30, 470) . Eight of the nineteen (42%) plt txns had CCIs < 7,500 and were indicative of unsuccessful txn; however, when compared with platelet txns of 5 days old or less during the same time period to the same pts, similar CCIs were obtained. In addition, these low CCI txns could be ascribed to pt co-morbid conditions at the time of txn. No adverse reactions were noted in any txn. Conclusions: Apheresis plts collected on the COBE Spectra using the single needle LRS procedure and stored for 7 days using ELP collection/storage bags demonstrate adequate in-vivo efficacy. Background Data on long-term survival of patients receiving autologous transfusions for surgery are limited. Better information would be useful in evaluating the cost-effectiveness of pre-operative autologous blood donation (PABD) and intraoperative autologous transfusion (IAT) compared to allogeneic transfusion (ALLO). Survival of Olmsted County (OC) surgical patients following autologous transfusion was examined. Methods We examined computer records of transfused patients from mid-1992 through 1998 to determine OC residency. Gender, age, number and type of products transfused, surgical service (SS), and Charlson comorbidity index (CI) were noted. Patients were counted once by including only the first transfusion episode in the time period. Vital status was determined by examining patient records for contact in the third quarter of 2003 or death. Tracing through public electronic databases was used for patients without a date of death in the medical record and for whom there had been no recent contact. Results Of 5313 OC residents transfused, 3574 underwent surgery (67%). Of the surgical patients, 288 (8%) received PABD, 1258 (35%) IAT, 156 (4%) PABD and IAT, and 1872 (52%) ALLO only. Ten-year survivals were significantly different with 81% for PABD, 65% for IAT, 83% for PABD and IAT, 34% for ALLO only. Characteristics differed significantly (Table) . SS responsible for most of patients receiving PABD were orthopedic (ortho) (63%), ENT (13%), and plastic (9%) surgery. SS for IAT patients were cardiac (66%), ortho (15%), and vascular (12%) surgery. For PABD and IAT, SS were ortho (78%), cardiac (15%), and thoracic (6%) surgery. For ALLO only SS were ortho (37%), general (19%), and cardiac (8%) surgery. Multiple Cox proportional hazards regression, adjusting for age, CI, SS, RBC dose, and blood product given, showed a significant benefit of autologous transfusion on survival (p < 0.001). Conclusion The receipt of any autologous blood was associated with significantly better survival after adjusting for age, CI, surgical service, and RBC dose.

Background: Dendritic cells (DCs), as potent antigen presenting cells have been successfully applied in some cancer immunotherapies. They can be cultured from peripheral blood monocytes isolated by immunoselection, density gradient separation and elutriation. We tested the capacity of monocytes isolated through a cord blood filter to differentiate into functional DCs; distinguishing this method as a quick alternative, for culture of at least some DC types, to more expensive or laborious selection methods. Methods: MNC apheresis units from healthy consented donors (n = 10) were passed through cord blood filters (Stem Quick TM E, Asahi Medical Co., Ltd, Tokyo, Japan). WBCs, attached to the filter membrane, were collected by sequential "back-flushing" with saline and cold dextran, and then adhered onto polystyrene (1-2 h, 37°C). Non-adherent lymphocytes (~97% CD3+) were removed and adherent cells (>85% CD14+ monocytes) cultured for 2-7 days in serum-free CellGro® media (CellGenix, Germany) with or without GM-CSF + IL4 (50 ng/ml) or GM-CSF + IL4 and a cocktail of proinflammatory cytokines (10 ng/ml IL6; 1 uM PGE2; 10 ng/ml IL1B; and 25 ng/ml TNFa), reported to make "fast" DCs in 2 days. Non-filtered, ficolled adherent cells were similarly treated and compared. Cultures were characterized by morphology, phenotype, cytokine production, T cell stimulatory capacity and migratory responses to chemokines. Results: As early as 2-4 days, unlike ficolled adherent cells, cells cultured from "filter" monocytes showed "typical" veiled morphology, expressed DC markers: CD40, CD80/86, HLADR, chemokine receptors (CCR5, CCR7), and the maturation marker, CD83, as well as stimulated strong mixed lymphocyte responses (1DC : 1T cell, 26213 ± 8000 cpm), presented both tetanus toxoid and influenza to autologous T cells (stimulation index = 8-10 above antigen-free controls), and migrated toward lymph node-derived chemokines, CCL19 and CCL21. Endogenous PGE2, required for efficient migration, was detected between days1-2 at levels equivalent to that of exogenously added PGE2 used to make "fast" DCs. IL6 was also detected at the exogenous dose while IL1B and TNFa were substantially lower (~0.3 ng/ml). Conclusions: Our studies demonstrate the feasibility of using filtration to isolate monocytes, which may promote the release of endogenous factors that contribute to "fast" DC differentiation. In this way, filtration may serve as a useful model to mimic in vivo transendothelial migration into lymphatic vessels. Moreover, this method offers a strong logistical advantage, as filters can be easily incorporated into closed system procedures. Background: Group B streptococcus (GBS) is a leading cause of life threatening infection in neonates. In 1996, the CDC recommended the routine prophylactic administration of antibiotics to mothers with suspected GBS infection. The impact of using umbilical cord blood (UCB) stem cells collected from GBS antigen positive maternal donors on transplant outcome is currently unknown. Study: GBS antigen testing results on maternal donors became available Jan 1998. Donor information was compared to product culture results for correlation. Engraftment and survival were examined to determine the effect on outcome following transplantation of cord blood units collected from GBS antigen-positive maternal donors treated with antibiotics where the cord blood culture results were negative. Results: Analysis of the donor database revealed that 1465 maternal donors of 7498 units collected from Jan 2003-Apr 2004 were Group B strep antigen-positive (19.5%). None of the corresponding UCB products demonstrated growth of GBS, most likely due to the effectiveness of intrapartum antibiotic treatment. Sixteen UCB products collected from Jan 1996-Apr 2004 grew GBS in culture. In compliance with program policy, all culture positive units were discarded. No maternal symptoms or documentation of antibiotic use was evident for 15 of these units (one had fever). Nine tested negative, 2 were not tested, and 5 were not asked (prior to 1998) . The infant of one of the 16 moms who tested negative became symptomatic and was treated for GBS. Thirty one products from GBS "colonized" (antigen-positive, culture-negative) donors were infused with a minimum follow-up time post transplant of 3 months. No product demonstrated growth of GBS after thaw in the transplant center. Furthermore, there was no incidence of GBS infection in recipients post transplant. Overall survival of the recipients was similar to outcomes in patients receiving cord blood products from GBS negative donors (p = 0.3134). Similarly, a significant difference with regard to time to engraftment was not detected between the two groups (p = 0.3656). Conclusion: Preliminary data demonstrates that utilizing UCB stem cells from maternal donors colonized with group B strep does not adversely affect engraftment kinetics or overall survival. Introduction: Immunocompetent Rh negative individuals exposed to Rh positive red cells form an anti-D in approximately 75-80% of cases. The potential to form anti-D in allogeneic Rh mismatched HPC transplants either from exposure of an Rh negative recipient to Rh positive donor RBCs and platelets or from donor derived RBCs and platelets following engraftment or, in the case of an Rh negative graft, being exposed to host derived RBCs and platelets is a theoretical concern. However, the exposure of an Rh negative immune system to Rh positive red cells due to the transplant is unavoidable (from either recipient or donor origin). We reviewed the transplants performed in our institution to determine the frequency of anti-D formation in such HPC transplants. Method: All allogeneic transplants performed between July 1993 and February 2004 were retrospectively reviewed. The Rh mismatched transplants were identified, the transfusion histories reviewed, and the formation of anti-D was noted. Results: There were 258 allogeneic transplants performed of which 41 were Rh mismatched transplants. Three transplants involved the same patient and an additional 3 patients had a pre-existing anti-D. In total there were 36 evaluable patients (22 related donor transplants and 14 unrelated donor transplants). There were 16 Rh negative recipients who received Rh positive HPC and 20 Rh positive recipients who received Rh negative HPC. One patient in each group formed an anti-D post-transplant (6.25% and 5.0% respectively). Both patients were the recipients of a related donor. The one Rh negative recipient who received an Rh positive transplant and formed an anti-D 1 month following transplant received Rh positive platelets but no Rh positive RBCs. The one Rh positive recipient who received an Rh negative transplant and formed an anti-D 6 months after transplant received both Rh positive RBCs and platelets. Overall, 15 patients received Rh positive RBCs and 21 patients received Rh negative RBCs exclusively. All patients received Rh positive single donor platelets. Discussion: This review demonstrates that the formation of anti-D in an Rh mismatched HPC transplant is an infrequent event. The overall frequency of anti-D formation was 2 in 36 patients or 5.5%.

S Wagner, A Myrup, M S Walker, American Red Cross, Rockville, MD Introduction: Blood outgrowth endothelial cells are thought to arise from very rare cells derived from circulating angioblasts that are present in the mononuclear fraction of blood. Recently, Lin and colleagues have successfully expanded endothelial cells from single colonies derived from precursor cells present in buffy coat into confluent monolayers on fibronectin-or collagen-coated polystyrene surfaces (J Clin Invest 2000;105:71-7). However, a method for sterile closed system culture of these cells has not yet been described. Here, we describe one closed system for culture of blood outgrowth endothelial cells after several unsuccessful attempts investigating the potential for their closed system propagation on different fibronectin-coated plastic substrates. Materials and methods: Mononuclear cells were enriched from 6 mononuclear apheresis units collected from normal donors by eluting strongly adherent cells from cord blood filters (Transfusion 2003;43S:53A, abstract SP35) and resuspending the eluted cells in EGM-2 medium which contains VEGF, bFGF, IGF, and EGR and 10% human serum. Approximately 2 ¥ 10 8 mononuclear cells were introduced into human fibronectin-coated sterile 250 cm 2 polycarbonate cassettes containing inlet and outlet tubing. Medium with serum and cytokines and trypsin with EDTA were sterilely introduced into standard blood transfer containers using a sterile connecting device with a custom-fabricated 0.2 mm sterile barrel filter connected to blood bank tubing. Supplemented medium or trypsin-EDTA were introduced and removed from cassettes using the sterile connecting device. Medium was changed every 2 days. Following expansion, cells were assayed by FACS analysis and ability to take up acetylated LDL. Results: Between 3 and 18 colonies of cells with "cobblestone" fibroblast-like morphology were observed in each cassette within 2 weeks of culture. Colonies were ready for passage to new cassettes within 3 to 4 weeks of culture. More than 2 ¥ 10 6 cells were harvested from the second cassette in 4 of the 6 cultures within 4 weeks, and cells displayed surface markers characteristic of endothelial cells including CD31 + , CD105 + and CD146 + . Cells were uniformly capable of acetylated LDL uptake. Blood outgrowth cells were HLA-ABC + , yet CD14 -, and CD45 -. In separate experiments, cells were CD34 + early in culture but gradually lost this marker with increasing passage. Conclusion: Blood outgrowth endothelial cells can be expanded in fibronectin coated polycarbonate cassettes using a closed system design.

M S Hossain, Winship Cancer Institute, Atlanta, GA; L Lezhava, Emory University School of Medicine, Atlanta, GA; E K Waller, Winship Cancer Institute, Atlanta, GA; J D Roback, Emory University School of Medicine, Atlanta, GA Background: We previously reported using a mouse allogeneic BMT model that lethal murine CMV (MCMV) infection at the time of transplant could be prevented by concurrent infusion of amotosalen-treated donor splenocytes. However, in the clinical setting, it would be preferable to administer adoptive immunotherapy at the time of transplant to provide antiviral coverage over the ensuing weeks. We now demonstrate that amotosalen-treated donor splenocytes, administered prophylactically at the time of transplant, persist in BMT recipients and prevent lethal MCMV disease following delayed viral challenge 7 days later. Methods: Irradiated (11 Gy) CB6F1(C57BL/6 ¥ BALB/c) (H-2b/d, CD45.2+, Thy1.2+) recipient mice received 5 ¥ 10Ÿ6 T cell depleted bone marrow cells from BA (H-2b, CD45.2+, Thy1.1+) donors. Some BMT recipients also received 10 ¥ 10Ÿ6 amotosalen-treated splenocytes from PepBoy (H-2b, CD45.1+, Thy1.2+) mice immunized with mCMV more than 60 days before. Mice were challenged with 2.5 ¥ 10Ÿ4 pfu mCMV on day 7 after transplant, a reproducibly lethal inoculum in the absence of adoptive immunotherapy. Donor T-cells were tracked by flow cytometry using congenic surface markers (CD45, Thy1). Antiviral CD8+ CTL were identified using tetramer and intracellular cytokine staining. Results: BMT mice without adoptive immunotherapy experienced 100% lethality within 15 days after delayed (day 7) MCMV challenge. In contrast, all recipients of amotosalen-treated splenocytes survived to day+ 100 after MCMV infection without evidence of GvHD. BMT recipients infused with 3 ¥ 10Ÿ6 untreated splenocytes also survived MCMV challenge, but later suffered from chronic GvHD. Amotosalen-treated T cells, isolated from the spleen, liver and peritoneal cavity of BMT recipients, responded similarly to viral infection based on expansion of tetramer-positive antiviral CTLs and expression of T cell activation markers CD69, CD25, and CD134. Ten days after MCMV infection, mice receiving 3 ¥ 10Ÿ6 prophylactic untreated splenocytes had no detectable thymocytes. In contrast, recipients of 10 ¥ 10Ÿ6 amotosalen-treated splenocytes showed robust thymopoiesis primarily due to bone marrow-derived T-cell precursors. Conclusion: Prophylactically administered adoptive immunotherapy with amotosalen-treated donor splenocytes effectively protects against delayed lethal MCMV infection. Activated antiviral T-cells home to end organ targets of MCMV replication. Administration of amotosalen-treated splenocytes also promotes thymopoiesis by bone marrow derived T-cells, shortening the period of immune reconstitution post-transplant.

Background: Since introduction into the U.S in 1999, West Nile virus (WNV) has spread rapidly to become the largest arbovirus epidemic in recorded history. Testing laboratories successfully implemented minipool (MP) nucleic acid technology (NAT) in 2003. However, concern about break-through transmissions missed by MP testing has prompted consideration of individual donation (ID) NAT during seasonal epidemics. The 2003 WNV NAT testing experience at Blood Systems Laboratories (BSL) provided the opportunity to prospectively evaluate policy options for 2004. Methods: We had access to the entire set of WNV MP-NAT screening results and the total number of donations screened by BSL each week for each of 79 blood centers from June 29 to November 22, 2003. We conducted analyses of potential strategies for initiating and discontinuing ID-NAT testing. Following preliminary analyses of over 12 strategies, we focused on 5 initiation strategies consisting of either counts of repeat reactive (RR) donations or weekly RR rates alone, or in combination. We added 3 discontinuation strategies to each initiation strategy: dropping below the initiation trigger alone, or coupled with the addition of a 1 or 2-week period of no yield. Results: Overall effectiveness of each of the five strategies at signaling the need to switch to ID-NAT for blood centers meeting the given trigger ranged from a high of 100% to a low of 57%. The addition of a no-yield requirement for 1 or 2weeks in order to discontinue ID-NAT substantially increased testing burden. Combined initiation and discontinuation strategies would have resulted in an ID-NAT testing burden between 10 and 50% for a minimum 10-week period extending to a maximum 20-week period before returing to MP-NAT testing of all donations. For our organization the most feasible ID-NAT initiation strategy was a trigger of 2 MP RR donations and a weekly MP RR rate of 1/1000, which, with these data, had an effectiveness of 81% and led to peak weekly ID-NAT testing of 20 to 40% of donations depending on the discontinuation rule. Conclusion: By assuming 2004 WNV activity is likely to be similar to 2003 activity and analyzing the consequences of a range of different ID-NAT trigger strategies, we characterized the balance between available testing capacity and maximum attainable safety. The validity of this approach will be monitored in 2004 to determine how appropriate it was in providing useful policy information.

M P Busch, L H Tobler, N Gefter, Blood Systems Research Institute, San Francisco, CA; S Caglioti, Blood Systems Laboratory, Tempe, AZ; D Todd, S A Glynn, Westat, Inc., Rockville, MD; H E Prince, Focus Technologies, Inc., Cypress, CA; S H Kleinman, Westat, Inc. and University of British Columbia, Victoria, BC, Canada Background: To better understand the dynamics of West Nile Virus (WNV) viremia and antibody seroconversion (SC) and assess a possible role for IgM screening, a study was conducted to ascertain the temporal relationship between nucleic acid amplification technology (NAT) yield and IgM SC during the 2003 epidemic. Methods: Donations from North Dakota were prospectively screened by mini-pool (MP) -NAT using 16-unit MPs and the Gen-Probe/Chiron transcription-mediated amplification (TMA) assay. Reactive MPs were resolved to reactive individual donations (ID) then confirmed by repeat WNV TMA, PCR and IgM testing of index units and donor followup specimens. IgM testing was performed retrospectively on frozen archived MP-NAT non-reactive specimens that were available from North Dakota using the Focus Technologies IgM EIA, with confirmation by repeat IgM and IgG testing by Focus Technologies. IgM rates were adjusted to account for sample availability. Results: The table summarizes the results of MP-NAT screening of all 7,073 donations and IgM screening of a representative subset of 3,979 donations from July 1 through Sept 27, 2003. Viremic donations were first detected the week of July 13, peaked at 1% from 8/17-8/30, and were no longer detected beyond Sept 13. IgM prevalence (weighted for sampling) increased progressively from the time of initial detection of NAT yield cases to a level of 5.0% several weeks beyond detectable viremia. Conclusion: IgM was not detected early in the epidemic and thus would not be useful as an early screening test. As expected, IgM rates peak several weeks after detection of peak viremia. ID-NAT is being performed on the study sample for correlation of RNA with IgM reactivity as the epidemic evolves and to estimate the value of IgM screening for detection of low-level viremia. Late in an epidemic, IgM testing would lead to significant rates of unit loss and donor deferral. Even in a highly affected region, most donors show no evidence of exposure and would be susceptible to infection in future years, indicating need for ongoing donor screening.

S Orton, FDA, Rockville, MD; S L Stramer, J Paolillo, American Red Cross, Rockville, MD; R Y Dodd, American Red Cross Holland Laboratory, Rockville, MD Background: Screening of US blood donors for WNV began in 2003, and included asking donors about fever with headache in the 7 days prior to donation and WNV RNA nucleic acid testing under IND. Donors were asked to complete a questionnaire as part of follow-up. Study Design: A case control study of donors with reactive WNV NAT results was conducted. Donors were asked to complete a comprehensive questionnaire covering demographics, symptomatology and dates of onset, travel, residence, and other viral diagnoses or vaccinations. Donors with confirmed WNV NAT results were classified as cases; donors with false positive WNV NAT results were classified as controls. Results: 846 donors with reactive WNV NAT results were invited to participate. 313 cases and 308 controls (50% of each) participated. Cases and controls were similar in age (48 years median for both), but cases (compared to controls) were more likely to be male (83% vs 47%), white (52% vs 44%), have high school or less education (32% vs 26%), and live in rural areas (69% vs 36%). Symptoms were stratified as occurring prior to donation or the day of/after donation. Fever, headache, eye pain, severe muscle pain, new rash and general weakness reported either pre-or post-donation were significantly more frequent (p-value < 0.001) among cases. Post-donation chills, swollen glands, joint and bone pain were also significantly more frequent among cases. At least one Background: Although thrombotic thrombocytopenic purpura (TTP) has been described following cardiovascular and abdominal surgery, the diagnosis of postoperative TTP is frequently delayed. Anemia and thrombocytopenia may be attributed to postoperative bleeding or platelet consumption. Several reported cases of postoperative TTP were treated with therapeutic plasma exchange (TPE). One-third of these cases died, from TTP or an underlying condition, but in 3 surviving patients, no relapse was reported at 19 months post treatment. We report a case of postoperative TTP refractory to TPE. Case Report: A 58 year-old African American female with peripheral vascular disease underwent left aorto-femoral bypass graft. One week after surgery she experienced tingling and numbness in both hands, followed by two episodes of slurred speech. When a facial droop developed she sought medical help. A right hemiparesis and bilateral subacute to chronic parietal infarcts on head CT were consistent with cerebrovascular accident. Laboratory tests showed thrombocytopenia and microangiopathic hemolytic anemia. VWF Cleaving protease (ADAMTS13) activity was low and ADAMTS 13 inhibitor was elevated (see table) . Over the next few days the patient became confused and disoriented. A diagnosis of TTP was made. She was treated with daily TPE of one plasma volume with cryoprecipitatereduced fresh frozen plasma. TPE was discontinued after 8 procedures and she appeared to be in neurologic and hematologic remission. However, 3 days later her LDH was 857 U/L. After 8 daily TPE and oral prednisone 1 mg/Kg/day, she again achieved remission. She was discharged on a steroid taper, but outpatient follow-up revealed rising LDH and dropping platelets, consistent with refractoriness. Her plasma showed low ADAMTS 13 activity and a very high inhibitor. She received 8 additional TPE, followed by splenectomy. At last follow-up one week post-splenectomy she was free of symptoms and signs of TTP.

preference. We wish to understand which instrument to recommend for such donors when maximum platelet collection is the goal. Methods Evaluate the platelet collection capabilities of Amicus and Accel. Retrospective data were collected on 31 donors who qualified to donate on both instruments. Each donated on Amicus and Accel, and typically completed both donations within a 3-5 week period. Results are presented in Table 1 . Table 1 : Results of Amicus and Accel Conclusions: As far as we know, this is the first reported case of refractory TTP following vascular surgery. The very low level of ADAMTS 13 activity and high inhibitor level both at presentation and at diagnosis of refractoriness may correlate with the absence of a sustained response to TPE. Splenectomy may be a good therapeutic option in postoperative refractory TTP, as in other cases of refractory TTP.

E M Senaldi, M Calvetto, Blood Center of New Jersey, East Orange, NJ Background Demand for platelet products from our area hospitals continues to increase. To optimize supply, we have instituted a triage system that matches donors and technology to maximize platelet collections in a 90 minute donation. For each donor, triage begins with a review of prior donations, frequency, technology preference, ABO type, donor weight and previous platelet count. The triage then factors in the platelet pre-count done onsite with a QBC machine during the health history. At the end of the history, the phlebotomist will suggest the technology and procedure. We use both the Amicus Separator © software v2.5, double needle ("Amicus") and the Trima Accel Automated Blood Component Collection System © software v5.0, single needle ("Accel"). The donor can accept or reject the recommendation. Some donors qualify to donate on either Amicus or Accel and have no Conclusions Amicus provides better platelet collection capabilities than Accel. We were able to target, and get, more platelet products per donor on Amicus than on Accel. Both our double split rate and overall split rate were much higher on Amicus. The split rate for triples was roughly equivalent. For donors who qualify to donate on either Amicus or Accel and have no preference for either, we would recommend Amicus when maximum platelet collection is the goal.

M J McAteer, R Langley, Gambro BCT, Inc, Lakewood, CO Introduction: A plateletpheresis donor's ability to donate multiple platelet products is dependent on the donor's platelet count, total blood volume, and the post procedure platelet count used to select the donor. Based on concerns for donor safety, it is common practice to limit the donor's post procedure platelet count to ≥100,000 per uL. How that number is determined becomes important when configuring an apheresis system's minimum platelet post count. The Trima platelet post count predict algorithm is based on a simple mass balance equation with the number of platelets in the donor's blood pool, and the number of platelets put into the collect bag, determined. This conservative approach does not take platelet mobilization into account , a phenomenon known to occur for most donors. The purpose of this study was to evaluate platelet mobilization during Trima platelet collection procedures; and, based on that value, provide a recommendation for where to configure the Trima post procedure platelet count. Methods: A series of clinical studies were done where the donor's actual measured post platelet count was compared to the Trima predicted post count. The measured values came from an automated cell counter analysis of donor blood taken following the collection procedure, including rinseback. In the first study, data from 409 Trima system donors were analyzed and the ratio of measured to predicted platelet post count calculated. In the second study, 79 Trima Accel plaelet donors were analyzed. Results: The ratio of measured versus predicted post procedure platelet count for the 409 Trima donors was 1.27 ± 0.14. The ratio for the 79 Trima Accel donors was 1.28 ± 0.16. Discussion: These results show that following Trima platelet collection the donors have, on average, 27-28% more platelets than predicted. A donor predicted to have an 80,000 per uL post count would actually have 101,600 to 102,400 per uL circulating platelets. Linear regression analysis of the data showed that less than 4% of the donors had ratio's of 1.0 or less. Very few donors with typical platelet counts approach the post count disqualification level, so the risk of over depletion is very small. Conclusion:

H L Groth, P B Contestable, P Hosimer, E Grogan, S Hess, E Woodward, H Warren, S R Lee, Ortho Clinical Diagnostics, Rochester, NY Background: Chagas' disease is caused by the parasite Trypanosoma cruzi and is endemic to regions of Latin America. The risk of transmission of T. cruzi through blood transfusion in the United States is a growing concern, due to increased immigration to the US. We have developed an ELISA for antibody to T. cruzi suitable for screening blood donations in a low prevalence setting. Methods: The Ortho screening assay for detection of antibody to T. cruzi uses a standard ELISA procedure in a microtiter plate coated with T. cruzi antigens prepared from lysed epimastigotes. The cutoff value for the assay is calculated by the mean absorbance of the Positive Calibrator at 492 nm multiplied by a constant. Clinical sensitivity was assessed in 755 putative Chagas' disease specimens from 12 countries. Specificity was assessed in 37,103 random US blood donor specimens, 500 matched serum/plasma samples, and 187 specimens with potentially interfering serologies (pre and post influenza vaccine, ANA, RF, schistosoma, malaria, syphilis, leishmania). Repeat reactive (RR) samples had radioimmunoprecipitation assay (RIPA) testing performed to confirm antibody status. Reproducibility was assessed by 3 operators running 3 plates per day for 3 days using a panel of 3 reactive samples and 1 non-reactive sample. Potential Of the 221 seropositive donors, 52 were repeat donors. Most donors only had two donations available for analysis, however, several donated at least three times during the 3-year period. Fourteen donors were initially seronegative and seroconverted upon further donation. Three of the donors who seroconverted later cleared the infection. Seventeen initially seropositive donors subsequently cleared the infection. Twenty-one donors were identified as seropositive and remained seropositive throughout. Conclusions: During the 3-year period, the seroprevalence of A. phagocytophilum was similar each year. Examination of repeat donations demonstrated new infections, antibody persistence over time, as well as clearance of antibodies and presumably the agent. It's unclear if antibody persistence or apparent acute infections pose a risk for transfusion transmission. Given the high seroprevalence rates and the agent's survival in stored blood, it's surprising only one transfusion case has been reported. With the use of leukoreduction, it's possible that A. phagocytophilum is removed prior to transfusion. The reported case of transfusion-transmitted HGE involved 2 units of non-leukoreduced RBCs. Further research is needed to define the threat A. phagocytophilum poses to blood safety. BACKGROUND: The recent occurrence of a variant Creutzfeldt-Jakob Disease (vCJD) in the United Kingdom has raised the question of the possibility of transmission of the causative agent by blood transfusion from infected individuals with no clinical symptoms of the disease. In the present study, we evaluated the performance characteristics of a prototype of a new leukocyte reduction filter for the removal of infectious prion from red cell concentrates (RCC). STUDY DESIGN: Endogenous Infectivity Study: A total pool of 500 mL of whole blood were collected into CPD anticoagulant from scrapie infected hamsters. The blood sample (500 mL) was processed into RCC. About 300 mL of the RCC in SAGM additive solution were filtered with a prion removal filter (Pall Medical, NY) . The levels of infectious prion in the RCC before and after filtration were measured with a Western blot assay. In Vitro Spiking Studies: RCC in AS3 or SAGM additive solutions were obtained from AABB accredited blood banks. Ten Percent (10%) scrapie brain homogenates (SBH) in buffered saline was prepared from brains of hamsters infected with 263 K hamster-adapted scrapie (PrPsc). About 30 mL of the SBH was added into 270 mL of RCC. The SBH-contaminated RCC was filtered at room temperature with a prion removal filter. The levels of SBH in the RCC were determined before and after filtration with a Western blot assay. In addition, about 50 micro liters of the pre and post filtration RCC were injected intracrebrally into healthy hamsters. Leukocyte Removal Efficiency: In addition to prion removal, the levels of residual leukocytes in the filtered RCC were measured with a flow cytometric method. RESULTS: Results from the endogenous infectivity study that simulates a possible clinical condition, showed complete removal of the infectious prion from full unit of RCC to a level that is below the limit of detection of the Western blot assay of about 2 logs. After about 300 days, none of the animals injected with filtered PrPsc-contaminated RCC have developed any clinical symptoms of the disease. In contrast, some of the hamsters that received unfil-tered RCC have developed the disease. The mean concentration of residual leukocytes in all the units (N = 20) processed with the prion removal was well below of 1.0 ¥ 10 6 leukocytes per unit. CONCLUSION: The present results show that prototype of a prion removal filter was effective in removing infectious prion from RCC prepared from scrapie contaminated WB below the limit of detection of Western blot assay. In addition, none of the hamsters injected with filtered RCC have developed the disease after 300 days. The use of this type of filter may help reduce the risk of vCJD through blood transfusion.

L Qu, Institute for Transfusion Medicine, Pittsburgh, PA; D Rowe, S Xu, University of Pittsburgh Graduate School of Public Health, Pittsburgh, PA; D Triulzi, Institute for Transfusion Medicine, Pittsburgh, PA BACKGROUND: Epstein Barr Virus (EBV) infects and establishes latency in B lymphocytes in majority of the adult population, including blood donors. Leukoreduction (LR) has been shown to reduce the number of CMV infected leukocytes from blood components. The current study was designed to measure the effect of LR on EBV infected B cells in RBC components. METHODS: Sixteen AS-5 RBC units were LR in the laboratory within 72 hours of collection according to manufacturer's instructions. Pre and post LR samples from each unit were assayed for EBV DNA. Lymphocytes were isolated by Ficoll density gradient centrifugation. On pre LR samples, B cells were further purified by CD-19 magnetic bead selection of the lymphocyte fraction. On post LR samples, the entire lymphocyte fraction was analyzed for EBV DNA. All samples were assayed for EBV genomic copy number by quantitative TaqMan® PCR. RESULTS: EBV genomes were detected in CD19+ B cells in 14/16 pre-LR RBC units. EBV genomic copy number in the units ranged from 0.18 to 96.84 per 10 5 CD19+ B lymphocytes representing approximately 45 to 24,210 total EBV genomes per bag. (Assumes 10 5 B cells per ml in normal donor blood, and RBC component contains half of the total lymphocytes from whole blood). LR rendered all but one unit EBV negative by PCR. The lone PCR positive unit after LR amplified only 1.2 EBV genome copies from lymphocytes recovered form the entire unit of RBC; this unit had the highest EBV viral load pre-LR (24,210 EBV genomes). These results indicate that a 4 log reduction of EBV genomic copy number can be achieved with LR of RBC units. CONCLUSIONS: LR reduces EBV genomes in RBC units by 4-logs and therefore renders most RBC units EBVnegative by sensitive PCR. Clinical studies are needed to determine to what extent EBV infectivity is reduced by LR.

J-P Allain, University of Cambridge, Cambridge, United Kingdom; N Etiz, Refik Saydam Hygiene Centre, Ankara, Turkey; A Parsyan, University of Cambridge, Cambridge, United Kingdom; D Candotti, National Blood Service, Cambridge, United Kingdom Background: To reduce the risk of infection with HE by transfusion, plasma pool screening and exclusion of pools with >10 4 IU/ml is implemented. However, persistent infection with lower viral load was reported and might constitute a safety risk in transfused blood components. HE persistence (viral DNA and specific IgG) was assessed in blood donations from the UK and Sub-Saharan Africa. Design and Methods: HE DNA was screened by nested-PCR in pools of 10 plasmas from 2,440 blood donations from the UK (1000), Ghana (1000), Malawi (80), and South Africa (360). Viral load was quantified by real-time Q-PCR in individual samples from positive pools. Specific IgM and IgG were screened with Biotrin kits. Complexed and free virus were separated by protein G chromatography. NS1-VP1 region was sequenced and phylogenetically analyzed. Results: Persistent EH DNA was detected in 0.5-1.3% and IgG in 62-85% of the donations. Persistent infection was characterized by low viral load (median: 558 IU/ml; range: 42 -1.3 ¥ 10 5 IU/ml), antibody-complexed virus, free specific IgG, and potentially infectious free virus (5-70% of total EH DNA). HE genotype 1 was prevalent in the UK, Malawi and South Africa. In contrast, only HE genotype 3 was prevalent in Ghana. HE genotype 3 had considerable genetic diversity clustering in two probable subtypes. EH genotype 1-based antibody assays failed to detect 38.5% of Ghanaian samples containing antibodies to HE genotype 3 but detected cases of genotype 3 persistent infection. Conclusions: Persistent human erythroviral infection is present in approximately 1% of blood donors. Persistence is characterized by low viral load and variable Background: The sickle cell hemolytic transfusion reaction (HTR) syndrome consists of an acute or delayed HTR along with manifestations of a sickle cell pain crisis and a drop in Hb below pretransfusion levels. The worsening anemia is due to destruction of donor RBCs, suppressed erythropoiesis, and possibly hyperhemolysis of autologous RBCs (bystander hemolysis). It is recommended that transfusions should be avoided, as they can lead to a further drop in Hb. Case Report: A 21 year old man with sickle cell disease and a history of transfusion was hospitalized for pain crisis and acute chest syndrome. His blood bank workup showed antibodies to Fy a , Jk b , and hr S and his DAT was negative. He was transfused 2 cross-match compatible units of RBCs that were negative for C, e, hr S , K, Fy a , Fy b , Jk b , and S. The pre-and posttrasfusion Hb levels were 7.5 and 8.3 g/dl. Three days later, his pain worsened, his Hb level dropped, and he had hemoglobinuria. IVIG, methylprednisolone, erythropoietin, and rituximab were administered. Fifteen days posttransfusion his Hb concentration dropped to 3.1 g/dl, with evidence of hemolysis of both donor and autologous RBCs, resulting in severe congestive heart failure and acute renal failure. No new antibodies were identified; a repeat DAT was positive only for complement, and a RBC eluate was negative. A decision was made to transfuse RBCs to overcome his circulatory failure. A combination of fluid overload, anemia, and the possibility of immune-mediated destruction of donor RBCs prompted us to design a novel isovolemic procedure to increase his Hb level without removing his own RBCs or causing fluid overload. This goal was achieved with a plasma-to-RBC exchange on a COBE Spectra cell separator primed with RBCs resuspended in normal saline and operated on a plasma exchange program, but substituting 3 cross-match compatible packed RBC units as the replacement instead of FFP. After the procedure, there was no evidence of hemolysis. Over the following six days, his heart failure resolved, his Hb concentration rose steadily to 7.5 g/dl, and he was transferred out of the ICU, and subsequently discharged home in good condition. Conclusions: Plasma-to-RBC replacement might be beneficial for patients with life-threatening anemia as a result of sickle cell HTR syndrome. This intervention may provide immediate improvement in oxygen-carrying capacity while conserving the patient's own RBCs. This case provides another report of successful transfusion in sickle cell HTR syndrome utilizing treatment with steroids, IVIG, and rituximab. Background: Henoch-Schonlein purpura (HSP) is the most common small vessel vasculitis in children. It is an immunoglobulin A-mediated disease affecting skin, joints, gastrointestinal tract (GI) and kidneys. The classic presentation of rash on lower extremities and buttock area is generally self-limited, requiring corticosteroids only for significant symptoms. Immunosuppressive therapy and plasma exchange (PE) have been reported for life-threatening complications. To our knowledge, this is the first report to describe successful outcome of PE for HSP and massive GI bleed. Case Report: The 15 year old patient with history of hemophilia A and high titer Factor VIII (F VIII) inhibitor levels was diagnosed with HSP one month prior to this admission. At that time, he presented with palpable purpura, abdominal pain, bloody diarrhea and arthralgia. The diagnosis was confirmed by skin biopsy showing deposition of IgA immune complexes in the small vessels. The patient improved and then relapsed and was admitted for further work-up and therapy. Four days prior to initiation of PE the patient was noted to have frank blood in his stool which rapidly progressed to massive bloody diarrhea despite administration of aminocaproic acid at 30 mg/kg every 6 hours, methylprednisolone at 0.5 mg/kg BID, octreotide to a maximum of 1 mcg/kg/hour and continuous recombinant F VIII infusion at 3 units/kg/hour. Tagged red blood cell (RBC) scan showed active bleeding near hepatic flexure. In the 3 days prior to initiation of PE, the patient required 675 mL, 1344 mL and 2137 mL of packed RBCs (pRBC) respectively with hematocrit values from 30.7% to 19.2% and 0, 1 and 4 units of apheresis platelets respectively to maintain a platelet count >50 k/ml. Single volume PE, the only alteration in therapy in an attempt to achieve hemostasis, was initiated with fresh frozen plasma (FFP). In the 24 hours subsequent to the procedure the patient required 510 mL of pRBCs and then no further pRBCs. Review of his laboratory data strongly suggested that the hemostasis was not due to replacement of his coagulation factors with FFP. At the time of the most severe bleeding, the platelet count was 82 k/mL, prothrombin time was elevated to 15.9 s (Reference 10.5-13.5 s), while his partial thromboplastin time, fibrinogen level and F VIII level were within reference range. The patient received 4 total daily PE procedures and then every other day procedures. His stool was guaiac negative after a total of 6 procedures. Conclusion: Evidence from our experience suggests that PE has a valuable therapeutic role in treatment of steroid-resistant HSP with massive gastrointestinal hemorrhage.

S A Woloskie, L Cooling, R D Davenport, University of Michigan, Ann Arbor, MI Background: Waldenstrom's macroglobulinemia is a rare B-cell malignancy characterized by a monoclonal IgM paraproteinemia. Complications can include bleeding, Raynaud's phenomenon, thrombosis, hyperviscosity syndrome, and cryoglobulinemia. We describe a patient with Waldenstrom's and a severe type I cryoglobulinemia that significantly hindered therapeutic plasmapheresis (TPEX). Case Report: A 45-year-old man presented with new Waldenstrom's macroglobulinemia, retinal vein thrombosis, and vision loss. Laboratory studies were remarkable for a severe type I monoclonal IgM cryoglobulinemia (cryocrit 98%), which required strict collection and transport of blood specimens at 37-40 C to prevent congealing. Emergent TPEX had been attempted at two outside facilities without success due to circuit occlusion. The patient was subsequently transferred to our facility for TPEX prior to starting chemotherapy. Our first attempt at TPEX was also unsuccessful despite aggressive anticoagulation, hydration, warming blankets and a blood warmer. Subsequent procedures were performed in a heated room (>34 C), pre-warmed prime and replacement fluids (37 C), and insulation of all access/return lines by wrapping lines with towels. In addition, forced-air warming blankets were placed over the patient and the apheresis machine. To prevent rouleaux, intial settings were hematocrit = 65%, inlet whole blood/anticoagulant ratio = 10 : 1, and inlet flow rate = 40 mL/min. Saline flushes were needed throughout the procedure to maintain separation and circuit patency. Over four hours, 9 L of blood were processed and 1300 mL plasma removed. A total of six procedures were completed with improved vision, decreased serum viscosity (>4.0 to 1.84 cps), IgM (>6 to 3.5 gm/dL), and cryocrit (98% to 75%). Conclusion: TPEX is an effective treament for Waldenstrom's macroglobulinemia, with even small reductions in IgM leading to reduced serum viscosity and clinical improvement. TPEX can be complicated, however, with IgM paraproteins possessing a high thermal amplitude. In our patient, the use of forced-air warming blankets was instrumental in preventing circuit loss due to cryoprecipitation. Raising the input hematocrit, coupled with aggressive anticoagulation and prewarmed solutions, also helped facilitate separation and maintain circuit patency.

Automated Red Cell Reduction in Polycythemia Using a Non-Therapeutic Apheresis Instrument P M Carey, S L Pinkard, R A Sacher, Hoxworth Blood Center, Cincinnati, OH INTRODUCTION: Polycythemia of primary or secondary etiology can cause significant ischemic medical problems for patients. Reducing the red cell volume without compromising the patient's oxygenation or compensatory mechanisms is one mode of management. We report our experience, using automated red cell collection (AR) with an approved apheresis instrument to provide therapeutic care in three patients. STUDY: Each patient agreed to AR after discussion of the risks and benefits. The instrument was operated without any programming changes. It was necessary to override the patient HCT because the program is set for a maximum donor HCT of 55%. Because of increased blood viscosity in these patients, access pressure alarms would sound at intervals. To resume flow, normal saline boluses were administered to help reduce the viscosity. As part of the routine AR procedure, 80% of the RBC collected volume was replaced as normal saline. This was in addition to the bolus volumes. CONCLUSION: The procedures lasted 22-25 minutes. Each patient had at least two AR procedures. The interval between procedures ranged from 12 h-12 d. The patient outcomes are summarized in the table. The patients were stable and tolerated the procedures without adverse events.

All patients exhibited a reduction in their red cell volume as reflected by the HCT. These reductions were 22.7%, 28.4% and 6.4%, respectively. Patient 3 did not exhibited the expected reduction based on a clinical management decision as the patient prepared for an operative procedure. Automated double red cell collection is a viable alternative in managing therapeutic phlebotomy in patients with a presenting hematocrit >55%. Background: Psoriasis is a common chronic inflammatory skin disease that affects approximately 2 per cent of the world's population; it is characterized by a marked inflammatory infiltrate and hyperproliferation of keratinocytes. The infiltrate is composed of skin-infiltrating T cells, predominantly of the memory phenotype, neutrophils, lining macrophages, and increased numbers of dendritic cells (DC) . There is evidence that T cells play a crucial role in the immunopathogenesis of this disease. Lymphocytapheresis (LCP) has been frequently used in the therapy of diseases which have a suspected cellular immune mechanism, and its therapeutic efficacy is probably ascribable to the depletion of activated lymphocytes. Case report: We performed a course of LCP, with the aim to remove lymphocytes, in a 54 female outpatient who had a 30 years' history of psoriasis vulgaris (PV) and had been treated only with topical corticosteroids, which produced only transient effects. Psoriasis area and severity index (PASI) before the LCP scheduled was 33. Five weekly LCP were performed , without any side-effects, by Cobe Spectra auto-PBSC (V 6.1) cell separator through MNC software; the collection hematocrit was adapted in order to select lymphocytes. The mean value of: a) lymphocytes number removed per each session was 3 ¥ 10 9 ; b) procedure time was 108 minutes; c) buffy coat volume depleted was 84.2 mL: all that processing an average of the patient's 1.2 total blood volume (3,726 blood mL). No significative variation due to LCP was noticed in the absolute leukocyte numbers and in the relative numbers of T, B, NK, antigen presenting cells, and monocytes in peripheral blood before and after each LCP, and one week after the last apheresis session. The skin lesions improved dramatically and, after the last LCP, PASI score was reduced to 8. No LCP side-effect was reported. Conclusions: Our results show that PV can be treated very effectively through LCP; these results are very different from the results of some authors who treated PV by granulocyte and monocyte adsorption apheresis. LCP mechanism of action has to be focused better in order to define more precisely the mechanisms underlying its therapeutic potential, which probably is not limited to cell-depletion, but targets the complex series of interactions between lymphocytes, antigen, and DC. Since most therapies that are currently available for psoriasis have doselimiting toxic effects, LCP could become a valuable therapeutic alternative. Moreover the rapid clearing of skin lesions is an important aspect of effective psoriasis management and may correlate with the patient's satisfaction with treatment.

BACKGROUND: Red blood cells regulate tissue circulation and O 2 delivery by releasing the vasodilator adenosine triphosphate (ATP) in response to hypoxia. When released extracellularly ATP is rapidly degraded to adenosine diphosphate (ADP) in the circulation by ectonucleotidases. ATP and ADP activate subtypes of the large P2 receptor family (15 subtypes). Here we show that ADP acting on P2Y 13 receptors on red blood cells serves as a negative feedback pathway for the inhibition of ATP release. METHODS: mRNA was quantified with real-time PCR. Western blot was used to detect P2 receptors with available antibodies. cAMP levels were determined with an enzyme immunoassay. ATP release was measured in incubated red blood cells using microdialysis and a luciferase assay. In a pig model, catheters were inserted through the carotid artery to place a catheter in the left coronary artery, and through the jugular vein to place a microdialysis probe in the coronary vein. 2-MeSADP was injected in the artery and ATP levels were measured in the coronary vein. RESULTS: mRNA of the ADP receptor P2Y 13 was highly expressed in human red blood cells and reticulocytes, whilst other ADP receptors were not. The stable ADP analogue 2-MeSADP decreased ATP release from red blood cells by inhibition of cAMP. The P2Y 12 and P2Y 13 receptor antagonist AR-C67085 (30 mM), but not the P2Y 1 blocker MRS2179, inhibited the effects of 2-MeSADP. At doses where AR-C67085 only blocks P2Y 12 (100 nM), it had no effect. AR-C67085 and the nucleotidase apyrase increased cAMP per se, indicating a constant cAMP inhibitory effect of endogenous extracellular ADP. 2-MeSADP reduced plasma ATP concentrations in an in vivo pig model. CONCLUSION: P2Y 13 is selectively expressed in human red blood cells. The ATP degradation product ADP inhibited ATP release by acting on this receptor. This negative feedback system could be important in the control of plasma ATP levels and tissue circulation. Because blood consists of approximately 40% red blood cells, containing a 1000-fold higher ATP concentration than plasma (mM vs. uM), even a minor release of ATP from the high intracellular concentrations could have major circulatory effects. A negative system may therefore be of great physiological importance to mitigate ATP release. In addition, this finding could be of interest for efforts to preserve intracellular ATP in red blood cells during storage. Background. Loss of phospholipid (PL) asymmetry in the membrane of red blood cells (RBC) results in exposure of phoshatidylserine (PS) and to subsequent removal from the circulation. In this study, we investigated during long-term storage of RBC two activities affecting PL asymmetry: the ATPdependent translocase (or flippase, transporting PS from outer to inner leaflet) and PL scrambling (which could move PS from inner to outer leaflet). Methods. RBC concentrates were stored in SAGM at 4°C for up to 7 weeks. Flippase was measured by determining the inward translocation of NBD-PS by flow cytometry. Scrambling activity can most conveniently be measured by determining the inward translocation of NBD-phosphatidylcholine. PS exposure was measured with AnnexinV-FITC. Results were correlated with cellular ATP levels. Results. Flippase activity started to decrease at day 28 of storage and reached 51 ± 3% (mean ± SEM, n = 4) of the control value after 7 weeks. Despite this decrease, PS exposure remained below 5%. Scrambling activity remained absent during storage. The decrease in flippase activity was correlated with the decline in cellular ATP, but rejuvenation of RBC after 7 weeks to increase ATP levels only partially restored flippase activity. Conclusions. Long-term storage of RBC does affect the ATP-dependent flippase due to the decline in cellular ATP and a decrease in catalytic activity. PS exposure remains low, probably due to the absence of scrambling activity.

L Thibault, A Beauséjour, M J De Grandmont, C Côté, J Perreault, G Dumas, J F Leblanc, R Lemieux, Héma-Québec, Sainte-Foy, PQ, Canada

Background. Whole blood-derived platelet concentrates (PC) must currently be prepared within 8 hours (8 h) after blood collection. In opposition, the buffy coat (BC) method, allows a 24-hour (24 h) hold (H) at room temperature (RT) before blood processing. The maximal 8 hH at RT is problematic for late day blood clinics since preparation of components must be done at night.We have evaluated the effect of a 24 hH of whole blood at RT on the quality of components prepared by the PRP method. Study design. The effect of a 24 h versus 8 hH at RT was studied using blood bags collected from 130 volunteers. To minimize variations, blood were simultaneously collected from 2 ABO-identical donors, immediately pooled and splitted. The bags were stored for either 6-8 h or 22-24 h at RT. Rapid cooling to RT of the 24 hH group was achieved with Compocool® plates as used in the BC method. Leucoreduced red blood cells (RBC), PC, fresh frozen plasma (FFP), cryoprecipitate (Cryo) and cryo supernatant (CryoS) components were prepared using the standard PRP method. Samples for testing were aseptically recovered at various times during storage and analyzed for 24 in vitro parameters routinely used to assess the quality of blood components. For RBC and PC, the 8 hH and 24 hH results were compared in each donor pair and were normalized to the result of the day 1 sample of the 8 hH group. Results. For RBC, the results of the two groups were not significantly different (p > 0.01) for weight, volume, glucose and sodium. Highly significant (p < 0.01) differences were observed in the 24 hH group for pH (-0.05 unit), 2,3-DPG (-36%) and lactate (+49%) at day 1, but those were reduced at day 6 and no longer apparent at day 21, 42 and 56. The residual leukocyte counts were statistically higher in the 24 hH group. For PC, the results were not statistically different for counts, glucose, and ESC. Significantly lower pH (-0.13 unit) and higher lactate (+53%) were observed in the 24 hH group at day 1 but the differences were no longer observed at day 5 and 7. PC of the 24 hH group expressed less CD62p (-9%) and were more responsive to thrombin (+21%) at day 7. For frozen FFP, Cryo and CryoS, no significant differences were observed in the two groups for FV, vWF, total proteins, Fg and FVIII (FFP and CryoS) . The FVIII content of Cryo was slighty lower (-11%) in the 24 hH group. Conclusions. The transient effects observed indicate that the quality of stored RBC and PC is not irreversibly affected by the prolonged 24 hHb at RT and that the metabolic reserve of the components is sufficient to correct the adverse effects within days. It thus appears that the 24 hH of the buffy coat method could also be acceptable for use in the PRP method. Background: Currently the FDA regulation allows for the freezing of red cells on day 6 and frozen storage of additive red cells from whole blood 24 hours after thawing and deglycerolization. However, red cells collected on the automated Trima blood collection system have not been approved for freezing, thawing and 24 hour post-storage. Objective: The purpose of this study was to compare the in vivo and in vitro results of leukoreduced (LR) Trima Accel apheresis (Test) RBC units to RBC units (Control) derived from whole blood (WB) after freezing and thawing. Methods: Twenty-one paired normal donors from two sites were randomized to donate a Trima RBC unit and a WB unit on two separate occasions with at least 8 weeks between donations. Trima units were LR using Trima integrated filters. WB units were LR with Pall filters. All units were kept at room temperature for no longer than 8 hours after collection. AS-3 was added to Trima units and AS-3 or AS-5 was added to the WB units. Units were stored at 2-6°C for six days and then frozen according to the Meryman method. After a minimum frozen storage of 28 days, units were thawed using the Cobe 2991 cell washer and resuspended in saline and returned to 2-6°C storage for one day. On Days 0, 6 (pre-Freeze), and one day post thaw samples were removed for in vitro assays. At the end of storage, in vivo 24-hour % recovery was measured using the 51 Cr single label method. Results: The 24-hour % recovery for one day post-thaw Trima pRBCs (89.6%) showed similar results to control 15A 2004-Vol. 44, Supplement pRBCs (89.5%) and were well above the FDA's recommendation of greater than or equal to 75%. ATP was significantly higher in the Trima pRBCs. Conclusion: WB and Trima RBC units frozen (on Day 6) and thawed were not associated with significant increases in electrolytes or unexpected hemolysis. Both WB and Trima RBCs frozen on Day 6 and stored for a minimum of 28 days are resilient to the freeze/thaw process and have acceptable in vitro and in vivo results. In Vitro Characteristics of Trima and WB units Post Thaw received oral dexamethasone (8 mg) plus a single subcutaneous injection of glycosylated G-CSF (Lenograstim) at medians of 1.5 (1.0-2.3) mg/kg (n = 43), 3.1 (2.4-4.1) mg/kg (n = 73), 6.0 (4.3-7.9) mg/kg (n = 123), and 12.0 (8.2-17.2) mg/kg (n = 21), and underwent neutrophil aphereses using the Spectra PMN program. We compared WBC and PMN mobilization and collection results and evaluated the severity and clinical significance of donor adverse reactions. Fifty-two neutropenic patients (29 children, 23 adults) underwent 271 neutrophil transfusions (GTX) every other day to maintain peripheral WBC levels constantly above 500/mL. RESULTS: Within the dose range 1.5/3/6 mg/kg each doubling step was associated with a 10-15% PMN increase in peripheral blood up to 32.8 (19.1-49. 2) ¥ 10 9 /L (6 mg/kg; p £ 0.00032) as well as in the neutrophil concentrate up to 79 (34-150) ¥ 10 9 /U (6 mg/kg; p £ 0.00042). A further doubling to 12mg/kg achieved neither better mobilization nor better apheresis results. The rate of clinically important adverse reactions increased already with the 6 mg/kg mobilization step. The GTX resulted in peak WBC increments to 3.8 (0.4-18.2) ¥ 10 9 /L (children) and 1.6 (0.3-9.4) ¥ 10 9 /L (adults), but in adults the WBC threshold of 500/mL was not continuously overcome. CONCLUSIONS: The most effective dose / response ratio for PMN mobilization was demonstrated in the 6 mg/kg Lenograstim group. In neutropenic adults GTX treatment on an every other day schedule may be ineffective.

A Grinev, S Daniel, A Machuca, DEETD/OBRR/CBER/FDA, Bethesda, MD; S L Stramer, American Red Cross, Gaithersburg, MD; I K Hewlett, M Rios, DEETD/OBRR/CBER/FDA, Bethesda, MD BACKGROUND: The objective of this study was to investigate genetic variation of WNV over the 2002 epidemic in the US. Variation in the viral genome may affect pathogenesis, possible treatment and interfere with screening and diagnostic assays. WNV first appeared in the US in 1999 and since then epidemics have reoccurred for 4 consecutive years. The US epidemics of 2002 and 2003 led to the largest meningoencephalitis outbreaks in the Western hemisphere and the largest WNV outbreaks ever reported. Preliminary molecular phylogenetic studies divide isolates of WNV into two lineages. Linage II strains are mostly found in Africa. Lineage I strains are more widely distributed and have been responsible for all the recent large outbreaks. Although immeasurable efforts have been directed to the study of WNV, very little is known about its potential for genomic variation. METHODS: Plasma samples from 7 blood donors, who were identified as WNV-NAT-positive during 2002 epidemic, were used for virus isolation in Vero cells. RNA was extracted from the isolates, reverse transcribed and the cDNA used for sequence analysis. One isolate was completely sequenced while 6 other isolates had all viral structural regions (5'NTR, Cap, preM, M and Env) sequenced. Sequence results were analyzed using Vector NTI software and compared to genetic sequences from the prototype WN-NY99 isolate. RESULTS: Analysis of the isolate that was entirely sequenced showed 20 nucleotide (nt) mutations plus one nt insertion at position 10497. Five of the 20 mutations were associated with amino acid (aa) substitutions (M22T; A52V; V449A; N684S; S2301G). Six of the 7 isolates from the 2002 epidemic shared two mutations (1442T>C [V449A] and 2466C>T [silent] ) in the envelope gene. One isolate had a single mutation (1487T>C silent) in the envelope gene that was not found in the other 6 isolates. CONCLUSION: The small number of nucleotide mutations in the envelope gene and the single conservative amino acid substitution in the envelope variable region suggest the absence of strong selective pressure and limited evolution of West Nile virus from 1999 to 2002. However, (1) some mutations lead to aa substitutions that could affect viral pathogenesis; (2) changes in nt and aa sequences may in the future affect sensitivity of screening and diagnostic assays.

Y Tang, C A Hapip, B Liu, S L Stramer, C T Fang, American Red Cross, Rockville, MD Background: In recent years, the frequency of West Nile virus (WNV) outbreaks in humans has increased, and these outbreaks have been associ-BACKGROUND: Flow cytometry (FC) is a rapid and sensitive method for quantitating residual WBCs (rWBC) in filtered blood components, but is typically complicated by the need for internal reference beads. We report that the Guava PCA-96[TM] capillary cytometer accurately quantifies rWBCs without reference beads. Furthermore, we describe an automated highthroughput workstation for staining samples that directly integrates the PCA-96 for data acquisition. METHODS: Whole blood samples from volunteer donors were double-leukoreduced for use as a diluent. WBCs were quantitated in non-leukoreduced samples from the same donors by automated hematology analyzer, and appropriate volumes were added to diluent to produce standard curves (50, 15, 3, 1, or 0 rWBC/microL). 15 rWBC/microL approximately corresponds to a blood component with 5 ¥ 10E6 rWBC. Samples were assayed using LeucoCOUNT (BD), or by an in-house procedure where RBCs were lysed and stained with propidium iodide (PI) or ViaCount Flex[TM](Guava). Addition of staining reagents was initially performed manually. Subsequently, samples were automatically stained in 96well plates on a Tecan Genesis workstation and read on an integrated PCA-96. The results were compared to counts obtained with the Leuco-COUNT kit read on a BD FACScan flow cytometer. RESULTS: The PCA-96 accurately determines sample volume assayed, potentially obviating the need for addition of internal reference beads to the samples. To validate the volume determinations, identical standard curve samples were tested using either LeucoCOUNT reagents (including beads) on a BD FACScan or the same reagent (without beads) on the PCA-96. The results were similar, and accurately approximated the expected rWBC counts (e.g., 14.2 vs. 15.4 at 15 rWBC/microL), although the PCA-96 produced significantly lower CVs (5.1 vs. 14.6; p = 0.002) at 15 rWBC/microL. Thus, volume determination on the PCA-96 can replace internal bead standards for normalizing rWBC counts. Similar results were obtained when standards were assayed in parallel by LeucoCOUNT (at 15 rWBC/microL: 15.3, CV = 13.2), PI (16.0, CV = 4.3), or Flex (17.1, CV = 12.6) on the PCA-96. Results were nearly identical when rWBC were counted by completely automated PI assay (at 15 rWBC/microL: 16.0, CV = 5.9). CONCLUSIONS: Low levels of rWBC are accurately quantitated on the PCA-96, and performance is not compromised by automated staining on an integrated workstation. This rapid, highthroughput system would be applicable to meeting any increase in QC requirements imposed on filtered components. We are currently optimizing the techniques for counting rWBCs in platelet components.

H-G Heuft, L Goudeva, Hannover Medical School, Hannover, Germany; N Schwella, University Clinic Marburg, Marburg, Germany; N Pulver, L Grigull, R Blasczyk, Hannover Medical School, Hannover, Germany OBJECTIVE: To evaluate the dose -response relationship of Lenograstim plus dexamethasone for neutrophil mobilization and collection for transfusion. STUDY DESIGN: In a prospective study, 260 healthy volunteers ated with a higher incidence of severe diseases. WNV nucleic acid sequence determination can not only be used to identify the WNV strains responsible for outbreaks but can also be used as a tool to study the relationships of the nucleic acid sequence mutation and the pathogenicity of WNV. Methods: WNV RNA was isolated and purified from the plasma of blood donors using the Qiagen Virus RNA Kit. A real-time TaqMan RT-PCR procedure was developed and used to confirm the WNV RNA in the samples and to determine its concentration. The sequence region coding for WNV envelope was amplified by RT-PCR using an appropriately designed primer pair. Amplified fragments were cloned into a plasmid DNA, pDrive (Qiagen), and then transformed into competent E. coli cells. Plasmid DNA containing the WNV sequence fragment was purified. The nucleic acid sequences were determined by Cycling Sequencing method using ABI Prism BigDye Terminator kit and read on an ABI Prism Genetic Analyzer. The NCI Vector NTI program was used to analyze the WNV nucleic acid sequences. Results: Nucleic acid sequence fragments were amplified and sequenced from plasma samples of 8 transfusion-transmitted implicated cases in 2002 and from plasma samples of 10 positive blood donors in 2003. Our data showed that all sequences of WNV isolated from the positive donors were most closely related to the WNV isolated from NY-99 flamingo. Within the 850 nucleotides sequenced (genome positions 1683 to 2532), the number of nucleotide differences that occurred between WNV NY-99 and the WNV in the plasma samples of the positive donors ranged from 2 to 10. Relative to the WNV sequence of NY-99 flamingo, the 2003 samples had more nucleotide substitutions (2 to 10) than the 2002 samples (2 to 6). Furthermore, most of the nucleotide differences were in the third base of a codon. Therefore, the frequency of the amino acid mutations was much lower than nucleotide mutations among different isolates. The amino acid residue differences that occurred between WN-NY 99 and WNV in tested samples ranged 0 to 3. The relationship of the pathogenicity of the WNV and nucleic acid sequence mutation remains to be determined. Conclusions: These results demonstrate that the WNV responsible for the US outbreaks of 2002 and 2003 in human belongs to one WNV strain. This strain was most closely related to WNV isolate NY-99, which itself is closely related to a strain identified in Israel in 1998.

S Zou, F Musavi, E P Notari IV, K E Fujii, R Y Dodd, for the ARCNET Research Group. American Red Cross, Rockville, MD Background: Health and risk history questions have been introduced into the pre-donation interview process to identify blood donors believed to pose a higher risk of infectious diseases to recipients. This study assesses the current impact of some of those questions. Methods: Donor deferral and donation data were extracted from a database of a large blood system in the US. Prevalence rates of hepatitis B surface antigen (HBsAg) and antibodies to human immunodeficiency virus (anti-HIV) and hepatitis C virus (anti-HCV) were obtained for donors who were temporarily deferred in [2000] [2001] and later returned to donate blood in [2000] [2001] [2002] [2003] . The deferral questions included a history of the following in the past 12 months: close contact with viral hepatitis or receipt of hepatitis B immune globulin; blood transfusion or organ/tissue transplant; tattoo, ear/body piercing, acupuncture, accidental needle-stick, contact with someone else's blood, or drug snorting; syphilis or gonorrhea; sex with a drug user; sex with someone who has taken money or drugs for sex since 1977; giving money or drugs for sex; sex with anyone who has taken clotting factor concentrates; sex with anyone who has had AIDS or tested positive for the AIDS virus; sex with a male who has had sex with another male since 1977. The results were compared with prevalence rates for blood units donated by first-time and repeat donors in 2002. Results: Among donors temporarily deferred in 2000-2001 due to histories identified above, 24.33% (18,634) returned and donated blood in 2000-2003. More older deferred donors returned to donate blood. With gender and age differences controlled, returning donors had a prevalence rate of anti-HCV (0.097%) higher than repeat donors (0.005%, p < 0.01) but lower than first-time donors (0.256%, p < 0.01). The prevalence rate of HBsAg among returned donors (0.054%) was comparable to that among first-time donors (0.070%, p = 0.49). The prevalence rate of anti-HIV among returned donors (0.005%) was between that among first-time donors (0.012%) and that among repeat donors (0.001%) but the result was inconclusive due to the low number of observations. Conclusions: Presenting donors temporarily deferred in 2000-2001 for potential risk of infectious diseases and who later returned to donate, as a group, do not appear to have higher infection rates of HIV, HBV or HCV than productive first-time donors.

J P Swanevelder, A D Heyns, South African National Blood Service, Roodepoort, South Africa Background: South Africa (SA) is severely affected by an escalating HIV epidemic. Amidst HIV infection estimates of 9.3% in youth (<25 yrs) and 15.5% in adults (25 yrs and older), the South African Blood Transfusion services (SANBS), strives to limit the residual risk of transfusing HIV to less than 1 in 100 000 red cells transfused. Methods: A Risk Management Programme (RMP) based on a Multivariate analysis of HIV prevalence in donation cohorts, was developed and implemented in June 1999, changing recruitment practices of the time. Demographics and donation history of the donor at the time of donating, determined a Risk Category (RC) assignment between I and. RC's are as follows, donations in cohorts with prevalence <0.0099% (RC I), 0.010%-0.0999% (RC II), 0.100%-0.999% (RC III) and 1.000% + (RC IV Based on the Risk Management Model, the potential donor population is profiled, and procurement targeted to optimize donations from RC's I and II. Education programmes promoting safe lifestyles and aiming to decrease HIV infection in prospective blood donors were introduced, supplemented by pre-donation one-one-one interviews aimed at excluding donors exposed to possible HIV infection. As per Blood Safety Policy, blood is processed and issued to patients according to the RC hierarchy. Products from RC III donations are only issued in emergency and if no lower risk blood is available, with the informed consent of the prescribing health practitioner, who assesses the risk-benefit of the transfusion. Results: 1998 Results: 2002 Conclusion: Implementation of RMP has significantly increased blood supply safety providing objective means to manage HIV transfusion risk. As SA is still in an expanding phase of the epidemic, the RMP is a dynamic programme which is continually evaluated and adjusted. The model may be applied in other countries with a high prevalence of HIV infection.

J P Aubuchon, L Herschel, J Roger, Dartmouth-Hitchcock Medical Center, Lebanon, NH; H Taylor, P Whitley, Eastern Virginia Medical School, Norfolk, VA; J Li, R Edrich, R P Goodrich, Navigant Biotechnologies, Inc, Lakewood, CO Background: Pathogen reduction technologies (PRTs) for platelet (plt) components offer a means to address continued viral transmission risks and imperfect bacterial detection systems. We investigated the efficacy of apheresis (Trima Accel, Gambro) plts treated with riboflavin (B2) + UV light (Mirasol TM , Navigant Biotechnologies) in a single-blind, crossover study in comparison to untreated plts. Methods: Normal subjects (n = 24) donated plts on two occasions ≥2 wks apart. Units were randomized to control or test arms, the latter receiving the addition of 28 mL of 500 mM B2 and exposure to 6.2 J/mL UV light in a specially designed illuminator over 8-10 min. Plts were stored for 5 d in usual conditions. Biochemical and hematologic analyses were performed pre-and post-illumination on Day 0 and at the end of storage. An aliquot of each unit was labeled with 111 In oxine (Br J Haematol 1983; 44:717-23) , and recovery and survival were determined from samples taken over the following 10 d using the COST program (multiple hit model). Observed recoveries were corrected for radiolabel elution in the injectate, plasma radioactivity, and residual label present 10 days after injection. Com-17A 2004-Vol. 44, Supplement parison was via a paired t-test. Results: All subjects tolerated the infusions well. Plt content of treated units was maintained from Day 0 (4.1 ± 0.4 ¥ 10 11 ) to Day 5 (4.0 ± 0.4 ¥ 10 11 ). Treatment with B2 + UV was associated with an increase in lactate production (control: 0.039 ± 0.008; test: 0.072 ± 0.015 mmol/10 12 cells/h; p < 0.001) with concomitant increases in glucose consumption. pH (control: 7.38 ± 0.07; test: 7.02 ± 0.10) and swirling were well maintained during 5 d of storage in all units. P-selectin expression increased from 17.6 ± 11.1% (Day 0) to 32.2 ± 15.1% (Day 5) in control units and from 23.8 ± 14.9 to 59.4 ± 23.5% in treated units (p < 0.001). Recovery of treated platelets (52.0 ± 18.3%) was reduced from that of control platelets (67.8 ± 13.4%), and survival was similarly shortened (4.8 ± 1.4 vs. 6.2 ± 1.2 d; p < 0.001). With the Day 10 correction, apparent recovery decreased by 1.3% for both groups, and survival shortened by 0.4 d. Conclusions: Treatment of platelets with B2 + UV to reduce the risks of viral transmission and bacterial contamination are associated with some alterations in parameters of clinical efficacy. However, the treatment effects noted here allow these PRT platelets to retain in vitro and in vivo capabilities that are similar to PRT and licensed platelet components that have been shown to have useful clinical applicability. The recovery, survival and metabolic properties of Mirasol TM platelets should provide sufficient hemostatic support in thrombocytopenia to justify patient clinical trials. , based on the association of amotosalen-HCI (Psoralen S-59) and UVA rays, as pathogen inactivation system. Clinical trial phase III (euroSPRITE) has demonstrated that Pathogen Inactivated PLTc obtained in this way maintain their vitality and function. The object of our study was to evaluate the clinical use of Pathogen Inactivated PLTc (IP) on two categories of patients not studied during phase III of euroSPRITE trial. METHODS: the trial began in April 2003 and enrolled 44 patients divided in two cohorts. The first cohort consisting of 28 thrombocytopenic cirrhotic patients (CIp) on a waiting list for liver transplant who needed PLT transfusion (PLT-t) before invasive procedures, divided in 2 random groups (cases and controls). 14 of those patients received IP and 14 received random PLT (CP). The second cohort consisted of 16 paediatric patients (CAp) divided in two random groups of 8, with congenital cyanogen cardiopathy, subjected to cardiac surgery who, during extracorporeal circulation, underwent a high level of PLT destruction and needed PLT-t as per protocol. Eight of those patients received IP and 8 received CP. In all patients the following were monitored: pre-and post-transfusion platelet count, Corrected Count Increment (CCI) at 1 h (t 1 ) and at 24 hrs (t 2 ) after transfusion, bleeding, presence of transfusional reactions and the need of more PLT-t in the following 2 weeks. Results are shown in the table. CCI means were comparable for both CAp and CIp transfused with IP or CP. Wilcoxon analysis shows no statistical difference in CCI (p < 0,05) on t 1 and t 2 for either groups of each cohort. No bleeding, no presence of transfusional reactions and no need for more PLT-t were observed. CONCLUSIONS: Preliminary data show that IP has demonstrated no significant difference in CCI and in post transfusion platelet count in the two cohorts of patients studied, together with a therapeutic efficacy similar to random CP, with the advantage of inactivating the pathogens of transfusional interest, reaching closer to 0 risk level. The follow-up at 1, 3, and 6 months to date has not shown any adverse or side effects. ABO Inc were asymptomatic in 11 (46%), severe in 12 (50%) and fatal in one (4%) case. AHTR were due to ABO Inc in 13 (31%), to autoantibodies in 6 (14%), to irregular antibodies in 18 (43%) and to unknown causes in 5 (12%) cases. AHTR were mild in 18 (43%), severe in 22 (52%) and fatal in 2 (5%) cases. The 2 main antibodies responsible for DHTR were anti-E (19%) and anti-Jk a (17%). DHTR were fatal in 1 (1.4%), severe in 5 (7.2%) and mild in 64 (91.4%) cases. Conclusion: Increasing rates for DSTR only reflect better reporting of this type of reaction. The slight downward trend in ABO Inc is encouraging but it remains a serious risk of transfusion and the use of newer technologies such as electronic patient and unit identification devices could help to reduce this risk.

L Zhou, D A Giacherio, L L W Cooling, R D Davenport, University of Michigan Hospitals, Ann Arbor, MI Background: Circulatory fluid overload (CFO) due to transfusion results from infusion of too large volume of blood products or infusion at a too fast rate for a compromised cardiovascular system. Symptoms and signs include dyspnea, orthopnea, systemic hypertension, jugular venous distension, S3 and rales. On CXR, cardiomegaly and pulmonary edema can be seen. However, these manifestations and CXR findings are not specific for CFO and some can also be associated with febrile non-hemolytic transfusion reaction or transfusion-related acute lung injury (TRALI). B-natriuretic peptide (BNP) is a 32-aa polypeptide secreted from the ventricles in response to ventricular volume expansion and pressure overload. BNP has been used as a diagnostic tool in the emergency department to improve the diagnosis of CHF. This study is to explore the usage of BNP in the differential diagnosis of transfusion-related CFO. Methods: Blood samples in EDTA tubes were obtained from patients presenting with symptoms suspicious for CFO during or after receiving transfusions. BNP Triage immunoassays were performed on these samples. BNP level from a pre-transfusion sample from the patient (usually a T & S sample) was determined at the same time. Samples from patients presenting with symptoms not associated with CFO are used for control. Diagnosis of transfusion-related CFO was TRANSFUSION 2004-Vol. 44 , Supplement based on a careful review of patient's symptoms and signs, CXR findings, and pertinent clinical history (eg. patient's fluid balance 8 hrs and/or 24 hrs prior to transfusion). Results: Whole blood BNP level remained stable for up to 72 hrs. A BNP test result was defined as positive if the post-transfusion/pre-transfusion ratio was ≥1.5 and the post-transfusion BNP level ≥100 pg/ml. We found that in 29 patients who had BNP levels determined, 17 were diagnosed as having CFO, and the remaining 12 having either febrile non-hemolytic transfusion reaction or no transfusion reactions. In 17 patients with CFO, 14 had positive BNP and 3 had negative results. In comparison, 11 of the 12 patients who did not have a diagnosis of CFO had negative BNP results. Overall, BNP test has a sensitivity and specificity of 82.4% and 91.7% respectively in diagnosis of transfusion-related CFO. Its positive predictive value is 93.3% and negative predictive value is 78.6%. Conclusion: Our study suggests that in patients who present symptoms suggestive of CFO during or after transfusion, BNP can be a useful adjunct marker in confirming volume overload as the cause of dyspnea/tachypnea and symptoms related to cardiovascular compromise. (center B; N = 151) . In center B we measured the Pb concentration (conc) in donor RBCs, platelet-rich plasma (PRP) and platelet-poor plasma (PPP) using a 1 ml aliquot of WB submitted from a CBC sample and two 1 ml aliquots by sterile docking from each donor's platelet collection. In center A, two aliquots of WB were collected: a red-top tube used for RBC Pb conc determination and a serum-separator tube from which the plasma Pb sample was drawn. Plasma Pb conc was used as an internal control for all donor samples. Based on the World Health Organization (WHO)'s allowable oral Pb intake (25 mg/kg/week) with 10% Pb absorption, an allowable intravenous Pb intake would be 0.36 mg/kg/day. Using a single, daily RBC transfusion dose of 20 cc/kg, we calculated an acceptable Pb conc for donor RBCs to be <1.8 mg/dl. Using a similar calculation for platelets (transfusion dose of 10 cc/kg), the acceptable donor platelet Pb conc is <3.6 mg/dl. Pb concs were measured using the Spectra AA600 graphite furnace atomic absorption method (Varian, Inc, Palo Alto, CA) with an analytic time per sample of 5 minutes and a direct cost of $4.16. Results: Pb concentrated in RBCs with RBC Pb concs ranging from 0.2-17.9 mg/dl with a mean of 1.9 mg/dl (+/-1.9 mg/dl). PPP had Pb concs ranging from 0.1-1.4 mg/dl (mean = 0.5 mg/dl) and PRP had Pb concs ranging from 0.1-1.0 mg/dl (mean = 0.49 mg/dl). Extrapolating from WHO recommendations, 62.5% (188/301) of donors would be acceptable for neonatal RBC transfusion. One hundred percent of platelets and plasma would be acceptable for neonatal transfusion. There was no statistically significant difference in Pb levels between the PRP and PPP from each donor (p = 0.28). Conclusions: In our region, donor testing is fast and inexpensive with nearly two-thirds of our donors having RBC Pb concentrations below the presumed toxic level for neonatal infusion. It is thus feasible to screen donors for Pb and use "low-leaded" blood for neonatal transfusion therapy. Platelet and plasma Pb concentrations were not significantly increased. Therefore, platelets and plasma would be exempt from testing and determination of transfusion product eligibility.

A Janowska, L A Marquez-Curtis, Canadian Blood Services, Edmonton, AB, Canada; M Z Ratajczak, James Graham Brown Cancer Center, Louisville, KY Background: Clinical and experimental data provide evidence that platelets contribute to cancer metastasis, although their mechanism(s) of action remain unclear. Nevertheless, platelet concentrates are frequently transfused to cancer patients suffering from thrombocytopenia after chemother-apy. Recently we reported that microvesicles and exosomes derived from activated platelets (PMV) transfer various surface receptors/adhesion molecules to target cells and modulate their biological responses. In this work, we hypothesized that PMV may interact with cancer cells, increasing their invasive potential. Methods: PMV were isolated from outdated platelet concentrates and pre-incubated with human breast cancer cells (invasive MDA-MB-231 and BT-549, and non-invasive T47D cell lines), and their binding to these cells was evaluated by flow cytometry. The effect of PMV on the invasive potential of these cells was determined by evaluating the expression of matrix metalloproteinases (MMPs) which degrade basement membranes and matrix and non-matrix substrates, cell chemoinvasion and proliferation (using RT-PCR, zymography, trans-Matrigel invasion, Western blotting and proliferation assays). In addition, the effects of PMV on interactions of tumor cells with stroma were analyzed using co-cultures of breast cancer cells with bone marrow fibroblastic cells. Results: We found that PMV (i) transferred platelet-derived CD41 to the surface of breast cancer cells; (ii) stimulated the secretion of MMP-9 and enhanced chemoinvasion across the reconstituted basement membrane Matrigel of invasive MDA-MB-231 and BT-549 cells; (iii) increased phosphorylation of the MAPK p42/44 and AKT signaling pathways; (iv) increased the proliferation of the non-invasive T47D cells; and finally, (iv) induced the activation of proMMP-2 constitutively secreted by fibroblastic cells in co-cultures. Conclusions: We found that PMV enhanced the in vitro invasive potential of breast cancer cells, and because concentrations of PMV are known to be higher in old platelet concentrates than in fresh ones, we recommend that cancer patients should preferably be transfused with fresh platelet concentrates only.

The Role of Platelet Derived Soluble CD40L (Scd40l; CD154) in Adverse Reactions to Platelet Transfusions N Blumberg, K F Gettings, C Turner, J M Heal, R P Phipps, University of Rochester, Rochester, NY Background: Reactions to platelet transfusions occur despite pre-storage leukoreduction. This suggests that platelet-derived mediators are involved. Platelets are the primary source of sCD40L, a key inflammatory mediator which induces PGE2 and COX-2 synthesis. Methods: This is a cohort study. Supernatants from pooled, post-storage leukoreduced, whole blood platelet concentrates were assayed by ELISA for white cell (IL-6, IL-8, MCP-1) and platelet-derived (sCD40L, RANTES) mediators of inflammation. These levels were correlated with adverse events, including rigors, changes in temperature, pulse and respiration. Results: 534 transfusions had evaluable data. There were 12 reported and 2 unreported reactions-10 febrile and 4 allergic. Reaction rates after post-storage leukoreduced platelet transfusions were low (2.6% of transfusions), and unreported reactions uncommon (0.36%). Only one patient had more than a single reaction. Transfusions with reactions had statistically significantly higher mean levels of infused IL-6 (2.3 fold higher; p = 0.005), IL-8 (2.2; p = 0.001), MCP-1 (2.6; p = 0.002) and sCD40L (1.24; p = 0.015), but not RANTES. (1.14; p = 0.22). The vast majority (>93% for each mediator) of patients infused with mediator levels in the highest quintile had no reactions. The mean change pre-to post-transfusion in pulse, respirations and temperature in the transfusions with no reactions was 0 to -1. Transfusions with reactions averaged + 8 beats/minute (p = 0.003 versus no reaction), +2 breathes/minute (p = 0.03) and +0.4 degrees C°(p = 0.015). When supernatant levels of all five mediators were summed, the reaction rates in the first through fifth quintiles increased proportionately from 1% to 7% (p = 0.027). All but one reaction occurred in patients with hematologic malignancies (13 reactions/380 transfusions; 3.4%) (p = 0.04 versus other diagnoses). The reaction rate in other clinical settings ranged from 0-1%. There were no reactions seen with platelets transfused in the operating room (n = 52) and only one out of 124 transfusions given to outpatients. Conclusions: These are the first data demonstrating that a plateletderived mediator, sCD40L, is associated with adverse events after transfusion. While virtually all patients with reactions received transfusions with high levels of white cell and/or platelet derived mediators, the vast majority (>93%) of transfusions with the highest mediator levels led to no reaction. Of note, inpatients with hematologic malignancies accounted for almost all reactions. We speculate that clinical factors, e.g., inflammation or leukopenia, may be key factors determining whether infused mediators cause reactions. Objective: To present incidence of and describe bacterial contaminations reported to the Quebec Hemovigilance System (QHS). Methods: Transfusion safety officers in hospitals investigated and reported adverse transfusion reactions to the Quebec Health Ministry. Bacterial contamination (BC) was certain if same bacteria grew in product and recipient cultures, BC was probable if product culture was positive but recipient blood culture was negative because recipient was on antibiotics or no culture was done (recipient had signs/symptoms of BC The 2 BC with apheresis platelets (AP) were before introduction of bacterial detection in AP (Mar 2003) . Four definite BC were with pooled platelets (PP) (3 S. Epidermidis and 1 S. Aureus), and one with AP (S. Epidermidis) and 2 with RBCs (P. Acnes and Staph. Coag.neg.). BC related to RBCs caused 1 death, those to FFP caused only minor reactions, those to AP caused only minor reactions and those to PP caused severe reactions in 7 cases, mild ones in 10 and 2 deaths for a death rate of 1 : 28,000 for PP. Gram + germs were involved in 93.5% of BC and caused all the severe reactions. Conclusion: BC is still an important problem of transfusion and PP presented the major risk in Quebec for the period 2000-2003. Derivation of the first 30 ml of blood at donation was implemented in Feb 2003, QHS could be used in the future to evaluate the effectiveness of such a measure.

J L Jacobson, E W Streun, R J Davey, New York Blood Center, New York, NY Background Bacterial platelet contamination is estimated to occur at an incidence of 1 : 1000 to 1 : 3000. 1 : 15,000 reactions is severe. 1 : 60,000 leads to mortality. From 1985-1999, bacterially contaminated blood products accounted for 77 of 694 transfusion-related fatalities reported to the FDA. We noted a cluster of bacterially contaminated products at our center in 2002. Methods We performed a retrospective review of all post-transfusion communications (PTCs) received from January 2002 through April 2004. Cases of reported transfusion-related bacterial contamination were analyzed. Results During calendar year 2002, 39 PTCs were recieved. 13 of 39 (33.3%) were due to suspected bacterial contamination. Of those 13 cases, 3 were related to apheresis platelets, 5 were related to whole blood derived platelets, and 5 were related to RBCs. Bacterially contaminated platelet transfusions were temporally related to the death of five patients. Two fatalities were related to a split apheresis platelet product, one half of which cultured Klebsiella pneumonia. The implicated whole blood derived platelet pools cultured Klebsiella oxytocia, Bacillus, and Klebsiella pneumonia. During calendar year 2003, 44 PTCs were received. 9 of 44 (20.4%) were due to suspected bacterial contamination. 3 were related to apheresis platelets, 2 were related to whole blood derived plateletes, 3 were related to RBCs, and 1 was related to FFP. No patient mortality related to bacterial contaminated products was reported in 2003. During the first quarter of 2004, no bacterial contamination related PTCs were received. Conclusions A cluster of bacterially contaminated blood products occured in 2002. We could not identify a specific cause for this cluster. In order to reduce the incidence of transfusion-related bacterial events, we conducted extensive retraining of phlebotomy personnel and instituted a new arm scrub. We also S56-040B Initial Results of Culturing Apheresis Platelets to Detect Bacterial Contamination L A Chambers, American Red Cross, Washington, DC Background: The American Association of Blood Banks' Standards require the use of methods to detect bacterial contamination in platelet components as a means to reduce the risk of septic transfusion reactions. Methods: In March 2004, an automated culture system was fully implemented in a large, multicenter blood program that performs approximately 8,000 apheresis platelet collections per week. Aliquot samples obtained 24 or more hours after collection were inoculated into aerobic culture media bottles which were incubated and monitored for evidence of carbon dioxide production as an indicator of bacterial growth. Labeling and distribution proceeded if the cultures were negative after 12 hours; however, culturing continued until outdate of the components. Ordinarily, positive bottles were tested to confirm the presence and characterize the nature of the bacteria; the associated products, if not distributed, were re-sampled and re-cultured prior to discard. Results: In the first 2 months (9 weeks), 50 out of approximately 72,000 inoculated bottles were positive (1 in 1440 tests):

The bacteria associated with the true positive results included various Staphylococci (n = 7), various Streptococci (n = 4), Citrobacter and Serratia. The donor of the unit containing Citrobacter reported an episode of diarrhea on the day following donation; otherwise, the donor follow-up was noncontributory. Conclusions: With the sampling and culture process used, a majority of positive results from quality control culturing of apheresis platelets were false positive. The frequency of detecting contaminated collections was approximately 1 in 5000. As expected, Streptococci and Staphylococci predominated. Potentially contributing donor factors were not often identifiable. When positive culture results were obtained after the component had been distributed and transfused, the organism specificities and lack of reactions in the recipients suggested that these were also false positive cultures. Background Septic reactions due to bacterially contaminated platelets are the most frequent transfusion-associated infectious risk in the USA today. Bacterial contamination of platelets is estimated to occur at an incidence of 1 : 1000 to 1 : 3000. Roughly 1 : 15,000 of the reactions are severe, and 1 : 60,000 lead to mortality. Our organization implemented bacterial detection on all apheresis platelets (SDP) in an attempt to improve the safety of SDP transfusions. Methods 4 ml aliquots taken from the parent bag of all SDPs were inoculated into aerobic culture bottles at least 24 hours after collection and monitored using the bioMerieux BacT/ALERT. If the bottles remained negative after 24 hours of incubation, the SDPs were labeled and made available to hospitals. The bottles continued to be monitored for the life of the product. Any bottle that became positive was sent along with the platelet product for confirmatory gram stain, culture, and identification. Results During the period from 10-13-2003 to 4-30-2004, 24 ,909 SDP samples were processed. 15 (0.06%) were intially positive on the BacT instrument. Confirmatory testing of the BacT bottle was positive in 9 (0.04%) and negative in 6 (0.02%). Confirmatory testing of the related SDP was positive in 5 (0.02%) and negative in 10 (0.04%). The 5 positive SPDs were true positives (TP), as both the bottle and the product cultured the same organism. There were 6 false positives (FP) which were negative in both the bottle and the product. In 4 cases, an organism was cultured from the bottle but not the product, indicating laboratory contamination. The time to detect a positive had a mean of 21.9 hrs (range:3.8 to 62.4 hrs) and a median of 23.3 hrs. 6 gram-positive and 2 gram-negative organisms were identified. Conclusions Microbial detection can be successfully utilized to perform quality control on SDPs. In our experience, 1 in 1527 BacT bottles was initially positive. 1 in 4982 proved to be a TP while 1 in 2491 was shown to be a FP or lab contamination. Although the majority of organisms will be identified within the first 24 hours of incubation, clinically significant bacteria will continue be identified beyond this period. Quality control testing for bacterial contamination is a major step foward in reducing the infectious risk associated with platelet transfusions.

Confirmation was not obtained in 12%. Detection time was less than 48 hrs in 87 cases, of which 15 were administered, including 2 B. cereus. In 208 samples detection was more than 48 hrs, of which 140 were transfused. None of the positively cultured RDP's, which were used for transfusion (n = 155) were involved in transfusion reactions in patients. However, two patients suffered from B. cereus septicaemia, after transfusion of RDP's. In both cases B. cereus was cultured from the platelet bags, whereas the samples cultured in the BacT/ALERT® remained negative. The applied bacterial screening guidelines prevented only 122/261 (47%) transfusions of RDP's, which were involved in confirmed positively cultered samples. Conclusions: The presently used policy of issueing platelets on a negative-todate basis resulted in a high percentage of recall procedures and a low prevention rate. In at least 2 cases detection of B. cereus, which was involved in 2 life-threatening TTI's, had failed. Other strategies should be applied to guarantee a higher degree of bacterial safety. was defined as an IP with negative pouch and platelet cultures. A true positive (TP) was an IP with positive pouch and platelet cultures. TP cases were followed by organism identification, phlebotomy and donor investigation. Results: The table below summarizes the bacterial detection data for this period. 96% of collected pheresis platelets were completely tested prerelease. When platelets were released with incomplete testing, no IP requiring product retrieval was observed. The overall IP rate was 0.5%, with highest monthly rates after addition of newly trained staff. Possible causes of FP were sampling technique (overfilling of the pouch, inadequate tube stripping) and high platelet numbers or platelet metabolic activity in the pheresis unit. Two TP cases were observed. One case was due to P. acnes, a skin contaminant. In the other, both skin contaminants (B. cerus and coagulase negative Staph.) and an enteric Gram-negative bacillus were identified. Investigation revealed asymptomatic urinary tract infection in the donor. One unit with coagulase negative Staph. was missed by BDS and transfused, causing a moderate febrile reaction in the recipient, who, fortunately, recovered rapidly. Background: Bacterial screening with BacT/ALERT® in the Southwest part of the Netherlands was implemented in October, 2001. Methods: All platelet products were cultured for seven days in the fully automated BacT/ALERT® system (bioMérieux Inc., Durham, NC). Platelets were issued on a "negative-to-date basis". After positive signals involved products were blocked and removed from the inventory. When involved and related products were already released, hospitals were notified and products were recalled as part of an active haemovigilance program. Positive cultures were confirmed and identified in an independent laboratory. Results: In thirty months of testing, 295 out of 39.342 (0.75%) samples were found positive. 172 (58.3%) random donor platelets (RDP's) were issued, of which 155 (90%) were already administered at the time of positive signals. Analysis identified mainly skin-derived flora (210/295; 71%), but also 15 times B. cereus (5%).

Conclusions: Implementation of pre-release platelet bacterial detection has been operationally feasible and indeed resulted in increased transfusion safety. The observed contamination rate is within the range reported in the literature. Decreased effective shelf life, however, has required rigorous platelet inventory management to meet demand. Adoption of enhanced bacterial detection methods would improve detection of true contaminations and inventory issues. on in-date panels. Use of expired rRBC is often needed to identify or exclude antibody specificities. Accurate results with these cells depend on the integrity of the antigens. There are no published studies documenting the stability of red cell antigens on the expired rRBC from commercially-prepared panels. Study Design: Twenty-three donor-derived red cell antibodies (Ab) of common specificities in the RH, KEL, FY, JK, MNS, and LE systems with initial reactivity *32122+ were tested against rRBC panels from four US commercial manufacturers. Saline tube tests (RT, 30 min 37 C, IgG) were performed on in-date rRBC and at 1 month intervals to 3 months past expiration. Reactivity with expired antigen positive (Ag+) and antigen negative (Ag-) rRBC was compared to the reaction obtained with in-date cells. Acceptable results were defined as reactivity with expired rRBC that was consistent with the reactivity of the same in-date rRBC. Results: Thirteen Ab with only AHG reactivity gave acceptable results with all cells tested. [D, c, E, K(2) , k, Jk a (2), Jk b , Fy a (2), S(2)]. Reactions at 3 mo. post-expiration were >1 grade weaker than the in-date result on at least 1 cell for 7 of the 13 Ab [c, K(2), Jk a , Jk b , Fy a , S]. A second anti-k showed variable reactivity at RT phase but all expired k+ cells were positive at AHG. Two examples of anti-M were reactive at all phases with in-date rRBC. The consistency of these reactions varied during the study period, but all expired M+ rRBC were reactive in at least one test phase. No M-rRBC were reactive. Unacceptable results (defined as 1 mo, 2 mo, or 3 mo-post expiration tests showing no reactivity with Ag+ rRBC at any phase or reactivity with Ag-rRBC) are found below.

RFLP and serologically phenotyped with human anti-Tc a , anti-Tc c and anti-Cr a . Cr a expression appeared normal but there was variable reactivity for Tc a . Two of the samples that reacted weaker with anti-Tc a , as compared to controls, were heterozygous (RL) by genotyping. Conclusion: The frequency of Tc b in Malians is not significantly different from that reported for African-Americans and, therefore, probably does not offer a major survival advantage against malaria in Africa. The apparent quantitative variation observed in some samples may either be inherited or acquired and deserves further study.

Crystal Structure of NNA7, an FAB Fragment That Recognizes the Glycopeptide N Blood Group Antigen S L Spitalnik, S-C Song, Columbia University, New York, NY; K Xie, J E Wedekind, University of Rochester, Rochester, NY Background: We are using the human MN blood group system to examine the fine specificity of glycopeptide-specific antibodies. We previously constructed a Fab family in which the anti-N N92 H chain Fd fragment was held constant while the L chain was "shuffled." The N92 and NNA7 Fabs have low and high affinity for N, respectively, although their L chains are 92% identical at the amino acid level. The NNA7-G91S Fab, in which amino acid 91 of the NNA7 L chain CDR3 was changed from Gly to Ser (as in N92) by site-directed mutagenesis, had low affinity for N. Thus, this subtle molecular change had a large effect on affinity. Determining the 3-D structures of these Fabs will provide important insights into how they recognize this clinically relevant glycopeptide antigen. Methods: NNA7 crystals were developed by the hanging-drop method, analyzed by X-ray diffraction, and the structure was solved. Results: Diffraction quality crystals were obtained in 0.1 M 2-morpholinoethanesulfonic acid (MES), 0.3 M (NH4)2SO4, pH 6.5, containing 24% (v/v) poly(ethylene) glycol monomethyl ether 5000 and 4-8 mM YCl3. Crystals belong to space group P212121, with unit cell dimensions of a = 57.9 Å, b = 77.2 Å, c = 118.5 Å, and with one molecule per asymmetric unit. The structure was solved by molecular replacement and refined to 1.83 Å resolution with an R-factor of 20.5% and free R-factor of 24.5%. A crescent-shaped channel formed at the interface of the H and L chain variable regions (12 Å wide and 9 Å deep) contains a single molecule of MES, which has structural homology to a monosaccharide. The MES buffer molecule interacts with H chain G33, W52, S53, N95, and Y100b. In addition, L chain G91 is found at the base of the putative antigen-binding site. The NNA7-G91S mutant was also crystallized and its structure is being solved. Conclusions: This is the first crystal structure of a glycopeptide-specific Fab. Future structural studies of NNA7 complexed with the N antigen will provide further insights into this interaction.

A Seltsam, Transfusion Medicine, Hannover Medical School, Hannover, Germany; C Das Gupta, Biotest AG, Dreieich, Germany; R Blasczyk, Transfusion Medicine, Hannover Medical School, Hannover, Germany Background: Until now, weak blood group A and B phenotypes have been correlated with single base substitutions occurring in the coding and noncoding sequences from exons 2 to 7 of the ABO gene. The resulting A or B glycosyltransferases are predicted to exhibit single amino acid changes and/or truncated or elongated C-termini, affecting the proteins' enzyme activities by altering the sugar binding site. Methods: An healthy blood donor from Germany diagnosed as having weak A antigen expression and relatives of him were subjected to extended ABO typing. Serologic investigations were performed using standard techniques and flow cytometry. The genetic basis of the ABO phenotypes was determined by PCR-SSP and subsequent sequence analysis of the complete sequence and two regulatory regions of the ABO gene. HeLa cells were used to transfect ABO expression plasmids. Results: The donor's red blood cells were typed as AweakB and his serum contained weakly reactive anti-A1 antibodies. Sequence analysis identified a single TAEC transition at position +2 of the ATG start codon of one allele in an otherwise ABO*A101/ABO*B101 genotype. This new point mutation resulted in amino acid exchange from methionine to threonine at position 1. ABO typing of the donor's children revealed the phenotype A1B and the genotype ABO*A101/ABO*B101 for the son and the phenotype A1 and the genotype ABO*A101/ABO*Avar for the daughter, demonstrating that the new polymorphism is located in an ABO*A allele. In the transfection studies, no decrease of A activity was observed on HeLa cells transfected with plasmids containing the variant ABO*A allele. However, a significant reduction Overall, results with expired rRBC were accurate in 100% of tests when double dose Ag+ rRBC were tested, 97% of tests with single dose Ag+ rRBC, and 98% of tests with Ag-rRBC. Conclusion: Most RBC Ag studied were detectable 3 mo after rRBC expiration. Variability was most often seen with single dose Ag+ RBC. The M and N Ag showed greatest variation. Individual cell variability can be countered by testing >1 double-dose Ag+ or Ag-rRBC. This study is the first to give data and fact-based recommendations on the use of expired rRBC. Background: The death of more than one million children in Africa per year can be attributed to cerebral malaria and severe anemia caused by Plasmodium falciparum infection. Previous studies have associated several blood groups with resistance to malaria invasion. These have been found because they showed differences in antigenic frequency between Caucasians and Blacks or because higher frequencies of a protective type were observed in Black Africans. Due to limited studies in African-Americans of the Cromer blood group carried on decay accelerating factor (DAF), we investigated the Tc a/b polymorphism in a malaria endemic area of West Africa to determine any association with malaria. Methods: A total of 239 West Africans from Mali were screened for the Tc a/b polymorphism using two methods. PCR was used to amplify exon 2 from 124 genomic DNA samples. The 300 bp product was digested with Stu1, which recognizes the Tc b allele, and electrophoresed on agarose gels. The remaining 115 samples were screened by immunoblotting of SDS-PAGE separated red cell membrane proteins and probing for both complement receptor one (CR1) and DAF using monoclonal antibodies J3D3 and 917-3G2 (anti-Tc a -like), respectively. Results: The genotype frequency of the Tc a (R18) allele was 0.98; while the Tc b (18L) allele was 0.02. The phenotype frequency of Tc b in Africans was 5.6% which does not differ from that reported for African-Americans (5.3%). Seven samples appeared to have low or no DAF on immunoblot when compared to the CR1 control. Those samples were further analyzed by PCR- TRANSFUSION 2004-Vol. 44, Supplement of A antigen expression (36-70 %) was detectable on HeLa cells co-transfected with plasmids specific for the variant ABO*A allele and the normal ABO*B101 allele compared to cells co-transfected with ABO*A101 and ABO*B101 constructs. A normal A activity was observed with an ABO*A101 construct where the codons encoding for Met1 and a potential start site at amino acid position 26 were artificially disrupted. Conclusion: The data provide evidence that a nearly full functional A transferase can be produced by alternative translation start sites in the transmembrane domain or stem region. The weak blood group A phenotype of the donor most likely resulted from competition between a normal B transferase and a N-truncated A transferase.

S64-040C DNA Analysis as an Essential Adjunct to the Identification of Knops System Antibodies C Lomas-Francis, K Hue-Roye, V I Powell, G Halverson, D Charles-Pierre, M E Reid, New York Blood Center, New York, NY Background: Identification of antibodies in the Knops blood group system is notoriously difficult due to the scarcity of reliable antibodies and bona fide RBCs of known type, and the variable expression of CD35 (CR1). Antibodies to Knops system antigens are clinically insignificant but can delay the identification of clinically significant antibodies. We investigated whether DNA analysis could be used to aid in antibody identification. Method: Hemagglutination with monoclonal anti-CD35 (E11 & MIMA133) and/or nucleotide (nt) sequencing were performed on a sample from the person who originally made anti-Kn a (KNO), and on patient samples whose serum was thought to contain anti-Knops antibodies. PCR-RFLP assays for McC (nt 4795) and Sl (nt 4828) polymorphisms were performed on 127 donors who self-identified as being African American. Results: KNO had AA in place of the common GG at nt 4708 (predicted to change Val1561Met), confirming the mutation in other Kn(a-) samples. The results of sequencing 33 patient DNA samples were compared to the antibody identification from historical records and are given in the table. The PCR-RFLP assays on the 127 donors showed that 25 were 4795 AA/4828 AA; 38 were AA/AG; 21 were AG/GG; 19 were AG/AG, 13 were AA/GG; and 11 were GG/GG, which corresponds to allele frequencies of 75.6% McC a and 24.4% McC b , and 42% Sl a and 58% Sl2. Thus, McC(a-) Sl(a-) occurred in 8.6% and McC(a+) Sl(a-) in 26% of samples. Tests with anti-CD35 revealed that low copy number can occur on Sl(a-) RBCs as well as on RBCs with the common Knops genotype. A panel of RBCs was selected from the 127 donors and used to test serum from 21 patients. Anti-Sl a was convincingly identified in 4 samples, and anti-McC a /Sl a in 5 samples. Conclusion: Use of RBCs from donors of known Knops genotype will aid in the correct identification of antibodies to antigens in the Knops blood group system. Clearly classified samples studied for their molecular basis are needed in order to alleviate the confusion caused by testing with human antibodies that were misidentified. SC1/SC2 were selected for analysis to permit the rapid identification of clinically relevant "minor" blood groups, and donors lacking high prevalence antigens. Primers for a single multiplex PCR reaction, and allele-specific oligonucleotides with variable 3¢-terminal nucleotide for elongation-mediated multiplex analysis of polymorphisms were designed. Color-encoded beads displaying the elongation probes were assembled into planar arrays of small footprint on semiconductor chips, permitting the instant imaging of fluorescent elongation products from the entire array. The array design was validated using DNA prepared from 50 samples and then used to analyze 433 samples from donors and patients of known partial phenotype. Results: The DOA/DOB array results agreed with manual PCR-RFLP assays in 150 of 150 samples. The COA/COB results agreed with phenotype results in 88 of 88 samples. For 252 of 253 FYA/FYB and GATA tests the array results agreed with the phenotype, and for 34 of 34 the array results agreed with manual PCR-RFLP assays. The one outlier typed FYB/FYB, GATA heterozygous (W/M) by both the bead array and manual PCR-RFLP methods, and FY nt 265 homozygous wild type by manual PCR-RFLP while the phenotype was Fy(a-b-) on two samples. The reason for the silencing of one of the FYB in this donor is unknown and the entire FY gene is being sequenced. Conclusion: The new bead array format facilitates reliable blood grouping by multiplex analysis of a custom SNP panel. The design flexibility and convenience of this platform enable reliable, high throughput screening of patients and donors with minimal reagent consumption.

Merging the Pathogenesis of Transfusion Related Acute Lung Injury: The Priming Activity of the 5 b (HNA-3) Antibody P Kopko, Blood Source, Sacremento, CA; B Curtis, Blood Center of Southeastern Wisconsin, Milwaukee, WI; M Kelher, N McLaughlin, C C Silliman, Bonfils Blood Center, Denver, CO Background: Transfusion related acute lung injury (TRALI) is a serious complication of blood administration. Two different pathogenetic mechanisms have been proposed. The first is the infusion of antibodies directed against specific HLA or granulocyte antigens resulting in complement activation, PMN sequestration, and TRALI. The second consists of two events: 1) a predisposing clinical condition resulting in PMN sequestration in the lungs followed by 2) the infusion of lipids from stored blood, which prime PMNs in vitro and activate PMNs adherent to pulmonary endothelium resulting in endothelial damage, capillary leak and TRALI. We hypothesize that antibodies directed against specific PMN antigens cause rapid PMN priming that is specific to the antigen they recognize. Methods: human PMNs were isolated from healthy 5 b (HNA-3+) and HNA-3-donors and preincubated with buffer or Fab' 2 fragments against the FC receptors (CD16, CD32 & CD64). PMNs were incubated for 5 min with 10% fresh plasma from healthy, antibody negative males (FP), plasma from 3 donors who had antibodies against HNA-3 and were implicated in TRALI reactions (HNA-3+), or 10% immune complexes (ICs) made from heated sera (+control). The maximal rate of O 2 -(nmol/min) was measured as reduction of cytochrome c at 550 nm. Results: Priming is expressed as augmentation of the fMLP-activated respiratory burst. The ICs effectively primed the PMN oxidase as compared to FP controls (* = p < 0.05) and were significantly inhibited by pre-treatment with the Fab' 2 fragements against Fc receptors (# = p < 0.05 as compared to IC). In addition, the three separate HNA-3+ plasmas significantly primed the PMN oxidase as compared to FP-treated controls. Pre-treatment with Fab' 2 fragments did not significantly augment respiratory burst of the FPtreated control PMNs nor did it significantly diminish the priming of the HNA-3+ plasma samples. Furthermore, the three HNA-3+ plasmas did not prime HNA-3-PMNs from 3 separate donors (results not shown). We conclude that antibodies directed against the HNA-3 locus rapidly and effectively prime HNA-3+ PMNs, priming which is specific and not due to activation of Fc receptors. Therefore, antibodies directed against specific granulocyte antigens and lipids from stored blood may cause TRALI through a common, final pathway of PMN activation.

M E Reid, P Vissavajjhala, New York Blood Center, New York, NY; G P Hashmi, T Shariff, M Seul, BioArray Solutions, Ltd., Warren, NJ; K Hue-Roye, D Charles-Pierre, C Lomas-Francis, A Chaudhuri, New York Blood Center, New York, NY Background: Historically, blood group antigen typing has been done by hemagglutination, which has limitations such as non-availability of certain commercial antibody reagents. The advent of cloning of blood group genes permits analyses of blood group antigens at the DNA level. The aim of this study was to develop and validate such an approach using a novel custom array format. BACKGROUND: Transfusion-Related Acute Lung Injury (TRALI), one of the most common noninfectious, life-threatening complications of blood transfusion, is associated with the administration of blood that contains anti-MHC (Major Histocompatibility Complex) antibodies. To date, theories regarding TRALI pathogenesis are based on in vitro animal studies and human clinical case studies. Lack of controlled, in vivo animal experiments has made it difficult to test theories of TRALI pathogenesis as well as possible therapeutic interventions. Our objective was to develop an in vivo murine model that shares features with human TRALI. METHODS: BALB/c (H-2 d ) mice were injected with 34-1-2s, a murine IgG2a anti-H-2 d (D/K) monoclonal antibody (mAb). Hypothermia (which reflects the development of shock) was measured by rectal thermography and dyspnea was measured by noninvasive barometric plethysmography over the next 2 hours. RESULTS: Hypothermia and dyspnea were rapidly induced in BALB/c mice but not in congenic H-2 k controls. Pathological changes in the lungs after anti-MHC mAb challenge included an increase in vascular permeability that led to pulmonary edema and protein leak into the alveolar spaces. In addition, a neutrophilic alveolar infiltrate developed within 2 hours of challenge. This evolved into a mononuclear alveolar infiltrate by 24 hours after challenge. Pretreatment with anti-FcgRII/III mAb to block FcgRII/III receptors, gadolinium to deplete macrophages, anti-granulocyte mAb to deplete granulocytes, or PAF, histamine, or leukotriene antagonists, inhibited the induction of hypothermia and dyspnea. CONCLUSIONS: These results suggest that IgG, FcgRIII, granulocytes, macrophages, PAF, histamine, and leukotrienes are important in anti-MHC antibody induction of shock and pulmonary dysfunction. This animal model should be useful for determining the pathogenic mechanisms of TRALI and for identifying potential therapeutic interventions. Objective: To present the adverse transfusion reactions (ATR) that were accompanied by respiratory signs or symptoms and trends in the incidence of transfusion-related acute lung injury (TRALI) reported to the Quebec Hemovigilance System (QHS) for the period 2000-2003. Methods: Transfusion safety officers in hospitals investigated and reported ATR to the Quebec Health Ministry. Signs, symptoms and diagnoses of ATR are reported in a standard format to the QHS. Decrease in O 2 saturation was taken as reported by hospitals even if no values were given. TRALI was defined as acute respiratory distress with pulmonary infiltrates without evidence of volume overload within six hours of transfusion. Rates per 10,000 transfused units were calculated using actual transfused units as per hospital monthly reports on utilization. Results: From Jan 2000 to Dec 2003, a total of 4,576 ATR possibly, probably or definitely associated with transfusion were reported by 61 hospitals representing 85% of transfusion activity in Quebec. Data for 2003 are still provisional because of reporting delays. The proportion of ATR with dyspnea represented 11.1% (507) declining from 19.9% in 2000 to 10.3% in 2003 (p < 0.01, X 2 for trend). The ATR with dyspnea as one of the symptoms were diagnosed as: febrile non hemolytic transfusion reaction (34.7%), volume overload (31.6%), major allergic reaction (18.1%), minor allergic reaction (12.4%) and TRALI (3.9%). Decrease in O 2 saturation was present in 2.6% of ATR, increasing from 1.9% in 2000 to 3.1% in 2003 (p < 0.01,X 2 for trend). There were 21 cases of TRALI. Rates per 10,000 units were:

Conclusion: A significant proportion of ATR were accompanied by respiratory signs or symptoms. The decrease in these proportions over the 4-year period could be explained by increasing reports of minor reactions with time. The increasing number of TRALI reported probably reflects better recognition of the syndrome. We however need a better definition of TRALI in QHS to better validate the cases and distinguish them from volume overload. Some mild cases of TRALI (if such cases exist) might have been hidden in cases of decreased O 2 saturation without respiratory distress.

Epidemiology of Transfusion Related Acute Lung Injury in Gifit, the French Hemovigilance Database. a Study of the French Hemovigilance Network P C Renaudier, M P Vo Mai, J M Azanowsky, P Breton, S Cheze, A Girard, L Hauser, J F Legras, D Rebibo, C Waller, N Ounnoughene, Unité d'Hémovigilance, LYON cedex 04, France Background: Transfusion Related Acute Lung Injury (TRALI) is a life-threatening complication of allogenic blood transfusion, manifested typically by a non-cardiogenic pulmonary edema. However, the magnitude of the risk of TRALI remains unknown at this time, all the more so as a variety of other clinically similar respiratory complications can be associated with transfusion. Objective: To describe TRALI observed in GIFIT, the French hemovigilance database. Population and methods: All adverse effects (AE) regardless of their severity should be notified on a normalized form to AFSSAPS, the French health product safety agency, as required by the French law. As the item << TRALI >> was explicitly present since 09/01/2001, we screened also the database with << pulmonary edema AND fluid overload excluded >>, or with comments including the words << TRALI >>, << anti-granulocyte antibodies >>, << non-cardiogenic pulmonary edema >>, << white lung syndrome >>. Results: From 07/01/1994 to 07/01/2003, 60,723 AE were registered in GIFIT, corresponding to about 25,000,000 labile blood products (LBP) transfused. 34 AE fulfilled the above criteria (1.3/1,000,000 LBP transfused). There were 17 men and 17 women, 12 men and 5 women were 60 years old or more. 32 patients presented with pulmonary symptoms, 28 with dyspnea, 21 with shivers and/or hyperthermia and 20 with a typical pulmonary edema. 6 TRALI were fatal, 24 were life-threatening and 4 were benign. 7 patients received only packed red cells, 2 only fresh frozen plasma and 15 only platelet concentrates; among the 10 remainders, 5 received platelets concentrates associated with an other LBP. Conclusion: (1) TRALI seems to be more frequent in elderly patients, as noted in other reports; (2) Platelets concentrates are the more frequently LBP involved in TRALI; (3) TRALI patients may have a worse prognosis than previously reported, as 20 % of them died; (4) However, besides the classical lifethreatening presentation, a benign form of TRALI seems to exist.

High Probability Transfusion-Related Acute Lung Injury Cases: Clinical Characteristics L A Chambers, American Red Cross, Washington, DC; M E Wissel, Central Ohio Region, American Red Cross, Columbus, OH; R J Benjamin, New England Region, American Red Cross, Dedham, MA Background: Transfusion-related acute lung injury (TRALI) can be characterized as an acute bilateral diffuse pulmonary process in the absence of right atrial hypertension, associated with significant hypoxia (e.g. O 2 saturation <90% on room air), that develops during, or within 6 hours following, transfusion. Until a diagnostic abnormality is identified, these clinical criteria can be used to identify high probability TRALI cases. Methods: All cases of potential TRALI received, investigated and closed in 2003 at a large blood system were reviewed by 3 physicians to identify those meeting the above clinical criteria. Thirty six high-probability TRALI cases involving 120 components were analyzed to characterize the patient demographics, diagnoses, presenting signs and involved component type(s). Results: Patient age (24-90, mean 61.3) and sex (18 male, 18 female) were similar to those for other transfusion recipients.

In 5 patients (14%), death was attributed to TRALI. Associated findings included a drop in systolic blood pressure of >30 mmHg (15/36), increase in systolic blood pressure >30 mmHg (13/36), rigors (12/36) and fever (9/36). Conclusions: Using specific clinical criteria to define high-probability TRALI cases, the at-risk group is demographically similar to other transfusion recipients. Cases are skewed toward FFP transfusions and diagnoses for which FFP transfusion would be indicated though red cell transfusion are not infrequently implicated. The most common underlying diagnosis is bleeding associated with coumadin therapy. Fever, rigors and hypotension are present in a minority of cases and hypertension is not rare, contrary to previous descriptions of the associated findings.

Use of BNP to Evaluate TRALI R J Bowman, N Laudi, D Aslan, A Burgher, University of Minnesota, Medical School, Minneapolis, MN Background. The differential diagnosis of respiratory failure associated with transfusion includes transfusion-related acute lung injury (TRALI) and volume overload-often in the setting of congestive heart failure (CHF). Brain natriuretic peptide (BNP) is a peptide hormone released by the cardiac ventricles and atria in response to volume or pressure overload. Elevated levels of BNP have proved highly predictive of dyspnea due to CHF (1). Case Report. A 65 year-old male experienced respiratory failure during transfusion of the second of two units of red blood cells. The transfusion was interrupted and a chest radiograph within 90 minutes revealed new, diffuse disease. Mechanical ventilation was required and at intubation, suction returned pink, frothy pulmonary fluid. Troponin elevation (3.66 mg/L) and serial ECGs confirmed new myocardial infarct, thought to have occurred after the reaction. The clinical team believed TRALI, rather than CHF, was the probable diagnosis because a previous echocardiogram (4 months prior) had shown a normal ejection fraction (60%), and the onset of dyspnea was so rapid. Over the ensuing two days, the patient exhibited diuresis (5100 mL), with marked improvement in chest radiograph and reduced oxygen requirements. We retrospectively evaluated BNP on refrigerated samples stored for five days. BNP (Biosite Triage® BNP assay) was 532 pg/mL (normal < 100) in a pre-transfusion sample taken for type and screen and 606 immediately post-transfusion. The patient improved with supportive therapy and was discharged home about two weeks later. Conclusions. While TRALI seemed to be a compelling initial diagnosis, significant diuresis and measurement of BNP levels on pre-and post-transfusion samples led to the conclusion that volume overload and CHF were the cause of the respiratory failure. Although BNP levels have not been measured in patients with TRALI, they may offer valuable diagnostic support to differentiate TRALI from volume overload seen in CHF. BNP can be measured on stored pretransfusion samples; results are readily available and, if elevated, may favor volume overload and CHF. An elevated BNP does not rule out the diagnosis of TRALI, since both CHF and TRALI may co-exist. While tests for white cell antibodies, protein content of edema fluid, interleukin 6, and neutrophil priming activity may favor the diagnosis of TRALI, additional analytes are needed. This case suggests, but does not establish, the utility of BNP to evaluate respiratory failure occurring during transfusion. Further assessments remain necessary. Background Previous investigations indicated that contaminating bacteria are predominantly present in the first mL's of a donation. We evaluated nationwide the effect of diversion of the first 20 mL's as well as various skin cleansing methods in the presence of 100% bacterial screening for buffycoat derived pooled platelet concentrates (PC), prepared from 5 whole blood collections. Methods PC were cultured with an automated system (BacT/Alert). Data on bacterial contamination of PC were acquired at two periods in time. From January-September 2002, skin cleansing was done with variable methods, from October 2002, standardised double cleansing with 70% isopropyl alcohol was performed. In addition, one blood bank used a sample bag (Composampler V), for diversion of the first 20 mL's over the whole period. Results A small, but significant difference (p = 0.03) before and after introduction of standardised double cleansing was found for standard whole blood collections. The percentage of contaminated pooled platelets was significantly reduced with the diversion bag (p < 0.001; Pearson Chi-Square); further reduction by standardised cleansing (from 0.5-0.36%) was not significant.

Conclusion The use of a deviation bag for collecting the first 20 mL for screening tests reduces the incidence of bacterial contamination in the final buffy-coat derived PC (from 5 individual donations) in routine practice. An additional effect of skin cleansing methods was found.

M E Brecher, S N Hay, University of North Carolina, Chapel Hill, NC; S J Rothenberg, bioMerieux, Durham, NC Background: Bacterial detection of platelet rich plasma (PRP) derived platelets presents unique challenges for countries that do not allow prestorage pooling. This study validated the BacT/ALERT® for use in testing pooled PRP-derived platelets with 9 contaminating organisms. Methods: Isolates of B. cereus, E. cloacae, E. coli, K. pneumoniae, S. aureus, S. epidermidis, S. marcescens, S. viridans and P. acnes were inoculated into 2 PRP-derived platelet pools (10 and 100 CFU/mL). Four mLs of each post bacterial inoculation sample were inoculated into both plastic aerobic and anaerobic bottles and 0.5 mL was plated onto blood agar. Results: All organisms (excluding P. acnes) were detected in 8.2 to 22.0/7.6 to 20.3 hours (10/100 CFU/mL) and the mean time to detection was 15.0/13.1 hours (10/100 CFU/mL). P. acnes was detected with the anaerobic bottles in a mean of 74.9/64.3 hours (10/100 CFU/mL). With E. cloacae, E. coli, K. pneumoniae, S. marcescens, and S. viridans detection with the anaerobic bottles was faster or equivalent to the detection with the aerobic bottles. This was most notable with S. viridans where the anaerobic bottle was reactive on average 21.6/10.8 hours (10/100 CFU/mL) faster than the aerobic bottle. Conclusions: This study validates the use of the BacT/ALERT system for the detection of bacteria in PRP-derived platelets in a pooled format. Overall, the use of a BacT/ALERT allowed the detection of pooled PRP-derived platelets inoculated with 9 bacteria at 10/100 CFU/mL in 7.6 to 22.0 hours (excluding P. acnes).

Background: The BacT/ALERT (bioMerieux, Durham NC) has been cleared by the FDA for the in run quality control of leukocyte reduced apheresis platelet and PRP products. We have previously validated the BacT/ALERT in 4 seeded platelet studies. We reviewed our experience with the recovery of bacteria with low inoculums (£5 CFU/mL). Methods: Retrospective review of 4 studies was conducted. Organism, recovered concentration and mean time to detection for standard aerobic bottles and standard anaerobic bottles were reviewed. The recovered concentrations were originally determined via quantitative culture on 5% sheep blood agar plates. Results: Time to reactivity and recovered concentration is summarized in the Table. In aggregate, data from 10 different organisms showed 100% recovery. Conclusion: The BacT/ALERT is capable of detecting positive bottles when the recovered concentrations are low (£5 CFU/mL) over a wide range of organisms. Table  1 . Recovered CFU/mL and mean time to detection for organisms in 4 studies standard culture (CFU/ml) and eBDS testing at various times after PC inoculation (0 hr) and pooling (24 hr) through day 7. eBDS samples were incubated at 35°C for 24 hrs and tested for bacteria by measuring % oxygen. In 8 studies, eBDS testing was also performed after a 16-hour incubation. Results: The table shows bacteria levels with corresponding eBDS detection results for the low inocula tests using 24-hr incubation. All PC samples tested positive for bacteria at all sampling times, and 11 of 12 pools tested positive immediately after pooling, with 6 having £2 CFU/ml S. epi. The one pool that did not test positive had undetectable S. epi levels by culture of the pool and of the incubated eBDS sample. Thus, probably no bacteria were included in this sample taken for testing. All subsequent pool samples tested positive, irrespective of storage time up to 7 days. All high inocula PCs and resulting pools tested positive at all sample times. Growth rates of S. epi were similar in the PCs and pools up to 7 days. All PC and pool samples taken at or after 24 hrs and incubated for only 16 hrs were detected.

* Studies 1-3 were in apheresis platelets. Study 4 was in pooled PRP platelets. ** Growth not expected in this anaerobic organism.

S Young, R Smith, S Perez, M McCray, TriCore Reference Laboratories, Albuquerque, NM; S Holme, T Lieu, E Nelson, Pall Medical, Covina, CA Background: The ability to store random donor platelet concentrates (PC) as a pool would provide significant benefits for transfusion services, including having a therapeutic dose immediately available in emergency situations, reducing pool outdating, and enabling a single sampling for bacteria testing instead of sampling all the individual PCs. The purpose of this study was to determine if sampling after pooling PCs affected the ability of the Pall eBDS to detect a bacterially contaminated product. eBDS is currently cleared for testing individual PC for bacteria and has a sensitivity level of approximately 1 CFU/ml. This study used a slow growing organism, Staphylococcus epidermidis (S. epi), as a worst case test. Study Design: Twelve PC were inoculated with low levels (1-12 CFU/ml) of S. epi, and 4 PC were inoculated with high levels (>50 CFU/ml). After 24 hrs, 10 ml of the inoculated PC were removed and mixed with 10 ml from each of 5 other noninoculated PC to form a "pool". Samples of PCs and pools were taken for * <1 = undetected by culture.

The results indicate that the sensitivity of the Pall eBDS to detect bacteria at low levels was not affected by testing samples of pooled PC products, or by using a shortened, 16 hr, sample incubation.

K-A T Nguyen, T Yamamoto, H Sandhu, Blood Centers of the Pacific, San Francisco, CA Background: The Pall second generation Enhanced Bacterial Detection System (eBDS) was recently FDA approved for detection of bacterial contamination in leukoreduced platelets. Relative to the first generation BDS, eBDS includes major modifications to augment bacterial growth and sensitivity while potentially improving specificity. This study evaluated eBDS performance compared to three other current bacterial detection methods. Methods: Pheresis platelets were inoculated with 6 organisms at levels of 0-10, 11-100, and >100 CFU/mL. <1 hour (hr) post-inoculation, 2 eBDS sample pouches were obtained and incubated at 35C for 16 or 24 hours, respectively (manufacturer recommends 24-30 hr). Samples were also obtained for aerobic Bio-Merieux BacT/Alert and pH testing. eBDS, BacT/Alert and pH testing were also performed after 24 hr post-inoculation and platelet storage at 20-24 C. Swirling was evaluated at 0 and 24 hr postinoculation. Bacterial quantitation was performed at inoculation and after 24 hr platelet storage. Results: eBDS detected E. coli, B. cereus, and K. pneumoniae at all inoculation levels, both 0 and 24 hr post-inoculation, even when pouches were incubated for 16 hr. All samples of S. epidermidis were also detected, except for 2 pouches obtained at 0 hr post-inoculation and incubated for 16 hr. Both strains of S. agalactiae tested were missed by eBDS at high levels of inoculation when the pouches were incubated for 24 hr. For samples missed by eBDS, however the %O2 was just slightly above the failure threshold. All samples were detected BacT/Alert. Swirling was observed in all inoculated platelets except for those with >10e3 CFU/mL B. cereus. pH was >6.2 in all tested platelets except for 1 unit with >10e5 CFU/mL K. pneumoniae. Conclusions: The current eBDS test represents a major improvement in sensitivity compared to the first generation BDS. S. agalactiae was missed under conditions favoring anaerobic growth. However, minor adjustment of the %O2 failure threshold would further enhance detection to a level equivalent to BacT/Alert in this study. Swirling and pH, in contrast, were relatively insensitive. Corp, Deerfield, IL) was initiated at the time of platelet pooling using pH meter IQ120 miniLab (IQ Scientific Instruments Inc, San Diego, CA). Based on literature review and assessment of pH values of RDPs from v-A and v-B in the week prior to implementation, platelets with pH ≥ 6.6 were considered acceptable for transfusion. To further validate this pH value, a 1-mL sample from all RDP pools was sent for MPC at the time of issuance. Data for RDP pH testing and cultures was compared for v-A and v-B from 03/01 to 04/30/04. Results: During this interval, we obtained 920 pre-storage leukocyte-reduced (LR) RDPs from v-A (mean age 3.6 d) and 774 non-LR RDPs from v-B (mean age 3.9 d). There were 2 pH failures from v-A and 32 from v-B (p < 0.0001). There was no correlation between pH failure and platelet age (p = 0.5146). The pH range, mean, median, and mode for the vendors are listed below.

726T>G mutation in exon 6 of DAF, which is predicted to encode Gln208 in the fourth SCR. Screening of DAF deletion mutants with the antibody verified that the antigenic determinant was located within SCR4. At nucleotide 726, the proband's husband was homozygous for T, and her parents, sister, brother and infant were all 726T/G. Conclusions: We describe a novel high prevalence Cromer antigen, which is characterized by His208 in DAF. The absence of the antigen is associated with Gln208. The antigen, which we are naming ZENA, is located between GUTI and UMC antigens in the primary sequence and has been assigned the ISBT provisional number 021.013. Background: although antibodies implicated in blood group antigen (BGA) alloimmunization have been studied for many years, little is known about helper T cell responses that drive their production. The aim of this work was to determine T cell antigenic determinants involved in T cell responses against BGA, and the subsequent patterns of stimulatory cytokine production. Study Design: the Kidd blood group antigen (Jk a ), a well known immunogenic antigen, has been selected. We investigated cytokine expression by helper T cells after stimulation by Jk a peptides. Peripheral mononuclear cells (PBMCs) were isolated from Jk a primed blood donors (typed Jka-b+) producing anti-Jk a antibodies and from Jka+ donors as negative controls. Four 16-mer peptides mimicking the Jk a sequence, 266-293, and overlapping each over by 12 amino acids were tested to map alloreactive T cell epitopes. A sensitive method, real time RT-PCR, was chosen to quantify Th1-and Th2-type cytokines (respectively IL-2 and IL-4) after Jk a peptide stimulations. Results: the data showed that all anti-Jk a alloimmunized donors responded in a specific and significant way to at least one Jk a peptide. No response was observed in the control population. A clear Th1/Th2 dichotomy in the cytokine response was observed in the alloimmunized donors. Among the 11 Jk a alloimmunized donors tested, 4 produced IL-2 alone and 7 produced IL-4 alone. Two of the four peptides tested were able to elicit a response (IL-2 or IL-4) in 9 among 11 of these donors. Sequences among these two peptides could represent a pool of immunodominant epitopes. Conclusion: these data indicate that Th1 and Th2 subsets are associated to the specific memory humoral immune response against Jk a BGA. The cytokine pattern (IL-2 or IL-4) is characteristic of the donor whatever the peptide tested. A better understanding of peptides and cytokines involved in T cell responses to Jk a protein could constitute a first step toward evaluation of peptide immunotherapy to prevent or treat anti-BGA alloimmunization. RBCs from a Portuguese woman typed Le(a+b-) and were non-reactive with anti-A, anti-B, monoclonal and lectin anti-H, but reacted weakly with 30% of human anti-H. FUT1 analysis revealed homozygosity for a novel missense mutation, 689A>C (Q230P). FUT2 analysis showed homozygosity for the common non-secretor mutation, 428G>A (W143X). RBCs from an Israeli donor typed Le(a-b+) and reacted weakly with two reagent anti-A,B but were non-reactive with anti-A,B from three group O donors. Her ABO genotype There was a significant difference (p < 0.0001) in pH values between v-A and v-B manifesting in 0.2% RDP unit wastage from v-A compared to 4% wastage from v-B. Paradoxically, despite the better pH quality for v-A, the only bacterially contaminated RDP over this time period was a 5 d old unit (pH 7.2) from v-A demonstrating 5.0 ¥ 10 6 CFU/mL S. bovis resulting in clinical sepsis. Conclusions: There is a statistically significant difference in pH of RDPs between v-A and v-B possibly related to prestorage leukocyte reduction by v-A but not by v-B. Additionally, a clinically significant case of bacterial contamination occurred that met the pH criteria at our institution. Further studies are needed to standardize the sensitivity and specificity of pH testing for platelet bacterial contamination as well as to investigate the role of LR in pH determination. Background: An alloantibody to a new high prevalence antigen was detected in the serum of a pregnant Canadian of Syrian descent. The antibody was reactive with all cells tested except Dr(a-) and IFC-RBCs. The corresponding antigen was resistant to treatment of reactive RBCs with ficin, papain, trypsin and AET but sensitive to a-chymotrypsin. These results suggested an antigen in the Cromer blood group system and existing antigens in this system were ruled out. The purpose of this study was to determine the molecular basis associated with the novel antigen. Methods: Sequence analyses were performed on genomic DNA prepared from the blood of the proband. Exons 2 to 6 of DAF, which encode the extracellular short consensus repeat (SCR) domains of DAF that carry antigens of the Cromer blood group system, were sequenced. In order to epitope map the antigen, Chinese hamster ovary stable transfectants expressing human DAF or mutants lacking specific DAF SCRs were screened by immunoblotting with an eluate containing the patient's antibody. Results: Sequencing showed a was A 1 A 2 . Sequencing revealed homozygosity for a novel FUT1 allele, 538C>T (Q180X); 1089T>G (A363A). FUT2 analysis showed homozygosity for the known silent mutation, 357C>T, previously associated with an active enzyme, present together with the novel mutation 278C>T (A93V). A third sample of Croatian origin was non-reactive with anti-A, anti-B, anti-H and typed Le(a+b-). FUT1 was homozygous for the nonsense mutation 826C>T (Q276X) and FUT2 was silenced by 428G>A (W143X). CONCLUSION: Three novel FUT1/FUT2 alleles were identified. The first, FUT1 689A>C, encodes a weakly active enzyme resulting in weak RBC H antigen. The second (FUT1 538C>T, 1089T>G) is expected to be inactive. Weak A antigen present on the RBCs of this donor is probably due to the third novel allele, in FUT2 (278C>T;A93V), since the Le(a-b+) phenotype of the RBCs shows the enzyme is active. Heterozygosity for FUT1 826C>T was described in a para-Bombay sample (Kelly et al., PNAS 1994; 91:5843) . Expression studies showed 826C>T to be an inactivating mutation. Homozygosity for FUT1 826C>T associated with an inactive FUT2 gene is consistent with the O h phenotype of the Croatian donor. Our results highlight the heterogeneity at the FUT1 and FUT2 loci in persons of the Bombay and para-Bombay phenotypes.

S T Huang, Z Huang, M B Marques, University of Alabama at Birmingham, Birmingham, AL BACKGROUND: Polyethylene glycol (PEG) can be covalently bound to the red blood cell (RBC) surface proteins to prevent antigen-antibody reactions, including those related to ABO. Since PEG binds nonspecifically, we hypothesize that it also blocks RBC GPI-linked proteins. If these proteins are blocked, complement activation will lead to RBC hemolysis mimicking PNH. In order to confirm our hypothesis, we used the acidified serum lysis (Ham) test, which is specific for PNH. METHODS: We used different concentrations of methoxy PEG-succinimidyl propionate (MPEG; MW:5,000) and branched PEG-succinimidyl propionate-albumin conjugates (XPEG; MW:20,000) to coat the RBC surface. MPEG was crosslinked with human RBCs at 20C for one hour. XPEG was first incubated with human RBCs for one minute, followed by the addition of 5% human serum albumin, and subsequent incubation at 20C for one hour. After conjugation, the PEG-treated RBCs were incubated with ABO group specific fresh acidified serum (AS) or heat inactivated (IS), at 37 C for 30 min. Acidified serum was prepared by adjusting the pH to 6.5 with 0.2N HCl, and heat inactivation of complement was achieved by incubating the acidified serum at 56 C for 30 minutes. After incubation of the PEG-coated RBCs in both sera, the degree of hemolysis was recorded by macroscopic observation and determination of free hemoglobin (OD541) spectrophotometrically in the supernatant. RESULTS: See table. Footnotes as follow: underlined results: gross hemolysis; 100% hemolysis: OD541: 2.586; control untreated RBCs: AS: 0.073; IS: 0.071 CONCLUSION: We proved that pegylated RBCs mimic PNH cells showing positive Ham test in the higher concentrations of PEG (5% MPEG, 7% XPEG or higher) in fresh acidified serum but not in heat-inactivated serum. This suggests that PEG conjugated the GPI-linked proteins. PNH RBCs which lack these proteins suffer hemolysis especially at night, when the pH is lower. Since these proteins are important to prevent activation of the alternative complement pathway, their blockage may cause spontaneous hemolysis if cells are incubated with ABO specific serum at 37 C, especially at low pH. Since 10% MPEG-treated RBCs also hemolyzed in the absence of complement (IS), we speculate that there is another mechanism, perhaps related to direct RBC membrane damage in this high concentration of PEG. sometimes identified at the same time. This study was undertaken to evaluate the serological nature and clinical behavior of autoantibodies and alloantibodies occurring in the same patient. METHODS: One hundred consecutive patients with warm autoantibodies were studied. Patient medical records, including clinical laboratory data, and all associated blood bank serological studies were reviewed. RESULTS: Fifty three of the 100 patients with warm autoantibodies had detectable alloantibodies. Of these, 24 had single specificities and 29 had multiple antibodies (range: 2-7). Antibodies of the following blood group systems were identified: Rh (40), Kell (21), MNS (7), Duffy (6), "high-titer, low avidity" (6), Kidd (5) and Lutheran (2). Evaluation of the sequence of antibody formation revealed that in 34/53 cases, the autoantibody and alloantibody(s) were detected together (i.e. unable to determine which type of antibody was formed first). In only 4 of these cases (11%) was the autoantibody associated with hemolysis. In 14 cases, autoantibodies were detected after alloantibodies were identified. In none of these cases was the autoantibody associated with hemolysis. In the remaining 5 cases, autoantibodies were identified before alloantibodies were detected. In 4 of these cases (80%), the autoantibody was hemolytic. The relationship between the sequence of antibody formation and hemolysis was significantly different by chi-square analysis (p < 0.005). Evaluation of the clinical behavior of all 100 autoantibodies revealed that 29 were hemolytic. Of these, only 8 (28%) had associated alloantibodies while 21 (72%) occurred alone. Of the 71 autoantibodies not associated with hemolysis, 45 (63%) were accompanied by alloantibodies while the remaining 26 (37%) occurred alone. The relationship between the presence/absence of accompanying alloantibodies and hemolysis was also significantly different by chi-square analysis (p < 0.005). CONCLUSIONS: Most autoantibodies are not clinically significant with respect to hemolysis. Autoantibodies that are formed after alloimmunization, or that are detected concurrently, are unlikely to be hemolytic. Autoantibodies formed prior to alloimmunization are more likely to result in hemolysis. These findings may be useful in predicting the clinical significance of warm autoantibodies and determining the extent to which patients should be followed for hemolysis. They may also provide insight into the immunological mechanisms that give rise to autoantibodies of differing pathogenicities.

West-Austrian DNA Polymorphism of ABO Alleles Causing Weak Blood Group a Expression C Gassner, M Bodner, M Mitterberger, E Rainer, S Kilga-Nogler, D Schoenitzer, Central Institute for Blood Transfusion and Immunological Department, Innsbruck, Austria Background. The reason for weak expression of A antigens, as found e.g. in A3, Aend, Ax and Aw, were shown to be in part caused by altered A alleles of the ABO gene locus. To investigate the DNA-polymorphism of these alleles in our population, 11 individuals of known weak ABO blood group A expression were investigated serologically and with molecular methods. Study Design and Methods. ABO serology was done using gel matrix testing and plate test with human polyclonal anti-A, anti-H and anti-A hel. Blood samples of all serologically predefined "weak As" (without further specification) showed a weak reaction with human polyclonal anti-A and/or anti-AB and practically negative reactions with anti-A hel. Eleven donors with ABO genotype A/O (e.g. presence of an A allele and one allele showing a deletion of G261 in exon 6 of the ABO gene) were selected for DNA sequencing of exons 2-7 and intron 6 of the A allele. Discrimination of the A allele long range PCR reverse (exon 2-6) and forward (exon 6-7) amplification was accomplished by sequence specific priming at G261. Results. Of 11 "weak A" alleles, 4 were identical to the DNA sequence of A101 (allele designation at www.bioc.aecom.yu.edu, AF134412) and 1 was identical to A102 (X84746al). Two other A alleles also showed an A101 specific coding sequence, but with an insertion of 2 independent As in intron 6 and a C285G substitution in exon 6, respectively. Another 4 were identical to A101 but with substitutions C467T and G829A and a deletion of C1061; the later 4 were identical to A302 (no accession, Barjas-Castro ML, 2000) beside the additional C467T. Conclusions. Since 4(5) out of 11 samples with "weak A" expression showed A alleles identical to A101, other reasons for weak A expression than blood group A allele polymorphism can not be excluded. However, 6(5) showed DNA variations compared to A101, with 4 of them representing a unique allele, comparable to A302. Weak A expression is relatively frequent in the investigated donor population with a minimum (allele)frequency of at least 0,00081 (e.g. frequency of "weakA"/O heterozygous individuals is about 1 among 1000). Therefore, evaluation of new (monoclonal) ABO test sera is of importance and could be optimized using a molecularly predefined panel of known "weak A" test cells. Background: Extension of platelet (Plt) storage from 5 to 7 days would reduce the potential for shortages and outdating. Storage of leukoreduced Amicus ® products in PL 2410 containers up to 8 days was evaluated using in vitro measures and in vivo radiolabeled recovery and survival (R/S) studies. Study Design: Sixty products encompassing a range of Plt concentrations (0.9-2.6 ¥ 10 6 /uL), Plts/container (1.4-5.0 ¥ 10 11 ) and plasma storage volumes (98-331 mL) were evaluated. Following collection from healthy subjects, products were equally divided and stored under standard conditions up to 6 or 8 days. Day 1, 6, & 8 studies included measurement of product volume, Plt count, mean Plt volume, white cell count, pH, pCO 2 , pO 2 , HCO 3 -, glucose, lactate, LDH, morphology score (maximum score 400), extent of shape change (ESC), hypotonic shock response (HSR), and Pselectin expression. Bacterial cultures were also performed. Using dual radiolabels ( 111 In and 51 Cr), R/S values for 6-day and 8-day stored Plts (N = 18) were compared. R/S were calculated using a multiple hit function of the COST program. Results: All Plts were culture negative throughout the storage period. Mean Day 6 Plt recovery was 53 ± 11% compared with 49 ± 11% for cells stored for 8 days. Survival values were 163 ± 31 and 154 ± 31 hours, respectively. Mean Day 1, 6 & 8 pHs (22°C) were 7.4 ± 0.1, 7.2 ± 0.2 and 7.0 ± 0.3, respectively. Only one evaluable unit had a pH value of less than 6.2 at Day 8. Additional results of in vitro testing are summarized below. Conclusion: Storage of Amicus Plts in PL 2410 plastic containers for 8 days appears acceptable as assessed by laboratory testing and in vivo radiolabeled R/S studies.

to compare R and S between the two radiolabels. Six data sets were analyzed in 6 different labs using their own COST programs to compare the applications' outputs. Results: Labeling after 8 d of storage with 51 Cr and 111 In gave indistinguishable results: R = 45.8 ± 10.5% vs. 48.6 ± 8.0%; S = 6.0 ± 1.2 d vs. 5.7 ± 1.5 d. After 7 d of storage, R = 52.9 ± 4.3% and S = 7.1 ± 1.9 d which were 89.4 ± 8.5% and 75.6 ± 17.0% of the values with fresh platelets. These results were not inferior to the limit of 0.67*R fresh and 0.5*S fresh . Multiple applications of the COST program yielded the same results with all commonly used survival models and the same extrapolated T 0 recovery. Conclusions: We conclude that: (1) Comparing stored plts to a fresh sample from the day of reinfusion of the "test" system provides a reasonable means of assessing plt recovery and survival; (2) A statistically rigorous means of establishing non-inferiority avoiding the use of ratios is feasible; (3) The COST program in use in multiple laboratories gives the same results despite minor modifications; (4) Radiolabeling with 51 Cr or 111 In provide similar results, at least to 8 days of storage. Mathematical manipulations of results of plt efficacy studies should be standardized. Background: Transfusion of allogeneic platelets is the mainstay of therapy for patients with life-threatening thrombocytopenia. However, donated platelets can only be stored for 5 days and are maintained at room temperature, increasing the risk of bacterial overgrowth. Developing a method to produce functional platelets in vitro would greatly advance transfusion therapy. During our studies to understand megakaryocyte development, we discovered that a Src kinase inhibitor, SU6656, induces cellular enlargement, polyploidization, and cytoplasmic fragmentation of several hematopoietic cell lines. Therefore, we tested the hypothesis that these fragments possess platelet-like activity. Methods: The megakaryocytic cell line UT7-TPO was cultured in thrombopoietin with or without the addition of SU6656. Because of evidence that phorbol esters can also induce megakaryocytic differentiation, UT7-TPO cells were also cultured in the presence of PMA. Cell morphology was monitored daily by light microscopy, and the cell surface expression of megakaryocytic markers CD41 and CD62 was assessed by flow cytometry. The cells were harvested on day 6, when the majority of cells had become polyploid and started shedding platelet-like fragments. These fragments were separated from intact cells by low-speed centrifugation followed by high-speed centrifugation. Next, they were fractionated using a discontinuous albumin density gradient column (0/2%/4% albumin). After 90 minutes at unit gravity, cell fractions were collected, examined under a light microscope, and pooled based on cell size and morphology. The pooled fragments were washed with phosphate buffered saline (PBS), resuspended in PBS containing 60% human plasma, and aggregation was measured using collagen, ADP, arachidonic acid and epinephrine as agonists. Results: After 24-96 h in culture with SU6656, UT7-TPO cells demonstrated a marked increase in surface markers CD41 and CD62, reduced cell number, and increased nuclear ploidy (including 32 and 64N). By 4 days, proplatelet-like processes began to form and cellular fragmentation was evident. These fragments demonstrated rapid and prolonged aggregation in response to all agonists but no spontaneous aggregation was observed. Fragments generated in the presence of PMA displayed similar but less prolonged aggregation.(PMA 3 min vs SU6566 > 10 min). Conclusion: These initial studies provide evidence that it may be possible to promote megakaryocytic differentiation and thrombopoiesis in vitro from cell lines. Although many studies remain to be done to better characterize these platelet-like fragments, the demonstration of physiologic function deserves further investigation.

L J Dumont, A Smart, J Rice, T Vandenbroeke, Gambro BCT, Inc., Lakewood, CO BACKGROUND: Production of lactic acid as a glycolysis end-product during in vitro storage of platelets leads to reduced pH and platelet damage. Increased glycolytic rates are generally attributed to a Pasteur shift at low O 2 levels. However, this does not explain all observations of increased rates of glycolysis in stored platelets. We conducted this study to evaluate the effect of CO 2 level on glycolytic rates in stored platelets. METHODS: Aphere-S85-040G Further Investigation of Use of Radiolabeling Studies to Document Platelet Efficacy J P Aubuchon, L Herschel, J Roger, Dartmouth-Hitchcock Medical Center, Lebanon, NH Background: Previous validations of a platelet (plt) efficacy criterion based on comparison to the radiolabeled recovery (R) and survival (S) of fresh plts from the subject (Transfusion 2004; 44:131-33; 36-41) used an aliquot of plts from the unit reinfused within 20 h. Concerns about standardization of techniques across labs and time prompted further investigations to (1) establish a protocol to use plts processed manually on the day of reinfusion of the stored ("test") plts; (2) apply a more rigorous equivalency evaluation rather than depend on a ratio to compare fresh and stored plts; (3) document similarity of results of the COST program modified to run on multiple platforms; (4) document similarity of results using 51 Cr and 111 In radiolabels in platelets stored 8 d. Methods: With IRB approval, plts from an apheresis unit (n = 11; Trima Accel, Gambro) were labeled with 51 Cr on Day 7 and compared to plts collected from the subject that day and separated via a PRP method in conical plastic tubes before labeling with 111 In. (Br J Haematol 1983; 44:717-23 ) R and S were determined using COST beginning with the 24 h point with correction for injectate elution, plasma radioactivity, and RBC labeling. Paired differences between fresh and stored plts' R and S were compared to that between the mean fresh result and a multiple (0.67*R; 0.5*S) to determine non-inferiority with the standard. In a different study, 12 subjects had apheresis plts stored for 8 d and separate aliquots labeled by 51 Cr and 111 In sis platelets (Trima Accel) were collected from 2 donors, pooled on day of collection, and divided between a standard 1L ELP storage bag (Control) and a modified bag with lower average gas permeability (Test). Six (6) matched pairs were stored at 22°C on a horizontal shaker up to 7 days. Aliquots were removed on days 0, 1, 5, and 7 for measurement of pH, pCO 2 , pO 2 , lactate, glucose, and platelet concentration using standard methods. Glucose consumption and lactate production rates were calculated as rates normalized to platelet content (mmol/10E12 plt/hour). RESULTS: Glycolytic rates established over the first day of storage were consistently maintained for all products over storage. As expected, Test products had higher CO 2 accumulation, glycolytic rate and pH decline. However, O 2 was not depleted to levels required for a Pasteur shift in metabolism (2 mmHg). The lowest pO 2 observed in the test bag (28 mmHg) was associated with a low glucose rate (17.8). The average paired differences between Control and Test were 12.7 ± 6.1 (mean ± SE) for glucose and 22.6 ± 12.1 for lactate. Lactate rate averaged 1.78 ± 0.10 (mean ± SE) times glucose rate, and was not different between Control and Test.

withdrew prior to their second procedure resulting in 30 evaluable products. All bacterial cultures were negative after 7 day incubation. Lactate and glucose rates established over the first day of storage were consistent over the 7 day storage period.

CONCLUSION: Reduction of gas exchange rate for platelet storage bags results in retention of CO 2 , reduction in pH, and increased glycolytic rates. The increased utilization of glucose is not explained by O 2 starvation. CO 2 may be acting to inhibit rate limiting decarboxylating enzymes in the mitochondria, directly slowing down oxidative phosphorylation, and indirectly upregulating glycolysis. This potential effect is confounded in this study with changes in pH, although the inverse relationship between pH and glycolytic rate is inconsistent with other literature reports. The potential CO 2 effect proposed here would need to be confirmed in a test system which removes the pH confounding.

L J Dumont, T Vandenbroeke, Gambro BCT, Inc., Lakewood, CO BACKGROUND: Glucose catabolism by glycolysis results in the production of lactic acid with an associated drop in pH for in vitro stored platelets. In preliminary experiments, the glycolytic rate (estimated by lactate production) varied widely between stored products (22-114 mmol/10E12 plt/hr). This study was conducted to evaluate the donor repeatability of lactate production rates for apheresis platelets stored in plasma. METHODS: Apheresis platelets were collected using Trima version 4 or Trima Accel. Study design was a full factorial, randomized trial with replicates run in two blocks. Independent variables were Trima device (2 devices), Trima software (v.4, Accel), and platelet concentration (1.5, 2.5 ¥ 10E6 plt/mL) in 200 mL per bag. Platelets were collected from 16 donors on two separate visits, and stored in 1L ELP bags at 22°C on a horizontal shaker up to 7 days. Aliquots were removed on days 0, 1, 5, and 7 for measurement of pH, pCO 2 , pO 2 , lactate, glucose, and platelet concentration using standard methods. Two mL aliquots were cultured on days 1 & 7 in BacT/ALERT aerobic bottles. Glucose consumption and lactate production rates were calculated as rates normalized to platelet content (mmol/10E12 plt/hour). RESULTS: Two donors There was a high correlation between repeat products from the same donor (intraclass correlation coefficient, r = 0.826). Lactate generation rates from 154 unique donors were fit to a gamma distribution from which we estimate 10% of donors have a basal rate greater than 65 mmol/10E12plt/hr in the ELP storage system. CONCLUSION: The basal rates of lactic acid generation are highly correlated with donor. This may explain previous observations of some donors' platelets not storing well. Additional study is needed to determine if these differences are from dietary sources (i.e., fuel utilization), basic differences in cellular energy machinery (e.g., enzyme polymorphisms), or other causes.

T L Manlove, C Lombardi, J Miripol, Terumo Medical Corporation, Somerset, NJ Background: The Terumo IMUFLEX ® WB-SP (Saving Platelets) Blood Bag System with Integral Whole Blood leukocyte reduction filter (WB-SP System) is currently undergoing in-vivo clinical trials in the United States (US). The study reported here was designed to confirm acceptable in-vitro performance of a new investigational lot of the WB-SP System, and to assure that the components prepared from this lot would conform to AABB standards. The standards state that leukocyte-reduced (LR) blood products must retain a minimum of 85% of the original product and leave a residual leukocyte count of < 5 ¥ 10 6 per container. In addition, 90% of the platelet concentrates prepared must contain ≥5.5 ¥ 10 10 platelets per container and all units must have a pH of ≥6.2 (at 22°C) at the end of storage. Study Design: Units of Whole Blood (WB) were collected into 500 mL Terumo bags with 70 mL CPD from 16 donors. All units were maintained and filtered in the WB-SP System at room temperature within 8 hours of collection. Filtration duration was recorded. The LR-WB units were further processed into red cell concentrates (LR-RCC), platelet concentrates (LR-PC) and plasma (LR-Plasma) using standard production procedures. First and second centrifugations (Sorvall RC3BP centrifuge with a H6000 rotor) were 3,000 rpm for 4 minutes, slow stop setting 8, and 4,100 for 5 minutes, respectively. The BD FACSCalibur TM with the LeucoCount TM reagent kit was used to enumerate Residual White Blood Cells (WBC). Platelet counts were obtained with the Abbott CellDyn 3500, and the ABL Radiometer 700 was used to measure pH. Results: [Mean ± Standard Deviation (n = 16)]

The WB recovery of all units satisfied US standards with 100% of the WB units retaining > 90% of the original product. The platelet counts and pH of all the LR-PC on day 6 met the US standards. All units met the US and European (<1 ¥ 10 6 ) standards for residual WBC. Conclusions: All units in this study demonstrated acceptable filtration duration and leukoreduction. These units met and exceeded the volume recovery and leukoreduction requirements. This study demonstrated that the Terumo IMUFLEX WB-SP System achieved leukoreduction of WB that met both the US and European standards and produced LR-PC that met AABB standards.

S90-040H Experience with an "Improved" Version of an Hbsag Assay and Reentry of Associated Deferred Donors S L Stramer, A G Wagner, A C Fisher, American Red Cross, Gaithersburg, MD; R L Estabrooks, American Red Cross, Charlotte National Testing & Reference Laboratories, Charlotte, NC; M Schroeder-Jenkins, American Red Cross, Washington, DC; R Y Dodd, American Red Cross Holland Laboratory, Rockville, MD; J P Brodsky, Quality Analytics, Inc., Riverwoods, IL Background: HBsAg testing has been performed since the 1970s using tests of increasing sensitivity. As sensitivity increases, specificity (of both the screening and confirmatory tests) should also improve but not deteriorate. Methods: On August, 2003 a system of 9 labs located at various US geographic regions converted from a System 2 to System 3 HBsAg ELISA (Ortho Clinical Diagnostics). The companion System 3 reagent was used for neturalization (neut) testing. Neut results for all repeat reactive (RR) samples were compared to anti-HBc results (Ortho). All RRs were tested by a 2nd HBsAg EIA with equal or improved sensitivity to System 3 (Bio-Rad HBsAg 3.0 shaker). Donors whose HBsAg RR results are unconfirmed are permitted reinstatement based on FDA guidance whereas donors with false-positive (pos) neut results (based on nonreactive, NR, anti-HBc and Bio-Rad HBsAg tests) will be eligible for reentry based on NR follow-up test results including HBV DNA. Conversion to another HBsAg assay (Abbott) occurred following the 6-month experience. Results: The System 3 assay was validated to have improved sensitivity (ad/ay HBsAg detection increased from 0.7 to 0.1 ng/mL or less), and a RR rate of 0.05%. For the 6 months following conversion, 3.58 million donations were screened with a mean RR rate of 0.24% (range by lot = 0.13-0.3%; by lab = 0.16-0.36%); all labs were following all test kit instructions with no violations in procedure noted. In total, 8524 RR donors were deferred. Following neut, 17.7% (1509) were neut pos for an HBsAg prevalence of 1 : 2614 vs a historic rate of 85% neut pos and an HBsAg prevalence of 1 : 6647. Of the 1509 neut pos, 644 (42.7%) were anti-HBc NR. Of the 644, 76.6% (493) tested NR by Bio-Rad HBsAg, for a "false" confirmed pos rate of 1 : 7262. The remaining 865 (57.3%) of the 1509 neut pos donors tested anti-HBc RR (vs 96.5% previously). Of those neut pos samples containing anti-HBc, 836 (96.7%) were RR by Bio-Rad HBsAg (1 : 4282). Donor reentry efforts for all lost donors will be implemented. Following conversion to the Abbott EIA, RR rates decreased to 0.02% matching historic rates of 1997. Conclusions: Test kits should be properly optimized for sensitivity and specificity prior to implemementation. High donor loss and expense results if large numbers of donors are required to be followed and managed using novel reentry schemes. Background: Based largely on data from the Transfusion-Transmitted Viruses Study showing that donors positive only for anti-HBc (total) were associated with hepatitis B virus (HBV) transmission, screening blood donors for anti-HBc was adopted in the US in 1991. Later studies of HBsAgnegative donors positive for anti-HBc (total) showed that a small percentage were actively infected with circulating HBV DNA in their blood. The great majority of these donors were not recently infected as shown by their negative test results for IgM anti-HBc. In early 2004, a donor presented who was deferred in 1983 after he was implicated in two cases of posttransfusion hepatitis B involving two different donations four months apart. The former donation was positive for anti-HBc (total) but negative for HBsAg (Abbott, Auszyme, Procedure C) and anti-HBs. However, RIA (Abbott, Ausria) testing for HBsAg with overnight incubation yielded a positive, lowlevel screen with neutralization. The latter donation was positive for anti-HBc (total) and anti-HBs (S/CO of 1.1) but negative for HBsAg by EIA and RIA using overnight incubations. Testing of 10 additional stored samples from this donor collected from 1982 through 1984 revealed that he was negative for HBsAg (EIA and RIA) but positive for anti-HBc (total) on all samples. The first 3 of 5 samples collected in an 11-month period after the second implicated donation were the only samples positive for anti-HBe. Two of the 5, the first and the last, were positive for anti-HBs with S/CO's of 4.0 and 1.0, respectively, and were the only other samples found with anti-HBs. The donor's contemporary sample was negative for HBsAg (Auszyme, Procedures B and C), positive for anti-HBc (total, Abbott and Ortho), and positive for anti-HBs (S/CO of 1.6). Ten ml of donor plasma were centrifuged as 1 ml aliquots, extracted and amplified in triplicate by PCR (Roche COBAS AmpliScreen HBV Test, Investigational). Of the 10 extracts (30 amplifications), 4 were positive (one of 3 amplifications positive). The estimated sensitivity of the assay (95% LOD) testing 3 extracts of a 1 ml sample is 6 copies/ml. Conclusion: HBV can persist for long periods in donors who are negative for HBsAg but positive for anti-HBc (total) with no, or low-level anti-HBs. Such donors would likely transmit HBV to patients if they were not interdicted by FDA-mandated donor screening for anti-HBc (total).

Yield Analysis on Current Anti-Hbc Donor Deferral Policy C T Fang, E P Notari IV, S L Stramer, R Y Dodd, American Red Cross, Rockville, MD Background: The current donor deferral policy of many blood organizations allows a donor repeat reactive, RR, in an anti-HBc EIA (HBsAg nonreactive, NR) for the first time, to be notified, entered into surveillance and allowed to donate again. However, if any future donation is again RR, the donor will be indefinitely deferred. The donation status and the number of transfusable donations (yield) by such donors were determined for a 3.5 year period. Methods: Donors who donated in 2000 and were anti-HBc EIA (Ortho) RR (Ortho System 2 HBsAg NR) for the first time since 1995 were selected for study. Anti-HBc results on their further donations through the end of 2003 were compared to the anti-HBc NR donor population of the same year. Results: For calendar year 2000, 6,508,729 units were collected from 3,994,979 donors. During this period, 17,193 (0.43%) donors were found to be anti-HBc EIA RR (HBsAg NR) for the first time. These donors were notified. By the end of 2003, 12,656 (73.6%) had not returned for subsequent donation. This nonreturn rate was significantly higher than that in the anti-HBc NR donor group (34.6%; p << 0.0001). For the 4,537 anti-HBc RR donors who donated again, 3,442 (75.9%) were anti-HBc RR on their next donation and thus indefinitely deferred. A second anti-HBc RR donation, following at least one NR donation after the first anti-HBc RR donation, was identified in 153 (3.4%) donors. These donors had given a total of 352 acceptable donations (mean 2.3 donations per donor, range 1-30) between two anti-HBc RR donations for a yield of 0.66 donations per person-year. Finally, the remaining 942 (20.7%) one-time anti-HBc RR donors had given a total of 4,144 anti-HBc NR donations (mean 4.4 donations per donor, range 1-67) for a yield of 1.26 donations per person-year. This rate is comparable to that for the anti-HBc NR donor group (1.46 donations per person-year). Overall, for a time period of 3.5 years, 4,496 transfusable donations were collected from a total of 17,193 first-time anti-HBc EIA RR donors for an overall yield of 0.07 donations per person-year. Conclusions: Our data indicate that the yield of successful donations from an anti-HBc donor deferral policy allowing first-time reactives to continue to donate is low. However, cost-effectiveness for the comparison of a first-time deferral and secondtime deferral policy is needed prior to recommending changes. This study highlights, again, the need for a donor reentry policy using a more sensitive and specific anti-HBc test than currently available in the US. obtained from a comprehensive data base of over 31.25 million donations and were analyzed by calendar quarter and by donor status. Results: Overall seroprevalence and RNA yield rate of HCV were respectively 69 and 0.435 per hundred thousand (pht) donations; the corresponding figures for HIV were 3.4 and 0.032 pht. No temporal trends could be discerned for HIV seroprevalence or RNA yield rate over the study. However, the seroprevalence of HCV declined significantly over the study period, from an overall rate of 85 pht in Q1 1999 to 43 pht in Q4 2003 (R 2 = 0.75, p < 0.001; omitting aberrant data from Q3 2001 due to post-9/11 donations). There was no discernable trend in the HCV RNA yield rate with the exception of lower rates when pools of 128 were used. There was a clear and statistically significant increase in the seroprevalence of HCV in the third quarter of every year among both first-time and repeat donors (p < 0.025 in all instances). No such pattern was seen for HCV RNA yield rates. Finally, aggregate HCV RNA yield rates for pools of 16 in the period Q1 2002 to Q4 2004 (0.562 pht) were greater than those for the period Q4 1999 to Q2 2002 (0.444 pht); however, these were not significantly different. Conclusions: There was no significant correlation between NAT yield rates and seroprevalence over time for either HIV or HCV. HCV seroprevalence showed an overall pattern of decline and a repeating pattern of increase during the third calendar quarter. This appears to be due to increased presentation of donors with prevalent, rather than incident infection since RNA yield rates were unchanged. NAT yields for both HIV and HCV in pools of 16 appeared consistent for the reporting period.

L Cooling, D Hwang, J A Shayman, University of Michigan Hospitals, Ann Arbor, MI Introduction: ABO, LKE and Lewis (Le) antigens (Ag) play a role in many epithelial cancers, including metastatic renal cell carcinoma (RCC), which is often characterized by a H+, LKE+, Le Xphenotype. Recently, H-Ag has been linked to resistance to apoptosis whereas LKE and H-Ags may facilitate adhesion and metastasis. All-trans retinoic acid (ATRA) is a differentiating agent capable of altering H, LKE and Le expression in many tumors. We therefore tested ATRA in the poorly-differentiated, metastatic RCC line, ACHN. Methods: ACHN cells were grown in media, dimethyl sulfoxide (DMSO) or DMSO+ATRA for 2, 3, 7, 14 and 21 days. Cells were analyzed by flow cytometry with lectins and monoclonal antibodies against sialic acid, H, Le X , Le a , LKE, I, and T Ags. Apoptosis was monitored by percent cloning efficiency after serum deprivation (72 hrs) and heat shock (44C, 30 min). ATRA effects on cell proliferation were examined by cell viability, cell activation, cell number and anchorage-independent colony growth in soft agar. Results: ACHN cells were LKE+, H+, Le a +, I+, T+ and Le X -. ATRA decreased H-Ag two-fold (87%-37%+, P < 0.0001), while increasing LKE (P < 0.05), Le a (P < 0.001) and I (P < 0.05) expression. Decreased H-Ag expression was accompanied by a >95% increase in sensitivity to serum deprivationinduced apoptosis (P < 0.001), mediated through the RA receptor. ATRA had no effect on heat shock-induced apoptosis, cell growth or viability (>95%). ATRA enhanced anchorage-independent colony growth (132%, P = 0.03). Conclusions: Contrary to expectations, ATRA did not induce a well-differentiated RCC phenotype (LeX+, LKE-). ATRA did decrease H expression, with increased sensitivity to apoptosis, similar to that described in colon carcinoma. The antineoplastic effects of ATRA include transrepression of APmediated gene activation. Interestingly, 11 AP transcription factor binding sites are present within the 5¢ upstream regions of the H-gene (FUT1) promotor and transcription start site (exon 7). These observations may reflect direct inhibition of H-gene expression by ATRA. There has been no systematic monitoring of NAT results over the entire period of use, including post-licensure. Methods: A large blood system has used the Chiron/GenProbe Procleix assay for all donations. Testing was performed in pools of 128 from March 1999 to Sept 1999 and pools of 16 thereafter. RNA yield was defined as an RNA confirmed-positive, antibody-negative sample; confirmation involved PCR (NGI) and follow up to donor seroconversion. All donations were also screened for anti-HCV (Ortho System 3) with repeat reactives confirmed by RIBA 3.0, and for anti-HIV-1/HIV-2 (Abbott 3A77) with repeat reactives confirmed by western blot (Calypte). Data representing all testing results were activity of A and B transferases. Methods: A panel of ABO expression plasmids was established containing the six known missense mutations of the coding sequence from exon 2 to 5 of the ABO gene. Among them were the rare mutations 46G>A, 190G>A and 203G>C and the common mutations 106G>T, 188G>A, and 220C>T which are characteristic for the ABO*O02 allele. Blood group A-specific plasmids with single or multiple mutations were constructed by site directed mutagenesis. HeLa cells were used to transfect ABO expression plasmids. Plasmids encoding for the blood groups A1, B and O as well as for defined A and B subgroups were used as controls. A and B antigen expression on transfectants was analyzed by flow cytometry using monoclonal antibodies. Results: Transfection of plasmids containing single missense mutations or combinations of them resulted in a decrease of A antigen expressing cells (AEC) in a range from 2 to 13% and of mean fluorescence intensity (MFI) in a range from 2 to 30% compared to HeLa cells transfected with the ABO*A101 control construct. A reduction of A activity in a range from 4 to 31% (AEC) and from 8 to 40% (MFI) was detectable on HeLa cells transfected with both a variant ABO*A allele and the normal ABO*B101 allele compared to cells co-transfected with ABO*A101 and ABO*B101 constructs. In the co-expression experiments, the lowest expression levels were comparable with the A activity of HeLa cells transfected with a control plasmid containing a blood group Ax-related mutation. The weakest values for A antigen expression were observed for plasmids with mutations causing amino acid changes in the stem region of the A transferase. Conclusion: Mutations in early exons, outside the enzyme's active site, can affect expression of ABO blood group antigens. In particular, in AB phenotypes weak A and B subgroups can arise from competition between normal A or B glycosyltransferases and variant forms that contain amino acid changes in the transmembranous domain and stem region.

S T Johnson, C L Piefer, J T Fueger, J L Gottschall, R H Aster, Blood Center of Southeastern Wisconsin, Inc., Milwaukee, WI Background: Patients with drug-induced immune hemolytic anemia can present with drug-independent as well as drug-dependent antibodies (DDA). An inappropriate diagnosis of warm autoimmune hemolytic anemia can be made when drug-independent antibody is present and the patient's DAT is strongly positive due to IgG. Drug-induced hemolysis serologically presenting as cold agglutinin syndrome has not been previously reported. Case Report: A 57 year-old male receiving ceftriaxone (2 g intravenously for 4 days) for severe bronchitis, developed hemolytic anemia. On admission his hemoglobin was 14 g/dl but dropped to 5 g/dl four days later despite transfusion of 4 units. On day 4 his DAT was 4+ due to C3 only. Eluate was negative. In the absence of drug, using a saline tube method, serum reacted 1+ with all red cells at immediate spin (IS), 2+ after a 30 minute 37C incubation and negative in the IAT. All RBCs reacted 4+ in ficin IAT and nonreactive in a PEG IAT in the absence of drug. Patient serum in the presence of ceftriaxone (1 mg/ml) reacted to a titer of 256 at RT, 1024 at 37C and 2 at IAT. Eluate was positive (2+) after 60 minute 37C in the presence of ceftriaxone (1 mg/ml). 2-ME treatment of serum reduced the titer to 1 at RT, 2 at 37C and 0 at IAT indicating the DDA was predominantly IgM in nature. In order to determine if ceftriaxone was present in the patient's serum it was dialyzed overnight using Spectra/Por2, 12,000-14,000 Dalton MWCO in Phosphate Buffered Saline, pH 7.3. The dialyzed serum was negative at all phases and when ceftriaxone was added back to the test the DDA was again detected. Conclusion: Initial serologic results were similar to those seen with cold agglutinin syndrome, i.e., C3 only positive DAT and cold-reactive antibody present. Drug present in patient's serum at time of presentation allowed for the cold reactive (IgM) drug-dependant antibody binding without addition of drug to the test. It is important to be aware of these findings and medication history when investigating patients experiencing hemolytic anemia to avoid misdiagnosis. Background: Red blood cell (RBC) alloimmunization can limit therapeutic options for patients who become alloimmunized to multiple or widely expressed antigens. A murine model would be useful to better identify which immune mechanisms could be interfered with to treat or prevent RBC alloimmunization. Study Design and Methods: Donor mice (B6CBA-Tg-Fy b ) that express the human Fy b antigen on RBC and white blood cells were donors of whole blood which was centrifuged to obtain RBC that were washed. In some donors the spleen was removed and a cellular suspension was made that was added to the RBC. The recipient mice were divided into two groups given weekly intravenous transfusions: Group I: C57BL6 mice receiving 0.3 mL washed RBC; Group II: C57BL6 or B6CBA mice receiving 0.3 mL washed RBC and spleen cells. Circulating titers of IgG and IgM were measured in recipient sera by flow cytometry using RBC from donor mice as target cells. Sera were diluted 1 : 4, 1 : 16 and 1 : 64 in phosphate buffered saline, 1% bovine serum albumin, and 0.02% azide. Sera were adsorbed with human Fy(a-b-) RBC in an attempt to remove mouse anti-human antibodies. Adsorbed sera was then tested against human Fy(a-b+)RBC. A third group of C57BL6 mice received two weekly transfusions of RBC and spleen cells and were then transfused with 0.3 mL of biotin-labeled RBC. Blood samples were taken at 24 hours, 3 and 7 days post transfusion. Flow cytometry was performed to measure the percentage of biotin-labeled RBC remaining. Results: There was no change in staining for IgM or IgG in Group I mice. In contrast, Group II mice demonstrated IgM and IgG responses (see table below). There was score 9/12 agglutination of adsorbed sera with human Fy(a-b+)RBC. Decreased survival of transfused RBC was evident in mice at 24 hours post transfusion. Conclusions: It is possible to elicit the formation of anti-Fy b by intravenous transfusion in mice that lack this antigen. Sensitization appears to require continued weekly stimulation; however the reason for the significant increase in staining for IgM and IgG three weeks after the last transfusion is not clear. The survival of transfused Fy b positive RBC is diminished in sensitized mice. This may be a useful model to study RBC alloimmunization; additional studies are needed to further characterize this model. Representative results at the 1 : 4 dilution for a Group II B6CBA mouse. S100-040I Apparent Homozygosity of the Rare dCxes Haplotype and Associated Antibody to a High Incidence Rh Antigen J Poole, J Banks, A Wilkes, IBGRL, Bristol, United Kingdom; I Skidmore, J Quimby, National Blood Service, Birmingham, United Kingdom Background: The low incidence Rh antigen, C x , is usually associated with the DCe gene complex in whites. A rare Rh haplotype expressing C x , c,e,V,VS but no D or C was described in Somalians 1 . We describe here a case of apparent homozygosity of the dC x ce s haplotype and production of an antibody to a high incidence Rh antigen. Case Study: A Somalian female (KM) with two live children, in week 29 of her fourth pregnancy, had a strong antibody in her serum reacting by IAT with all RBC's except Rh null (¥3) and her own. RBC's with the following rare Rh phenotypes were incompatible: -D-/-D-(¥2), .D./.D., VS/VS, Rh:-51(MAR) [C w /C w ¥3], hr s -, hr B -. C x /C x cells were unavailable for testing. Absorption studies with 'antigen matched' cells removed the Rh-related antibody and additional antibodies were excluded, including anti-D. The chemiluminescence test, which measures the monocyte response to sensitised cells, was positive for the prediction of transfusion outcome and for HDN indicating a clinically significant antibody. Her RBC's had the Rh phenotype D-C-C w -C x + c+ E-e+ ce(weak +) V+ VS+ hr s + hr B + Rh:17, 29, 46, -51, Rh51-like positive (made in C w /C w ). A 3 kg baby girl was born at term with a strong positive DAT. The haemoglobin dropped from 16.6 g/dl to 10.6 g/dl over two weeks. Bilirubin levels were elevated to a maximum 240 mmol/l at 3 days but other parameters were normal. She was treated with phototherapy and discharged from hospital at 8 days. An autologous unit was taken from KM at 36 weeks but was not required for mother or baby. Conclusions: We conclude that KM has the probable Rh genotype dC x ce s /dC x ce s which gives rise to lack of a high incidence antigen. The antigen is lacking on her own RBC's and Rh null as judged by the reactivity of the antibody which she produced in response to pregnancy. Her RBC's were Rh:-51 but anti-Rh51 was excluded as Rh:-51 RBC's were strongly incompatible. This is the first example of apparent homozygosity of a rare Rh gene found only in Somalians and an associated antibody to a high incidence Rh antigen which caused mild HDN. Background: ABO genotyping is a powerful tool for resolving blood group discrepancies. The database of nucleotide mutations is extensive and both intron and exon nucleotide polymorphisms contribute to the expression of weak A and B phenotypes. We report an interesting case whose phenotype did not correlate to the genotype and the discovery of a novel B allele, which creates an amino acid substitution in an unusual location of the galactosaminyltransferase. Materials and Methods: Serological testing of the red cells from a 34 year old woman demonstrated an ABO discrepancy. PCR amplification and direct sequencing of exons 6 and 7 was performed using standard protocols. Results: Forward typing with murine monoclonal reagents: Anti-A: 4+, anti-B: 2+. Repeat testing with human anti-B: weakly reactive. Reverse typing: A 1 cells: 0, B cells: 2+. Broad spectrum DAT: negative. No unexpected antibodies were found. Analysis of ABO gene exons 6 and 7 revealed a consensus A 2 allele and a novel B allele characterized by a single nucleotide polymorphism at 556 A>G, which predicts a Met186Val amino acid substitution. Discussion: Amino acid 186 is located within a loop of amino acids adjacent to the active site of the enzyme. This series of approximately 15 amino acids appears disordered in the crystal structure of both blood group A and B [N-acetyl]galactosaminyltransferases (GTA, GTB). Although in some other glycosyltransferases, the binding of the donor substrate caused similar loops to change their orientation and assume an orderly configuration, the same has not been seen in the GTA and GTB crystal structures. The exact position of these amino acids during catalysis is currently not known. Only two mutations within the disordered loop have been reported and in these cases, weak B phenotypes have been observed. In a human model of GTB expressed in E. coli, an amino acid substitution within the disordered loop Arg188Ser virtually eliminated this enzyme's ability to transfer UDP-galactose (k cat = 4.0 ¥ 10 -6 s -1 ) with UDP-Gal donor [M. Palcic and R. Polakowski, unpublished observations]. k cat of normal GTB with UDP-Gal donor = 5.0 s -1 . The substitution of Met's long aliphatic chain with the smaller, bulkier side chain of valine might prevent the adoption of the proper orientation of the disordered loop during catalysis and thus reduce this novel GTB's ability to transfer galactose. Defining the role of this loop is important for a complete understanding of the mechanism of catalysis of the GTA and GTB enzymes. Kinetic studies of this donor's enzyme might help to explain the weak B phenotype. Background: Transfusion-associated, hematopoietic microchimerism (TA-MC) can persist in trauma patients. We recently reported TA-MC at shortterm follow-up in 24 out of 50 subjects (48%) who received at least two units of non-leukoreduced (nLR) packed RBCs and in 7 out of 13 subjects (54%) who received leukoreduced (LR) units. We now report results on long-term follow-up from 16 out of the 63 subjects recruited for this study (12 nLR and 4 LR). Materials and Methods: Blood from subjects was collected prior to hospital discharge. Two follow-up samples were collected thereafter (see table) . For some subjects, pre-transfusion samples were available. MC was determined by amplification of non-recipient HLA DR alleles (DR1, 3, 4, 7, 8, 9, 10, 11, 12, 13, 15, 16) using quantitative real-time PCR. We used sequence specific primers targeting specific alleles and SYBR green for detection of amplified products. Amplification targeting the HLA DQ-a and the major allele were also performed to quantify DNA input of each sample. Results: Long-term follow-up samples from 5 of the 16 patients showed persistent TA-MC. For these 5 subjects, 2 had available pre-transfusion samples which showed no donor derived cells. At discharge, 4 of the 5 had low levels, or undetectable, donor derived cells. All five showed persistence of donor cells, with MC ranging from 0.45 to 2.66 percent of the recipients' circulating WBC at the last follow-up. Based on analysis of donor specimens, MC from 3 (25%) of the subjects was linked to the nLR transfusion group (n = 12) and MC from the other 2 (50%) subjects to the LR transfusion group (n = 4). Quantification of donor derived DNA (genome copies per mL) in recipient's blood Sensitivity: 10 copies/mL; NA: Not Available; Post-tx: Post-transfusion; Post-Disch.: Post-Discharge, a period from discharge to indicated follow-up; d: days; m: months; %WBC: Percent peripheral WBC.

Conclusions: TA-MC may occur in transfused trauma patients. High-level long-term MC may develop in 30% of such subjects. LR does not eliminate MC.

C Matsumoto, T Araki, R Sobata, T Matsuda, H Okazaki, T Juji, Japanese Red Cross Central Blood Center, Tokyo, Japan; M Takanashi, Tokyo Metropolitan Red Cross Blood Center, Tokyo, Japan Background: Nonhemolytic transfusion reactions are often associated with allergic reactions. Mast cells are activated in allergic reactions. The mast cell activation process has three outcomes: degranulation, the de novo synthesis of leukotrienes, and the synthesis and secretion of cytokines and chemokines. We studied mRNA levels of cytokines in cord-blood-derived mast cells (CBDMCs) exposed to plasma samples implicated in allergic transfusion reactions (ATRs). Methods: To generate CBDMCs, mononuclear cells isolated from cord blood were incubated for more than 11 weeks in the presence of interleukin (IL)-6 and stem cell factor. IL-4 was added in the final week. To examine activated CBDMCs, they were sensitized with serum samples from an atopic patient and stimulated with an anti-IgE antibody. Some CBDMCs were exposed to plasma samples for one hour. RNA samples prepared from these CBDMCs were amplified by one-step RT-TaqMan PCR, using primer pairs for IL-3,-4,-5,-6,-8,-10,-13, and-18 (Assayson-Demand Gene Expression Products; ABI). Results: An increase in mRNA level was observed for all the cytokines investigated, except IL-4, after activation with atopic serum samples and the anti-IgE antibody. IL-8 mRNA level increased particularly. The CBDMCs were exposed to plasma samples from donated blood implicated in 35 ATR cases, and IL-8 mRNA was detected in these CBDMCs. The plasma sample, which had been isolated from a platelet concentrate implicated in a case of anaphylactic shock, was found to be the most effective in increasing IL-8 mRNA level. This plasma sample was fractionated by gel filtration chromatography. Fraction 6, corresponding to a protein of about 300 kDa, markedly increased IL-8 mRNA level in CBDMCs. Fraction 6 also induced histamine release from CBDMCs. Fraction 6 that had been previously incubated at 56°C for 30 minutes, or at 37°C in the presence of trypsin did not induce the increase in IL-8 mRNA level in CBDMCs. Conclusions: Some proteins of about 300 kDa in the blood used for transfusion may contribute to the development of ATRs by directly activating mast cells. Background: TRALI is a serious non-infectious transfusion risk traditionally characterized by dyspnea, hypoxemia, fever, chills, hypotension, tachycardia, and CXR infiltrates occurring within 6 hours of transfusion. We retrospectively analyzed the clinical and lab characteristics of TRALI reactions reported by transfusion service physicians. Methods: Patient and TRALI reaction histories were reported to the blood center by the hospital physician. Anti-HLA and anti-granulocyte testing of blood donors was performed on consenting donors potentially implicated by their history of pregnancy or transfusion. Donors with positive lab results were permanently deferred and the case considered confirmed for this report. Patient testing was not performed. Results:From Jan 2002 to Aug 2003, approximately 440,000 WB units were collected. 18 suspected TRALI reactions were reported to the blood center. HLA class I and/or II antibodies were detected in 5/17 (30%) of implicated donors. No granulocyte-specific antibodies were identified. All + HLA antibodies were in multiparous female donors. 1 donor refused testing. 9/18 reactions were associated with FFP alone or in combination with other components. 8/18 cases received 4 or more products in a 24hour period. There were no transfusion related fatalities among these reactions. Hospital physicians characterized TRALI reactions as follows: reaction type. The proportion of female donors was assessed for all components and for units implicated in transfusion reactions. Comparisons of risk were made using Odds Ratio (OR) statistics using SPSS software (SPSS Inc. Chicago, Ill.) . Results: 125,927 RBC and 27,058 SDP units were transfused in the study period. Of these 58,969 (46.8%) RBC and 8873 (32.8%) SDP units were from female donors. Hemolytic and septic reactions were too infrequent to allow further analysis. The relative likelihood of adverse reactions implicating a female versus a male donor (where OR = 1 implies male risk = female risk) were as shown. Male donors were slightly more frequently implicated in RBC "Other" reactions. 36 SDP donors (27 male, 9 female) were identified that were involved in more than one (2-3) reaction over the 5-year period. Discussion: Female donors were not more frequently implicated than male donors in mild or moderate transfusion reactions or repeat reactions in our institution, suggesting that alloantibodies are not a common cause of such events. * One with hypertension.All TRALI cases involved dyspnea. 5 confirmed TRALI reactions reported other symptoms: 3/5 cases had fever (v. 5/12 unconfirmed cases), 3/5 had chills (v. 2/12), 4/5 had hypoxemia (v. 5/12), 3/5 had CXR findings (v. 2/12) and 3/5 occurred in <6 hrs (v. 5/12).

Conclusions: Approximately a third of clinically diagnosed TRALI cases were confirmed over a 20 months. Information on suspected TRALI cases from hospitals was often incomplete. Lack of a uniform approach to diagnosis in the community (i.e. reliance on dyspnea alone) led to over-diagnosis and over-reporting. Other diagnoses may mimic TRALI, thus a standard protocol for reporting could lead to improved understanding of this adverse transfusion event. Background: Approximately 10-20% of female donors harbor alloantibodies against HLA and/or neutrophil-specific antigens that have been implicated in transfusion related acute lung injury (TRALI). Recent reports of lookback investigations into donors linked to fatal TRALI cases suggest that these antibodies may also cause mild or moderate adverse events interpreted as common febrile, allergic or atypical reactions. Given that transfusion reactions are reported following 0.1-1% of transfusions, we asked whether female donors potentially harboring alloantibodies were disproportionately involved? Methods: We performed a retrospective analysis of adverse reactions linked to single-unit red blood cell (RBC) or single donor apheresis platelet (SDP) transfusions performed at a major academic hospital over a five-year period. Reactions were investigated at the time of each report and classified as hemolytic, septic, febrile non-hemolytic, allergic or "probably not transfusion related (other)". A database was created of the RBC and SDP components supplied to the transfusion service over the corresponding time period from the local regional, and an in-house blood donor center, including the unit number, product type, donor gender and adverse Background: Transfusion-related acute lung injury (TRALI) can be characterized as an acute bilateral diffuse pulmonary process in the absence of right atrial hypertension, associated with significant hypoxia (e.g. O 2 saturation <90% on room air), that develops during, or within 6 hours following, transfusion. Until a diagnostic abnormality is identified, these clinical criteria can be used to identify high probability TRALI cases. Methods: All cases of potential TRALI received, investigated and closed in 2003 at a large blood system were reviewed by 3 physicians to identify those meeting the above clinical criteria. The results of the investigation and testing of the 120 donors involved in 36 high-probability TRALI cases were analyzed. In most investigations, transfused donors and parous female donors were tested for HLA class I, HLA class II and granulocyte-reactive antibodies by any of several methodologies. Recipient testing to confirm a crossmatch, or antigen-antibody, incompatibility was not performed. Results: Sixty donors were tested and 32 (53%) were positive for one or more leukocyte-reactive antibodies, all of whom were parous females. These donors were associated with 26 (72%) of the 36 cases:

Where performed and reported, most HLA class II (8/11), but not HLA class I (3/10), antibodies were broadly reactive (>70%); the specificities were varied. Of the 10 cases for which an antibody-positive donor was not located, 3 involved only FFP transfusion and 7 involved transfusion of red cells with a mean storage age of 17 days. Conclusions: Using specific clinical criteria to define high-probability TRALI cases, a parous female donor with leukocyte-reactive antibodies can be detected in a majority of cases. HLA class I and class II antibodies of a variety of specificities are most often detected; however, the class I antibodies tend to be less broadly reactive. When an antibody-positive donor is not located, the case often involves transfusion of stored red cells. Background: Transfusion-related acute lung injury (TRALI) can be a severe, life-threatening complication of transfusion. TRALI has been associated with blood components that contain antibodies directed against white cell antigens. The presence of antibody in the donor or recipient, corresponding to 1 or 2 cognate antigen(s) in the other member of the donor-recipient pair, has been demonstrated in many TRALI cases. We report a case of TRALI in which the donor possessed antibody directed against 3 cognate HLA antigens in the recipient. Case Report: An 82 year old female was admitted for lower extremity weakness and back pain. She underwent emergency L2 laminectomy and epidural biopsy for spinal abscess. All cultures were negative. She received a unit of apheresis platelets 14 days after surgery. After infusion of 150 mL, the patient became dyspneic, hypertensive and tachycardic. Oxygen saturation fell to 50%, prior to the patient becoming apneic. She was intubated and placed on mechanical ventilation. Laboratory investigation included HLA typing by serology (Class I) and PCR-SSP (Class II) and antibody screening with identification by flow cytometry (FlowPRA Screen/FlowPRA Specific Test, One Lambda, Canoga Park, CA). The donor possessed antibodies to HLA A3, B7, DR4, DR14 and DR9. The recipient had antibody to HLA B18. The recipient's HLA type was HLA A3; B7,41; DR4,11. Therefore, the implicated donor possessed antibodies directed against 3 antigens present in the recipient-HLA A3, B7 and DR4. Conclusion: TRALI can be associated with the presence of antibodies directed against multiple cognate antigens in the other member of the donor-recipient pair. A thorough laboratory investigation of TRALI should include testing of the donor and the recipient for concurrence of the implicated antigen(s)antibody(ies). Background: Monitoring of vital sign value (VSV) changes in patients undergoing blood component therapy is an integral part in the quality management of the transfusion process. Changes in VSV often serve as heralds to potential adverse outcomes experienced by patients. Little data exist, however, as to the magnitude of change necessary to prompt clinical consideration of such events. In this regard knowledge of what happens to VSV in patients undergoing uncomplicated transfusions (UCT) would be of interest. In a previous pilot study (Transfusion Vol. 42, Supplement S78-040B, 2002) we conducted a simple univariate analysis of VSV from a series of patients having UCT and observed minimal variation in VSV from baseline.

Here we extend those studies by expanding our cohort of patients, and analyze the data obtained using a more rigorous statistical approach. Methods: A retrospective chart review was performed as previously described. VSV analyzed included temperature (Temp, °C), pulse as beats per minute (Pulse, BMP), respiratory rate as respirations per minute (RR, RPM), systolic blood pressure (S-BP, mmHg) and diastolic blood pressure (D-BP, mmHg). Products examined included packed red blood cells (RBC), FFP, whole blood derived PLT (WBD PLT), and apheresis PLT (AP PLT). In contrast to our previous study, the cohort for this analysis was restricted to a single patient/single transfusion entry so that over-representational bias from inclusion of a particular single patient or group of patients in the data pool would not occur. Data were assembled into an Excel worksheet for further statistical analysis. Calculations for mean (m), SD, 95% CI and level of significance via a paired student's t-test were performed using Minitab 12.1. Values of p £ 0.05 were considered significant. Results: (see Table 1 ).

Conclusions: In this expanded cohort of UCT patients, little variation in vital sign values was observed. Although small statistically significant changes in the pulse and systolic blood pressure were identified in patients receiving RBC products, the magnitude of change did not appear to be clinically significant. Further studies comparing VSV changes with respect to various patient sub-populations, for example gender, age, cardiovascular disease, and patients experiencing suspected transfusion reactions, may be of interest.

A Ohsaka, K Abe, T Ohsawa, Juntendo University School of Medicine, Tokyo, Japan; N Ohbayashi, Olympus Systems, Tokyo, Japan Background: ABO-mismatched blood transfusions attributable to inadequate identification of the patient or the blood unit remain the most common serious hazard of transfusion. The final bedside check required to insure that the blood unit is intended for the patient is prerequisite to avoid incorrect transfusions. Methods: We have developed a network computerassisted blood transfusion management system connecting with a novel barcoded patient-blood unit identification system and automated blood testing instruments. The identification system is composed of 1) a hand-held computer that reads barcodes of the patient's wristband, the blood unit, the compatibility report form and the compatibility label attached to the blood pack, 2) barcode printers for patient's wristband, compatibility report form and compatibility label, and 3) a network computer-assisted blood transfusion management system capable of receiving patient's transfusion data from a bedside hand-held computer. To ensure that the blood unit is intended for the patient, we have performed verification procedures three times during transfusion process, i.e., the timing with the blood unit is 1) delivered from the transfusion laboratory, 2) received at operating rooms or inpatient wards, and 3) transfused at bedside. Results: This system allowed us to eliminate handwritten report forms and manual procedures in multiple steps of transfusion process and to monitor bedside verification procedures in real time at the transfusion laboratory. Following the introduction of the barcoded patient-blood unit identification system toward the all inpatient wards and operating rooms (July 2002), approximately 20,000 blood components have been transfused without mistransfusion. Conclusions: The network computer-assisted blood transfusion management system may be useful for avoidance of incorrect transfusion attributable to human errors and for improvement of transfusion safety, when the patient-blood unit identification system is connected. Further evaluation is needed to establish whether the barcoded patient-blood unit identification system could have a role in increasing transfusion safety. by comparing current typing results to a historical record; 12 (33%) by the service recognizing the error and alerting blood bank; 1 (3%) by blood bank staff recognizing a discrepancy other than blood type. Of the 23 MCS detected by discrepant typing results, 16 cases represented an incorrect initial type; 11 of these (69%) could have resulted in an ABO incompatible transfusion (1.1 transfusions per year). 9 of the 23 discrepant results were detected by CT; 2 (5.5%) represented an incorrect initial type and could have resulted in an ABO incompatible transfusion (0.2 transfusion per year). Of the remaining 7 cases detected by CT, the CT sample was the MCS. All MCS were attributable to personnel not adhering to patient identification protocols. Conclusions: In the absence of reliable electronic methods for bedside patient identification, avoidance of ABO incompatible transfusion is primarily dependent on the detection of MCS through comparison of a current blood type result with a previous type. Use of a CT, although also subject to error, is an effective way to create a historical type and an additional means of preventing ABO incompatible transfusions.

passive anti-A could be detected only in Case 2. Retrospective anti-A titers on the implicated group O PP at 22C and at the anti-human globulin (AHG) phase are shown below.

D Fauzie, R S Shirey, S Thoman, D Bensen-Kennedy, K E King, Johns Hopkins Hospital, Baltimore, MD Introduction: Anecdotal cases of severe HTR due to passively-acquired ABO antibodies in plateletpheresis products (PP) have been reported in nongroup O recipients of group O PP. Some have suggested that anti-A titers should be determined on group O PP, and that PP with high titer (>64) anti-A should not be transfused to group A recipients. We studied the rate of HTR in non-group O recipients of group O PP and the efficacy of anti-A titer determinations in preventing HTR. Materials and Methods: A retrospective study of all adult patients transfused with PP between Jan 2000 and Dec 2003 in a tertiary care hospital was conducted using computerized data to review ABO types of recipients, ABO types and volumes of PP transfused and all transfusion reactions (TRs) reported. Results: A total of 51,566 PP were transfused in the 4 year period. There was little or no risk of HTR due to passive ABO antibodies in 47,750 (92.6%) transfused PP, because 1) the PP were not group O PP and/or were ABO compatible with the recipient, or 2) the plasma was removed from the PP before transfusion, or 3) the recipient typed as group O at the time of PP transfusion due to group O BMT or group O RBC transfusions. There were 3,816 PP (7.4%) group O PP transfused to non-group O recipients with the potential to cause HTR due to passive ABO antibodies. The frequency of TRs between the 2 groups (i.e., PP without risk of HTR versus PP with risk of HTR) were: 795 TRs/47750 PP (1.66%) versus 99 TRs/3816 PP (2.59%). The 99 TRs in the group at risk for HTR due to passive ABO antibodies were evaluated as follows: 52 Allergic, 35 Febrile Non Hemolytic, 10 Atypical and 2 HTR. Both HTR cases involved group A recipients with hemoglobin decreases of 1 to 2 g/dL after group O PP transfusions. Both HTR had positive DATs (IgG and/or C3); Conclusions: We conclude that: 1) The rate of HTR in non-group O patients receiving group O PP is 2 : 3816 (0.05%) and 2) PP with low anti-A titers (<64) can cause HTR. Determination of anti-A titers in group O PP before transfusion might have prevented one of the HTR (Case 1) due to passive anti-A, but would not have prevented HTR in Case 2.

A Salam, M Hosain, K Sazama, B Lichtiger, A Narvios, M.D. Anderson Cancer Center, Houston, TX Background: Platelet transfusion is an important component of support for cancer patients with severe thrombocytopenia. Two preparations are available: whole blood derived platelets (WBDP) and apheresis platelets (AP). When selecting between WBDPs and APs, patient risks such as alloimmunization, infectious complications, and transfusion reactions, as well as management concerns including costs and availability, should be considered. Which kind of platelet (WBDP vs. AP) should be used for transfusion in cancer patients remains in dispute. This study analyzes the incidence, frequency and severity of transfusion reactions in cancer patients to determine if these data would provide a compelling reason for using APs (that are more expensive and difficult to obtain) over WBDPs (that are inexpensive and abundant). Methods: A retrospective analysis was done of all cancer patients who experienced adverse reactions during and immediately following infusion of both WBDP and AP in a 14-month period (August 2002 -September 2003 . Transfusion reactions were evaluated on the basis of age, sex, underlying diagnosis, number and types of reactions. Results: 141 adverse reactions to platelet infusions were analysed; 18 of 3969 transfusions of APs and 123 of 18,509 transfusions of WBDPs (in pools averaging 7 WBDP/dose). These 141 adverse reactions occurred during 22,478 transfusion events for an overall frequency of 1 in 159 infusions (0.6%). All patients were premedicated with Benadryl and Tylenol and/or Hydrocortisone. There were 68 males with a mean age of 43.9 years (range 7-77), and 73 females mean age of 48.4 (range 11-74), with no significant mean age difference (2 sided p = 0.14) in the overall reactions. Similarly, there was no significant difference in males and females with reactions who had received either WBDP (59 and 64) or AP (9 and 9). Reactions after AP were 1 in 220 transfusions and after WBDP pools were 1 in 150. Among 22,478 platelet transfusion events, WBDP (in pools averaging 7) were 1.5 times more likely to cause an adverse reaction than an AP, but this difference was not statistically significant with a 95% CI of 0.89-2.56 (p = 0.12). The most frequent reactions were considered minor such as hives and itching (81), chills (46), shortness of breath (27), and fever (18). There was no bacterial growth in all the samples submitted for culture. Conclusions: In this retrospective study, there was only a slightly higher, but not statistically significant, frequency of adverse reactions related to transfusion of pools of WBDP as compared to AP, with no differences noted in males vs. females of similar ages. More studies are needed to understand these findings.

J Schwartz, T Hamilton, P Cosgrove, D Strauss, K Grima, New York Blood Center, New York, NY Background: In Nov. 2003, we switched from single donor platelets (SDP) to random donor platelets (RDP) because of cost concerns. In March 2004, we began to test all RDP for bacteria using a urine reagent dipstick. We compared 6 month period reaction rates to platelet transfusions using mostly SDP to a 6 month period when we were using predominantly RDP. Methods: Reaction rates were compiled from the transfusion reaction reports that are returned to the Transfusion Service. Reaction rates from 11/02 through 4/03 (>90% SDP) were compared to reaction rates from 11/03 through 4/04 when predominately RDP were used. Bacterial testing was performed using Chemstrip 7 (Roche) urine reagent dipsticks. The method was validated using Multistix 10 SGG (Bayer) and the published procedure for RDP testing. Chemstrip 7 dipsticks were selected because it was easier to interpret the results. Acceptable products had to meet these criteria: pH > 7.0 and glucose > 250. All RDP were tested upon receipt in the Transfusion Service. Products that failed to meet criteria were returned to the issuing facility where they were sent for bacterial culture. Associated products that had not been distributed were quarantined until the culture results were known. Results: Bacterial testing: we had tested 4282 RDP in 3-4/04. 60 (1.4%) failed to meet criteria: 35 due to glucose, 7 due to pH and 18 because of both. Most of the failing products (36/60) were 5 days old of age. 51(89.5%) of the 57 products that were returned to the issuing center were culture negative; 6 are still pending. Reaction rates: between 11/02-4/03 we had 2492 platelet transfusions with 92.3% SDP. We had 17 (0.68%) reactions; 13 in SDP and 4 in RDP. Between 11/03-4/04 we had 2756 platelet transfusions with 73% RDP. We had 37 (1.34%) reactions, 35 of them in RDP. Most of the reactions in both groups consisted of fever and chills Comparing RDP reactions on monthly basis before and after bacterial testing revealed decrease in reaction rates from a range of 1.5-4.8% to a range of 0.72-0.88%. Conclusions: Reaction rates increase significantly when RDP are substituted for SDP despite the use of leukodepletion filters. Surrogate testing for bacterial contamination has poor specificity for bacterial contamination and results in excessive returns to the issuing center which must either culture these products to identify true contamination or recall all associated blood products which results in waste. A decreased reaction rate was an unexpected finding after bacterial testing was introduced. Eliminating these products from platelet pools may reduce the incidence of reactions because low pH and low glucose indirectly correlate with high levels of cytokines in the plasma of these products. Further studies are needed.

S M Fritzel, L Clough, M Wahabuddin, Y Kebede, C Spurlock, C R Santos, Inova Blood Donor Services, Annandale, VA Background: After Gambro BCT upgraded Trima® v. 4 to v. 5 Accel TM , several centers, including this one, noted several white blood cell (WBC) failures and notified FDA. In August, 2003, Gambro released a letter to customers advising all Trima® Accel TM v. 5.0 users to monitor the product WBC count to detect increased WBC content. According to this letter, the system is responding to specific physiological characteristics of individual donors, causing the leukocyte reduction system (LRS TM ) chamber to over-fill with WBCs that then spill into the platelet bag. The letter also states that a software patch will be available in the future. Trima® v. 5.01 patch installed in March 2004. Methods: Donors with WBC failures on Trima® v. 5.0 were asked to return and donate on Trima® v. 5.01 for comparison of products. Three donors who had repeated failures on v. 5.0 returned and donated on 5.01 at least once. The label advisory given by the system and the WBC count by Nageotte result for each donation were both recorded. Results: The three donors' last several consecutive donations are seen in the table.

In every case, Trima® v. 5.01 detected a condition that led to a "count WBC" product advisory message; v. 5.0 did not except in the case of spillover. Conclusions: Selected platelet donors' products are not leukoreduced by Trima® Accel TM v.5.0 and v.5.01. Trima® v.5.01 patch will advise operators to check WBC on most, if not all, of these products. These donors tend to have relatively large total blood volume (>5750 mL), from whom the most blood volume is processed and product volume removed. Background: The accuracy of an apheresis platelet count, (platelets/unit volume), and the timing of obtaining the sample for that count are important as it affects two products after splitting. This study was performed to determine whether sampling apheresis products at 24 hrs provided an accurate platelet count compared to sampling shortly after collection and if it accurately predicted the counts of each split product. Methods: Donors were selected based on their history of donating products that split. Collections were either from Amicus (n = 36) or Trima (n = 40). Platelet counts were obtained at 0-2 hrs (0 hr), 4-4.5 hrs (4 hr), 20-30 hrs (24 hr) and on both split products, if split. Ratios of 0/4 hr, 4/24 hr and 0/24 hr were calculated. A ratio of 1 indicates that the measurements are statistically the same. A ratio < 1 indicates that the numerator is less than the denominator and >1 the opposite. Results: There was no statistical difference between Amicus 0/4 and 4/24 hr and Trima 4/24 hr platelet count ratios. Since the Amicus and Trima 4 and 24 hr platelet counts were the same, we conclude that they are the "correct" counts. The ratios for the remaining combinations, Amicus 0/24 hr ratio (1.02) and Trima 0/24 hr and 0/4 hr ratios (1.05, 1.06) were significantly greater than 1. Although the split rate is reduced using the 24 hr count, decreasing from 92% to 83% for Amicus and from 90% to 75% for Trima, the number of non-standard split products with <3.0 ¥ 10 11 platelets is also reduced. For products that split based on the 0 hr count but would not have split if the 24 hr count had been used, 3/4 Amicus and 2/10 Trima splits contained <3.0 ¥ 10 11 platelets. Of the 8 remaining Trima products, 4 contained 3.0 ¥ 10 11 and 4 contained 3.1 ¥ 10 11 platelets. Platelet counts between the original unit and the splits were considered to be equivalent if the ratios were within 10% of the perfect value of 1. With 95% statistical confidence, all ratios were within10% when the 0 hr and 24 hr counts were compared to the split counts. The 24 hr counts were slightly closer to the split counts for Amicus (3-5% difference) and much closer for Trima (-0.3-1% difference) than the 0 hr counts (5-7% difference for both technologies). Conclusions: The 4 and 24 hr platelet counts appear to be "correct" for both Amicus and Trima. The 0 hr samples exhibited a higher platelet count than the 24 hr, particularly for Trima and to a lesser extent for Amicus. Although there was less than a 10% difference between the counts in the splits compared to the original unit at 0 and 24 hrs, the 24 hr counts were closer to the splits.

S M Fritzel, L Clough, Inova Blood Donor Services, Annandale, VA Background: Donor center upgraded Trima® v. 4 to v. 5 Accel TM . During licensure validation, single donor platelet components (SDP) with white blood cell (WBC) contamination were observed. Methods: The distributions of four collection parameters from SDP draws resulting in WBC >1 ¥ 10 6 by Nageotte (total 38) were analyzed retrospectively: instrument ID, donor ID, donor total blood volume (TBV), and product volume. The WBC count by TRANSFUSION 2004-Vol. 44, Supplement Coulter® Ac·T diff from the product 1 : 5 dilution sample was considered as well. All procedures performed using Trima® v. 5.0. For each procedure, the product advisory message read either label WBC < 1 ¥ 10 6 (82%) or had WBC spillover characteristic of an leukocyte reduction system (LRS®) chamber saturation event (18%). Results: There was an average of 5 collections per month failing leukoreduction below 1 ¥ 10 6 in a six month period. 20 different donors gave the 38 donations; 8 of the donors gave more than once. The events were evenly distributed among the Trima® instruments. The greater the donor TBV, the greater number of events seen, until the final TBV category representing an exceptionally large individual rarely seen in the apheresis donor pool (Table 1) . Double product splits represented the majority of events seen (Table 1) , followed by triples, then singles. In 24% of the donations, the WBC count was not higher than the background count. Conclusions: WBC contamination of SDP product from LRS chamber saturation is not correlated to season or individual instrument. WBC count of the product dilution sample is not an accurate measure of product contamination, but is helpful in the absence of other methods. Greater donor TBV, as well as other not yet determined physiological factors, and greater blood volume processed are probably contributing to these saturation events. White blood cell saturation of the LRS® chamber of the Trima® Accel TM v. 5 disposable set causing WBC contamination of the SDP product is an unexpected phenomenon of this upgrade. In many cases the contamination is not detected by either the instrument software programming or by the makeshift monitor of the WBC count from the product 1 : 5 dilution sample. In the era of 100% SDP bacterial detection by O 2 use testing, the market cannot bear the liability cost of non-leukoreduced SDP.

Procedure time: 23 minutes average Donors reacted positively to the short procedure time and tolerance of the procedure was very good. All donors would come back and would like the idea of donating the same amount of red cells by only 2 yearly visits instead of 4. Conclusion: The overall quality control of the red cell concentrates met the European and Swiss Guidelines. Since the acceptance of the procedure by the donors was very good, we implemented the double red cell procedure using ALYX.

J Lin, C Tzeng, L Liu, M Yi, J Lyou, C Yung, Transfusion Medicine, Taipei Veterans General Hospital, National Yang-Ming University, Taipei, Taiwan Republic of China Background: The default collection rate of granulocyte component is 3 mL per minute for COBE Spectra blood cell separator. If we processed 7 L at blood flow rate of 50 mL/minutes during leukapheresis, the volume of granulocyte component will be 420 mL. For the safety of a donor with total blood volume around 4500 mL who donated 3 consecutive days, we should decreased the collection rate during relative large volume leukapheresis or decrease the volume of processed blood. In this study, we evaluate the effect of collection rate on the yield of granulocyte collection. Methods: We retrospectively analyzed factors predicting the yields of granulocyte apheresis in 147 aphereses from 60 healthy donors (45 males and 15 females). Donors received 300 mg of granulocyte-colony-stimulating factor (G-CSF) subcutaneously and 10 mg of dexamethasone intramuscularly 12 hours before leukapheresis. Leukapheresis was performed on the COBE Spectra blood cell separator. Data on granulocyte collections were divided into 2 group on the basis of the collection rate for the apheresis procedure: Group I: collection rate of 1-2 mL/minute, Group II: collection rate of 2-3 mL/min. Red blood cell sedimentation was facilitated by hydroxyethyl starch (pentastarch); pentastarch: blood = 1 : 13 vol/vol. Results: Median age, body weight and total blood volume of donors were 36.5 years, (range 13-57), 67 kg (range 45-110), and 4503 mL (range: 2998-6470, 95% CI: 4297-4526), respectively. For each collection, 3.2 to 7 L of anticoagulated blood were processed at a flow-rate of 45-55 mL/min. The yields of leukaphereses per L of anticoagulated blood processed were shown in Table 1 . On multivariate analysis, variables that positively influence granulocyte yields were pre-leukapheresis white cell counts (odds ratio(OR): 2.34, 95% CI: 1.08-5.07; p = 0.031), donor's total blood volume (OR: 5.35, 95% CI: 2.40-11.92; p < 0.001). Variable that had a significant negative impact on the granulocyte yield was collection rate of 1-2 mL/minute (OR: 4.69, 95% CI: 1.94-11.32; p = 0.001). Conclusion: Granulocyte apheresis at a collection rate of 1-2 mL/minute rather than collection rate of 2-3 mL/minute results in a significant decrease in the granulocyte collection efficiency and granulocyte yield. Background: Donors are less and less willing to donate whole blood 4 times a year. A procedure allowing one to collect the same amount of red cells in only two sessions could be a great advantage for the donor, as well as for the blood center. After succesful introduction of the double red cell procedure of Trima (Gambro Inc.), we validated the double red cell collection (DRBC) system using the procedure of ALYX (Baxter, Transfusion Therapies). Methods: 100 donors (regular whole blood donors and apheresis donors) donated 2 units of red cells (RBCs) using the ALYX system according to the manufacturer's instructions. RBCs were leucoreduced automatically during the procedure. Data were collected on procedure time, donor acceptance and quality control of red cells [leucocyte count, hemoglobin (Hb), hematocrit (Hct)] according to the routinely performed quality control parameters. In a subset of 30 units the following storage parameters were analysed: Glucose Background: Transfusion medicine continues to struggle against blood supply shortages while donor pool suffering a constraint all over the world. Automated collection of 2 red blood cell units (2¥RBC) promises to alleviate this issue. Unfortunately, due to regulation restrictions on weight, height, Hct and Hb, not every donor eligible for whole blood donation meets the requirements for 2¥RBC donation. We studied donor eligibility rates and willingness for automated 2¥RBC collection in 2 populations under European and American guidelines. Methods: Two populations studied: A group with 606 donors living at ≥1,000 m in Tyrol, Austria (T) and another group with 571 donors living at the sea level in Mallorca, Spain (M). Donors filled a questionnaire about their willingness to donate 2¥RBC. A hemogram was performed on each donor (Coulter STKS). Donation records were also reviewed. Eligibility for 2¥RBC automated collection on both populations according to CE guideline: Both sexes >70 kg, >42% Hct and >14 g/dL Hb and AABB standards: >40% Hct, ≥167.6 cm for women and ≥155.4 cm for men, and weight depending on sex and RBC volume donated (women ≥68 kg for 360 mL and ≥79.5 kg for 400 mL; and men ≥59 kg for 360 mL, ≥68 kg for 400 mL and ≥79.5 kg for 420 mL) was evaluated. Data collected were analyzed using SPSS statistics program. Results: No difference on weight, height and age between groups. The 36% donors were eligible for 2 ¥ RBC under the CE guidelines and a range of 29% to 60% was observed for AABB standards. Significant fewer women met both guidelines and fewer donors where eligible at the sea level (1% women eligible in M versus 6% in T; 49% men eligible in M versus 72% in T, p < 0.0001). Tyrol donors previous donation history was 1.5 units lower (9.7) than M donors (11.2). The 60% of all donors were willing to donate 2¥RBC, and 25% of these donors met both CE and AABB requirements. Most donors willing to donate 2¥RBC had previously been only whole blood donors (≥75%), and the remaining were new donors (≥15%) and apheresis donors (£10%). Conclusion: Blood Banks could take advantage of the 25% donors meeting the current requirements for 2¥RBC dose, and willing to do it. Automated collection promotion strategies could increase donor willingness up to 36% simultaneously eligible for CE standards. Less frequent donors should be involved in 2xRBC as they show higher eligibility rates and collection is more efficient. Donor inexperience with automated collection systems does not appear to involve a restrain to double RBC automated donation in whole blood donors.

N M Meyer, American Red Cross, Washington, DC; L Kline, American Red Cross, Rockville, MD; C Brisson, Baxter Healthcare Corporation, Round Lake, IL ALYX Abstract BACKGROUND: Blood banks today are faced with increasing manufacturing costs, accelerated regulatory pressures, and a shrinking donor base. One way to meet these demands can be the use of an efficient, low cost, lightweight, multiple component collection system that could be optimized for use in both fixed site and mobile collection settings. Such a device could help optimize blood inventory levels and reduce costs while increasing the availability of collected blood units. The ALYX Automated Blood Collection System (Baxter Healthcare Corporation) is a single arm apheresis system that uses centrifugation to separate the red blood cells (RBC) from the plasma, add the additive solution and pump them through a leukoreduction filter to produce 2 leukoreduced RBC. The plasma is returned to the donor with sufficient 0.9% saline to compensate for volume loss. STUDY DESIGN: Three blood centers participated in an operational trial to evaluate ALYX. A 3 day training class was conducted by Baxter which included collecting RBC from donors. Operational data collected included; donor and staff satisfaction, effectiveness of recruitment materials, ease of instrument transport, procedure and leukoreduction time, procedures failures, donor reaction type and rate. Product quality data collected included; red cell recovery, residual WBCs, and RBC volume. RESULTS: A total of 371 procedures were performed (92% male and 8% female). The mean (±1SD) procedure time was 26 ± 4 minutes and filtration time 6.90 ± 0.57 minutes. There were 14 (3.8%) donor reactions, all light, of which only 3 (0.8%) resulted in discontinuation of the procedure. A total of 31 (8.1%) procedures could not be completed with 22 (5.9%) of them due to poor venous access. The absolute RBC volume was 160 ± 10 mL, RBC product volume 300 ± 13 mL, total Hgb 55.6 ± 3.6 g/unit, %RBC recovery 89 ± 2% and WBC <3.5-9 ¥ 10 5 /unit. Ease of transport using the transport cart was rated positive to very positive by 100% of staff, recruitment materials rated effective to very effective by 75% of staff and 78% (8% did not respond) rated their overall experience as positive to very positive. Donors rated their experience as positive to very positive. CONCLUSIONS: Based on the analysis of the operational and product data and operator and donor comments the results are favorable to support implementation of the ALYX Blood Collection System at both fixed and mobile operations. Background: Our Baxter Amicus separator for platelet collection has been upgraded to Amicus Crescendo with software (SW) version 2.5. After validation, performance was compared to the previous SW version 2.4. Study Design: Platelet routine collections were performed with 3 instruments (AD1, AD2, AD3) in 2 periods (SW 2.4: from Nov. 2000 to Nov. 2001, n = 106; SW 2.5: from June 2002 to April 2003, n = 106). 14 donors were involved in both periods. AD1 collected 33 units, AD2 35 and AD3 38 in each period. In contrast to SW 2.4, platelet collection with SW 2.5 also involved plasma collection (58.5% of procedures). The following parameters were evaluated: procedure time, whole blood volume processed, product volume, platelet concentration and actual platelet yield. The effect of the factors donor, target yield, equipment and software version on the previous parameters was assessed by SPSS statistical program analysis. Results: A significant effect of all factors over most parameters was observed. Using SW2.5, procedure time was reduced by 13.2% without plasma collection and even 20.9% with plasma collection compared to SW 2.4 (p < 0.0001). Platelet yield/min increased by 17.6% with SW 2.5 and plasma collection. Significantly less whole blood volume was processed with SW 2.5 (3,567 ± 610 ml vs 4.125 ± 563 ml). Platelet yield was 0.33 ¥ 10 11 lower with SW 2.5 (p = 0.01) and deviation from target with 2.4% was much lower than for SW 2.4 (25.5%; used estimated marginal means). * 1 = 4 ¥ 10 9 Plts/min; 1 p < 0.0008 target vs actual yield; 2 p > 0.25 target vs actual yield; 3 p < 0.0001 SW 2.4 vs SW 2.5.

Conclusion: Performance of the Amicus Crescendo separator with software 2.5 was superior to SW 2.4. This was demonstrated by a significant higher accuracy of the actual vs target yield. Concurrent plasma collection with the Amicus Crescendo separator may have contributed to the increased relative efficiency which resulted in a decreased procedure time. Background: When studying rare diseases, such as thrombotic thrombocytopenic purpura (TTP), a multicenter design is essential to identify a large cohort. However to date, experience with prospective multicenter TTP studies is limited. Logistical problems inherent in multi-center research are compounded by institutional differences, HIPAA, and variable administrative regulations among sites. To facilitate coordination of a multi-site NIH funded epidemiologic, clinical, and laboratory-based study of TTP, a four part centralized operational design was implemented and its progress is reviewed herein. Methods: The Study of the Surveillance Epidemiology and Risk Factors of TTP (SERF TTP) began in Aug 2002 with a goal of evaluating 300 newly diagnosed TTP patients and 600 matched controls in comprehensive clinical, epidemiologic, and laboratory based evaluations. Centralized coordination was organized into four core areas: epidemiology/data management, laboratory/genetic testing, clinical sites/plasmapheresis centers, and control screening/recruitment. Herein, we review the complex institutional approval process, which has occurred at 19 therapeutic apheresis sites. Data included in this preliminary report include: elapsed time was determined for IRB approval with HIPAA authorization, and subcontract completion; other review processes required by specific institutions, the frequency of IRB modifications, the number of sites actively enrolling participants, and a review of enrolled patient demographics. Results: Sixteen (84%) participating sites are active, one declined participation due to complexities in HIPAA compliance and two chose not to continue. Ten sites (63%) have IRB approval, 2 sites had multiple IRBs, 3 (19%) are pending and 3 (19%) have not completed IRB submission. Mean elapsed time for IRB approval, including modifications required for 50% of initial reviews, was 12 weeks (range 1.5 to 28 wks). Mean time for completion of subcontract agreements was 15 weeks (range: 6 to 24 wks). Five sites are now actively enrolling patients. To date, 15 patients (80% female) with an initial diagnosis of TTP have been enrolled. Mean age 38.6 years (range 20-61). Conclusion: Centralized coordination can facilitate establishment of a multicenter prospective study of the rare illness, TTP. For the SERF TTP Study, this coordination has facilitated the collaboration of heterogeneous group of study sites, while providing the flexibility to deal with local IRB, PHI, and administrative issues. With sites actively enrolling patients, the primary scientific goals of the research can begin to be realized. Background Dendritic cells (DCs) are the most potent antigen-presenting cells of the immune system and therefore increasingly used in adoptive tumor therapy. DCs can be generated from peripheral monocytes usually collected by mononuclear cell (MNC) apheresis. We investigated the quality of autologous MNCs collected with either the Amicus device (Baxter), or AS.TEC device (Fresenius) assessed by the differentiation of monocytes into DCs. Patients and Methods Patients suffering from malignancies (renal cell cancer = 6, melanoma stage III/IV = 34 and others = 10) underwent leukocyte apheresis either with the Amicus (n = 27, processed blood volume was 10.236 L ± 895) or the AS.TEC (n = 23, processed blood volume was 9953 L ± 1522.). MNCs of the products were isolated by density gradient centrifugation, subsequently washed and cultured according to standard protocols for 6 days. Thereafter DCs were pulsed with tumor antigen and cultured for another 3 days in medium containing proinflammatory cytokines which induces DC maturation. Mature DCs were characterized by FACS analysis. All blood cell and differential counts were performed with the Sysmex F-820 device (Kobe). For statistical analysis the Mann-Whitney U test was applied. Results No differences were found in the monocyte content of either apheresis product (p = 0.07) and the overall yield of MNCs (p = 0.4). The total amount of cells harvested at day 6 of culture was 567 ¥ 10 Ÿ 6 ± 291 for Amicus vs 517 ¥ 10 Ÿ 6 ± 323 for AS.TEC, respectively, (p = 0.5). Also the final yield of mature DCs (evaluable of 25 Amicus collections and 11 AS.TEC collections) revealed no significant differences: AMICUS: 138 ¥ 10 Ÿ 6 cells vs AS.TEC 185 ¥ 10 Ÿ 6 cells (p = 0.5). The yield of mature DCs in relation to the number of originally harvested monocytes was 7,2% in case of Amicus and 8.4% when the AS.TEC device was used. Conclusion Amicus and AS.TEC are equivalent in providing mononuclear cells for the generation of DCs.

L R Kagen, S J Graminske, R C Cushing, M A Gavin, J W Adamson, Blood Center of Southeastern Wisconsin, Milwaukee, WI; N Rugg, T Greenwalt, Hoxworth Blood Center, Cincinnati, OH; M Garcia-Lien, M McAteer, Gambro BCT, Lakewood, CO BACKGROUND: Freezing blood for long term storage is performed to secure rare type donors, military inventory and for special needs patients. Specific anticoagulant and preservative solution combinations for apheresis red blood cell (RBC) collection have been introduced to assist Blood Centers in their role in supporting patients with frozen components. This study compared in vitro results of Trima Accel (Gambro) apheresis (Test) RBC units collected in ACD-A/AS-3 to RBC units (Control) derived from whole blood (WB) collected in CPD/Adsol or Optisol after freezing and thawing. METHODS: Forty-one un-paired normal donors were randomized to donate a Trima single or double RBC unit or WB unit at two investigational sites. This study evaluated the in vitro characteristics of non-leukoreduced (NLR) Trima RBC units and RBCs derived from WB units. All units were kept at room temperature for no longer than 8 hours after collection. AS-3 was added to Trima units and AS-3 or Adsol was added to the WB units. Units were stored at 2-6°C for six days and then frozen according to the Meryman -65°C method. The units were held frozen for ≥28 days then thawed and deglycerolized using the Cobe 2991 Cell Processor, resuspended in saline and returned to 2-6°C storage for one day. On Days 0, 6 (pre-freeze), and one day post thaw, samples were removed for in vitro assays. Units were evaluated for volume, count, Hct, Hb, pH, pO 2 , pCO 2 , hemolysis, Na + , K + , glucose, osmotic fragility, and ATP at all sampling periods. RESULTS: Trima and WB units had differences in pH, ATP concentration and glucose concentration after frozen storage for at least 28 days. There were no differences between Trima and WB RBC units in volume, RBC count, hemolysis, osmotic fragility, supernatant Na + , and supernatant K + after thaw.

CONCLUSION: The data suggests that the NLR ACD-A/AS-3 RBC units frozen (on Day 6) and stored frozen for a minimum of 28 days are resilient to the freeze/thaw process and show acceptable in vitro results.

Background: Multiple sclerosis (MS) patients often seek alternative therapy for acute exacerbation after unsatisfactory results with medications such as steroids, IVIg, and immunosuppressive agents. The benefit of therapeutic plasmapheresis (TP) for MS has not been well established. Methods: We reviewed the medical records of 18 MS patients who received a total of 165 TP procedures at SLRHC from February 2001 through April 2004. Results: MS patients were referred for TP after developing acute exacerbation unresponsive to other medical treatment. One patient, intolerant to steroids and IVIg, had received prior TP at another institution. The patients' ages ranged from 32-66 years and 15 were female. The time from MS diagnosis to first presentation to our service ranged from 1 month to 26 years. Presenting symptoms included lower extremity weakness (16 patients), gait disturbance and imbalance (11 patients), upper extremity weakness (9 patients), spasticity (7 patients), visual impairment (5 patients), parasthesias (4 patients), back pain (2 patients), neck pain (2 patients), and generalized pain (2 patients). TP's were performed using a COBE Spectra continuous apheresis device via an internal jugular or femoral catheter, and treatment con-41A 2004-Vol. 44, Supplement sisted of 5-7 single-volume TP's (mean replacement volume = 2739 mL) with 5% albumin and saline replacement over 10-14 days, except in 3 patients in which treatment was discontinued after 2-4 TP's. One patient received 18 TP courses (total 93 TP's); all others received 1 TP course. Subjectively, 2 patients reported complete resolution of symptoms, 9 reported partial response to some or all symptoms, and 6 reported no response. Objectively, neurology and rehab medicine services described 2 complete responders, 10 partial responders, and 5 with no response. Response in 1 patient who received only 2 TP's could not be adequately evaluated. One patient responded repeatedly to 18 TP courses. There were no significant differences between responders and nonresponders in terms of age, symptoms, and time from diagnosis to first TP. TP complications in this group ranged from mild to moderate and included hypocalcemia requiring IV calcium gluconate (5 patients), catheter sepsis (2 patients), hypofibrinogenemia requiring FFP (2 patients), urticaria (2 patients), air embolism during catheter removal (1 patient), and symptomatic anemia (1 patient). Conclusion: TP is at least partially effective for many patients with MS exacerbation unresponsive to other therapy and is generally well tolerated. However, it is currently difficult to predict which MS patients will respond. Repeated TP treatments may be of benefit for some MS patients. Background: The preparation of plasma in parallel to platelets by cell separator donations should be possible. The donors would not suffer from frequently drawn additional blood volumes and donor laboratory testing had to be performed only once. We investigated whether this approach could serve as an additional source for fresh frozen plasma. Methods: Fresh frozen plasma was prepared over a period of about 2.5 years from 70 plateletpheresis donations using four Baxter Amicus Crescendo cell separators with software 2.5. and 27 donations with Gambro BCT Trima-System separators. Plasma volume was limited to a maximum of 220 ml in order not to exceed the limit of 650 ml total volume. Remaining platelets, leukocytes and RBC were determined and coaglutation parameters were surveyed. Results: The mean plasma volume obtained by the Amicus Crescendo separator was 183.6 ml ± 14.1 (129-218 ml) including two samples which did not reach the cut off of 160 ml, while the mean plasma volume of the Trima separator was 205.2 ml ± 11.6 (179-221 ml). Rest leukocytes were £1/ml and RBC were 5.6 ¥ 10 6 /L (mean) in each of the 97 plasma collections. The mean rest platelet concentration of 5.5 ¥ 10 9 /L (2-22 ¥ 10 9 /L) was under the limit set at 20 ¥ 10 9 /L in all but one separation. 12 plasmas, 6 from Amicus Crescendo and 6 from Trima separators, tested for coagulation parameters within shelf life surveys up to now over one year in three months intervals, were within permitted limits. As a positive side effect the total time per donation reduced from 58.7 ± 12 min to 53.5 ± 9 min when plasma was collected in addition to platelets with the Amicus instruments. Collection time was not evaluated for the Trima separators. Conclusions: The simultaneous isolation of platelets and plasma by cell separator donations is a promising trial to save costs, time and human resources. While the quality of the plasma is excellent, the varying volumes of plasma need to be corrected by process optimisation. This aspect is under further investigation.

Background: To enhance a large multi-center system's use of automated collection equipment, the Alyx Component Collection System (Alyx) was piloted and validated in 2004. The Alyx system is used to collect two units of leukoreduced red blood cells. The leukoreduction is integrated in the collection process; therefore further manipulation of the component in the lab is reduced. Study Design: The evaluation was conducted concurrently at two sites within the system. The first site was used for completion of training and collections conducted in a fixed location. The second site conducted collections in a mobile environment. Product quality control testing was completed at the main facility's laboratory. Acceptable product quality control results were used to determine pilot success. Collections were conducted using vendor approved collection nomagram. Blood center staff, donor, and center customer surveys were also conducted. Results: Fixed Site Collections: 102 units tested Background: Pregnancy has been associated with thrombotic thrombocytopenic pupura (TTP), which may be difficult to distinguish from HELLP syndrome or idiopathic thrombocytopenic purpura (ITP). Early diagnosis of TTP and treatment with therapeutic plasma exchange (TPE) has reduced the rate of mortality. Obstetric patients may exhibit any of the adverse reactions associated with TPE. This obstetrical patient exhibited severe allergic reactions during TPE for recurrent, pregnancy-associated TTP. Case Report: The patient was a 16 yr-old G 1 P 0 woman with persistent headaches and hypertension at 25 weeks gestation. Lab values were: Hgb 4, plts 15 K, LDH 8678, BUN 45, Cr 2.7. Caesarian section was performed at 35 weeks for fetal cardiac decelerations. Initial diagnosis was HELLP syndrome and she received RBCs and plts. There were schistocytes on peripheral smear, hematuria, and icterus. Daily TPE was initiated for TTP. On day 7, she had improved and was discharged. Eighteen days later, she presented with plts 87 K, LD 3944, and schistocytes. She received 8 daily TPE, tapered to QOD on an outpatient basis. For hives and itching, she was pretreated with diphenhydramine, solumedrol, and H2-blockers. Eosinophils were elevated at 10.5%. She received tapered dose prednisone. She improved with vincristine and did not require additional treatment. Four years later, at 28 weeks gestation, she presented with recurrent TTP and received 37 TPE in 4 months. TPE was complicated by severe allergic reactions manifested by confluent, erythematous rash; scleral edema; and dry cough. Immunoglobulin A levels were normal. Eosinophils were 10.3%. She required pretreatment with solumedrol, diphenhydramine, and H2-blockers. On one occasion, she received albuterol. She was treated with TPE, vincristine, and rituximab. She improved and is currently in remission with plts 280 K one month after her last TPE. Conclusions: We have described a patient with recurrent, pregnancy-associated TTP. She required several cycles of TPE after each delivery. Her second pregnancy was associated with recurrent, refractory TTP complicated by severe allergic reactions during TPE. She required solumedrol, diphenhydramine, H2-blockers, and albuterol. Recurrent TTP may be difficult to treat. Allergic reactions associated with TPE may increase in severity requiring treatment with multiple agents. It is important in patients with repeated allergic reactions to rule out the possibility of Immunoglobulin A deficiency. A multi-disciplinary approach is advised to prevent complications of allergic reaction and to implement treatment with chemotherapeutic agents to further reduce the risk of anaphylactic shock in these patients. Background: Granulocyte collections were usually performed using 30 mL of 46.7 percent trisodium citrate added to 500 mL of hydroxyethyl starch (Pentastarch). When 46.7 percent trisodium citrate was not available, we evaluated the efficacy and safety of the use of heparin as anticoagulant for collecting granulocytes. Methods: We prospectively studied 102 consecutive granulocyte collections in healthy 42 blood donors (male/female: 33/9, mean age: 34 years, age range: 18-57 years). Seven donors donated once, 10 twice, and 25 three times. Donors received 300 mg of granulocyte-colonystimulating factor (G-CSF) subcutaneously and 10 mg of dexamethasone intramuscularly 12 hours before leukapheresis. All leukapheresis procedures were performed using the COBE Spectra blood cell separator. Sedimenting agent was prepared by adding 5,000 units of heparin to a single 500 mL bottle of 6% pentastarch. Blood to pentastarch ratio was 13:1. Approximately 7 L of whole blood was processed during each procedure. Results: After mobilization with G-CSF and dexamethasone, the median peripheral white blood cell counts of the donors were 34,650/mL (95% confidence interval: 33,644/mL-37,178/mL). Yields per leukapheresis are presented as mean ± standard deviation (SD) in Table 1 . No donor experienced bleeding complication. Each granulocyte concentrate contained 33.4 ± 4.4 mL of pentastarch and 334 ± 44 units of heparin. No blood clot was found in the granulocyte concentrates. No bleeding complication was evident for recipients of these granulocyte concentrates. TABLE 1. Total cells and volume per granulocyte concentrate All the units produced had a content of leukocytes <1 ¥ 10 6 . Conclusions: Trima Accel allows to improve the efficiency of PP procedures as it has been shown through the data of our experience, which point out: 1. Reduction of the volume processed. 2. Reduction of the procedure time length with consequent reduction of the anticoagulant given to, and a higher confort for, the donor. 3. Same yield of PLT. 4. Leukoreduction of all units according to the standards in force.

The Impact of a 25 Apheresis Technology Substitution on a Regional Blood Center

The impact of a major apheresis technology change from 25 COBE® Spectra to Gambro BCT Trima® Accel TM machines was evaluated for purposes of platelet production over a three month period. Methods: The indices examined included collection times, split rate, number of imports, frequency of donation, and donor satisfaction. This data reflects the first full quarter of experience following completion of the conversion. Donor satisfaction was assessed via a survey of repeat donors (n = 238) immediately following an apheresis procedure on the Trima. The donation frequency compared first quarter 2004 to the same quarter of the previous year.

Conclusion: Heparin added in pentastarch is an efficacious and safe agent for centrifugation leukapheresis.

Productive Plateletpheresis: Trima Accel Vs Trima G Liumbruno, P E Centoni, S Ceretelli, R Ristori, P Molfettini, S Cremoni, M L Sodini, Apheresis Dep-Immunohematology & Transfusion Service, Livorno, Italy Background: Trima Accel (V 5) cell separator represents an evolution in the technology of productive plateletpheresis (PP), in comparison to the previous Trima (V 4.02) version; in fact it has been equipped with a single stage separation chamber joined to the LRS leukodepletion technology and it has a new software. All that aims to raise the collection efficiency of PP, yielding a superior quantity of platelets (PLT) in less time, or the same quantity at a lower speed in the same time space. Methods: We retrospectively analyzed data of PP procedures carried out in the last 6 months, in order to select a cohort of donors who had performed PP with both versions of Trima and for whom we had scheduled, in both cases, a procedure with the same forecast of PLT yield (4.5 ¥ 10 11 ); this with the aim to annull the variables linked to donor total blood volume, PLT pre-count, etc.. Results: 25 male donors cohort has been selected. In the following table, the results of the analysis we carried out, are beeing shown.

Split rates continue to trend upward each month from a baseline of 28% to a current 42%. Survey results indicated that 86.1% (n = 205) of the donors would be willing to donate at a frequency of once a month or more as a result of a positive experience on the Trima® Accel TM . Conclusion: The financial impact of this technology change includes a significant increase in cost of Trima® Accel TM disposables compared to the Spectra. However, these expenses are nearly, but not completely, offset by the increased number of split products, decreased number of imports, and decreased technical time. While the frequency of donation has remained static in the first quarter, the donor satisfaction survey and decreased donation time suggest that donation frequency will increase over time. Therefore, the full financial impact has not been entirely realized as a result of continued positive trending, limited number of applications validated, and relative immaturity of this technology to our center.

Southampton Apheresis Unit is one of the smallest units in NBS running 4 cell separators with an annual target (04/05) of collecting 4778 "adult theraputic units" of platelets in 2707 procedures. In June '02 the unit was reequipped with GAMBRO Trima version 4 replacing the GAMBRO Spectra machines which were used previously. Although TRIMA 4 introduced touch screen controls and was 10 minutes faster than Spectra per run, it also brought inherent problems. The move from dual to single arm collection meant we started experiencing haematoma's caused by the high-pressure returns. The needle set was now vital as we were unable to 'rinseback' via the flexi-cannula which meant, following a failed venepuncture, a donor was deferred for 8 weeks. Also as the return was 'pulsed' every 15 seconds, it meant that concentrations of ACD-A used for anti coagulant were higher which increased Citrate toxicity. "Accel" was in use in Dublin and the author first went on a fact finding visit. With the data collected, a local "validation team", comprising of the Unit Nursing staff together with Quality and Processing Managers was set up. The plan was to upgrade one TRIMA 4 to TRIMA "Accel. This was carried out 29/08 03 and on 02/09/03 we commenced the validation with GAMBRO providing training and on site support. Phase 1 of the validation required 3 runs per day to a total of 125 runs, and 43A 2004-Vol. 44, Supplement then, following a successful completion of Phase 1, we proceeded with Phase 2, converting the remaining machines individually conducting 25 runs per machine as a quality check between each conversion. The validation results! Between Sept and Dec 03 TRIMA "Accel" produced an average yield per donation of 6.19 ¥ 10 9 against a previous high of 5.99 ¥ 10 9 . Our double doses dropped fro 121 to 99 however our triple doses output increased from 21 to 55 over the same time scale and the average time per run reduced from 82.62 to 69.69 minutes. Other major advantages were "Accel" has a variable pressure sensor so that high-pressure returns were made more "gentle" reducing the incidence of haematoma. In conclusion, TRIMA "Accel" has increased our product yield and saved 20 minutes per donation. Machine settings can now be changed immediately without incurring any time penalties, Platelet aggregation is seldom a problem and the amount of Citrate used is reduced so less toxicity problems. The donors like the new machines because of the timesaving and as the machines are more effective they allowed more donors to be able to donate platelets. It is interesting to note that the clinic went from running at approximetly 80% of its target, to over 110% in 6 months after introduction of "Accel".

The Use of Double Erythrocytapheresis for Autologous Donation: Initial Results from an Ongoing Study E Rombout-Sestrienkova, P Teeuwen, A Nellisen, E Reuser, E Francken, P Van Noord, Sanquin Blood Bank Southeast Region, Maastricht, Netherlands Introduction: To explore the possibility of a more efficient type of autologous blood collection (compared to WB donation) and to moni-tor the possible level of stress of this type of a procedure, consecu-tive patients referred for autologous donations were invited to under-go double erythrocytapheresis with a full compensation of saline sol. Materials and methods: Blood pressure and pulse rate before and immediately after the procedure were measured as stress indicators for first time donors. In a similar fashion Hb, Ht and Ery's were measured. We looked for correlates between donor characteristics and initial values and changes in Hb and Ht. We also monitor all donor complaints and procedural complications. Results: We have evaluated 23 first donors (14 female and 9 male; mean age 64 years) who underwent a total of 33 procedures. Judged by the initial blood pressure and pulse rate and their changes during the procedure, this method did not seem be a stressful event for these donors (table1).The pre-cytapheresis haemoglobin content strongly correlated with the BMI (r = 0,53, p < 0,01). Also the decline in number of erythrocytes per procedure correlated with the BMI of the donor (r = 0.60 p < 0,01). Table 1 : Donor characteristics and haematological parameters:

harvests a high number of lymphocytes would allow the gradual dose therapy without submitting the donor to repeated procedures, provided we freeze the cells in several aliquots to be used later. The purpose of this study is to assess how many procedures yielded the number of lymphocytes required for transfusing gradually increasing doses. Methods Fifty-nine apheresis data performed in 41 donors were assessed. We used the Cobe Spectra® blood cell separator set on mononuclear cell program. Color-gram® was kept at about 3%.Collection duration was set up in accordance with the safety parameters default aiming at least one and a half donor's total blood volume (TBV) but usually two and a half or three TBV. The product of the collection was carried to the cryopreservation department in order to quantify the collected cells.Lymphocyte yield was measured by automated differential counting and also by CD3 positive cells counting by using flow cytometry analysis. Depending on the amount of lymphocytes, one fraction was delivered directly to patients' transfusion and the remainder was cryopreserved in aliquots according to the medical plans. Results The mean blood volume processed was 11250 ml-standard deviation (SD) 2497-and the mean leukocyte number per bag was 97,87 ¥ 10 6 /ml with a minimum of 42 ¥ 10 6 /ml and a maximum of 522 ¥ 10 6 /ml. Mean percentage of CD3+ cells per bag was 52,91%-SD 13,27. Regarding patients' weight, their mean was 60,8 Kg with extremes of 11,5 up to 87 Kg. Mean CD3+/Kg yield was 1,6 ¥ 10 8 /Kg-SD 0,78-with extremes of 0,49 up to 4,74 ¥ 10 8 /Kg. Conclusions Key factors determining the CD3+/Kg yield are the blood volume processed, a smooth running of the collection and the patients' weight, but nevertheless the number of apheresis to be carried out depends on the therapeutic strategy. Considering a starting point dose of 1 ¥ 10 8 CD3+/Kg only in 10 cases (16,9%), one more harvest would be necessary and only in 15 cases (25,4%) there would be cryopreserved lymphocytes available for a second infusion. If the starting point were 1 ¥ 10 7 CD3+/Kg then all cases could be handled with a single harvest and 100% of the patients would have aliquots stored for further transfusions. Procedures for clinical trials and those not categorized were counted separately. Results: A dramatic decrease in TA procedures was reversed in 2003, due to increased TA for TTP. Most procedures performed were for category I indications. These cases were mainly hematologic. TA in renal disease increased due to usage for humoral rejection of kidney allografts, a category III indication. No category IV procedures were performed during the study period. Protocol procedures were not done after 2001, causing the drop in the "other" disease category.

In 5 (15%) procedures donors had complaints. Three of those were a slight citrate reaction and two of those concern dizziness shortly after the procedure. All complaints disappeared without interference. Conclusions: Double erythrocytapheresis seems a feasible, well tole-rated, safe and efficient procedure to collect autologous erythrocytes. This method may thus provide an alternative for pre-operative whole blood collections. The cost-effectiveness of both procedures will be compared in due course. Conclusions: The majority of TA procedures at our hospital were performed for category I indications and hematologic disease. No TA was performed for category IV indications. A 5-year downward trend in TA procedures was reversed in 2003. Total TA utilization is not uniformly predictable at our hospital. Background: Sickle Cell Disease (SCD) and Thrombotic Thrombocytopenic Purpura (TTP) are un-common complications of pregnancy and seldom is pregnancy complicated by both. We describe such a patient. Case Report: A 22-year-old gravida 3 para 2-0-0-2 woman at 34 weeks gestation with SCD presented in pain crisis, was admitted and given morphine and intravenous fluids. Laboratory findings included thrombocytopenia, (Plt 64 K/mm 3 ), hemolytic anemia with schistocytes (Hgb 7.3 g/dL) and lactate dehydrogenase (LDH) of 1556 U/L (Ref 313-618). Other liver function tests were normal. The patient was diagnosed with TTP, transfused with 3 units RBCs, and daily one plasma volume exchanges (PE) using cryosupernatant as replacement fluid begun. After 9 PEs, platelets had increased to 135 K, LDH was 674 and PEs were rapidly weaned. She did well for 2 weeks, returned at 37 weeks gestation in early labor and 2 days later delivered a healthy baby. The patient's platelets decreased to 87 K and LDH was 2064. Diagnosed with relapsed TTP, she received daily PE and prednisone. Her clinical course was complicated by catheter infection but platelets increased and plateaued at >200 K with normal LDH after completing 12 PE. No further relapses have been reported. Conclusion: SCD and TTP are well-known complications of pregnancy associated with significant fetal as well as maternal morbidity. This woman's response to standard therapy was excellent and there was no associated infant morbidity.

C F Danielson, E M Skipworth, L W Jackson, J Parks, Clarian Health Partners, Inc., Indianapolis, IN; R Abonour, Indiana University Medical Center, Indianapolis, IN Background: AABB Standards for Hematopoietic Progenitor Cell and Cellular Product Services, 3 rd Edition, requires that autologous donors of hematopoietic progenitor cells have their history evaluated and be examined by their physician within 72 hours of scheduled donation. At our institution the Blood Bank/Apheresis physician is responsible for ensuring that the donor is in an acceptable state of health to undergo the collection. We here report the use of a brief form that we developed to facilitate this process. Methods: Patients are evaluated by the Hematopoietic Progenitor Cell Transplant Physician before being accepted for transplant. If transplant is planned a copy of the History and Physical from this visit is sent to Apheresis. Then, depending upon the disease, the patient may receive chemotherapy and/or G-CSF prior to presenting for HPC collection. When leukocyte counts are acceptable for collection the patient presents to the Apheresis area and is evaluated by the Blood Bank/Apheresis physician who is responsible for ensuring that the donor is in an acceptable state of health to undergo the collection. Results are recorded on the institution's basic "History-Physical-Progress Notes" which have been pre-stamped with the points required in the "Leukocyte Pre-Collection Note" (prompts shown underlined below). The form consists of 4 sections: 1) Space for date, patient description and reason for visit. 2) Subjective: "Has your health changed since you saw Dr. ____ on ____?" 3) Objective: (a) Vital signs (b) Allergies (c) Medications (d) Labs-CBC, Other (e) Physical exam-heart, lungs, other 4) Impression and Plan: Each section following the prompt contains sufficient space to record the needed data. Conclusion: The process, including use of this form, has been well received by the Blood Bank physicians. The process ensures that records of the patient's visit to their transplant physician are readily available to the Blood Bank physician. Any changes in the patient's condition since that visit can quickly be identified and addressed. Use of the form ensures that each patient is evaluated in a fairly standardized manner and the pre-stamped prompts reduce documentation time.

N Bandarenko, K F Stagg, S N Hay, M E Brecher, University of North Carolina, Chapel Hill, NC Background: Treatment of Thrombotic Thrombocytopenia Purpura (TTP) with daily therapeutic plasma exchange (TPE) may be accompanied by a variety of adjunctive interventions including steroids, anti-platelet agents, vincristine, and most recently rituximab. Rituximab, a murine/human monoclonal antibody directed against CD20 antigen on B-lymphocytes, is primarily used for treatment of non-Hodgkins lymphomas and occasionally to manage autoimmune diseases as well as TTP. Because of severe and fatal infusion reactions including heart failure, rituximab carries a boxed warning. Case Report: A 20 yo female presented with diarrhea, abdominal pain, fever, thrombocytopenia (plts = 37), schistocytes, elevated creatinine (Cr = 3.2) and serum LDH of 8011 U/L. With a diagnosis of TTP, she underwent 18 consecutive daily TPE with 1.0 to 1.5 plasma volumes replaced with cryopoor plasma (CPP). During the fourth TPE she experienced an anaphylactic reaction (chest tightness and dysphagia) that resolved after 75 mg antihistamine IV. TPE replacement was adjusted to a combination of 5% albumin and CPP to minimize plasma exposure and further severe allergic complications. With antihistamine premedication, she tolerated the next 14 TPE. Serum LDH plateaued at 1500-2000 U/L after an initial steady improvement. Her plt count reached 150 k after 12 TPE then gradually declined to 36 k despite continuing TPE for 6 additional days. At this point she had 18 TPE procedures along with corticosteroids 1 mg/kg/day and hemodialysis for renal failure. Because of the refractory behavior of her disease, rituximab was administered. TPE was held electively to allow time for drug distribution. Within 36 hours of rituximab infusion, she developed acute respiratory failure, chest pain, bloody sputum, hypotension and was moved to intensive care for management of biventricular cardiogenic shock. Ejection fraction was 5-10% by echocardiography. Once hemodynamically stable in the ICU, TPE was resumed (one day skipped). Her plt count quickly reached 241 k, and LDH dropped to normal following 4 more daily TPE (total of 22 TPE procedures). Her heart failure had completely resolved and she was discharged with continuing hemodialysis for renal failure. Rituximab was added to her medical record as a drug allergy. Conclusion: Refractory TTP can respond favorably to rituximab when used as an adjunct. However, interventions to avoid the potential mortality of TTP can also carry significant risk as illustrated by the severe sudden life threatening reaction of our patient. Use of rituximab for refractory TTP should follow a careful assessment of potential benefits and risks particularly in the absence of a standard definition of refractory TTP.

S Ohtani, Y Kanoh, T Akahoshi, Kitasato University School of Medicine, Sagamihara Kanagawa, Japan; Y Okuyama, K Hiruma, Tokyo Metropolitan Komagome Hospital, Tokyo, Japan; K Ohara, Kitasato University School of Medicine, Sagamihara Kanagawa, Japan Background: Leukocyte-reduced plasma has increasingly been requested for transfusion and for fractionation. Also, the Japanese government recently decided to implement a universal leukocyte reduction program for all blood products including plasma. Now, it has become necessary to establish more effective systems to collect leukocyte-reduced plasma. This study evaluated plasma collection on a new procedure that uses a high-separation (HS) core bowl for plasma apheresis. Materials and methods: We compared the characteristics of plasma products collected by the HS core bowl (n = 10) and standard core bowl (n = 5) using the Superlite apheresis machine (Haemonetics ® , Braintree MA) without leuko-reduction filters. The volume of plasma collected each cycle was controlled based on the donor hematocrit. Biochemical parameters, coagulation factor activities, and blood cell counts of plasma products were measured. The number of residual leukocytes was counted by flow cytometry methods. Results: Measurement values (mean ± SD) are summarized in Table 1 . Leukocyte counts using the HS core bowl decreased significantly (p < 0.05). Conclusion: We found that the plasma apheresis system using the HS core bowl can effectively prepare leukocytereduced plasma without using a leuko-reduction filter. There was no significant difference in biochemical parameters between the plasma products prepared by the HS core bowl and the standard bowl. 

T Hundhausen, A Doescher, DRK Blood Donation Service NSTOB, Oldenburg, Germany; V Franck, Baxter Healthcare Corporation, Nivelles, Belgium; T Volgmann, W Gebauer, E K Petershofen, T Mueller, F Schunter, DRK Blood Donation Service NSTOB, Oldenburg, Germany

Background: Blood component collection by apheresis is considered as means to optimise donor pool utilisation, product standardisation and quality. The present paper evaluates the new mobile apheresis device ALYX with regard to donor comfort, safety and RBC quality as compared to standard RBC. Methods: 36 experienced male blood donors were enrolled for two blood donations in a span of 10 weeks: one standard RBC donation, and the other a double-dose RBC donation collected with ALYX. Side effects during apheresis and the following five days were assessed by standardised questionnaire. Clinical data including heart rate, blood pressure, Hb and several coagulation parameters were taken pre and post donation. One bag of each ALYX donation was irradiated. RBC qualitity was assessed by paired comparison of a range of critical parameters over a span of 49 days. Results: 34 procedures were completed successfully. 94% of donors felt no physical impairment and would be willing to donate with ALYX on a regular basis. 70% were free of ACD-induced side effects. No serious side effects were observed and vital parameters remained basically unchanged. Apart from an elevated thrombin-antithrombin-complex (TAT) concentration usually also seen after standard donations, coagulation parameters remained normal. In one case, however, hemoglobin concentration post donation fell below 11.0 g/dl. ALYX RBC fullfill all European quality requirements for at least 42 days. As regards final volume and hematocrit ALYX RBC were less variable than standard RBC. In contrast, standard RBC had a lower supernatant total protein concentration, a higher pH and a higher intracellular ATP concentration during storage. The damage of irradiated ALYX RBC was reflected in higher hemolysis rate, extracellular potassium concentration and MCV increase and a faster intracellular ATP decrease compared to nonirradiated units just as in standard RBC. Conclusion: Among experienced blood donors ALYX is well received and proves to be safe and feasible. Side effects are similar to those known from other apheresis devices. Hemoglobin concentration seems to be a critical parameter, as some European countries require a Hb of at least 11.0 g/dl after donation. As regards red cell quality, superior standardisation is the major argument in favor of ALYX RBC, while pH and intracellular ATP concentration are advantageous for standard RBC. The decision whether or not to introduce the ALYX component collection system should focus on other issues than RBC quality.

H Mohr, Blood Center of the German Red Cross Chapters of NSTOB, Springe, Germany; A Bayer, Blood Center of the German Red Cross Chapters of NSTOB, Gera, Germany; T Müller, Blood Center of the German Red Cross Chapters of NSTOB, Springe, Germany Background: It has been postulated that pooling of platelets from whole blood collections of multiple donors may increase the risk of bacterial contamination. It was investigated how the preparation steps for buffy-coat derived platelets affect the survival of bacteria. Methods: 9 different bacteria strains were used (see in the "Results"-section). Blood donations (approx. 6-8 h after phlebotomia) were spiked with 1,000-20,000 colony forming units of either strain and stored for 8 h at room temperature. The isolated buffy coats were kept at the same temperature for another 4-6 h. PC in storage medium PAS-IIIM containing approx. 30% plasma were prepared from pools of 4 buffy coats. At each step of the preparation process bacterial titres were estimated by colony assay. Results: The following 5 bacteria strains were inactivated during storage in blood or buffy coats: Enterobacter cloacae, Escherichia coli, Pseudomonas aeruginosa, Serratia marcescens and Yersinia enterocolitica. The inactivation rates estimated ranged between ≥3.32 and ≥4.62 log 10. The corresponding PC remained sterile during storage at 22 ± 2°C for up to 7 days. Bacillus cereus, Staphylococcus aureus and Staphylococcus epidermidis did not multiply during storage in blood and buffy coats. Compared to their initial values in spiked blood the bacterial titres in the resulting PC were reduced by 1.7, 2.1 and 2.4 log 10, respectively. Klebsiella (K.) pneumoniae was the only bacterial strain that grew in blood and buffy coats. Multiplication rate was approx. 1 log 10 within 12 h. The bacterial titre was reduced by approx. 2 log 10 during preparation of the PC. Considering the different volumes of buffy coats and PC (60 vs. 300 ml), the total amount of bacteria was reduced by approx. 1.3 log 10 by the preparation procedure. Compared to the spiked blood donations the overall reduction factor for K. pneumoniae was about 1 log 10. Conclusions: Both, blood and buffy coats are a bad growth medium for most bacteria. With the exception of K. pneumoniae all strains tested did not multiply during storage. In addition, bacteria were significantly eliminated by the preparation procedure for PC. Compared with spiked blood the overall inactivation/elimination rates for the 9 bacteria strains used ranged between approx. 1 and = 4.6 log 10.

J C Yu, C Chong, M A Cortus, C Le, S Holme, Pall Medical, Covina, CA Background: Sepsis and death caused by transfusion of RBC contaminated with bacteria continues to be an issue in transfusion medicine. A new enhanced Bacteria Detection System (Pall eBDS) for use in platelet concentrates (PC) has been developed. Detection is based on % oxygen in the sample pouch after 35°C incubation. The feasibility of using this system in leukoreduced RCC has been explored. The performance of eBDS was investigated with 8 bacteria ( Table 1 ) that were selected based on clinical sepsis or morbidity reported in literature. Study Design: Bacteria were inoculated into RCC units at room temperature with two target levels, 10 and 100 CFU/mL. Immediately after mixing, samples were taken for CFU/mL and eBDS testing. RCC units were then stored at 4°C and samples were taken at different times during six weeks of storage to determine bacteria growth and detection with eBDS. The eBDS pouches were incubated at 35°C with agitation for 48 hrs. Results and Discussion: Rapid bacteria growth occurred with Yersinia enterocolitica reaching > 1 ¥ 10 9 CFU/mL in RCC after two weeks. (Table 1 ) For S. epidermidis, P. aeruginosa, and P. putida the bacteria levels slowly decreased during storage, while Bacillus and Klebsiella species had no viable cells present after two weeks. The same results were seen at the target spiking level of 10 CFU/mL. Over 270 samples were tested with 8 bacteria during the study. The % oxygen levels in eBDS pouches sampled from all units that still had viable bacteria were below 8.4 % for both 10 and 100 CFU/mL target spiking levels. The control % oxygen was above 14.1% for all 25 units of fresh or stored RCC with no bacteria.

Conclusion: These findings suggest that Pall eBDS can be used for detection of bacteria in leukoreduced RCC with sensitivity level of 10 CFU/mL or less using a 48 hr. incubation time at 35°C with agitation. The detection with Pall eBDS was unrelated to the storage time and temperature of the sample with viable bacteria present.

M A Cortus, C-Y Chong, C Le, J Chen Yu, S Holme, Pall Medical, Covina, CA Background: The Pall eBDS has recently been cleared by the FDA for testing of the presence of bacteria in platelet concentrates (PC) stored in plasma. The purpose of this study was to evaluate bacteria growth in pooled buffy coat (BC) PC stored in T-Sol, and the detection using the Pall eBDS. Method: Six leukoreduced BC-PC pools were prepared by standard procedures. Each pool was divided into six CLX storage bags. One bag per pool was inoculated with 1-15 CFU/mL of Staphylococcus epidermidis ATCC# One bag served as non-inoculated control. After mixing and after 24 hours of platelet product storage at 22°C, each platelet bag was sampled for culture (CFU/mL) and for eBDS. The eBDS sample pouches were incubated for 24 hours at 35°C for measurement of % oxygen. Bacteria growth in the BC-PC, and % oxygen levels in the pouches after incubation were then compared to corresponding historical data with plasma PC. Results: Similar mean inoculation levels were used for both T-Sol and plasma PC. Little or no growth after 24 hrs. was observed with S. epidermidis in T-Sol PC, similar to what had been observed with plasma PC. Slow growth (<1 log) was observed with S. agalactiae although significantly lower in T-Sol PC (p < 0.05). Log growth (>1 log) was observed with S. aureus with significantly higher growth in plasma. P. aeruginosa and S. choleraesuis frequently autosterilized in plasma PC, and showed an overall better growth in T-Sol PC. This is apparently the result of the bacteriacidal effects of plasma on the gram negative organisms. Mean pouch % oxygen was similar for the two types of PC with %O 2 < 2% for all organisms except S. agalactiae, which had higher %O 2 values in T-Sol samples (mean 3.4 ± 1.1% (stdev) for plasma PC vs. 9.0% ± 0.6% for T-Sol PC). Non-inoculated control % oxygen levels was higher (p < 0.05) for T-Sol BC-PC (n = 41): mean 19.0 ± 0.6 % vs. 16.9 ± 1.9% for plasma PC (n = 713). In view of this the eBDS pass/fail threshold for platelets stored in additive solution has been set to 12.5 % versus 9.4 % with platelets stored in plasma. Conclusion: These data suggest that although bacteria growth levels and the resulting % oxygen levels differ with platelets stored in additive solution as compared to plasma the detection with Pall eBDS is not significantly affected using a pass/fail threshold of 12.5%.

Background: Bacterial contamination of platelets is a significant cause of transfusion-associated morbidity and mortality. The survival of bacteria in platelet pools stored in plasma is greater than in those stored in platelet additive solution (PAS). The aim of this study was to evaluate the Pall enhanced Bacterial Detection System (eBDS) on platelet pools in PAS. eBDS utilises bacteria oxygen consumption as a marker for detecting the presence of aerobic and facultative anaerobic bacteria. Study design: Leukocyte depleted pooled PAS platelet units made from 4 buffy coats were split into 6 bags. Five bags were inoculated with approximately 300 colony forming units (CFU) of 5 different laboratory type strain bacteria resulting in a concentration of 1-15 CFU/mL. One bag was a negative control. Samples were taken immediately into eBDS pouches and also after 24 h incubation at room-temperature These pouches were incubated on a shaking agitator at 35°C. The oxygen level was measured in the eBDS pouches after 24 or 30 h incubation. If the oxygen level was less than 12.5% it was recorded as a sign of bacteria growth. 5 replicates were performed for the following: Enterobacter cloacae, Escherichia coli, Staphylococcus aureus, Salmonella choleraesuis, Streptococcus agalactiae, Serratia marcescens, Klebsiella pneumoniae, Pseudomonas aeruginosa, Staphylococcus epidermidis. 4 replicates were performed for Bacillus cereus. Results: In the majority of cases (42/49), the Pall eBDS detected the presence of bacteria when samples were taken immediately after inoculation and incubated for 24 hours. 4 of the pouches that did not detect positive were with B.cereus at 1 cfu/mL (no CFU in pouches). The eBDS also detected bacteria in the majority of inoculated units that were further stored for 24 hours (47/49). The 2 cases with no detection (S. choleraesuis) were associated with a decreased bacteria count in the platelet unit after 24 hours. Conclusion: Detecting bacteria in platelet pools stored in PAS by measurement of gaseous oxygen appears to be reliable. The eBDS is very easy to handle and useful for routine use. The failed detection of B. cereus and S. choleraesuis might have been due to very low bacteria counts or processes contributing to autosterilisation. Further studies will be done with these bacteria with a slightly higher inoculum.

T Mohammadi, R Pietersz, Sanquin Blood Bank North West Region, Amsterdam, Netherlands; C Vandenbroucke-Grauls, P Savelkoul, Department of Medical Microbiology and Infection Control, VU University Medical Center, Amsterdam, Netherlands; H Reesink, Sanquin Blood Bank North West Region, Amsterdam, Netherlands Background: platelet concentrates (PCs) are frequently associated with transfusion-transmitted infections. To assess the presence of bacteria in PCs a rapid and dependable method is needed. Based on real-time PCR technology, a broad-range 16S rDNA assay was developed. After optimization, the assay was validated and its performance compared to an automated culture system to determine its usefulness for bacterial routine screening of PCs. Methods: the validation was performed with leukoreduced PCs derived from 5 buffy coats in plasma. The presence of bacteria in these PCs was routinely assessed in an automated culture system (BacT/Alert). Culturing was conducted until a positive signal was detected or for up to seven days when remaining negative. The PCR assay was performed with DNA extracted from the same cultured samples. DNA extraction was done with a fully automated method (MagNA Pure). PCR amplification was performed with a set of universal primers and probe targeting eubacterial 16S rDNA. PCR results were compared with those obtained by culturing. Results: a total of 2146 samples were tested by both the PCR assay and culture system. Eighteen samples were found to be contaminated by both methods. Identified microorganisms included Propionibacterium spp., Staphylococcus spp., Bacillus spp., Micrococcus spp. and Peptostreptococcus spp. Two samples generated an amplification signal in the PCR but remained negative in the BacT/Alert. This discordance is probably due to the detection of DNA from non-viable bacteria by PCR that can not be detected by culturing. Conclusion: these results indicate that the PCR assay enables rapid, accurate and sensitive detection of bacterial DNA in PCs. Hence, this novel method may be a valuable improvement in monitoring bacterial contamination of PCs.

S M Ramirez-Arcos, I Martincic, C Greco, Canadian Blood Services, Ottawa, ON, Canada Development of Molecular Assays for the Detection and Identification of Bloodborne Bacteria in Apheresis Platelets in Canada Background. Bacterial contamination of blood products is currently the major microbiological cause of transfusion-associated morbidity and mortality. Thus, there is an urgent need for the development of efficient methods to detect and identify bacterial contaminants in blood products. We have recently implemented the BacT/ALERT system (Biomérieux) for bacterial detection in apheresis platelets. Although we found that this system is highly sensitive, it does not provide bacterial identification, which is the purpose of this study. Using state-of-the-art technology, such as real time-polymerase chain reaction (RT-PCR) and liquid arrays (LiquiChip system), we have implemented molecular methods for the detection and identification of Staphylococcus epidermidis from platelet samples. Although S. epidermidis is part of the normal skin flora, it can cause life-threatening diseases in immunocompromised people, who are usually the recipients of blood product transfusions. Study design. Platelet units were spiked with S. epidermidis at final concentrations of 10 and 100 cfu/ml. Samples from spiked platelets were taken daily over a period of seven days (from day 0 to day 6) for inoculation into BacT/ALERT aerobic culture bottles, colony counting, multireagent strip testing, Gram staining, and molecular detection of S. epidermidis. Cell division genes were used as DNA targets for PCR standardization. The molecular methods were evaluated for their specificity for S. epidermidis by determining their crossreactivity with other Staphylococcus species. Results. We showed that novel technology can be implemented for the detection and identification of S. epidermidis in platelets. Cell division genes proved to be good DNA targets, providing a specificity of 100% to the developed methods. Although the BacT/ALERT system is highly sensitive, with a detection level of <100 colony forming units (cfu)/ml, other methods (e.g. Gram staining) showed very low sensitivity with ranges of detection of ~10 7 cfu/ml. Novel implemented methods during this study allowed for the detection of S. epidermidis at concentrations of <10 5 cfu/ml. Conclusions. Our results show that molecular methods are time-and cost-effective alternatives for specific detection of S. epidermidis and other bacterial pathogens in platelets. Although further standardization is needed to improve the sensitivity of these methods, our data 47A 2004-Vol. 44, Supplement provides the basis for the implementation of multiplex-based detection and identification of bloodborne bacteria.

B Jaitner, B Phelps, D Chien, Y-L Fong, Chiron Corporation, Emeryville, CA Background: Real-time PCR instruments from Applied Biosystems (7900HT), Cepheid (Smartcycler), Corbett Research (Rotor-Gene), MJ Research (Opticon) and Stratagene (M ¥ 4000) were compared with regard to detection sensitivity, linearity, uniformity and precision. Methods: Realtime PCR with SybrGreen detection of genomic DNA from Staphylococcus aureus was used as a model system to evaluate the performance of each instrument. The genomic DNA was tested in twofold serial dilutions, from a single copy to approx. 2100 genomic copies. Results and Conclusions: Using SybrGreen detection of Staphylococcus aureus genomic DNA as model system, one genomic copy was reproducibly detected by the 7900HT and the Opticon. Three genomic copies were detected with the M ¥ 4000 and Rotor-Gene and seven with the Smartcycler, respectively, in a reproducible manner. For the latter instruments, the reproducible detection of lower genomic copy numbers was confined by the formation of primer dimers. The earliest detection of one genomic copy of Staphylococcus aureus was observed in the 7900HT with a Ct value of 31.5 +/-1.279 (% CV). The instrument precision for the Tm value was best for the Smartcycler with a Tm of 85°C +/-0.160 (% CV), the Opticon with a Tm of 83.9°C +/-0.220, and the 7900HT with a Tm of 84.5°C +/-0.265. The amplification and detection was linear over a range of 10 to 2100 genomic copies for all instruments with a R2 value of 0.99 for the 7900HT, Opticon, and Rotor-Gene and a R2 value of 0.98 for the M ¥ 4000 and the Smartcycler. The detailed comparison will be presented in the poster.

L A Beausang, A Levin, V Kovalenko, Immunetics, Inc., Boston, MA Background. On March 1, 2004, AABB standards mandated that United States blood centers commence testing all platelet units for bacterial contamination. This new standard was based on the significant risks to transfusion patients associated with contaminated platelet units. Approximately 4 million platelet units are transfused per year in the U.S., of which up to 4000 are potentially contaminated. We have developed a rapid, sensitive, and specific assay for the detection of bacteria in platelet units, and have now shown the sensitivity of the assay for both gram-positive and gram-negative species at concentrations approaching those detected by culture systems. Methods. The assay, which is based on the detection of peptidoglycan, a component of all bacteria, is simple: the method can be carried out in single tubes and color change read visually, or in a 96 well plate by a standard ELISA plate reader at 450 nm. The assay can be either quantitative or qualitative, and color development is directly proportional to peptidoglycan concentration. To increase sensitivity, new sample extraction reagents and procedure were devised. Sensitivity was studied with two bacterial species: S. marcescens (ATCC 43862) and S. epidermidis (ATCC 49134). The latter is a common but relatively slow growing pathogen. Colony forming units for each species were determined by plating serial dilutions. Platelets were spiked with each organism at 10, 100, and 1000 CFU/ml. Platelet samples were extracted and tested according to the assay protocol. Negative controls comprised similarly treated platelets without added bacteria. Results. The peptidoglycan-based bacterial detection assay was improved through an extraction procedure, which totally eliminated the inhibitory activity associated with plasma in platelet preparations; the procedure also clarified the platelet suspension, eliminating interference in reading the result. Extraction also increased the release of peptidoglycan from bacteria in a form which was more amenable to detection, resulting in higher assay sensitivity. In multiple assay runs, both S. marcescens and S. epidermidis were detected at concentrations down to 100 CFU/ml. This represents a 10-fold improvement in sensitivity compared with the pre-extraction protocol. Conclusions. Our assay based on the detection of peptidoglycan presents a rapid and costeffective approach to screening platelet units for bacterial contamination. The peptidoglycan test has now been shown to detect gram-positive and gram-negative species at concentrations in the range detected by lengthier culture procedures in current use. As such, it may be useful in ensuring the safety of platelet units and ultimately in extending viability.

R Gonzales, L Dunham, O Mark, P Woodell, A Hall, R Vanbrakle, K Sconce, H Seymour, C Chan, L Feige, S Beardsley, D Craft, Walter Reed Army Medical Center, Washington, DC Background: The most common transfusion-associated infectious risk in the United States today is bacterial contamination of platelet components. Effective 1 March 2004, the American Association of Blood Banks (AABB) requires a combination of strategies both to limit the initial inoculation of bacteria into the blood component and to detect the subsequent growth of bacteria during room temperature platelet storage. Study Design: We designed a protocol in accordance with AABB Association Bulletin #03-12 to validate the ability to detect bacterial contamination of platelet products using the VersaTREK (TREK Diagnostic Systems, Cleveland, OH) blood culture detection system. Leukocyte reduced apheresis products were collected from eligible donors using the MCS+LN 9000 Haemonetics (Braintree, MA). A total of 16 units were processed according to standard procedures and placed in a platelet incubator for 24 hours. Units were inoculated with either Coagulase-negative Staphylococcus or Escherichia coli at two different dilutions, ranging between 5 and 300 colony forming units (cfus) per unit. Two 9 ml aliquots were collected at 12-hour intervals for the first two days and daily afterwards for a total of 7 days. Each aliquot was withdrawn from each well mixed product using a Sterile Connecting Device (Terumo, Elkton, MD). Using aseptic technique, 4 mls of each aliquot was inoculated into each of two aerobic and two anaerobic blood culture bottles. One ml was diluted and plated onto blood agar plates (BAP) for colony counts to assess unit concentration of bacteria at each aliquot time. Results: For all units that received at least a 50 cfu initial bacterial inoculum, all 12 hour or later aliquots yielded a positive VersaTREK blood culture for all bottles inoculated, regardless of which bacteria was used. Organism identification was based on Gram stain and colony morphology. All positive bottles were detected by the VersaTREK within 24 hours of aliquot inoculation, time to detection ranged from 3.2-23.7 hours. Colony counts of all unit aliquots demonstrated 20-70 cfus/ml in all units by 24 hours. Conclusions: With an initial inoculation of 50 or more bacteria per unit, platelet units incubated at room temperature for 12 or more hours resulted in enough bacterial growth to generate a positive blood culture upon all aliquot subcultures. Combined with the fact that all blood culture bottles yielded detectable growth within 24 hours of inoculation, our data suggests that the VersaTREK blood culture system is sensitive to extremely small numbers of organisms and meets the intent of the AABB guidelines to reduce the transfusion of bacterially contaminated platelets.

E B Kahwash Sr., J Leonard, M Redmon, J Garr, J Snyder, W B Lockwood, University of Louisville, Louisville, KY Introduction: A procedure to detect the presence of bacterial contamination in PLTs is required by the AABB. The BACTEC continuous monitoring blood culture instrument (Becton Dickinson, Cockeysville, MD) for monitoring blood product QC is not approved by the FDA. This study was conducted to validate the BACTEC or BacT/ALERT (bioMerieux, Inc, Durham, NC) in eight local hospitals. Methods: Thirty swirling-positive, 1-day-old leukoreduced whole blood derived PLTs were combined into 5 equal pools. One aerobic uninoculated adult BACTEC bottle per pool served as a negative control. Each pool was aseptically inoculated through the access port with 10 3 CFU/ml of five common platelet contaminants (Table 1) . Each inoculum was aliquoted aseptically into two aerobic pediatric or adult bottles (3.5 mL and 7.0 mL aliquots, respectively). A Sterile Connection Device (TSCD, Terumo Medical Corp., Somerset, NJ) and transfer set (Charter Medical, Ltd, Winston-Salem, NC) were used in the aliquoting procedure. Mistakenly, one adult BacT/ALERT bottle was cultured with the pediatric aliquot (3.5 mL) instead of the adult aliquot of 7 mL. Each hospital processed their respective BACTEC or BacT/ALERT samples, which were monitored continuously for 5 days until time to initial bacterial growth was determined. Results: All negative control BACTEC cultures showed no growth at 5 days. The results of the BACTEC and BacT/ALERT cultures, both pediatric and adult bottles, are shown in Table 1 . All cultures were positive within 19 hours, including the adult BacT/ALERT bottle mistakenly processed with the pediatric aliquot. The earliest growth was Bacillus spp. (approximately 8 hours) in both BacT/ALERT and BACTEC, while the latest was Pseudomonas aeruginosa species (approximately 17 hours). However, 3 of 5 BACTEC and 2 of 3 BacT/ALERT cultures inoculated with Staphylococcus epidermidis grew TRANSFUSION 2004-Vol. 44, Supplement Acinetobacter believed to be a contaminate from originally impure samples or contaminated at inoculation. Conclusions: Only the BacT/ALERT system and Pall Bacterial Detection System (BDS, Pall Corporation, East Hills, NY) are FDA approved for QC of platelet components. Our study showed that BACTEC, which is not currently FDA approved for platelet QC, is equivalent in sensitivity and specificity. Table 1 : Bacterial Detection Validation Summary BACTEC: n = 5, BacT/ALERT: n = 3 passed visual inspection. While a low pH measurement does not provide direct correlation to the presence of bacteria, there clearly is an inverse relationship between the two measurements. This inverse relationship makes pH determination an acceptable screening tool.

S N Hay, M E Brecher, University of North Carolina, Chapel Hill, NC Background: Glucometry and pH analysis employing analyzers or multireagent strips have previously been validated for the detection of bacterial growth in whole blood derived platelet concentrates (Pall/MedSep CLX-CP2D and Baxter PL-732 CPD). This study validated the use of glucometry and pH determination in Terumo XT612 CPD platelet concentrates. Methods: Twenty Terumo XT612 CPD platelet concentrates were analyzed for pH, glucose and swirling on day 6 of storage (outdated). Three platelet concentrates were inoculated with Bacillus cereus, Staphylococcus epidermidis, Staphylococcus aureus or sterile saline (total n = 12) on day 1 of storage and were analyzed for pH, glucose, swirling and CFU/mL daily (days 1-8 of storage). Results: Based on glucose and pH determinations on outdated day 6 platelets and sterile controls, cutoffs of pH < 6.8 (analyzer), glucose <200 mg/dL (analyzer), pH < 7 (multireagent strip), glucose £100 mg/dL (multireagent strip) and absence of swirling were established. B. cereus was detected 1-3 days, S. aureus 3-4 days and S. epidermidis 4-5 days after inoculation at approximate concentrations of 10 7 CFU/mL. Swirling was the least reliable of the parameters studied. Glucose measured with an analyzer correlated with the reagent glucose (R 2 = 0.67) but had a bias, with lower estimates with the reagent strips. Similarly, pH measured with an analyzer correlated with the reagent pH (R 2 = 0.83) with a bias, with lower estimates with the reagent strips. Conclusion: This study validates the use of pH and glucose determinations as rapid and inexpensive methods to detect high levels of bacterial contamination in Terumo XT612 CPD platelet concentrates.

L Amorim, M E E D Lopes, J F Oliveira, C Mello, E Santos, I Azevedo, HEMORIO, Rio de Janeiro, Brazil BACKGROUND: Bacterial detection is now mandatory in platelet concentrates (PC), according to AABB standards. However, the cost-effectivenness and feasability of methods used to perform this task are still very controversial. We present here a comparison between two different methods for bacterial detection. METHODS: During a 50 day period, we tested for bacterial contamination all the PC released for transfusion. The bacterial detection tests were performed immediately before transfusion, at least 24 hours after the collection, using the urine dipstick for pH and glucose measurement. pH higher than 7.0 and/or glucose level higher than 250 mg/dL were required for PC release. If they do not fullfil these criteria, the PC were discarded and a sample for bacterial and fungus culture was collected. If the PC passed the test, 6 units were pooled, and a pool sample was also collected for culture. The culture was performed by the BactAlert method. RESULTS: A total of 2,821 PC were tested by urine dipsticks. From these, 225 (7.97%) failed, due to low pH. Only one (0.44%) of these 225 PC tested positive for bacteria (a Gram-positive rod). 432 culture of 6 units pools were performed. Seven pools (2.07% of the pools) showed a postive result (3 cases of Gram-positive rod, 2 cases of Staphylococcus warnerii, 1 case of Staphylococcus capitis and one case of Staphylococcus coagulase negative). The sensibility and specificity of the pH measurement for bacteria detection were, respectively, 12.5% and 92%. CONCLUSIONS: The rate of PC disposal due to a low pH was very high (7.9%), but the test sensitivity was very low. More sensitive tests should be employed, in order to improve the cost-benefit ratio. Study Design: Direct bacterial detection methods for random donor platelets (RDP) are limited by logistic and cost constraints. Many transfusion services have turned to indirect measurements such as pH to comply with AABB Standard 5.1.5.1. Our facility selected pH measurement and developed a validation plan to ensure acceptability. Leukoreduced and nonleukoreduced RDP were inoculated with S. aureus, B. cereus, and E. aerogenes on day 1 at approximately 100 CFU/ml. Prior to inoculation, sterile samples were collected and cultured (negative control). Each RDP was sampled on days 1-5 for a pH determination and dilutional culture. pH determinations were completed with a handheld pH meter. In addition, a routine blood culture was initiated on day 1 for each RDP. Results: Sterile cultures showed no growth at 5 days. Blood cultures taken immediately after product inoculation showed growth at 8.7-14.3 hours. Correlation of pH and CFU are shown in Tables I and II. Conclusion: pH and CFU count results suggest that as colony counts increase, pH values decline. Given the bactericidal effect of the plasma in each RDP coupled with growth requirements of each bacterium, it was not possible to determine a minimal or standard CFU count that results in an unacceptable pH (<7.0). Unacceptable pH values were obtained with CFU counts in the range of 10 8 -10 11 . pH testing is likely to be sensitive to lower levels of bacteria, however, the rapid growth phase (5-6 logs in 1 day for E. aerogenes; 2-3 logs in 1 day for B. cereus) precluded measuring the bacterial concentration at the time the pH dropped below 7. The data also suggest that as bacteria begin to die, the pH can increase (B. cereus and E. aerogenes). The contaminated RDP with an acceptable pH would not have 49A 2004-Vol. 44, Supplement tive of this study was to test platelet units by plate culture and the Platelet PGD Test, a rapid immunoassay-based test, to determine the frequency of bacterial contamination in platelet concentrates. METHODS AND RESULTS: Aseptic samples were collected in a bio-hood from in-date (d2-5, N = 443) and out-date (day 6-7, N = 2610) platelet units which were obtained from 4 different collection sites in the US. For plate culture, samples were pooled approximately 10-fold, 100 ul pooled sample was spread on TSA plates and plates were incubated for up to 5 days at 37C. Analytical sensitivity of the culture method as run was approximately 10e3 CFU/ml which correlates to the analytical detection limit of the Platelet PDG Test. Aseptically-collected samples were also analyzed by the Platelet PGD Test. For the PGD test, the sample migrates through the device by capillary action wherein antibodies specific for conserved Gram positive or Gram negative bacterial antigens would be captured and detected if bacteria were present in the sample. The time to result is approximately 20 minutes after sample addition, and test results are determined by visual examination. During the screening, 2/3053 (0.066%) samples were identified as PGD Test and culture positive. API identity testing indicated the isolates were Staph. epidermidis and Staph. warneri. These isolates were compared to other S. epidermidis (ATCC 12228, 14990, 29887, 35547) and S. warneri (ATCC 17917) strains for rates of growth in platelets, heat-inactivated plasma recovered from platelets and TSB media to determine if the phenotype of the platelet isolates was unique. Doubling times and survival rates in platelets did not differ (p > 0.05) among the ATCC and platelet isolate strains. For the S. warneri-contaminated unit, follow-up testing of the linked red cell unit by the collection site did confirm the presence of a Gram positive Staphylococcus when cultured on approximately day 10 of storage. CONCLUSIONS: The observed frequency of contamination of this study is consistent with literature citations of the incidence of bacterial contamination in platelet concentrates. Identification of contaminations of linked platelet and red cell units illustrates the potential for bacterial contamination in packed red cell units, even relatively early in their storage life.

J Hall, C Lajoie, Verax Biomedical Incorporated, Worcester, MA BACKGROUND: To improve the detection of platelet concentrates contaminated with bacteria, a variety of simple test methods, including visual inspection, Gram stain and/or pH measurements, are being evaluated in transfusion centers. The objective of this study was to compare the use of pH measurement and Platelet PGD Test in the detection of bacteria after inoculation in platelets. METHODS AND RESULTS: Six platelet units were sampled aseptically and aliquoted into growth tubes which were left uninoculated or inoculated with K. pneumoniae, E. coli, or S. epidermidis. At 24 hr intervals, tubes were aseptically sampled, and samples were tested for changes in pH and analyzed by the Platelet PGD Test, an immunoassay rapid test used to detect bacterial contamination. For the PGD test, the sample migrates through the device by capillary action wherein antibodies specific for conserved Gram positive or Gram negative bacterial antigens would be captured and detected if bacteria were present in the sample. The time to result is approximately 20 minutes after sample addition, and test results are determined by visual examination. Over the duration of the growth study (up to 5d), no changes in pH of uninoculated platelets were observed, though there was a slight variability in the starting pH of individual units. Frequency of pH change for samples over the growth period was: 100% (6/6) for K. pneumoniae, 100% (6/6) for E. coli, and 0% (0/6) for S. epidermidis inoculated samples. However, only 5 of 6 of K. pneumoniae samples supporting bacterial growth consistently dropped below pH 6.0 while all E. coli samples did. Positive test results for the Platelet PGD Test preceded a detectable change in pH by 16 to >48 hr in 100% of samples supporting bacterial growth. Additionally, the Platelet PGD test detected bacteria in all samples which supported growth yet failed to demonstrate pH changes. Changes in sample appearance (opacity, turbidity) were best correlated to samples dropping below pH 6. CONCLUSIONS: Results of this study suggest testing pH of platelet units to detect bacterial contamination would identify some rapidly growing Gram negative species, but detectability would be slower and at lower frequency than that observed by a more sensitive immunoassay-based Platelet PDG test.

J Hall, D Laverda, Verax Biomedical Incorporated, Worcester, MA BACKGROUND: Bacterial sepsis accounts for significant morbidity and mortality following transfusion of contaminated blood products. Though sensitive tests that can detect contaminating bacteria at very low levels are being developed, little is known about how bacteria grow in platelets, the degree to which they are susceptible to the antimicrobial effects of plasma, and whether susceptible strains can acquire resistance within the time span of platelet concentrate shelf life. This study investigates the influence that plasma resistance has on the growth kinetics of bacteria. METHODS: Cellfree plasma supernatants from whole blood-derived platelet concentrates were divided into portions that were either heat-inactivated (HI+) or left intact (non-HI). Growth at 22C was measured by optical density, and bacteria concentration was determined using standard curve equations. RESULTS: A range of growth characteristics were observed. 1) Serratia marcescens and Bacillus cereus grew readily in all plasmas tested; 2) Staphylococcus epidermidis grew in 5/6 plasmas; and 3) Escherichia coli grew in only 1/6 plasmas. Heat-inactivation of the plasma did not significantly change the kinetics of S. marcescens, B. cereus, or S. epidermidis growth. In contrast, E. coli was able to grow in all HI+ plasmas. After an extended incubation in one of the non-HI plasmas (plasma R), an isolate of E. coli was recovered that was now capable of early growth in non-HI plasma R and late growth in other non-HI plasmas. Further passage in non-HI plasma R resulted in a more resistant strain that grew in all plasmas with growth in non-HI plasma R equivalent to its HI+ counterpart. E. coli passaged in HI+ serum eventually developed a more modest resistance and took longer to reach peak concentrations. CONCLUSIONS: Growth in platelets is a dynamic process influenced by bacteria strain, nutritional requirements, and plasma constituents and may appear late in the period of platelet concentrate shelf life. Susceptible bacteria populations could initially be suppressed and not become evident without extended storage. Because of this variability, detection systems that rely on sampling closer to the time of product release may have an advantage in detecting these bacteria.

R Goodson, G Timoteo, J Yi, A Tabrizi-Wright, B Phelps, D Chien, Y-L Fong, Chiron Corporation, Emeryville, CA Background: Bacterial contamination of blood components is currently the most frequent infectious complication of blood component use, with an estimated incidence of 1000 to 3000 platelet units, which is associated with 100-150 deaths per year, and 1 in 30,000 RBC units in the U.S. 1 The risk of receiving a bacterial-contaminated blood product is estimated to be 50-to 250-fold higher than the combined risks of transfusion-mediated infectious of HIV 1/2, HCV, HBV and HTLV I/II. Due to the risk of bacterial contamination, the shelf life of platelets is limited to five days in the U.S. In 2003, the AABB recommended to implement a bacteria detection test for platelet components by March 2004. To develop a test for bacterial detection in platelets, platelet samples are usually artificially spiked with known amounts of bacteria as the study materials to model the contaminated platelets. In order to better control the desired amounts of each species of bacteria spiked into the modeled platelet samples, we need to have a very accurate method to predict and calculate the microbial loads for each species prior to spiking into platelet samples. In this study we investigated and established methods to accurately predict the correlation of optical density at 600 nm (OD600) to the actual plate counts, colony forming units per milliliter (CFU/ml), for each of the 16 Gram positive, Gram negative bacteria and the yeast Candida alibicans that are most prevalent to contaminated platelets. Materials and Methods: Bacterial strains were obtained from the Chiron Master Culture Collection or the ATCC. Culture media was prepared using DIFCO reagents. OD600 measurements were made using an Eppendorf BioPhotometer. Cells were grown at their optimal growth conditions. Cultures were sampled at time intervals consistent with their doubling times, and OD600 was measured. Immediately following OD600 measurements, samples were diluted and plated on sheep blood agar plates to obtain countable colonies. The data were plotted as OD600 verses time for the growth curves, and OD 600 verses colony counts for OD600 to CFU/ml curves. Results And Conclusion: Among the 17 microbes investigated, the correlation of OD600 to the actual plate counts (CFU/ml) for some microbial species are significantly different. Using the method we developed, we are able to consistently predict plate counts (CFU/ml) within 1-2 fold variations from OD600 measurements for each of the microbes studied. Background: In 1994 a large multi center blood collection facility compared the two-step Blood Donor Prep Kit (70% alcohol Sepp® and a 2% iodine tincture Frepp®) to 10% povidone-iodine. Positive skin cultures were lowest using the Blood Donor Prep Kit. As a result of this validation, the blood center switched donor arm prep agents. In 2003 the US Food and Drug Administration (FDA) issued "Options for Arm Preparation", which detailed FDA approved arm preparation procedures. The blood center conducted a Donor Arm Prep validation in 3 facilities comparing two of the four approved arm prep procedures. The objective was to determine the efficiency of 2% chlorhexidine in 70% isopropyl alcohol (ChloraPrep® 1.5 mL Frepp® Applicator) to their current two-step Blood Donor Prep Kit. Study Design: The validation was divided into two parts. First, a site was selected to do a focused validation that required one specifically trained individual to do side-by-side scrubs and collect skin cultures for colony counts on 40 different volunteers/potential donors. The second part of the validation occurred in two additional sites where several different individuals were trained to do sideby-side scrubs and collect skin cultures for colony counts on 100 different volunteers/potential donors per site. RESULTS: of 99.0 was obtained. These results compared favourably with studies reported from the National Blood Service, UK, using other common disinfectants such as Medi-Flex. Most donors found the smell or green colour of the disinfectant to be acceptable (99% and 91% respectively), and 16 donors (3%) reported transient skin irritation. The majority of donors (91%) either preferred or did not object to the new disinfectant replacing the current preparation. Staff found the preparation to spill easily resulting in staining of clothing or environment in up to 78% of cases. A revised packet containing less disinfectant (1.5 mL rather than 2.2 mL) was found to be more acceptable to staff. DISCUSSION AND CONCLUSION: The Australian Red Cross Blood Service had been searching for a standard national disinfectant to replace the several preparations currently in use for donor arm preparation prior to venesection. The disinfectant Persist Plus TM containing 1.5 mL per swab was found to be suitable both from the bacteriological and end-user acceptance points of view.

S Ribault, L Grave, A Faucon, P Nannini, I Besson-Faure, Hemosystem, Marseille, France Background: Bacterial contamination remains one of the major risks associated with blood product transfusion. The kinetics of bacterial growth in red blood cells concentrates (RBCC) are particular due to the storage conditions at 4°C. Scansystem TM was studied as a valuable tool for detection of bacteria during the 42 days storage period of RBCC. Methods: RBCC were inoculated with Staphylococcus epidermidis, Staphylococcus aureus, Streptococcus agalactiae, Bacillus cereus, Klebsiella oxytoca or Yersinia enterocolitica to give a final concentration of 10 CFU/mL (number of bacteria in the inoculum was established by plate counts). Spiked samples were stored during 42 days at 4°C and a 5 mL sample was taken every 3 days for immediate unitary Scansystem TM analysis. At day 14 and day 28, another sample was analysed within a pool of 20 samples. Five mL of the pool was incubated for 45 minutes in a reagent mix to remove the red blood cells. In the second step (45 minutes) bacteria were fluorescent stained before being concentrated by filtration on a black membrane. The membrane was then analysed by laser scanning with the Scansystem TM and bacteria enumerated. Results were compared to the reference system of standard colony counts. Results: Bacterial growth was observed with Y. enterocolitica and K. oxytoca from day 6 and day 21 respectively. All the other strains did not grow but remained alive and active (with the exception of S. epidermidis which died after day 3). The day 14 and day 28 samples allowed the reliable detection of a single contaminated sample within a pool of up to 20 test samples for all the strains that were still alive. No false positive result was encountered in this study. Conclusions: The Scansystem TM analyser allowed the fast, sensitive and reproducible detection of bacteria in red blood cell concentrates stored at 4°C. Moreover, the system was sensitive enough for the simultaneous testing of pool of up to 20 samples. The method therefore lends itself to incorporation into high sample through-put screening programmes.

S Ribault, A Faucon, I Besson-Faure, Hemosystem, Marseille, France Background: Bacterial contamination remains one of the major risks associated with blood product transfusion. The Scansystem TM has been developed for the detection of bacteria in apheresis and whole blood derived platelet concentrates as well as red blood cells concentrates (RBCC). One of the important features of the system is the red blood cells elimination step, enabling the detection of low bacteria concentrations by solid phase cytometry. Methods: RBCC were inoculated with Staphylococcus epidermidis, Staphylococcus aureus, Streptococcus agalactiae, Bacillus cereus, Escherichia coli, Pseudomonas aeruginosa, Klebsiella oxytoca or Yersinia enterocolitica to give a range of final concentrations from 1 to 10 4 CFU/mL (number of bacteria in the inoculum was established by plate counts). Spiked samples were incubated at 37°C for 18 hours in small volume transfer bags. The sensitivity of the system was tested on pools of 10 or 20 samples only one of which was 'spiked sample'. Five mL of the pool was incubated for 45 minutes in a reagent mix to remove the red blood cells. In the second step (45 minutes) bacteria were fluorescent stained before being concentrated by filtration on a black membrane. The membrane was then analysed by laser scanning with the Scansystem TM and bacteria enumerated. Results * Discrepancies in sampling errors and collection area contamination.

Conclusion: The 2% chlorhexidine in 70% isopropyl (CHG/IPA) alcohol solution demonstrated efficacy comparable to the Blood Donor Prep Kit. The CHG/IPA prep offers the following advantages; 1) a one-step prep procedure, therefore reducing donor arm preparation time and 2) an alternative prep for donors allergic to iodine.

P-Y Wong, V Colville, V L White, H M Walker, R A Morris, E M Wood, Australian Red Cross Blood Service, South Melbourne, Australia BACKGROUND: The Australian Red Cross Blood Service was considering a number of strategies to minimise the bacterial contamination rate in blood collected from donors. We believed the use of an effective disinfectant in preparing the blood donor arm before needle insertion was most important. A study was performed in 2003 to evaluate the effectiveness of a new disinfectant for this purpose. STUDY DESIGN: A prospective study of bacteria present on blood donors' arms before and after disinfection with a disposable applicator containing 1% chlorhexidine gluconate with 75% ethanol (Persist Plus TM , supplier Becton Dickinson). Staff and donor feedback on the applicator was obtained using a questionnaire. SUBJECTS AND METHODS: Permission was asked randomly from 200 donors each from three types of blood collection venues-fixed, mobile, and regional (rural) to be enrolled in the study. The antecubital fossa of the non-donating arm was tested by a direct swabbing and plating technique pre-and post-disinfection with the antiseptic Persist Plus TM . Normal venesection and blood collection then occurred from the opposite arm. Data entry and analysis of the questionnaires was performed using Epi-Info 2002. RESULTS: Pre-disinfection, 56% of 616 donor arms had colony counts of <5 cfu/plate and 64% had counts of <10 cfu/plate. After disinfection with Persist Plus TM , 99% of donor arms had counts of <5 cfu/plate and 99.5% had counts of <10 cfu/plate. There was no significant difference in the bacterial counts from donors in different blood collection venues. The mean colony count for all donors postdisinfection was 0.39. Overall, the percentage reduction in bacterial counts 51A 2004-Vol. 44, Supplement were compared to the reference system of standard colony counts. Results: Irrespective of ABO blood group, red blood cell number was reduced by greater than 3 Logs. Once red cell depleted, bacteria in the supernatant could be efficiently recovered and concentrated by filtration, allowing the reliable detection of a single contaminated sample within a pool of up to 20 samples. The system was shown to be sensitive down to as low as 1 CFU/mL in the original sample (with the exception of S. epidermidis 5 CFUs/mL). No false positive result was encountered in this study. Conclusions: The combination of the red cell depletion step, new staining method together with the Scansystem TM analyser allowed the fast, sensitive and reproducible detection of low concentrations of bacteria in RBCC (down to 1 CFU/mL). Moreover, the system was sensitive enough for the simultaneous testing of pools of up to 20 samples. The method therefore lends itself to incorporation into high sample through-put screening programmes.

L Grave, A Raimondo, P Nannini, I Besson-Faure, Hemosystem, Marseille, France Background: The Scansystem TM bacteria detection device allows sensitive, specific and rapid bacteria detection in Leukocyte Reduced Apheresis Platelets (LRAP). The device was developed for testing a mix of 3 samples from 3 different LRAP. In order to extend its use to the bacteria detection in Whole-Blood Platelet concentrates (WBPC), samples from 6 unidoses were mixed and tested together. Methods: Unidoses of Leukocyte Reduced WBPC were inoculated with either Gram-or Gram+ bacteria at 10 CFU/mL. 24 and 30 hours after inoculation, 1 mL of the contaminated unidose was sampled and mixed with 2 to 5 mL of sterile products mimiking a mix of 3 to 6 samples. Then, 3 mL of the mix was immediately processed with the Scansystem TM bacterial detection device. Results: At 30 hours time point, bacteria were detected in 100% of tests. In addition, at 24 hours, the detection was positive for all tests including one sample spiked with rapid growth bacteria. The number of samples (3 to 6) in the 3 mL mix had no influence on bacteria detection. Conclusions: The Scansystem TM bacteria detection device allows sensitive bacteria detection for 3 to 6 samples from wholeblood platelet unidoses tested together as a mix. Performances are similar than those established for Apheresis Platelets i.e. 100% detection at 30 hours. This sampling procedure would allow a single bacteria detection test for several (3 to 6) unidoses of WBPC instead of 3 to 6 single tests. It could also be extended to the bacteria detection in non leukocyte reduced platelet products.

L Grave, F Olivieri, P Nannini, I Besson-Faure, Hemosystem, Marseille, France Background: Platelets can be stored in additive solution with benefits for human transfusion. The Scansystem TM bacteria detection device allows sensitive, specific and rapid bacteria detection in platelet products (PC) and its use could be extended to the testing of platelets stored in additive solution (PCsa). Methods: 2 to 4 days old PCsa and PC were obtained from the Blood Centers of Francfort (Germany) and Lyon (France), respectively. PCsa were prepared with 30% plasma and 70% additive solution (T-SOL, Baxter). PC and PCas were spiked with either Gram-(E.coli; K.oxytoca; S.marcescens) or Gram+ (S.epidermidis; B.cereus; S.aureus) bacteria at 10 to 10 3 CFU/mL. Samples were analysed immediately or 24 and 30 hr after spiking by Scansystem TM bacteria detection device and by culture method for detection and numeration of bacteria. Results: There was no interference between the additive solution and Scansystem TM Platelet Kit. The removal of plasma proteins led to a better labelling and detection of bacteria in PCsa compared to PC. Growth of most bacteria strains was significantly enhanced in PCsa leading to their detection 24 hr after an initial spiking as low as 10 CFU/mL. At 30 hrs after platelet contamination, bacteria were detected in 100% of platelet products stored either in plasma or in additive solution. Conclusions: The Scansystem TM bacteria detection device can be used for platelet products stored either in plasma or in additive solution. Clinical studies are currently being performed to assess the impact of additive media on the growth of clinical bacteria strains. They will precisely define the optimal timing of PCas routine testing by Scansystem TM bacteria detection device.

Y Choo, L Rudon, Z Czajkowska, I Rankin, C Whitsett, Department of Pathology, Mount Sinai School of Medicine, New York, NY Background. The risk of developing sepsis from transfusion of pooled platelet concentrates (PCs) is greater than the risk from single donor platelets. Many blood centers have implemented bacterial testing of single donor platelets but few have implemented comparable testing of platelet concentrate. Sixty-seven percent of platelet transfusions at our hospital are from PCs purchased from blood centers. These PCs are not leukoreduced. Measurement of pH and glucose on PCs has previously been demonstrated to be useful in identifying contaminated platelet products. Methods. PCs were inoculated in a laminar flow hood with heavy concentrations of S. aureus, S. epidermidis, B. cereus, and K. pneumoniae. Glucose and pH determinations were performed with Chemstrip TM (Roche) and S/P pH Indicator Strips TM (Allegiance) previously validated with a glucometer and pH meter, respectively. Baseline glucose and pH determinations were made prior to inoculation and daily thereafter. Based on the literature, concentrates with a pH > 6.9 and glucose ≥ 250 mg/dl were considered acceptable for transfusion. Beginning March 1, 2004 all PCs were tested prior to pooling. Products failing either the glucose or pH cutoff were excluded from the pool, and sent for aerobic and anaerobic cultures. Results. Of twenty seven PCs inoculated with bacteria, only 7/27 PCs had pH £ 6.9 (six pH 6.9, one pH 6.3) when glucose was lower than 250 mg/dl. pH was 7.2 in 20 PCs. Our PC screening results are summarized in the Table. During the same period there were five febrile non-hemolytic transfusion reactions reported after transfusions of PCs. Cultures from four pools were negative and the fifth was not cultured. Table. PC screening by glucose and pH measurements.

Conclusions. It is feasible to screen PCs by glucose and pH dipsticks in a large transfusion service. Changes in glucose levels appear earlier than changes in pH in PL732 bags. The number of PCs discarded using these screening criteria is acceptable (0.04%). The negative culture results from the PCs implicated in febrile transfusion reactions support the usefulness of this technique in identifying PCs with heavy bacterial contamination.

K E Puca, T C Boyer, C J Grygny, J L Gottschall, Blood Center of Southeastern Wisconsin, Inc., Milwaukee, WI Background: Measurement of pH is an acceptable procedure to meet AABB Standard 5.1.5.1 for the detection of bacterial contamination in platelet components. Use of pH strips is one way to meet this Standard. While cost-effective, this method is limited by a low sensitivity and high falsepositive rate. We present our experience in implementing a blood centerfocused program for detection of bacterial contamintion using pH paper in whole blood-derived platelets (WBPLC). Methods: Using pH paper (Hydrion Papers, Micro Essential Laboratory Inc. Brooklyn, NY) WBPLC are tested every 8 hours to ensure that an adequate supply of tested WBPLC are available to fill orders in a timely manner. The number and attribute (i.e. ABO/Rh type, CMV status) of WBPLC tested are determined daily based on the current available inventory of apheresis platelets. On average, 12 WBPLC are tested every 8 hours. Quality control testing of the pH paper is performed with known pH Buffer Solutions, 6.0 and 7.0 (Fisher Scientific, Fair Lawn, NJ) per shift. WBPLC are considered "acceptable" or negative if the pH is ≥7.0. If the products are not used within 8 hours of testing, they must be retested to be available for inventory or new products must be tested. If a WBPLC is found to have a pH < 7.0, a sample from the WBPLC is cultured using the bioMerieux BacT/ALERT system. The culture is held for 5 days, after which time the product is discarded. Results: After 2 months of testing, 42 of 961 WBPLC tested for pH have been "unacceptable" or found to have a pH < 7.0. None of these 42 WBPLC were culture-positive by BacT/ALERT, sampling bags. Four mL of product from each sampling bag was inoculated into an aerobic culture bottle and incubated for 5 days in accordance with manufacturer's recommendations. All culture bottles that gave a positive signal, their matching sampling bags and the original product (if available) were sent to a microbiology reference lab for further analysis. A Gram stain and subculture onto suitable agar media (for aerobic and anaerobic culture) were performed. A sample is considered BacT/ALERT true positive (TP) if the same organism is recovered from the bottle and the original platelet bag. A sample is considered false positive (FP) if NO growth is observed after subculture of the bottle, sampling bag and product. A sample is considered discrepant/contaminant if organisms are recovered from the BacT/ALERT bottle but NOT from the corresponding bags upon subculture. Results: Five samples were TP. Eleven samples were INITIALLY considered FP. Of these, 4 samples were determined to be positive due to an erroneous signal caused by a faulty cell in the BacT/ALERT. Seven samples were considered discrepant/contaminants since no organisms were recovered upon sub-culture of the matching bags.

Of the 961 WBPLC, the number of samples taken from each product for pH testing ranged from 1 to 10. Of the 42 WBPLC with pH < 7.0, 23 (55%) occurred on the first or second sample. Conclusion: To date, we have not detected any true-positives using pH as a method for bacterial detection in WBPLC. Our false positive rate of 4.4% was expected, and confirms that testing with pH paper has a high false-positive rate. WBPLC with a shelf-life of 3 days or more had a higher incidence of "unacceptable" pH. Pre-testing of WBPLC inventory is an efficient way of ensuring a readily available supply of platelets when needed.

G Boothe, T A Renner, J L Hupp, R M Giulitto, S L Wilkinson, Hoxworth Blood Center, Cincinnati, OH; C Bunch, Pall Medical Corporation, East Hills, NY Background: Pall® eBDS (Med Sep, Covina, CA) was implemented for 100% quality control testing of all Pheresis/Random Donor Platelets (Platelets). Concerns included: 1) delays in product release; 2) decreased shelf-life leading to increased wastage; and 3) high false positive rate, even though initial studies reported significant decreases in this rate. We report our experiences regarding testing and inventory management issues that we have observed with the implementation of eBDS at our facility. Methods: Diversion occurs with Apheresis collections but not Whole Blood collections. Sampling and testing is performed in accordance with manufacturer recommendations. Platelets are sampled 24 hours post manufacture (Day 0). Platelets are batched in groups of 10 and assigned a batch ID number. Sterile docking of the Pall® eBDS Sample Set to each Platelet routinely begins at 1200 Tues.-Sun. Each batch takes 1 hr. to complete. Batches are placed in segregated areas on climate-controlled (35 C) flat-bed agitators. Batch ID, location, batch incubation start time and temperature are recorded. Batch testing must be completed between 24-30 hrs. post sampling time. Batch testing averages 10 min. Result recording is electronic, as is download to the mainframe. Results: Initially, product release delays were seen due to occasional lack of sampling documentation (start times), repeat sampling due to under/over-fills (Pheresis only) and incomplete donor registration affecting start time assignment. These delays have been minimized since startup. Inventory release occurs early day 3, leaving >2 day shelf-life. Pheresis platelet wastage has initially increased (4.4% pre-eBDS; 6.5% post-eBDS). Random platelet wastage has initially decreased (23.1% pre-eBDS; 20.0% post-eBDS). Testing has identified 5 positives out of 4789 (0.1%) units sampled. One false positive (1/1008; 0.1%) has been observed with Pheresis testing. One false positive and 3 true positives (4/3781; 0.1%) has been observed with Random testing. All true positives contained coagulase-negative Staphylococcus. Conclusions: Operational delays have been minimized as the eBDS process continues. Delay to day 3 may adversely influence Pheresis wastage. Our positive rate is consistent with the manufacturer's target rate of 0.1%, and reflects the cumulative positive rate reported by other users (7/11,000; 0.06%).

R Fayed, C Taylor, S Linauts, Puget Sound Blood Center, Seattle, WA Background: On June 02, 2003, the BacT/ALERT 3D Signature was implemented for the detection of aerobic microorganisms in apheresis units at Puget Sound Blood Center. This evaluation presents results obtained during one year of bacterial screening. Methods: Apheresis platelets were collected into Trima platelet bags. Units were stored at 20-24°C for at least 24 hrs before sterile docking and transferring 10 mL of product to Charter Medical 39 units were lost due to technical difficulties during docking and transferring into sampling bags. Conclusion: The data demonstrate that the BacT/ALERT can be used to detect common microorganisms found in platelets within a reasonable detection time. This technique was able to identify 5 true positive samples, thus preventing them from being transfused into patients. A discrepant/contaminant status can occur due to accidental introduction of laboratory contaminants into the culture bottle or to defective media. Further work with the manufacturer is needed to decrease these media problems as well as the number of false positives.

J A Rios, J Norton, R J Benjamin, New England Region, American Red Cross, Dedham, MA BACKGROUND: Bacterial contamination is a common cause of blood recipient adverse reactions. Both the AABB and CAP require blood centers or hospital transfusion services to implement methods to limit bacterial contamination of platelets. METHODS: All platelets collected by apheresis were tested using bioMerioux BacT/Alert system using only an aerobic bottle. A sample from the primary platelet bag was obtained at least 24 hours after collection for inoculation into a BacT bottle, and the bottle was monitored for CO 2 production until the outdate of the platelets. Positive BacT bottles and their associated apheresis platelets were sent to a microbiology lab for identification. The corresponding apheresis platelets were not used for transfusion. RESULTS: During March and April 2004 our blood center collected 6,225 apheresis platelets using the Cobe Spectra, Cobe Trima, and Baxter Amicus cell separators. Seven donations were associated with a positive bacterial screening result. In five donations (71%) the initial positive bacterial screening results were classified as false positives based on negative culture results of the BacT bottle and/or the associated single donor platelets. In one donation (14%) the initial positive bacterial screening test result was classified as a true positive based on the microbiology culture results of the BacT bottle and the associated apheresis platelets (Staph coag negative). In another donation the initial positive bacterial screening test result (14%) was associated with a positive microbiology test result in the BacT bottle (Corynebacterium species). The corresponding apheresis platelets were transfused without adverse events in the recipients. Bacterial detection resulted in a 12 hour delay in release of platelets and required 3.5 extra FTE. Confirmatory Culture Results of Initial Positive BacT Cultures 53A 2004-Vol. 44, Supplement CONCLUSION: While bacterial screening with the BacT/Alert system can be used to prevent septic transfusion reactions, the system is associated with high rate of false positive test results. The organisms associated with contaminated apheresis platelets are normal skin flora. The rate of bacterial contamination in apheresis platelets is 1/6,225.

Background: In order to minimize the chances of contamination, a sampling device is generally used for inoculation of platelet sample into a BacT/ALERT culture bottle. As a result, it becomes difficult to control the 4 mL sample volume as required in the package insert. We studied the effects of sample volume on detection time using apheresis platelets spiked with nine different bacteria. Methods: Less than two day old leukocyte-reduced apheresis platelets were spiked with one of nine freshly cultured bacteria (Escherichia coli, Serratia marcescens, Staphylococcus epidermidis, Staphylococcus aureus, Enterobacter cloacae, Klebsiella oxytoca, Bacillus cereus, Streptococcus mitis, and Propionibacterium acnes) at two levels (13-75 CFU/mL and 87-727 CFU/mL). Within one hour, 2, 4, 6, and 8 mL of the inoculated platelets were aseptically injected into BacT/ALERT aerobic culture bottles. All bottles were placed in the BacT/ALERT incubator for incubation up to 7 days or until positive. In addition, standard agar plate cultures were simultaneously set up to determine spiking concentrations. Aliquots from each positive culture bottle were further plate cultured for Gram staining, bacteria isolation and identification to verify bacterial species. Results: P. acnes, an anaerobe, showed no growth in aerobic culture bottles. All culture bottles inoculated with platelets spiked with E. coli, S. marcescens, S. aureus, E. cloacae, K. oxytoca, and B. cereus gave a positive signal between 9.2 and 12.9 hours, while culture bottles with S. epidermidis and S. mitis required 15.3 to 22.4 hours for a positive result. Although, in general, the time to positivity was reduced slightly with higher volume input, overall there were no statistically significant differences among different sample volumes for both spiking levels (p > 0.4 by Repeat ANOVA F-test). The greatest differences were observed between the 2 mL and the 4 mL inputs with a reduced mean detection time of 15-17 minutes. Conclusion: Our data indicate that overall there is no statistically significant difference in detection or time to positivity among platelet sample volume between 2.0 and 8.0 mL inoculated into BacT/ALERT aerobic culture bottles for the nine bacteria studied.

K B Peck, K D Horn, G E Tegtmeier, J E Menitove, Community Blood Center, Kansas City, MO Introduction: The AABB Standards (22 nd Edition) require methods to limit and detect bacterial contamination of all platelet components. To address the latter issue, we implemented bacterial testing of Platelets Pheresis products using the BioMerieux BacTAlert system on September 2, 2003. We report the results of testing through April 30, 2004. Methods: Platelets Pheresis products are tested for platelet count on day of collection, held for 24 hours, and sampled for bacterial detection.(Units which have been sampled are labeled and released for distribution). Units are divided into batches according to collection times; several units are sampled under a laminar flow hood at two hour intervals. We connect a sampling device with an attached syringe to each unit using a sterile connection device. We remove four mLs of well-mixed platelets into the syringe and transfer the contents to an aerobic culture bottle placed subsequently into the BacTAlert. We monitor the BacTAlert 24/7 via a remote computer link to the distribu-tion department. When a positive occurs, the computer system automatically prints a recall report for use by distribution in obtaining the positive product. Positive bottles and the product, when available, are sent to a local hospital for STAT gram stain, culture and organism identification. Results: Six "positive" cultures have occurred among 7,315 apheresis collections. In all but the last case, we retrieved the products prior to transfusion. Two positives were device false-positives, and two others appear to have been technique-related, yielding a false-positive rate of 1 : 1828. The two remaining positives (E. coli and S. epidermidis) appear to be donor-related, an isolation rate of 1 : 3658. Platelet outdates have not increased since the implementation of bacterial testing.

Conclusion: Using the BioMerieux BacTAlert system, we interdicted 3 units (of 11,200 units) likely to have transmitted bacteria. Bacterial testing using the procedure established at our Center reduces risk of transfusion associated-sepsis. The procedure, carried out by well-trained technicians, has resulted in a very low level of erroneous positives and has not adversely affected platelet outdating. Background: With the introduction of in run quality control culturing of platelet products for bacterial contamination at centralized sites, distribution of the products often occurs with "negative to date" bacterial testing results. If a culture was to become positive, the collection center will need to notify the receiving hospital in a timely manner in order to facilitate product withdrawal or physician notification. We validated a remote notification software system BacT/Notify (bioMerieux, Durham NC) that connects to a BacT/ALERT Microbial Detection System (bioMerieux, Durham NC) which allows automated notification of positive culture results via fax, email or pager as well as access to most recent available culture status through a secure web server. Study Design: The BacT/Notify system was linked to the BacT/ALERT (BTA) instrument and a simulated Blood Bank Information System (BIS). The BacT/Notify was connected via a network cable to the hospital email system and internet and to an analog phone line for faxing capability and remote maintenance. Culture results were generated by the BTA by loading un-inoculated plastic culture bottles and manually changing the status to positive or negative as required. Four to 8 accessions (each accession included one aerobic and one anaerobic culture bottle) were loaded daily to obtain 20 runs (one run = 4 accessions). The BacT/Notify system sent notifications of positive culture results to remote clients as email messages and faxes and stored the data in a database accessible via a secure web server. Results: The BacT/Notify received the culture results from the BTA and sent the culture results to the simulated BIS unaltered. Distribution of positive culture results to remote clients as email messages and faxes occurred as expected. All culture results (including negative to date and final results) were available for access via a secure web page. The system logged all interactions between BacT/Notify and the remote clients, including transmission of positive results, the receipt of positive results by the remote clients, and access of the web page. Conclusion: The BacT/Notify system allows remote client sites to be notified quickly of a positive culture result from a platelet product and allows access to the most recent culture status prior to issue of a product. The BacT/Notify facilitates timely withdrawal of the culture positive product from inventory or notification of clinicians if the product has been transfused. 

Background: The BacT/ALERT microbial detection system utilizes a color sensor in the culture bottle to monitor the presence of carbon dioxide produced as a result of bacterial growth and metabolism of substrates in the culture medium. However, due to safety and other logistical concerns, a known positive sample containing live bacteria for routine quality control purposes may be inappropriate. We have evaluated two types of material which can potentially be used for this purpose: 1) preserved bacterial culture and 2) sterile carbonated fluid. Methods: 1) Quanti-Cult ® (Remel Inc., Lenexa, KS) products containing specified number of preserved Escherichia coli or Staphylococcus epidermidis were evaluated for their bacterial viabilities and counts (CFU) with the standard agar plate method. Sterile PBS was added to the rehydration fluid at different time points to determine if there was any effect on the detection time by the BacT/ALERT system. 2) A commercial carbonated drink (e.g. Sprite) was used to evaluate the effects of injection volume, time and method of culture bottle inoculation on the time for a positive signal in the BacT/ALERT system. Results: 1) Warm, sterile PBS could be added to Quanti-Cult rehydration fluid either before or after the rehydration of preserved bacteria. It was found that bacteria were viable for at least one hour after rehydration with colony counts meeting the product specification of 10-100 CFU. The detection time by the BacT/ALERT was comparable to the culture bottles which were inoculated with platelets spiked with the same amount of freshly cultured E. coli (~12 hours) or S. epidermidis (~18 hours). 2) For carbonated fluid, an inoculation volume of 2-4 mL was required for a positive signal. More importantly, a minimum pre-incubation time of 90 minutes for the culture bottles was required prior to inoculation. The pre-incubated bottle might be inoculated while in the incubator. Alternatively, the bottle might be removed for inoculation; however, it must be replaced into the same position within 10 minutes of inoculation. Conclusions: Our data indicate that 1) the commercial preserved bacteria are suitable for routine QC use to check on the growth promotion of culture medium and the detection system; 2) carbonated drink can be used to artificially generate a positive reaction in the culture bottle; however, a standard operating procedure must be established and followed for a reliable outcome.

S Wendel, L E Morato, R Fontão-Wendel, Hospital Sirio Libanes Blood Bank, Sao Paulo, Brazil Background: Bacterial contamination of blood products is a concern worldwide. In the past 7 years, we did a routine screening for all packed red cell (PRC) and single donor platelet concentrates (SD-PLT) using BacT/ALERT ® aerobic culture bottles. However the bottle formulation has been changed, which contributed to a decrease in the rate of contamination. Methods: data were split in two phases: I (Jan/97-Apr/01, Pedi-BacT TM bottles) and II (May/01-Mar/04, BacT/ALERT ® PF bottle). For each component and phase, a 2.0 ml aliquot is inoculated within 4-6 hours after collection and incubated for at least 24 hours. If culture is negative, component is labeled as negative to date and released for transfusion, while the 7-day incubation was maintained. All initially positive cultures were retested and if still positive, the agent was identified. Results: For Phase I, 31556 PRC and 5541 SD-PLT were tested, with positive signal in 88 PRC (0.28%) and 20 SD-PLT (0.36%). For Phase II, 18594 PRC and 3464 SD-PLT were tested, with positive signal in only 3 PRC (0.016% OR = 17.3 IC95 5.7 to 85.6) and 1 SD-PLT (0.029% OR = 12.5 IC95 2.0 to 520.0). For Phase I the contamination was mainly by Propionibacterium sp (n = 102, mean detection time = 107 hours), with one pathogenic bacteria (S. epidermidis-PRC, +signal = 24 hours); the remaining 5 cases were represented by doubtful pathogenic agents (bacteria/mean detection time: S.hominis/17 h; S.sacharelyticus/72 h; coagulase negative Staphylococcus sp./96 h; C.amycolatum/168 h). For phase II, S. epidermidis (2 cases-PRC) and S.aureus (1 case-SD-PLT) were detected with mean incubation time of 24 and 13 hours, respectively. Conclusion: The incidence rate of bacterial contamination dramatically decreased solely on changing inoculation bottles (0.28% ¥ 0.016%-PRC, and 0.36% ¥ 0.029%-SD-PLT, p < 0.001). We wonder whether bacterial contamination incidence is really low, or are we facing false negative results because of the short delay time for inoculation and/or small inoculum. Anyway, we showed a low, albeit consistent positive bacterial contamination, which has been promptly prevented by this measure. As from this study, we extended the inoculation period for a minimum of 12 and 24 hours for PRC and SD-PLT, respectively, but data are not available yet.

R R Gammon, D G Richards, South Florida Blood Banks, Lake Park, FL Background: As of March 1, 2004 the AABB Standards required, as part of the accreditation process, that there be a method to limit and detect bacterial contamination in all platelet components. We attempted to determine the incidence of bacterial contamination of platelets collected by apheresis at our blood center after implementation of bacterial detection. Study Design: We used a culture method that detected CO 2 production, a marker for bacterial growth, to determine contamination of leukocyte reduced platelets collected by apheresis. Platelets were stored at 20-24 C with continuous gentle agitation for 24 hours post-collection. Samples were taken at the same time that the plateletpheresis was split and then inoculated under aseptic conditions at a laboratory bench top into aerobic culture bottles. Cultures were incubated for four days (until expiration); platelets with negative cultures were released for transfusion after 24 hours. If a culture became positive after the platelet was released, the consignee was notified and any nontransfused split products were discarded. Reports of transfusion reactions from consignees were monitored. Donor service personnel were in-serviced on strict adherence to blood center policies and the clinical significance of properly preparing the antecubital region before phlebotomy the week prior to implementation of bacterial detection. Results: From 3/1/04 to 4/15/04, 1,029 leukocyte reduced platelet donations collected by apheresis were tested. 2/1029 (0.19%) had positive cultures at a mean of 23.0 hours incubation (range 18.8 to 27.2 hours). From the products with positive results, one was transfused, the remainder was discarded. For the time period reviewed, there were no reports of transfusion reactions due to bacterial contamination of any of the 2,152 plateletpheresis components that were tested and released. Conclusions: We report rates of bacterial contamination of 1 per 1,076 plateletpheresis components produced which is consistent with published results. 1 We did not note excessive contamination rates without the use of a laminar flow hood during culture inoculation. Staff education shortly before testing implementation may have minimized bacterial contamination and resultant product discard. 1. AABB Association Bulletin 03-12.

C A Conry-Cantilena, J W Gibble, Greater Chesapeake and Potomac American Red Cross, Baltimore, MD Background: Automated culture of all plateletpheresis products allows detection and isolation of contaminating organisms. This report describes isolation of a potentially pathogenic organism and interdiction of the platelet product. Case report: A 43 yo female repeat volunteer plateletpheresis donor underwent routine screening and uneventfully donated a 554 ml platelet/plasma product from a 5 liter WB processed procedure on the Amicus blood cell separator. Per SOP, an aliquot of the product was tested for bacterial contamination using an automated culture system. At 16.7 hours, the central monitoring system alarmed. The culture bottle grew betahemolytic Group C streptococci. On follow-up, the donor was not aware of any sources of an infection at the time of her donation, felt well, though reported a medical history of mitral valve prolapse. She worked as a large animal handler, often helping with horses, and had treated a horse's infected wound with her bare hands within 7 days prior to her platelet donation. She noted that the skin on her hands was often cracked and fissured. Her repeat blood culture 2 weeks after the donation was sterile. Conclusion: It is likely that this unusual organism was not a microbial contaminant but that donor had beta-hemolytic Group C streptococci bacteremia at the time of her platelet donation. This organism is implicated in wound infections and equine mastitis in veterinary medicine. It may cause pyogenic infections in humans. Bacterial culture was effective in interdicting this infected platelet component. Case Report A healthy 50 year old male, on routine blood work had a WBC of 5.3 ¥ 10 9 /mm 3 with 54% blasts. Diagnosis was AML with M1 subtype with standard cytogenetic study showing a male karyotype with trisomy 11. Induction chemotherapy was given with Daunorubicin and Cytarabine (3 + 7) on ECOG protocol. He developed neutropenic fevers Day 7 with initial blood and urine cultures being negative. He was on fluconazole anti-fungal prophylaxis. His Day 14 Bone marrow biopsy showed residual leukemia, he was re-induced with same chemotherapy which caused prolonged neutropenia. On hospital day # 30 (day 9 of re-induction chemotherapy), he developed scattered, varying size non-blanching erythematous lesions on hands and feet and low-grade fever. White count was 0.7 ¥ 10 9 /mm 3 . A skin biopsy showed infectious vasculitis with vaso-invasive fungus. Blood cultures grew Fusarium species. Despite standard voriconazole therapy for 4 days, his lesions progressed with positive blood cultures with new lesions occurred diffusely on his face, arms, chest, abdomen and legs. Total body CT scan was unremarkable. Failure of treatment in this prolonged neutropenic patient (WBC 0.4 ¥ 10 9 /mm 3 ) prompted initiation of granulocyte infusions with GM-CSF (5 mcg/kg) along with liposomal amphoterecin B (Ambisome at a dose of 5 mg/kg). Daily granulocyte infusions followed by GM-CSF and Ambisone 12 hours later was given for 6 days. Granulocytes obtained by prednisone stimulation of donor had average yield of 1.75 ¥ 10 10 cells (range 1.3 to 3.2). Results Treatment was tolerated with no significant pulmonary compromise or side-effects. His skin lesions stabilized two days post therapy and no new lesions seen after 6 days. Subsequent blood cultures were negative and white count was 1.9. Patient was discharged home on daily IV Ambisome on hospital day # 53 with remission of the leukemia. The GM-CSF was discontinued prior to discharge when the WBC was 15.3 (ANC > 10,000.) Conclusion Use of Amphotericin B with Granulocyte transfusion is controversial due to possibility of pulmonary compromise. Use of GM-CSF is not common in such cases due to potential negative effects. However when standard anti-microbial modalities fail consideration should be given to use of this triad of therapy in immunocompromised patients with rapidly progressive potential fatal fusarium infection.

A Insunza, I Romon, L Gonzalez-Ponte, V Hermosa, Banco de Sangre y Tejidos de Cantabria, Santander, Spain BACKGROUND: Cryopreserved PBPC infusion is associated with adverse events that can be life-threatening and even fatal, but information from large series is lacking. We have prospectively recorded our experience in the last 5 years. STUDY DESIGN: 275 infusions of autologous (248) or allogeneic (27) PBPCs cryopreserved with 10% DMSO were performed in 263 patients. After premedication with corticosteroids and antihistamines, cells were thawed and infused through a standard transfusion set at a controlled flow rate. Signs and symptoms, blood arterial pressure (BP) and heart rate were recorded after each bag. For multivariate analysis Logistic Regression was applied. RESULTS: The median volume infused was 420 ml (40-1640) and the median flow rate 14.1 ml/min (3.5-30). Bad taste (59%), throat itching (51%) and cough (34%) were the most common effects. Nausea (12%), cold sensation (9%), hot sensation in the face (9%), vomiting (8%) and abdominal pain (3%) were less common. Diarrhea, flushing and dyspnea were observed in less than 2% of patients each. Cardiovascular changes are shown in the following table

No patient had bradycardia below 50 beats per minute and only 2 below 60 beats per minute. There were few clinically significant (one case of bronchospasm and one case of syncope on the next day possibly related to arrhytmia) and no life-threatening events. Side effects and cardiovascular changes were not associated with the volume infused. In multivariate analysis the most significant associations (p < 0.01) were bad taste with lower age, hot sensation in the face, nausea and vomiting with female sex, increase in blood pressure greater than 10% with red cell concentration in the product and decrease in systolic blood pressure greater than 10% with G-CSF mobilization. Slowing the flow rate was usually effective in relieving symptoms. CONCLUSION: With simple precautions and simple clinical monitoring, cryopreserved PBPC infusion is a safe procedure with a very low incidence of serious adverse events. Side effects are not associated with the volume infused but patient age and sex and composition of the infused product seem to play a role.

Background DLI aiming to prevent rejection after allogeneic bone marrow transplantation (BMT) has fewer cases reported, although with successful outcomes. Ommen syndrome is a form of severe combined immunodeficiency with clinical symptoms present as early as a few weeks of life. Key symptoms are erythematous rash (ER), hepatosplenomegaly, lymphadenopathy accompanied by recurrent infections and alopecia. Laboratorial findings include B-lymphocyte counts decreased whereas T-lymphocyte counts are increased. It is fatal if untreated. Early BMT is required. Case report A nine-month-old boy was admitted. He had ER in the neck area when he was one month old. Two weeks later he was hospitalized due to severe stomatitis and dehydration. When he was two months old he was hospitalized again because of dermatitis and acute otitis media. He was not gaining weight and when he was 3 and a half months old he had bronchopneumonia and sepsis. In his fifth month he had an endocarditis. At this time, immunodeficiency was suspected. Upon admission, the physical examination was remarkable for ER, hepatosplenomegaly, and axillary lymphadenopathy. Immunophenotyping analysis showed lymphocyte NK CD56+CD2+CD16(-)CD57(-): 14%; lymphocyte B CD19+CD22+CD45RA+HLA-DR+CD10(-):

7.0%; lymphocyte T CD2+CD3+CD5+CD7+HLA-DR+: 79%. T cells were divided into CD4+:0.4%; CD8+:30%; CD4+CD8+: 0.1%; CD4 (-)CD8(-): 69.4%. The diagnosis of Omenn syndrome was the one that fit better. Due to the severe nature of the disease, it was decided that a BMT should to be carried out. There was a twelve-year-old son, compatible, and a non-myeloablative conditioning regimen was performed by using bussulfan 8 mg/kg and fludarabine 30 mg/m 2 /day for 5 days. Cyclosporin was used for prophylactic therapy against graft versus host disease (GVHD). Around day +300 after the BMT, a chimerism assay (amplification of locus VNTRs/STRs) showed a reduced number of donor cells. Moreover, the patient's skin lesion relapsed. Because of indications of graft rejection, a DLI was planned. Harvesting was performed in a COBE Spectra Cell Separator. Volume processed was 5000 ml, collected volume was 139 ml, leukocytes collected was 80 ¥ 10 6 . CD3+/kg was 4.74 ¥ 10 8 . 1 ¥ 10 5 CD3+/kg was infused. After 3 months, a chimerism assay demonstrated about 55-65% of nucleated cells proceeding from the donor. Two years later, the patient shows a good general health condition, without recurrent infections and gaining weight. Conclusion The challenge of holding on a chimerism, which permits reconstitution of the immune system without development of severe GVHD, is a goal to be reached. However, DLI gives evidence of efficacy against graft rejection after nonmyeloablative BMT. Background: Although histocompatibility differences are thought to initiate the anti-host response to the graft, a syndrome with pathology identical to graft-vs-host disease (GVHD) can spontaneously occur after autologous haematopoietic stem cell transplantation (AHSCT). This rare "autoagression TRANSFUSION 2004-Vol. 44, Supplement syndrome" may also be induced by the administration of cyclosporin A (CyA) after AHSCT. It has been reported that 33% of series of patients developed CyA-induced autologous GVHD. This fact occurred in 5 out of 12 patients with Multiple Myeloma conditioned with melphalan. It was significantly more frequent in patients older than 33 years and a significant negative association was found with HLA-A10 and HLA-B16. Both spontaneous and CyAinduced autologous GVHD are generally self-limited and mostly confined to the skin. We report a patient undergoing AHSCT for Multiple Myeloma in whom a severe GVHD has developed spontaneously Case report: A 64-year old man was diagnosed with IgG lambda Multiple Myeloma. Treatment with six cycles of VBCMP/VBAD chemotherapy resulted in partial respose, and peripheral blood progenitor cells were harvested and cryopreserved. The patient received conditioning therapy with high-dose melphalan (200 mg/m2). A total of 2 ¥ 10 6 CD34+ cells were transplanted and, apart from neutropenic fever, there were no complications during the early post-transplant period. Filtered and irradiated blood components including 2 units of RBCs, a five random-donor units of platelets and an apheresis platelet concentrate were transfused before haematopoietic recovery was achieved on day +12 (neutrophil count of 1 ¥ 10 9 /L and platelet count above 20 ¥ 10 9 /L).Three weeks following the transplant he developed a widespread pruritic erythematous rash including the palms of the hands and sole of the feet, clinically suggestive of grade III acute GVHD. Furthermore, the patient had a massive diarrhea. A skin biopsy confirmed the diagnosis. The patient was started on immunosuppressive treatment with oral methylprednisolone (2 mg/kg/day) which led to gradual improvement of the symptoms. Since blood products administered pre and during transplant procedure were irradiated and the analysis of microsatellite DNA showed a unique allelic profile, the diagnosis of spontaneous GVHD was confirmed. The patient's HLA type was A*68 B*07,53. Conclusions: This unusual case of spontaneous GVHD after AHSCT presents all the risk factors identified in cases of CyA-induced GVHD leading to the supposition that CyA only enhances the failure of the regenerating immune system to acquire tolerance to self MHC antigens in the post-transplant period. It is a collaborative program between several institutions aiming to collect, process, and store cord blood units, in accordance to national and international (NETCORD) requirements. The processing of its units is completely automatized, using a cellular separation and concentration system (SEPAX-Biosafe). Cryopreservation of cord blood units (CBU) is done using a controlled freezing and storage system with liquid nitrogen at -196°C (Bioarchive System TG3626); the CBU contain 10% DMSO, 1% Dextran 40 and 0.8% hespan as cryoprotectants. The validation studies include: initial and final nucleated cell count, CD34+ determination, flow cytometry cell viability (FACScalibur, BD), as well as serologic and microbiologic studies, ABO/Rh and HLA. The search for transplant units is done based on HLA (A, B, DR, DQ), the number of nucleated cells (NC) and the viability of CD34+ cells. OBJECTIVE. To present and evaluate the results of the first transplant done in Mexico with mexicans donor and receptor, from the time we obtained the unit until the follow-up of the transplant results. MATERIAL AND METHODS. A 10 year-old male patient, weighing 35 kg, with acute lymphoblastic T cell leukemia (L2 FAB), in his third complete remision, entered the INP transplant program. We looked for CBU at the CNTS and found a unit with 3/6 antigen match between donor and recipient (considering broad A, B, DR and DQ high-resolution alleles), with total NC of 6 ¥ 10 8, 1.1 ¥ 106 viable CD34+ cells in 25 ml. We started outpatient preparation on day -8 before the transplant, with external TBI, cyclophosphamide, and ATG. He was hospitalized two days before the transplant in an isolated room with HEPA filters. The CBU of 25 ml was thawed before the infusion and a v/v dilution with Rheomacrodex-albumin solution. We obtained a final volume of 50 ml, which was infused to the patient through a central catheter. RESULTS We followed the patient with CBC and chimeric DNA salting out tests(see table) . On day +20 we performed chimeric DNA salting studies in the recipient and found a mixed chimerism (60% QM12-40% QM13). The patient was discharged on day +30. Outpatient follow-up on day +57 showed that the patient was in excellent state of health and in complete remission, without evidence of GVHD.

W G Ward, J Gardelli, J Pellom, J Sweetenham, University of Colorado Hospital, Denver, CO Background: Allogeneic peripheral blood progenitor cells (PBPCs) collected by apheresis are used to treat various high-risk malignancies. To obtain the optimal number of cells for infusion, two consecutive days of donor apheresis collections are usually required. PBPCs obtained on the first day of collection are stored overnight, and infused the following day with freshly collected PBPCs. This study was undertaken to determine the effect on cell recovery and function of storing PBPCs overnight using a methodology of product dilution and refrigeration, and to provide a retrospective validation of our institution's procedure. Methods: PBPCs and extra plasma from HLAmatched donors were collected using the COBE Spectra. The first collection was weighed, and samples removed for assaying sterility, WBC, viability by trypan blue, CFU-GM and CD34/7AAD by flow cytometry. Products were diluted to 1 ¥ 10 8 WBC/ml using donor plasma or Isolyte, and stored in a 1-6°C monitored refrigerator overnight in the original collection bag. The next day, prior to infusion, samples were again removed for similar testing. After infusion, infusion reactions, product sterility cultures and engraftment were monitored for each patient. Results:

There were no adverse reactions to product infusion, all pre-infusion product sterility cultures were negative, and engraftment times were within expected ranges. Conclusions: Allogeneic PBPC products may be safely stored overnight and infused with other PBPC products without significantly changing important cell counts or function, or patient outcome.

E J Uhlmann, D M Lublin, D S Sempek, J F Dipersio, L T Goodnough, Washington University, St Louis, MO Background: Delayed engraftment is a significant risk for patients undergoing autologous stem cell transplant. For QI, we identified laboratory factors that influence stem cell viability and outcomes. Methods: We analyzed patients undergoing stem cell collection and infusion who had delayed or failed engraftment in 2003 (n = 9) and 14 controls. This initial analysis showed a strong effect of pre-freeze processing time (PFPT) of the product on engraftment. We therefore prospectively evaluated the effects of PFPT and WBC concentration on post-thaw CD34+ cell recovery. Patients (n = 41) were included in the study if collection goal (>2.5E6/kg CD34+ cells) was met with one apheresis. We retrospectively analyzed PFPT and neutrophil recovery time (ANC ≥ 500) in 241 patients. Patients (n = 35) were in the delayed group if ANC ≥ 500 time was ≥ 19 days post-transplant (mean 13.7 + 5.5), with others in the control group (n = 206). Results: Delayed or failed engraftment was associated with prolonged PFPT: 442 minutes vs. 226 minutes for controls (p = 0.001). Product cell concentrations in the two groups were 2.87E8/l and 1.88E8/l (p = 0.124), respectively. In 41 patients studied prospectively, we found a negative association of CD34+ cell recovery with PFPT (R = 0.25), with 70% and 50% progenitor cell survival occurring after 72 minutes and 230 minutes of PFPT, respectively. Total nucleated cell recovery was high during these conditions (mean 91%) but no correlation with CD34+ cell recovery. In our review of 241 patients, we found that the average PFPT for controls was 223 minutes, while for delayed engrafters it was 337 minutes (p < 0.01). Product nucleated cell concentration in the two groups were 1.96E8/l and 2.54E8/l (p = 0.04). There was no difference in the total pre-freeze CD34+ cell yield (8.47E6/kg and 1.0E7/kg, p = 0.46) Conclusions: Prolonged pre-freeze processing time results in loss of progenitor cell viability, and engraftment failure or delay. Progenitor cell collection products should be frozen in less than 2 hours in order to ensure greater than 70% post-thaw recovery. Elevated total nucleated cell concentration is also associated with loss of progenitor cell viability, therefore cell concentrations greater than 2E8/l should be avoided. Background. An optimal system that would enable the liquid storage of hematopoietic stem and progenitor cells (SPC) for several days or weeks could facilitate organisation of autologous or allogenic transplantation procedures by avoiding the necessity of cryopreservation in some cases. Low (3%) O 2 concentration maintains primitive stem cells in liquid cultures at 37°C without affecting the committed progenitors (Ivanovic et al, Stem Cells, in press). Also, more physiologic CO 2 concentrations could improve the cell survival. We then compared the maintenance of CD34+ cells mobilized by G-CSF (6 patients with lymphoma or solid tumors) stored at +4°C in air (21% O 2 , 0.05% CO 2 ) and in gas mixture with 3% O 2 and 6% CO 2 (GM). Study Design: Two hundred thousands of CD34+ cells purified from cytapheresis products by using Myltenyi Mini-Macs technique were stored in 4 ml of liquidstorage medium (S.3, Stem a, Saint Clement, France) in closed flasks with 75 ml of air or GM for 3 and 5 days. The GM was obtained in incubator insert chambers by using PROOX and PROCO 2 gas controllers (Biospherix, Redfield, NY). The flasks were exposed 45 minutes to GM or to air at 20°C before being rapidly closed, placed at 4°C and stored for 3 or 5 days. Results: The number of viable CD34+ cells (trypan blue test) and "non apoptotic cells" (flowcytometry (FC) analysis: Ann V-, PI-), decreased in air (40% of initial values at day 5), but were well maintained in GM (80% of initial values). Onset of apoptosis (Ann V+, PI-) and of cell death (Ann V+ PI+) was evident at day-3 and -5, respectively, only in air. The absolute number of cells in G0 phase of cycle (FC analysis: PI/Ki67) was maintained (day-3) and increased (day-5) in GM, while it was 6-7 times lower in air than in GM after 5 days. The CFU-GM doubled at day-3 and at day-5 with GM, while, they decreased by 2 times at day 5 in air. BFU-E were maintained at day-3 in both conditions, and decreased at day-5 (50% and 70% with respect to the Day-0 values for air and GM, respectively). CFU-Mix were well maintained in air, and their number increased three-fold at day-3 and day-5 in GM. Conclusion: Liquid storage of CD34+ cells at 4°C with low O 2 (3%) and high CO 2 (6%) concentrations could significantly improve their maintenance. Furthermore, it enhances simultaneously their clonogenic activity and entering in G0, a phenomenon never described before. More experiments are necessary to elucidate this phenomenon, to study the maintenance of more primitive stem cells (SRC) in these conditions and to adjust better the concentrations of O 2 and CO 2 in order to optimise the liquid storage of CD34+ cells. Background: Leukocyte depletion of blood products is mandatory in France. Used leukocyte-depletion filters (LDF) containing hundreds of millions of WBC, are discarded. Steady-state blood contains low quantities of stem and progenitor cells that are most likely retained on LDF. We developed a technique to recover leukocytes from the LDF, to purify the CD34+ cells and to assay their clonogenic potential, capacity of expansion in liquid cultures, and the capacity to be induced into all three hematopoietic lineages: erythroid, myeloid, and megakaryocytic. Methods: After leukocyte depletion of whole blood LDFs connected with the empty collection bags were used (<24 h after blood sampling). The cells were washed from the filters into a transfer bag that was then welded and centrifuged to isolate a buffy-coat. The plateletrich supernatant was discarded, the buffy-coat recovered and MNC (mononuclear cells) isolated on ficoll density gradient. The CD34+ cells were then purified by using Miltenyi magnetic bead columns. Results: One LDF contains 447 ± 53 ¥ 10 3 CD34+ cells and 158 ±15.8 6 ¥ 10 3 CFU-GM (45.5/10 5 MNC) (n = 40, donor age = 36.4 ± 13.8 years). There were no differences in yield of CD34+ cells and CFU-GM with respect to the age of donors (18-62 years). The number of CFU-GM/100 ml of the donor blood was always correlated with the number of CD34+ cells. The LDF-derived CD34+ cells and CFU-GM could be expanded ex-vivo by culturing the MNC for 7 days in a serum-free medium supplemented by cytokines (G-CSF, SCF, MGDF (100 ng/ml each)). However, the expansion factor was very heterogeneous for various individual samples, reaching 60 fold in same cases. In differentiation experiments, the purified CD34+ cells recovered from the filters differentiated in megakaryocytes (SCF + Tpo), hemoglibinized erythroid cells (SCF + Epo), eosinophyls (SCF + IL3) and monocytes and granulocytes (SCF + GM-CSF or G-CSF). Conclusion: It is possible to recover functionally preserved CD34+ cells from LDF. These cells differentiate at least in three myeloid lineages and could be expanded ex vivo. We are currently optimizing conditions of ex-vivo expansion of purified LDF-recovered steady state CD34+ cells and evaluating their capacity to engraft NOD/SCID mice. It is possible to combine several cryopreserved samples of MNC or CD34+ cells of the same regular donor. Thus the banking of LDF cells might be interesting for autologous and allogeneic cell therapy as well as for ex vivo RBC production.

P P Young, A A Teleron, Vanderbilt University Medical Center, Nashville, TN; B Carlson, American Red Cross, Tennessee Valley Region, Nashville, TN Background: Neovascularization in tumors, wounds, and sites of ischemic injury occur by both angiogenesis (proliferation from pre-existing vessels) and by vasculogenesis (in situ differentiation of endothelial cells from circulating endothelial progenitor cells [EPCs] ). EPCs can be obtained from bone marrow (BM), cord blood, or by ex vivo expansion of human peripheral blood (PB). The ease of obtaining human PB-EPCs has led many recent studies to utilize PB-EPCs. The ability to obtain large numbers of PB-EPCs would greatly facilitate characterization to further our understanding of EPC biology and their application in cellular gene therapy. Methods: We isolated PBMNCs from whole blood or from the material obtained from back-flushing WBC reduction filters followed by density gradient centrifugation. The cells from both sources were then cultured separately under defined conditions to quantify EPC yield from each source. Expanded cells were characterized as EPCs by their ability to take up acetylated LDL and stain with UEA-1 lectin. Results: PBMNCs derived from whole blood had a similar CD34 + /lincell percentage as compared to material obtained from WBC filters. Clusters of spindle shaped cells with endothelial morphology were apparent by day 3. Analysis of the cells on day 7 revealed that all of the spindle-shaped cells expanded from both sources took up acetylated LDL and stained with UEA-1 lectin. The yield of EPCs per million PBMNCs plated was ~3-fold higher from fresh blood as compared to WBC filter. However, since greater numbers of PBMNCs were obtained from each filter, the average number of EPCs obtained from one filter was 5.4 million versus 0.4 million EPCs from one tube of fresh blood (~14 fold yield). Conclusion: Using blood donor WBC reduction filters provides a safe, inexpensive, and readily available source for large numbers of PBMNCs from which cultureexpanded EPCs can be generated for further study. Background: Adult bone marrow mesenchymal stem cells and progenitors (BM MSC/P) constitute an attractive source of potential transplantable cells for therapy of skeletal diseases. A murine model of BM MSC/P might be helpful in the design of specific approaches to correct specific genetic and skeletal diseases by gene delivery after MSC/P transplantation. Methods: We investigated the feasibility of isolation, ex vivo expansion, transduction and transplantation of BM MSC/P in a murine experimental model. For isolation and ex vivo expansion, we compared the method of BM MSC/P isolation based on plastic adherence and culture under defined conditions followed by macrophage depletion (method A), with a hematopoietic/ endothelial-depleted (CD45-/Ter119-/CD11b-/CD34-/CD31-) bone marrow population (method B). Multilineage differentiation of MSC/P into osteoblasts, chondrocytes and adipocytes was assessed by tissue specific staining. Transduction of MSC/P was performed with the retroviral MIEG3 vector expressing Enhanced Green Fluorescent Protein (EGFP). For analysis of skeletal tissue engraftment, we transplanted: 1) whole BM cells from smooth muscle a-actin/EYFP (SMAA8-EYFP) or smooth muscle gactin/EGFP (SMGA13.7-EGFP) transgenic mice, or 2) ex vivo expanded MSC/P expressing either an a-actin-driven EYFP or EGFP driven by a constitutively active retrovirus promoter (MIEG3-transduced). A total of either 10 7 BM cells or 2 ¥ 10 5 MSC/P were injected intravenously via the tail vein of 7 Gy-irradiated syngeneic recipients. Results: We observed that method A yielded higher colony-forming-units of fibroblast type (CFU-F) content. MSC/P isolated and expanded by method A were easily transduced using retroviral vectors and grew exponentially until reaching greater than 10 15fold and 80-fold cell and CFU-F expansion, respectively. Exponentially growing MSC/P could be differentiated into adipocytes, osteoblasts and chondrocytes and we observed that transplantation of SMAA8-EYFP but not SMGA13.7-EGFP transgenic BM cells resulted in patches of spindled cells surrounding pre-or post-sinusoidal vessels in the BM of non-transgenic recipient mice. Conclusions: We have demonstrated the feasibility of the expansion and transduction of MSC/P in a murine model and BM stroma transplantation by systemic infusion. These data confirm that MSC/P comply with the major requirements of a cell vector for gene therapy: efficient propagation and transduction ex vivo, together with the ability to be genetically loaded and transplanted.

J C Zimring, T E Chadwick, C D Hillyer, J D Roback, Emory University School of Medicine, Atlanta, GA BACKGROUND: In murine models, a number of strategies have been developed to induce allospecific tolerance to prevent transplant rejection, including transfusion based strategies of immunomodulation. However, protocols that work well in mice can be ineffective in primate populations. For example, costimulatory blockade results in long-term allograft acceptance in rodents but gives variable and/or limited benefits in primates. One explanation is that mediators of inflammation, which are absent in pathogen free murine environments, are common in humans and non-human primates. Consistent with this notion, viral infection and/or activators of innate immunity such as lipopolysaccharide (LPS) interfere with tolerance induction by costimulatory blockade in mice. In contrast, transfusion based adoptive immunotherapy using veto cells induces allotolerance and increased graft survival in both mice and primates. Based upon this difference, we hypothesized that veto cells would not be inhibited by mediators of inflammation. METHODS: Highly active veto cells from C57BL/6 (H-2 b ) mice were generated in tissue culture by stimulating with third party C3H (H-2 k ) splenocytes in a mixed lymphocyte reaction (MLR). Veto activity was then assayed in a second MLR consisting of BALB/c (H-2 d ) responders stimulated with C57BL/6 stimulators. To test the effect of activators of innate immunity in this system, titrations of LPS, double stranded RNA, IFN-g or TNF-a were added to the veto assay cultures and the effect on veto activity was measured. Independent assays were performed to confirm the biological activity of the tested molecules.

RESULTS: The ability of veto cells to inhibit alloimmunity was not significantly altered by any of the above inflammatory molecules. There was no significant effect of LPS, dsRNA, or IFN-g over a 25-fold range of concentrations on veto activity. Likewise TNF-a had no effect when adequate veto cells were present, but did modestly inhibit the veto effect when very few veto cells were added. The inflammatory mediators themselves (in the absence of veto cells) did not suppress lytic activity. DISCUSSION: These findings indicate that common inflammatory molecules do not reverse veto based tolerance under conditions reported to inhibit tolerance induction by other approaches. These findings may help to explain why veto cell based therapies appear to succeed in primate populations and underscore a distinct advantage of veto cell based approaches, especially in the setting of ill or inflamed patients. Ongoing studies of transfusion based tolerance induction using veto cells are in progress. 

shown to regulate immune responses both in vivo and in vitro. Extracorporeal photopheresis (ECP) involves the clinical reinfusion of peripheral blood leukocytes that are undergoing apoptosis following exposure ex vivo to 8methoxypsoralen (8-MOP) and UVA light. ECP is approved for the palliative treatment of cutaneous T cell lymphoma and has been reported to have utility in immune-mediated inflammatory diseases such as graft versus host disease, transplantation rejection, atopic dermatitis and autoimmune diseases such as rheumatoid arthritis, scleroderma and Crohn's disease. We have evidence to suggest that ECP therapy may modulate host dendritic cell function and induce regulatory T cell generation. When co-incubated with ECP-treated cells, activated dendritic cells produce reduced levels of proinflammatory cytokines, such as IL-12, while TGFb levels are modestly increased. Activation of CD4 + T cells in the presence of allogeneic dendritic cells and ECP-treated cells promotes generation of a population of T cells that can suppress proliferation of naive syngeneic T cells, as well as, suppress IFNg production. To confirm these findings in vivo, we employed a murine contact hypersensitivity model. ECP-treated or control leukocytes from mice sensitized with the hapten dinitrofluorobenzene (DNFB) were injected intravenously into naïve recipients. Compared to controls, mice that received ECP-treated cells demonstrated significantly less ear swelling following sensitization and challenge with DNFB. Suppression of ear swelling was specific for DNFB and cell-mediated, as demonstrated by the ability to transfer DNFB tolerance to naïve mice, which could appropriately respond to the unrelated hapten oxazalone. Transfer of this tolerance was abrogated by depletion of either CD4 + or CD25 + T cell populations. Collectively, these results suggest that the delivery of ECP-treated cells promotes generation of regulatory T cells that are capable of modulating immune responses. Regulatory T cells have been implicated in the control of GvHD, and as previously reported, ECP has demonstrated beneficial activity in GvHD (Blood 92(9) 1988 p3098). As a result, international phase II clinical trials in GvHD patients are currently underway to assess the efficacy of photopheresis in these patients.

M Chevrier, C Proulx, Héma-Québec, Sainte-Foy, PQ, Canada; L Peltier, Héma-Québec, Montréal, PQ, Canada; I St-Amour, M-E Nolin, Héma-Québec, Sainte-Foy, PQ, Canada Background: Cord blood (CB) is routinely used as a source of human stem/progenitor hematopoietic cells (CD34 + cells) for transplantation. One of the most clinically relevant criteria of quality of collected cord blood units is the absolute count of viable CD34 + cells. For that purpose, we validated Stem-kit TM for the qualification of CB samples on a BD FACS flow cytometer in preparation for the first CB bank in Québec. Methods: The Stem-kit TM method was evaluated for the accurate counting of viable CD34 + cells. To perform the linearity assays, blood samples were spiked with known numbers of KG-1a cells (CD34 + cell line). Mixtures of heat (55°C)-killed KG1a cells and viable KG-1a were used to evaluate the viability. The cell count and viability were confirmed by Trypan blue exclusion. Then, the Stem-kit TM 59A 2004-Vol. 44, Supplement was evaluated with unprocessed and reduced CB (buffy coat). CB units were centrifuged to produce a buffy coat layer and processed using OptiPress II to separate the blood into plasma, buffy coat and red blood cell concentrate. Results: Results showed a very good correlation for the viability and number of KG-1a cells in comparison with Trypan blue cell counts. The linearity of the numeration of CD34 + cells revealed a dynamic range of 0-3000 cells/mL with a detection limit of 2 cells/ml and a quantification limit of 6 cells/ml. The comparison between the unprocessed CB and buffy coat demonstrated the ability of the Stem-kit TM to evaluate accurately the number of viable CD34 + cells in buffy coats. Furthermore, the acquisition and analysis of the samples is facilitated by the fact that the quantity of cells is increased proportionally to the volume reduction. Also, results showed that the reduction process preserved the number and viability of the CD34 + cells present in the starting CB samples. Conclusions: The use of Stem-kit TM provides a sensitive and rapid method to accurately determine the number of viable CD34 + cells in CB samples and can thus be used for quality control of CB units for banking. The use of Stem-kit will reduce significantly the time required for the qualification of CB and thus improves the banking process. Background. Recent studies suggest that circulating bone marrow (BM)derived cells may contribute to both endo-and exogenous regeneration, the most unexpected, if controversial, way being through transdifferentiation into differentiated non-hematopoietic cells, including insulin-producing pancreatic beta cells. Such (stem cell) plasticity seems to be driven by cell/organ injury. We have observed donor-derived pancreatic islet cells after BM-transplantation (BMT) to otherwise healthy mice. In such mice, the BMT-conditioning regimen (lethal whole body radiation, WBR) in itself leads to reduced islet insulin expression, which corrects 1-2 weeks after BMT. Methods. Female C57BL/6 mice, 8-10 week old, were treated with Streptozotocin (STZ, 200 mg/kg i.p.). Blood glucose was measured using test strips and a blood glucose meter (Accu-Chek R , Roche, IN). The majority of mice developed overt diabetes (Glc > 350 mg/dL) within a week, but a subset of mice ("subdiabetic") were more resistant. The subdiabetic, and a few completely STZ-resistant mice, underwent lethal WBR (11 Gy) 11 days after the STZinjection. In another set of experiments, mice surviving >40 days after STZinjection (long-term surviving), that were subdiabetic (either throughout, or after return to a subdiabetic state following a period of overt diabetes), were treated with either lethal or sublethal (5.5 Gy) WBR. Results. The Table  shows the development of blood glucose levels in one mouse with overtly diabetic, one with subdiabetic, and one with no STZ-response. Recombinant factor VIIa (rFVIIa) is approved for the treatment of bleeding in hemophiliacs with inhibitors. However, rFVIIa has also been used to achieve hemostasis in patients with uncontrolled hemorrhage due to a variety of etiologies. Here we report our experience using rFVIIa in patients with hemorrhagic complications upon removal from CPB. Methods: Use of rFVIIa for non-hemophiliacs required approval by our transfusion service. We followed 7 consecutive patients undergoing cardiac surgery requiring CPB who received rFVIIa intraoperatively for refractory bleeding. Data were collected by retrospective chart review and physician interview, and included diagnosis, blood product utilization, dose of rFVIIa, and effect on hemostasis. Cost analysis was also performed. Results: Seven patients received rFVIIa for refractory bleeding after CPB: 6 males and 1 female. Indications for surgery were: ventricular-assist device placements (3), cardiac transplant (2), and valve replacement (2). All patients had oozing and/or coagulopathic bleeding refractory to protamine and standard therapy. Patients received one 90 ug/kg dose of rFVIIa rounded to the closest 1.2 mg vial. Use of rFVIIa led to a significant decrease in blood product utilization. No thrombotic events were noted. The average costs per patient of providing blood products pre-and post rFVIIa were $5819 and $471, respectively. Recombinant factor VIIa dosing averaged $7714 per patient. It should be noted that our cost estimates for blood products were conservative and indirect costs were not taken into account.

In one experiment, blood glucose levels in six subdiabetic mice ranged from 204-286 on day 11, and from 397-600 on day 19 (8 days after irradiation). Without BM-rescue, mice start to die 10-12 days after lethal WBR. Several long-term surviving, subdiabetic STZ-treated mice became overtly diabetic after treatment with lethal or sublethal WBR. Conclusion. WBR appears to negatively impact the insulin reserve of pancreatic islets. This in vivo effect is enhanced in already-compromised islets (STZ-treated mice). Combining carefully graded and timed insults with WBR and STZ, respectively, may constitute a novel diabetic model, optimized for the study of BM stem cellbased regeneration and role of radiation-induced injury as a regeneration stimulus.

Conclusions: The decrease in blood product utilization suggests that rFVIIa can be used to treat refractory bleeding in patients who undergo cardiopulmonary bypass. Although no thrombotic complications were noted in this small cohort of patients, larger prospective studies are needed to evaluate the safety of rFVIIa in patients undergoing cardiac surgery. Cost analysis revealed that the cost of rFVIIa was largely offset by the decrease in blood product utilization.

D M Bensen-Kennedy, R S Shirey, K King, P Ness, Johns Hopkins Hospital, Baltimore, MD Introduction: T-activation is enzymatic exposure of the Thomsen-Friedenreich antigen (T antigen) on RBCs by bacterial and viral neuraminidases. Since T polyagglutinins (anti-T) are present in all normal adult sera and in polyclonal typing reagents, T polyagglutination (T Poly) may be recognized due to ABO typing discrepancies. T Poly may not be detected using newer monoclonal typing reagents. The transfusion of plasma products to patients with T Poly has been associated with in vivo hemolysis. We present a case of T Poly associated with severe hemolysis and bleeding in the face of necrotizing enterocolitis (NEC) where plasma transfusion was avoided with the use of recombinant factor VIIa (rFVIIa). Case report: A 1219 g male newborn with osteogenesis imperfecta was delivered by caesarian section at 34 4/7 weeks due to intrauterine growth retardation. He typed as group A Pos with negative antibody screening tests and negative DATs. The infant was hospitalized for 16 days and then sent to an outside facility. On day of life (DOL) 20 the infant's HCT dropped from 31% to 22% without bleeding. By DOL 21, he had decreased feeding and urine cultures were positive for group D Enterococcus. Antibiotics were started and group O RBCs were transfused. 60A SCIENTIFIC SECTION TRANSFUSION 2004-Vol. 44, Supplement On DOL 22, the infant developed bloody stools and NEC was diagnosed. On DOL 23, the infant was readmitted for emergency bowel resection with transfusion of multiple units of group O RBCs and ABO compatible plasma and platelets. He developed frank hemolysis with excessive bleeding from multiple sites, and a transfusion reaction evaluation was requested. The pre and post transfusion (Tx) samples revealed marked hemoglobinemia, but the serologic workup was unremarkable. In light of the clinical setting, testing for T Poly was performed. T Poly was identified in all samples by lectin tests; the T Poly RBCs reacted with normal pooled group AB plasma to a titer of 16 (pre Tx) and 64 (post Tx) at 22C. Additional surgery was managed with washed RBCs and rFVIIa yielding transient improvement in hemostasis. His clinical course ultimately worsened and he died following withdrawal of support. Conclusion: 1) This case underscores the importance of clinically recognizing patients at risk for T Poly. Since T Poly is no longer detected by routine preTx testing, the responsibility for suspecting and initiating appropriate testing has shifted from the blood bank to the clinician. 2) The clinical significance of T Poly and Tx of plasma products remains controversial. Avoidance of plasma Tx is advisable, 3) In this clinical setting, the use of rFVIIa may offer a treatment alternative to plasma Tx.

T G Hirose, A Huff, A K Nobori, Good Samaritan Hospital, Los Angeles, CA INTRODUCTION Protime (PT) and activated Partial Thromboplastin Time (aPTT) with associated 1 : 1 Mixing Studies are used to diagnose disorders of altered bleeding, clotting, and the therapeutic effects of anticoagulants. PURPOSE New anticoagulants exert their effects at different points of the coagulation cascade with resulting alterations in PT, aPTT mixing studies that are generally unknown. Therefore, this study was designed to measure the in vitro effects of a variety of anticoagulants: unfractionated heparin (UFH), low molecular weight heparin (LMWH), pentasaccharide, and direct antithrombin inhibitor. MATERIALS AND METHOD Standard therapeutic concentrations and suprapharmacologic concentrations were calculated based on serial drug dilution studies. Unfractionated Heparin, Enoxaparin, Fondaparinux, Lepirudin, and Argatroban were mixed with normal saline from stock drug vials utilizing Gilson micropipettes. Standard PT, aPTT and 1 : 1 Mixing assays were run on a Sysmex CA-6000 automated coagulation instrument in the same manner as routine patient testing. DATA

A Odeleye, A Presley, M Passwater, P Clark, P D Mintz, University of Virginia Health System, Charlottesville, VA BACKGROUND: Rattlesnake venom-induced thrombocytopenia (VIT) can occur in the absence of significantly abnormal coagulation studies in patients treated with Crotalidae Polyvalent Immune Fab (Ovine; CroFab TM ). Ovine, a mixture of 4 different monospecific antivenins (Crotalus atrox, Crotalus adamanteus, Crotalus scutulatus, Agkistrodon piscivorus), works by binding and neutralizing venom toxins; previous studies have indicated Ovine has a shorter in vivo half-life than crotalid venoms. The mechanism of persistent thrombocytopenia is unknown, but it has been suggested VIT is caused by platelet aggregation and sequestration at the site of the snake bite due to venom proteins. CASE REPORTS: Review of one medical center's records from 5/02-8/03 revealed 33 patients transferred to the emergency department after rattlesnake bites. 31 of these, presenting with benign symptoms requiring wound irrigation, tetanus toxoid injection, and multiple vials of Cro-Fab TM injection, were summarily discharged with no apparent sequelae. 2 patients required protracted hospital stays and transfusion support due to persistent thrombocytopenia. Patient 1 presented with respiratory compromise due to preglottic edema requiring intubation and respiratory support, petechiae, and extensive ecchymosis of the left upper extremity. Presenting platelet count was 206 k/mL, which fell to < 10 k/mL in 28 hours, with mildly elevated PT and INR that normalized over 2 days with transfusion of 6 units of FFP. 5 doses of platelets (4 each with 5 pooled whole blood-derived [WBD] units, and 1 apheresis unit) were also given between days 2 and 5 (Table) . Patient 2 presented with extensive ecchymosis of left upper extremity and a platelet count of <10 k/mL in spite of repeated doses of antivenin, with normal coagulation parameters. Transfusion of 19 doses of pooled platelets (88 WBD) over 7 days did not result in normalization of the platelet counts (Table) . There was no laboratory evidence of alloimmunization to platelet transfusion in either patient. The 1st patient was discharged with a platelet count of 44 k/mL; the 2nd self-discharged with a platelet count of 38 k/mL. CONCLUSIONS: Persistent thrombocytopenia may occur after a rattlesnake bite. Although CroFab TM can minimize venom damage, physicians should remain vigilant of the possibility of severe thrombocytopenia and hemorrhage.

RESULTS UFH is generally thought to correct with 1 : 1 Mixing Studies in therapeutic concentrations. This study demonstrates that for the instrument and reagents used, incomplete aPTT correction may occur suggesting residual heparin influence at this drug dose. This is an important finding as incomplete correction may be interpreted as an inhibitor effect rather than an anticoagulant effect. Enoxaparin does not demonstrate this type of elevation for 1 : 1 aPTT Mixing Studies in therapeutic concentrations. Argatroban and Lepirudin exhibit inhibitor-like effects for 1 : 1 aPTT Mixing Studies. High doses of Fondaparinux will prolong the aPTT but does not appear to prolong the Protime to the same degree. This suggests that either the Protime reagent is less sensitive to the effects of high dose Fondaparinux or a novel pathway may exist leading to thrombin generation. CONCLUSION This study demonstrates that the new generation anticoagulants are able to influence routine coagulation assays. Following confirmation utilizing in vivo assays, we hope that this data will provide new reference points against which the diagnosis of specific coagulation disorders can be based. erolization and 60 ± 5% after glycerolization. This investigational study assessed the high separation (HS) bowl to automate glycerolization and eliminate the need for blood bank centrifuges. Study Design and Methods: Units of 500 ml non-leukoreduced CPD whole blood were stored at 4 C for up to six days prior to glycerolization (non-rejuvenated) or stored at 4 C for up to 10 days and modified with a solution of pyruvate, inosine, phosphate and adenine (Rejuvesol, rejuvenated). The RBCs were glycerolized using the ACP215 with the HS bowl, frozen for at least 2 weeks at -80 C, deglycerolized in the ACP215 with the 325 ml bowl then stored in AS-3 at 4 C for 21 days. Results: Using this system with the HS bowl, RBCs in CPD whole blood with hematocrits of 38 to 45% were glycerolized in 50 minutes. The manual double centrifugation procedures required 2 1 / 2 hours to perform. The only technical aid needed was the sterile attachment of the RBC unit to the glycerolization set using a sterile connector device. The freeze-thaw recovery for RBC units was 98 ± 1% and the freeze-thaw-wash recovery was 87 ± 6%. Following storage of the RBC at 4 C in AS-3 for 21 days hemolysis was 1.0 ± 0.6%; extracellular K+ level was 26 mEq/L; RBC K+ level was 65% of the value on the day of deglycerolization. In Vitro Storage Data (mean ± sd, n = 10 non-rejuvenated-NRJ, 15 rejuvenated units-RJ) Conclusion: The frozen leukoreduced ACD-A/Adsol RBC components met the standards for mean RBC recovery of >80% after deglycerolization, mean radiolabeled RBC in vivo recovery of ≥75%, and mean hemolysis of less than 1%. The ACD-A/Adsol RBC units are suitable for transfusion. To evaluate if the increase in % hemolysis was directly related to the additive solution, RBCs were frozen and stored in PVC bags with AS-3 or AS-5 additive solutions after deglycerolization. METHODS: Four units of whole blood were collected in CPD, leukoreduced, packed, pooled, and thoroughly mixed in order to reduce natural variability in various individual donors. They were then separated into four aliquots of approximately equal volume. After addition of AS-5, units were stored for 6 days at 4°C. Glycerolization was carried out on day 6 using the Haemonetics ACP 215. Cells were frozen for a minimum of six days in the same PVC bags and deglycerolized using the same device. At the end of deglycerolization, 100 mL of AS-5 (2 units) or AS-3 (2 units) were added. All units were then stored at 4°C for 28 days. Each unit was sampled and tested for supernatant Hb at day 0, 7, 14, 21, and 28 after deglycerolization. A total of 3 pools with four donors per pool were evaluated. RESULTS: A statistically significant increase in % hemolysis after deglycerolization was seen when AS-5 was used as the additive solution when compared to AS-3. The mean average % hemolysis at day 28 was 5.6% with AS-5 and 0.7 with AS-3. Testing in triplicate of 0.9% saline-0.2% dextrose (the last washing solution before addition of additive solution), AS-5, and AS-3 resulted in average osmolarity values of 296, 358, and 287 mOsm/kg, respectively. CONCLUSIONS: The increased osmolarity of AS-5 when compared to 0.9% saline-0.2% dextrose and AS-3 likely results in increased hemolysis by causing osmotic shrinkage of RBCs. Other factors such as key differences between AS-5 and AS-3 composition, can not be ruled out. BACKGROUND: Reports of high bag breakage rates with PVC bags during thawing of RBCs has prompted a search for a better plastic bag that can withstand the severity of freezing, shipping, and thawing prior to deglycerolization. Previous results have shown that: 1) a plastic composed of ethylene vinyl acetate (EVA) performs better than other plastics when external forces are applied at the rigorously low dry ice shipping temperatures 1 ; 2) RBCs can be frozen in EVA bags with no adverse effects on RBC viability during post-thawed storage 2 . In order to evaluate the effects of EVA plastics on RBCs throughout the entire storage period, RBCs were stored in EVA bags from the day of collection until 28 days after deglycerolization. METHODS: Four units of blood were collected in CPD, leukoreduced, packed, pooled and, after thorough mixing, separated into four aliquots of approximately equal volume. After addition of AS-5, two units were stored in PVC bags and two in EVA bags for 6 days at 4°C. Glycerolization was carried out on day 6 using the Haemonetics ACP 215. Cells were frozen for Conclusion: RBC processed in the automated functionally closed Haemonetics ACP215 with the HS bowl had similar quality to RBC processed by the non-automated procedure using the ACP215 device. Although further studies are required to validate the advancement, this automated functionally closed procedure significantly reduced the technical requirements and time needed for the glycerolization process. Background: Red blood cell (RBC) units with rare blood types as well as autologous units may be stored up to 10 years when frozen. However, freezing RBC products anticoagulated with ACD-A and stored in Adsol Red Cell Preservation Solution has not been previously studied. This study evaluated the effect of freezing and thawing ACD-A/Adsol RBCs components using standard conditions. Study Design: This study was performed at three clinical sites. Twenty-three RBC units were collected concurrently with platelets using the Amicus ® Single Needle procedure. ACD-A RBC products were leukoreduced using the Amicus Red Cell Set and stored in Adsol Preservation Solution for six days at 1-6°C and then glycerolized using the high concentration glycerol method. The units were frozen and stored at a minimum of -65°C for at least thirty days. At the end of the storage period, products were thawed and deglycerolized using the Cobe 2991 Blood Cell Processor. In vitro testing was performed after collection on Day 0 or 1 and at the end of storage. Parameters evaluated in vitro were RBC count, unit volume, plasma hemoglobin, total hemoglobin, spun hematocrit, plasma potassium, pH, ATP, osmolality, sterility, and RBC recovery following deglycerolization. Autologous 51 Cr radiolabeled RBC recovery was assessed in vivo at the end of storage. Results: Data from the primary assessments (Mean ± SD) are summarized in the table below. a minimum of six days in the same PVC and EVA bags and deglycerolized using the same device. At the end of deglycerolization, 100 mL of AS-5 were added. All units were then stored at 4°C in the same type of plastic for 28 days. Each unit was sampled and tested for supernatant Hb, ATP, and 2-3 DPG before freezing at day 6 of liquid storage, and at day 0, 7, 14, 21, and 28 after deglycerolization. A total of five pools with four donors per pool were evaluated. RESULTS: No statistically significant difference was seen in ATP and 2-3 DPG values. A marked increase in % hemolysis after deglycerolization was seen with EVA bags when compared to PVC bags. Even PVC bags had an average hemolysis of over 3% at 28 days after deglycerolization. However, there was no statistically significant difference between PVC and EVA bags during the 6-day storage at 4°C and % hemolysis was well below 1%. CONCLUSIONS: EVA bags may be used for liquid storage of RBCs before freezing but not for post-thawed storage of RBCs. The increased % hemolysis seen after deglycerolization, even with PVC bags, may be the result of the additive solution. 

Refractometer and Osmometer to Estimate Glycerol in Thawed, Glycerolized RBC C R Valeri, A Gennouai, J P Lane, G Ragno, Naval Blood Research Laboratory, Boston, MA Background: Measurements for the quality of red blood cells frozen with 40% W/V glycerol and stored at -80 C include freeze-thaw and freeze-thawwash recovery values, extracellular potassium, concentration of glycerol in the supernatant of the thawed glycerolized RBC and the supernatant of the deglycerolized RBC, percent hemolysis in the deglycerolized RBCs, and bacteriologic cultures. Study Design and Methods: This study was done to measure the glycerol concentration in the supernatant of the thawed glycerolized RBC and the freeze-thaw recovery of the RBC. In addition, glycerol solutions were prepared at concentrations of 34 gm% to 46 gm%. The glycerol concentration and refraction were measured in the hand-held Palm Abbe refractometer and the osmolality was measured using the osmometer. Results: Concentrations of glycerol ranging from 34 to 46 gm% correlated to the glycerol concentration and refraction measured using the Palm Abbe refractometer and the osmolality measured by osmometry. In 103 units of RBC frozen with 40% W/V glycerol at -80 C for a mean of 16 years, the refraction and glycerol concentration of the supernatant of thawed, glycerolized RBC measured using the Palm-Abbe refractometer and the osmolality measured by osmometry estimated the glycerol concentration. In the 103 units studied, no correlation was observed between freeze-thaw hemolysis and refraction (%) in the supernatant of the thawed glycerolized RBC. n = 15). All units were stored for 28 days post-irradiation at 2-6°C. Sample collection and in vitro assays were performed before GI and at the end of storage. Volume, RBC count, Hct, ATP, Hb, supernatant Hb, % hemolysis, pH, and supernatant K + values were measured at all time points. The in vivo 24 hour RBC % recovery and the ATP % recovery were determined at the end of storage. Results: Average % radiolabeled RBC in vivo recovery for units in Group 1 and 2 with GI on Day 1 and Day 14 were 81.7% and 80.1%, respectively. Mean ATP % Recovery was 87.5% and 74.1% after storage for Groups 1 and 2, respectively. Hemolysis was <1% for all units at the end of storage. There were no significant increases in K + compared to previously reported studies performed in Adsol RBC units. The means for all test results were within expected ranges.

Conclusion: The hand-held Palm Abbe refractometer can be used to estimate the glycerol concentration in the supernatant of the thawed, glycerolized RBC. Background: Gamma irradiation (GI) is performed on cellular blood units prior to transfusion to prevent transfusion-associated graft-versus-host disease (GVHD). This study evaluated ACD-A/Adsol RBCs collected using the Amicus Separator, subjected to GI and stored under standard conditions. Methods: Thirty normal subjects donated concurrent RBC / platelet units on the Amicus Separator at two investigational sites. RBCs were collected in ACD-A with 100 mL of Adsol added and were leukoreduced at room temperature within 8 hours of collection. Units were divided into two groups and GI with 2500 cGy on either Day 1 (Group 1; n = 15) or Day 14 (Group 2;

Conclusion: Leukoreduced ACD-A/Adsol RBC units irradiated with 2500 cGy on Day 1 or Day 14 and stored for an additional 28 days exceeded the FDA standard of 75% for in vivo recovery. All units met regulatory standards and are acceptable for transfusion.

N Rugg, T J Greenwalt, A D Joines, J F Gormas, P G Pratt, Hoxworth Blood Center, Cincinnati, OH; S Janik, W J Lee, B Brantigan, Baxter Healthcare Corporation, Round Lake, IL Background: Irradiation of blood components for immunocompromised patients has been used for over 30 years. The AABB and FDA recommend a minimum gamma irradiation dose of 25 Gy to the central portion of the blood bag. This study evaluated the effect of gamma irradiation on leukoreduced RBC products collected using the Amicus Separator with ACD-A anticoagulant and stored in Adsol preservative solution. Study Design & Methods: Healthy volunteers who met FDA and AABB guidelines for apheresis donation participated in this open label study and donated a unit of RBCs using the Amicus Separator. The RBCs were irradiated at a 30 Gy dosage using the Gammacell 3000 Elan (Nordion International Inc. Ontario, Canada) for 6 minutes and 51 seconds. Group A (n = 7) products were irradiated on day 1 and stored at 1-6°C for 28 days. Group B (n = 6) products were irradiated on day 14 and stored at 1-6°C for 28 days post irradiation. Samples for in vitro testing were taken prior to irradiation, and at the end of storage. In vitro testing included: potassium, pH, plasma hemoglobin and ATP levels. At the end of storage in vivo 24-hour % recovery was measured using the 51 Cr single label method. Results: The results demonstrated that leukoreduced ACD-A/Adsol RBC products gamma irradiated at 30 Gy on day 1 and stored for 28 days, met the end of storage standards for mean % hemolysis and mean radiolabeled RBC in vivo recovery, and are suitable for transfusion. The data also shows that leukoreduced ACD-A/Adsol RBC products gamma irradiated using 30 Gy on day 14 and then stored for 28 days do not meet the standard for mean radiolabeled RBC in vivo recovery. Relevant data are shown in the table. BACKGROUND: It is a well accepted practice to gamma irradiate (GI) all cellular blood components prior to transfusion to immunocompromised recipients to prevent transfusion-associated graft-versus-host (TA-GVHD). The purpose of this study was to compare the results of GI on Trima RBC Accel Apheresis (Gambro) units drawn in ACD-A anticoagulant and RBCs derived from whole blood (WB) units drawn in CPD. STUDY DESIGN: This study was performed in two parts at three independent investigational sites. Part A of this study evaluated the in vivo and in vitro characteristics of leukoreduced (LR) Trima RBC (Test) units and RBCs derived from standard WB (Control) units collected from paired donors. Twenty-four paired normal donors were randomized to donate a Trima unit and a WB unit on two separate occasions with at least 8 weeks between donations. Part B evaluated in vitro characteristics of non-leukoreduced (NLR) Trima RBC units and WB units collected from different donors. Forty-four un-paired normal donors were randomized to donate a Trima single or double RBC unit or a WB unit on one occasion. All units collected in both Part A and B were kept at room temperature for no longer than 8 hours. AS-3 was added to Trima units and either Adsol or Optisol additive solutions were added to WB units. All units were treated with 2500 cGy GI on the 14 th day of storage and then returned to storage for 28 more days. All units were stored for a total of 42 days at 2-6°C. Samples were removed for testing on the day of collection (Day 0), day of GI prior to GI (Day 14) and on the Day 42. RESULTS: The ATP concentration and pH reported after storage were not different between Trima and WB units. Trima units had lower hemolysis (%), lower release of potassium and lower osmotic fragility levels than their paired WB units. Single label 24 hour RBC recovery was 82.1 ± 8.1% for Trima units and 76.7 ± 8.1 % for WB units (p £ 0.05).

The Effect of G-Radiation in Red Blood Cells: Implication on Ionic Homeostasis L A Barbosa, O C Moreira, V H Oliveira, A A C Araujo, J A Mignaco, C F L Fontes, UFRJ, Rio de Janeiro, Brazil; C M Nogueira, Blood Transfusion Unit-University Hospital-UFRJ, Rio de Janeiro, Brazil

Background: There is a growing demand for irradiated cellular blood products for transfusion services. The storage of previously irradiated red blood cells (RBC) can generate some damage to RBC membranes. It has been demonstrated that the levels of extra cellular K + are elevated in stored irradiated RBC. Active extrusion of Na + and inward transport of K + mediated by Na + ,K + -ATPase are responsible for the low Na + content and the high K + content of human RBCs. The outward K + gradient and inward Na + gradient generated by Na + ,K + -ATPase activity are then used by gradient-driven system to move other ions against a gradient. In this study our goal is to evaluate the g-radiation effects on supernatant ionic concentrations, Na + ,K + -ATPase and Ca +2 -ATPase activities in RBC. Methods: Blood from health subjects was collected, at the blood transfusion unit, in AS-1 bags according to standard donor room procedures. RBCs were separated from platelets and plasma by centrifugation, followed by a treatment of g-radiation (25 Gy) and maintained at 4°C. With the course of post-irradiation period (1, 14 and 24 days), aliquots were collected in sterile conditions and cells were submitted to centrifugation to obtain supernatant. These supernatants were submitted to measurement of pH and external K + , Na + , by atomic absorption spectrometry. RBCs plasma membrane ghosts were prepared from pellet fraction (Rega et al, 1979) to measure Na + ,K + -ATPase and Ca +2 -ATPase activity (Norby et al, 1981; Clark and Carafoli, 1983) . As a control RBC, obtained in same conditions, not submitted to g-radiation, was used. Results: The K + concentration was progressively increased with ageing in all blood bag samples, but the irradiation effect on day 1 increased about two folds the external K + , from 4.0 mmol/L (control) to 9.8-9.4 mmol/L (irradiated samples). A decrease of Na + and pH levels was also observed, but these changes were not related to a direct irradiation effect. No changes were observed in Ca +2 -ATPase activity, averaging 10.5 (control) to 11.4-13.1 nmol Pi.mg -1 .min -1 (irradiated samples). However, a dramatic decrease in Na + ,K + -ATPase activity was observed, from 49.8 (control) to 8.98-9.51 nmol Pi.mg -1 .min -1 (irradiated samples). The difference between control and irradiated samples remained constant through the time course of the experiment. Conclusions: The changes observed in ion levels suggest that g-irradiation affected plasma membrane permeability. The observed inhibition of Na + , K + -ATPase activity could be one important parameter in irradiated RBCs to determine expiration date for transfusion procedures.

The Effect of Irradiation on Platelet Concentrates, Stored for 7 Days P F Van der Meer, R N I Pietersz, L Scholtalbers, L A E Liefting, Sanquin Blood Bank North West Region, Amsterdam, Netherlands Background: The storage of white cell (WBC)-reduced platelet concentrates (PCs) can be extended from 5 to 7 days provided bacterial screening is performed. Irradiation up to 50 Gray (Gy) does not affect platelet quality, but the effect of pre storage irradiation with subsequent storage up to 7 days is not known. Method: Two WBC reduced PCs, each made from 5 buffy coats and a unit of plasma, were pooled and divided into pool A (control group) and pool B (study group). PCs in group B were irradiated immediately after preparation with 25 Gy. PCs in both groups A and B were stored on a continuous flat bed shaker at 20-24°C. Swirl, pH and CD62 expression were determined on day 1, 6 and 8. Twelve experiments were performed and compared with a paired t-test, p < 0.05 was considered significant. Results: See table (day 8 values; mean ± SD; n = 12). Pooling and dividing of the PCs was successful with respect to volume and platelet number. On day 8, the pH in group B is slightly lower than in group A, but the difference is not significant. In group A, pH was <6.8 in 3/12 PCs, versus 5/12 in group B. The CD62 expression in irradiated PCs is not significantly higher than in non-irradiated PCs. Conclusion: Irradiation had no significant effect on platelet quality when stored for up to 7 days after blood collection.

CONCLUSION: Trima units GI on Day 14 and stored for an additional 28 days exceeded the minimum 75% requirement for in vivo recovery measured by single labeling technique. The differences found between Trima and WB units in total RBC count, supernatant Hb, Hemolysis, K + , Na + , Glucose, Osmotic Fragility and in vivo recovery are within the expected ranges. RBCs collected with the ACD-A/AS-3 Trima Accel System treated with 2500 cGy GI meet the applicable regulatory standards. 

J Monson, A Bracey, R Radovancevic, St. Luke's Episcopal Hospital, Houston, TX; C Wong, Gulf Coast Regional Blood Center, Houston, TX Background: AABB Standards require that Fresh Frozen Plasma (FFP) components are thawed at 30-37°C and expired in 24 hours if not transfused. Thawed plasma not used by patients represents a loss in revenue as well as resource. Increased blood-related costs in healthcare have recently become a concern. One way to contain these costs is to maximize resource utilization. The goal of this research is to evaluate the combined effects of thawing FFP at 37°C and 45°C and the prolonged storage of liquid plasma at 1-6°C on the coagulation Factors V, VII, and VIII. Study Design and Methods: Ten pairs of group O plasma were prepared; one set for 37°C thawing and the other, 45°C thawing. In order to duplicate the average volume of plasma procured from a single whole blood donation, plasma samples from two donors were pooled and then divided into two separate bags, each containing approximately 200-250 mL. The prepared plasma samples were frozen and stored at less than -18°C. One set was thawed at 37°C and the other set at 45°C. Thawed plasma was stored at 1-6°C as liquid plasma. Samples were obtained on day 0, 5, 10, 15, and 20 and assayed for clotting Factors V, VII, and VIII. Results: Factor V and VIII activity decreased over time, as expected, in both treatment groups. No difference in activity was noted between the groups (p > 0.12); however, thawing time was diminished by 20%. Residual Factor V activity was 61.3 ± 15% and Factor VIII activity was 62.7 ± 15% on storage day 10. There was no change in Factor VII activity over the storage interval. Conclusion: Although plasma clotting factor levels are reduced in storage, the therapeutic levels of Factors V and VIII are maintained in liquid plasma stored up to 10 days at 1-6°C. Thawing of FFP at 45°C decreased the thawing time but did not affect the factor activity of Factors V, VII, and VIII. Background Thawing of fresh frozen plasma (FFP) to critically ill patients with demand on short delivery time continues to be an important working operation in blood service practice. With current techniques, the thawing procedure takes about 20-30 minutes. In the present study a new technology using radio fields, Radio field Thawing Device (RTD), was applied for thawing FFP. The main advantage of this technology is that the thawing time is substantially reduced. The new technique was compared to an established method using heated air technology (HAT). Parameters subjected to assessment were temperature after thawing and analysis of factor V (FV), factor VIII (FVIII), protein S (PS) and plasma protein electrophoresis. Methods Five plasma donors donated plasma by apheresis (Haemonetics PCS2) until approximately 600 ml was donated. Each donation was aliquoted into two equal units. Protein S was assessed by LIA technique and FVIII and FV with one-way clotting assays. The time for thawing was set to two time periods, one for each thawing technique. Samples were collected at the following time points; before freezing (B.F.), after thawing (A.T.), one week after thawing (1 w) and two weeks after thawing (2 w). After thawing the plasma was stored at 4 ± 2°C. The power of the RTD was set to 600 W at a frequency of 135.5 MHz. Results Plasma volume (ml) (Mean ± SD) 249.6 ± 2.3 HAT, 251.4 ± 2.1 RTD Temperature (°C) (Mean ± SD) Before -78.4 ± 1.7 HAT, -78.4 ± 1.7 RTD After +12.6 ± 5.6 HAT, +17.0 ± 6.5 RTD Thawing time (min) (Mean ± SD) 23.4 ± 0.5 HAT, 7.8 ± 0.4 RTD FV (kE/L) (Mean ± SD) HAT 1.34 ± 0.21 B.F., 1.35 ± 0.21 A.T., 1.12 ± 0.21 1 w, 1.02 ± 0.21 2 w RTD 1.34 ± 0.21 B.F., 1.32 ± 0.20 A.T., 1.12 ± 0.21 1 w, 0.97 ± 0.23 2 w FVIII (kIE/L) (Mean ± SD) HAT 1.34 ± 0.43 B.F., 1.21 ± 0.31 A.T., 0.59 ± 0.17 1 w, 0.55 ± 0.16 2 w RTD 1.34 ± 0.43 B.F., 1.26 ± 0.32 A.T., 0.57 ± 0.15 1 w, 0.53 ± 0.17 2 w PS (kIE/L) (Mean ± SD) HAT 0.91 ± 0.13 B.F., 0.91 ± 0.06 A.T., 0.87 ± 0.11 1 w, 0.87 ± 0.08 2 w RTD 0.91 ± 0.13 B.F., 0.86 ± 0.11 A.T., 0.86 ± 0.10 1 w, 0.89 ± 0.08 2 w B.F., before freezing; A.T., after thawing; 1 w, one week; 2 w, two weeks; HAT, heated air tech-nology; RTD, Radio field Thawing Device; FV, factor V; FVIII, factor VIII; PS, protein S No changes were found in the plasma protein electrophoresis pattern. Conclusion The new technology for thawing FFP implies a significant reduction of the thawing time for FFP. The impact of thawing and storage on FV, FVIII, PS and plasma protein electrophoresis does not significantly differ between HAT and RTD.

H C Pleasant, J J Julleis, Blood Systems, Scottsdale, AZ Background: As a result of increased automated collections with a corresponding decrease in whole blood collections, a need was identified to allow the manufacture of Cryoprecipitate AHF from Fresh Frozen Plasma collected by apheresis technology. No apheresis collection device manufacturer has obtained approval for the manufacture of the product from plasma collected using their machine. The evaluation was designed to validate that Cryoprecipitated AHF prepared from Fresh Frozen Plasma collected by apheresis technology meets the established requirements for Factor VIII and Fibrinogen. Method: Three locations, which currently manufacture Cryoprecipitated AHF from whole blood plasma, manufactured cryoprecipitate products from apheresis plasma and submitted to a centralized testing facility for Factor VIII and fibrinogen testing. Containers were sterile connected to the original collection container to divide the plasma into components weighing between 196 mL and 267 mL and accommodate the cryoprecipitate manufacturing process. The cryo rich apheresis plasma was frozen within 6 hours of collection, thawed using a circulating Cryo Thaw Bath at 1-6C, centrifuged and separated using the same process used for whole blood derived Cryoprecipitated AHF. The Fresh Frozen Plasma, Pheresis was collected using machines from two manufacturers. Components submitted for testing were tested individually. The established acceptance criteria for each component are:

Results: Factor VIII levels met established requirements. Fibrinogen levels were significantly below requirements on 33% (16/48) of the components tested. No difference in results was seen between apheresis machines. The average results of Cryoprecipitated AHF from the apheresis plasma evaluation is compared to whole blood derived Cryoprecipitated AHF in the table below.

Conclusions: The low Fibrinogen levels were unexpected and more investigation is needed to determine the cause. The process has not been implemented. Background: Individuals with Hemophilia B require replacement therapy with recombinant or plasma-derived coagulation factor IX (fIX). More benefit per injected dose might be obtained if fIX clearance could be slowed. The contribution of overall size to recombinant fIX clearance was investigated using deletion mutagenesis, enzymatic modification, and genetic fusion to albumin. Methods: The previously described fIX expression plasmid LNMbeta-IXIL which contains both the entire human fIX cDNA, and a 428 nucleotide engineered mini-intron I, was altered using PCR and other standard methods of DNA manipulation, and transferred into pCDNA3.1 (Invitrogen). Plasmids encoding wild-type human fIX (hufIX), hufIX delta 155-177, and hufIXHSA, were produced, validated via DNA sequencing, and used to The deletion mutant lacked most of its activation peptide, including all sites of post-translational modification, while the fusion protein contained the 415 residues of hufIX joined to the 584 residues of human serum albumin via a hexglycyl-monovalanyl spacer. Recombinant hufIX proteins secreted from engineered CHO cells were purified by immunoaffinity chromatography and characterised with respect to in vitro clotting activity and in vivo survival in fIX knockout mice. Outdated commercial recombinant hufIX (Benefix, Wyeth) was also tested, with or without enzymatic deglycosylation by peptide: N-glycosidase F (PNGase F). Results: The specific activity (SA) of wild-type hufIX (WT) and Benefix did not differ, while hufIX delta 155-177 had >80% WT SA and hufIXHSA <20% WT SA. Following intravenous injection in fIX knockout mice, recoveries did not differ for WT and, but were lower for hufIXHSA. The terminal catabolic half-life of delta 155-177 was reduced while that of hufIXHSA was increased two-fold relative to WT. The observed area under the clearance curve (AUC) was reduced 10-fold for delta 155-177 versus WT, but only 2-fold for hufIXHSA. Enzymatic removal of the Nlinked glycans in Benefix accelerated its clearance, but not to the same degree as deleting residues 155-177. Conclusions: Reducing the molecular volume of fIX by reducing its polypeptide chain length and/or its degree of post-translational modification accelerates its clearance, while fusion to albumin slows its terminal clearance. The latter gain in terms of improving fIX as a blood product is offset by losses in fIX procoagulant activity and initial recovery.

S Picker, S Stürner, L Oustianskaja, B S Gathof, University of Cologne, Cologne, Germany BACKGROUND: Leukodepletion of blood products prevents leukocyteinduced transfusion complications. Filtered whole blood (WB) is already widely used for autologous donations. If component-like quality parameters can be achieved by prestorage leukodepletion of WB, filtered WB could be rational also for homologous transfusion, particularly in instances where red cell and blood volume deficit are simultaneously present. DESIGN: 16 units of WB in CPDA donated by healthy volunteers were leukofiltered prestorage and stored for up to 49 days. Nonfiltered WB units in CPDA (n = 16) and filtered red blood cell concentrates in SAGM (RBC, n = 14) served as controls. Several haematological, biochemical and coagulatory parameters were determined. RESULTS: Apart from significant differences in haematocrit (56.2 ± 3.6 vs. 37.9 ± 3.9%), and plasma concentrations of free haemoglobin (93.1 ± 37.8 vs. 57.8 ± 24.3 g/dL), K + (38.9 ± 5.3 vs. 31.5 ± 4.3 mmol/L), and ATP (2.7 ± 0.2 vs. 1.6 ± 0.4 mmol/g Hb) with higher levels in RBC, no remarkable differences were observed with haemolysis (0.23 ± 0.07% vs. 0.31 ± 0.13) and pH value (6.63 ± 0.03 vs. 6.62 ± 0.02) between filtered RBC and WB at storage end. Lack of leukodepletion manifested in significantly higher rates of haemolysis (0.44 ± 0.21%), free haemoglobin (89.6 ± 43.5 g/dL), and lower pH values (6.56 ± 0.04). Until storage day 42 filtered WB units had haemostatically effective amounts (% of the initial mean values) of stable (F XI 97.5%) and labile clotting factors (F V 92.9%, F VIII 69.2%), and inhibitors (AT III 89.0%) while markers of activated coagulation had not increased. CONCLUSION: Our data indicate a quality of filtered WB comparable to filtered RBC. Furthermore, prestorage leukoreduction of WB preserves coagulation activity until 42 days of storage. Leukodepleted WB could be an interesting option to facilitate and economize blood supply, especially for surgical or trauma patients. The clinical suitability remains to be investigated by means of clinical studies. Backround: The transfusion of platelet concentrates is one possibility to treat acute bleeding. Until now the quality of a concentrate is set in volume, platelet count, pH, swirling and sterility. But there are still many unsuccess-ful platelet transfusions in patients. Another quality criterion to improve the quality of the concentrates might be to test the adhesion of the platelets in low shear rates via fibrinogen in vitro with the platelet adhesion assay (PADA) from HaemoSys GmbH Jena. We compared pool preparation versus apheresis preparation of platelet concentrates on different days of storage. Methods: 20 apheresis concentrates of the Trima ® Gambro BCT and 20 pool concentrates out of blood of 4 whole blood donors are tested on day 0, day 2, day 4 and day 6. PADA is performed according to the instructions of the manufacturer. Platelets of the sample to be measured adhere to added special test particles. As control, a sample without test particles is used. In sample and control the platelet count is measured. The measure of PADA is the adhesion index (AI) with a value of 50 ± 20 (mean value ±2SD) in citrated blood of whole blood donors. The P-LCR is measured by the Sysmex ® . Additionally the citrate concentration is measured in both concentrates. Results: The average AI of the apheresis concentrates on day 0 is 45 and it drops down to 12 continuously until day 6. The pool concentrates are more stable: on day 0 the AI is 34 and on day 6 it is still 25. With 17.36% on day 0 and 17.83% on day 6 the P-LCR is also unaltered in the pool concentrate. In the apheresis concentrate the P-LCR increase from 12.7% on day 0 to 24.2% on day 6. The citrate concentration is nearly the same in both types of preparation. Conclusions: According to the test the adhesion function of the pool concentrates is more stable, lower on day 0 of storage but much better on day 6. The increasing P-LCR in the apheresis concentrates can be a sign of loss of function. Pursuant to the PADA the apheresis concentrates should not be transfused after day 4 and the transfusion of pool concentrates after day 5 of storage can be discussed. This has to be verified in further studies, possibly in transfusion success in patients.

G Stiegler, Clinic for Blood Group Serology and Transfusion Medicine, Vienna, Austria; V Mayaudon, Baxter R&D Europe S.C.R.L., Nivelles, Belgium; G Leitner, P Höcker, Clinic for Blood Group Serology and Transfusion Medicine, Vienna, Austria Background Apheresis Platelets have to be suspended in 53-68% additive solution (InterSol TM , Baxter) when the pathogen reduction treatment with the Intercept TM Blood System is applied. When cell separators without automatic addition of InterSol TM are used, the plasma volume of the platelet concentrates has to be reduced by centrifugation (Intercept TM Preparation Set). Subsequently, platelets are resuspended in InterSol TM before the photochemical treatment is performed. Methods First, this study evaluated the performance of the Intercept TM Preparation Set and Blood System with 12 apheresis platelet concentrates produced with the ComTec TM (Fresenius, Germany). Second, platelets' in vitro characteristic (pH, glucose, lactate, LDH, hypotonic shock response, aggregation with collagen) were followed over an extended storage period of 7 days.

Results are given in mean and (standard deviation), *p < 0.02 versus previous evaluation step (Wilcoxon Matched Pairs for day 2, and Friedman Anova for the days 2-7). Conclusion The Intercept TM Preparation Set used in combination with ComTec TM apheresis platelets is capable of producing platelet units meeting the requirements for Intercept TM Blood System. The whole processing steps of Intercept TM Preparation Set and Blood System cause a mean platelet dose loss of 16.5% which has to be considered for reaching satisfying platelet increments after transfusion. A gradual decline 66A SCIENTIFIC SECTION TRANSFUSION 2004-Vol. 44, Supplement of platelets' in vitro characteristics is observed during storage of 7 days. The pH levels decrease after the treatment and during storage but well comply with international quality guidelines.

H Ohto, S Ezuki, Fukushima Medical University, Fukushima, Japan; T Ito, Kawasumi Laboratories, Tokyo, Japan; K Kawabata, Fukushima Medical University, Fukushima, Japan Background: The short shelf-life of platelets makes it difficult to stockpile them and this is a problem especially for countries with ageing populations. This study evaluated a newly developed platelet bag (PO-80, Kawasumi, Japan) highly permeable to oxygen to extend the expiration period to 7 days or more. Methods: Single-donor apheresis platelets (4.3 ¥ 10 11 mean) from donors with the same blood type were pooled, divided equally into two bags, PO-80 and PL2410 (control, Baxter), which have 2660 and 2024 ml/m 2 ·day·atm of oxygen permeability respectively, and stored for up to 9 days at room temperature with horizontal agitation. On days 0, 1, 3, 5, 7 and 9 of storage, swirling, platelet counts, mean platelet volume, pO 2 , pCO 2 , pH, glucose, lactate, %HSR, aggregation with ADP and collagen, and P-selectin expression were measured. Five experiments were done. Results: The swirling pattern was preserved better for up to 9 days in PO-80 (5/5) than in control (4/5) bags. The mean consumption rate of glucose in PO-80 and the control were 36.1 and 35.0 (mM/10 12 Plt/Hr), respectively. Aggressive drop of glucose (0 mmol/L) with prominent lactate accumulation (18.2 mmol/L) was observed on day 9 in one of five control bags. The pO 2 level in control bags dropped more significantlly by 29% (55.2 mmHg) on day 3 than in PO-80 bags (73.1 mmHg) compared with the initial level (78.3 mmHg) (p < 0.001). These results suggest that aerobic metabolism of high concentration platelets was maintained better with greater O 2 consumption and less lactate generation in PO-80 bags than in control bags. The %HSR and aggregation decreased gradually in a similar manner in both bags until day 7, and became a detrimental defect in one of five control bags on day 9. P-selectin expression was slightly higher in control bags than in PO-80 bags with no statistical difference. In one control bag P-selectin expression reached to 84% and was accompanied by a loss of swirling. Conclusion: The biochemical and functional characteristics of platelets at a high concentration were preserved better for 9 days when stored in a bag with higher oxygen permeability (PO-80) than in a marketed second generation bag.

C Stebler, Red Cross Blood transfusion Center, Basel, Switzerland; T Paglieroni, P Holland, Blood Source, Sacramento, CA Background: Pathogen inactivation minimizes bacterial growth of contaminated platelet components. The INTERCEPT (Amotosalen HCL) pathogen inactivation system is currently used in combination with PAS III (P) storage solution (35% plasma and 65% PAS III). Studies have suggested that a modified PAS III solution (MP) containing magnesium and potassium may improve the quality and shelf life of stored platelets. This study compared storage parameters in pathogen-inactivated apheresis platelets stored in two different PAS III solutions (P and MP). Methods: Ten leukoreduced-platelet units (single needle apheresis procedure, Amicus cell separator, Baxter Fenwal) were collected and split. P solution was added to one portion, MP to the other. Platelets were then processed with the INTERCEPT blood system for platelets and stored for 8 days. The following parameters were analysed on day 0, 1, 3, 6 and 8: pH, Lactate, Glucose, HCO 3 , and the platelet parameters extent of shape change (ESC) and hypotonic shock response (HSR) plus CD63 and CD62P expression. The BACTAlert system was used to detect bacterial contamination. Results: All units collected had less than 1 ¥ 10 6 residual WBC/unit and no bacterial growth was detected after 8 days storage. Storage Parameters in Pathogen-Reduced Apheresis Platelets Stored in Two Different Storage Solutions (P and MP) Results = mean of 10 units; * p< 0.05 MP when compared against P.

Conclusion: Paired studies of pathogen-inactivated apheresis platelets stored in modified PAS III show better in vitro storage parameters than in standard PAS III. Most differences are statistically significant up to day 6, but on day 8 the results are essentially equal.

J Ringwald, V Weisbach, R Zimmermann, E Strasser, S Achenbach, University Hospital Erlangen, Germany, Erlangen, Germany; M Antoon, Gambro BCT, Zaventem, Belgium; R Eckstein, University Hospital Erlangen, Germany, Erlangen, Germany Background: Apheresis PLT concentrates in PLT additive solutions (PAS-APCs) have become the focus of interest. Most studies about PAS-APCs concentrated on the influence of different PAS on PLT quality. With increasing use of PAS-APCs questions arise on possible impact of manipulations on PLT quality. We compared in vitro parameters of PAS-APCs with different hold time prior to addition of two PAS. Methods: 20 blood donors underwent two PLT apheresis with Trima ACCEL (interval >/= 2 weeks) collecting on each occasion an APC at a concentration of 4000 ¥ 10e3 PLT/ml in 150 ml plasma which was split into 2 single products containing 3 ¥ 10e11 PLTs. Either 140 ml of PAS II (donation 1) or PAS IIIM (donation 2) was added after 2 or 8 h hold time resulting in PAS-APCs of 225 ml with concentration of 1400 ¥ 10e3 PLT/ml and PAS proportion of 65%. APCs were stored for 7 days. Samples were taken on days 1, 5 and 7. Lactate, glucose, MPV, pH, LDH, CD62-expression (CD62), and hypotonic shock response (HSR) were tested. Results: CD62, HSR, and MPV did not differ significantly between hold times of 2 h or 8 h at day 5 and 7. Further data are given in table 1. PLTs stored in PASIIIM showed lower levels for CD62, LDH, and pH at day 5 and 7 and lower daily lactate increase per 106 PLT during storage than PLT in PAS II. HSR values were higher for PAS IIIM at day 5 and 7 but differed significantly only at day 7. Conclusions: PLTs stored in PAS IIIM maintain better function and cell integrity than PLTs in PAS II. Although difference in hold time of 6 h does not have an impact on PLT function during storage there seems to be an influence on PLT metabolism as PLT with a 2 h hold time showed higher lactate production and glucose utilisation. 

S Miyata, S Yamamoto, H Yoshimura, T Kawai, T Sano, National Cardiocascular Center, Suita, Japan; H Ohto, Fukushima Medical University, Fukushima, Japan; M Sugimoto, Nara Medical University, Kashihara, Japan Background: In Japan, the storage of platelet concentrates (PCs) is still restricted to 3 days, even though the limitation in the supply of PCs has become a serious problem recently with the aging of Japanese society. To increase the storage time from 3 to 5 days, or longer, evaluation of the quality of PCs stored for 5 days and over will become crucial. The quality of PCs depends mainly on the incidence of bacterial contamination and the function of stored platelets. For the latter, however, it is still unclear what kind of the in vitro test optimally predicts the in vivo behavior of transfused, stored platelets. Methods: We observed real-time processes of thrombus formation of stored platelets on a collagen surface under shear stress conditions utilizing a parallel plate flow chamber, which reproduces physiologic blood flow conditions in vitro. To mimic transfusion process, platelet-depleted blood (platelet count <20,000/mL) was made from whole blood, drawn from a blood type-matched healthy volunteer, by removing platelet rich plasma (PRP). PC and the volunteer's platelet poor plasma (PPP) were added to plateletdepleted blood to give final hematocrit and platelet concentration of 45% and 150,000/mL, respectively. PRP or PPP was added instead of PC as a positive or negative control, respectively. The reconstituted whole blood was aspirated across the collagen-coated coverslip assembled into a parallel plate flow chamber. Real-time processes of platelet thrombus formation on the collagen-coated surface was observed under both high and low shear stress conditions. After 10 minutes of perfusion, a series of z-sectional images of platelet thrombi was captured throughout the height of thrombi. To calculate thrombus height and volume, a 3D image of entire thrombi was reconstructed by a 3D measurement software. Results: The volume and height of stored platelet thrombi tend to decrease as a function of the storage time, although initial platelet adhesion on the surface was comparable to positive controls independent of storage time up to 7 days. The height of 5-day-stored platelet thrombi was about half that of 1-day-stored platelet thrombi. The process of 7-day-stored platelet thrombus growth was damaged in a manner similar to that in patients treated with low-dose aspirin. Conclusions: Our results suggest that long time storage may cause irreversible damage to the activation process of platelets to some extent. This storage lesion may become a problem in using 7-day-stored PCs for the treatment of massive bleeding in surgical patients, although it may not become evident for the prophylactic use in patients with chemotherapyinduced thrombocytopenia.

V Leytin, D J Allen, A Gwozdz, B Garvey, J Freedman, St. Michael's Hospital, Toronto, ON, Canada BACKGROUND: Role of P-selectin (CD62) and glycoprotein (GP) Iba in posttransfusion clearance of platelet concentrates (PCs) is unclear. STUDY DESIGN AND METHODS: Platelet activation in vitro was determined by flow cytometry using anti-CD62 and anti-GPIba antibodies. PC clearance in vivo was evaluated in an animal model using rabbits with inhibited reticuloendothelial system, as measured by 0.5-hour (R 0.5 ), 24-hour (R 24 ) and total (R S ) platelet recoveries, and survival time (ST). Correlations were analysed between in vitro assays of platelet activation and in vivo clearance of conventional (Day 2-5), outdated (Day 7-8) and refrigerated PCs. RESULTS: Binding of anti-CD62 was significantly increased and of anti-GPIb decreased in outdated and refrigerated PCs compared to conventional PCs. Negative correlation was observed between in vitro anti-CD62-binding and their fast (R 0.5 ) clearance but not with delayed (R 24 and ST) clearance. In contrast, anti-GPIba-binding showed positive correlations with delayed but not with fast platelet clearance. Overall (R S ) clearance correlated better with anti-GPIba-than with anti-CD62-binding. CD62 density on the platelet surface was decreased after PC transfusion, whereas GPIb density remained unchanged. CONCLUSION: These data suggest that CD62 exposure on the platelet surface during PC storage triggers fast CD62-mediated PC clearance, whereas in vitro GPIba changes are involved in delayed GPIbamediated PC clearance.

Background: Several automatic systems exist for the fractionation of whole blood based on the buffy coat (BC) method. Methods: We analyzed the efficacy of two automatic fractionation systems in the routine preparation of leucoreduced pooled platelet concentrates (PC) using the BC top & bottom method. Four hundred fifty mL of whole blood were collected from donors in quadruple "top & bottom" blood bag systems (Baxter Fenwal, La Châtre, France). Blood bags were centrifuged at 5000 g for 10 min and plasma and red blood cell concentrates were then removed, leaving the BC in the primary whole blood bag. In a first period of the study (January 1 through July 15), blood fractionation was performed with a specially adapted Compomat G3 (NPBI, the Netherlands), and in a second (up to December 31) separation was performed using an Optipress II (Baxter). After blood screening, 5 BC of the same ABO group were pooled with a platelet additive solution (T-Sol, Baxter). The pool was centrifuged at 400 g for 7 minutes in a specially designed centrifuge adaptor. The supernatant was transferred through a Sepacell filter to a 1.3 L bag for storage. Immediately after preparation of pooled PC, a sample was removed and collected in EDTA for counting in a hematological analyzer (MDII, Coulter). Data were analyzed with SPSS v. 11.5.1 statistical package.

Conclusion: Platelet content in PC routinely prepared from whole blood donations using the "top & bottom" BC removal method is higher with Optipress II than with Compomat G3, although the final volume of the product is greater.

T L Yamamoto, K A Nguyen, N Hirschler, Blood Centers of the Pacific, San Francisco, CA Introduction: Recent standards require procedures to limit and detect bacterial contamination in all platelet components. Current FDA-licensed bacterial detection methods are approved only for leukoreduced platelets (LR plt). The Leukotrap RCPL inline filter system for the production of LR red blood cells and whole blood (WB) derived plt was evaluated in a blood center setting. We describe the evaluation of the RCPL system for LR plt production. Materials and Methods: 53 WB units were collected with the RCPL sets. Baseline WBC and plt count were performed on the WB units. Plt-rich plasma was produced using manufacturer recommendations and LR using the inline filter. Post-production WBC and plt counts, volume, and pH were measured. Acceptance criteria, based on AABB standards, are listed in Table 1 . Results: The results are summarized in Table 2 . All acceptance criteria were met except for the plt yield, which failed by 3%. Possible root causes for yield failure were investigated. The plt yield did not correlated with centrifuge serial number or settings, individual staff technique, or RCPL filter performance. There was an average of 8.8% reduction in plt yield from Day 1 to Day 5. However, this was not the major cause. The most significant factor was the positive, although not absolute, correlation between the pre-LR and post-LR plt count. When WB units with plt counts of ≥190,000/uL were used, plt yields met criteria. Conclusion: LR plt produced with the Leukotrap RCPL System were did not meet acceptable yield criteria. The major factors were the high requirement for plt yield and initial lower plt counts in the WB units, rather than filter performance or process failure. Background The OrbiSac BC System is a new method that involves automated standardized processing of leukoreduced PCs from BCs. Annually, our blood center handles 70.000 units of whole blood and produces 9.000 BC derived PCs. We validated the system in regard to functionality and handling as well as platelet yield and the leukoreduction efficiency. In addition, we assessed metabolic and platelet activation parameters during storage. Methods Four randomized BCs and 300 ml Platelet Additive Solution (T-Sol) were sterile connected to the OrbiSac BC disposable, after which the pooling, centrifugation, platelet expression and leukoreduction were automatically performed in the OrbiSac BC System. The OrbiSac method was compared with our manual preparation method. This involves pooling of four randomized BCs together with 300 ml T-Sol into a bag with a leukoreduction filter. After centrifugation at 480 g for 5 min, the PC was transferred to the storage bag by using an Optipress II. Results Functionality and handling Two technicians produced 18 and 51 OrbiSac-PCs, respectively. One found the system improved the handling, the other saw no real improvement in handling compared to the manual method. Recovery of platelets and leukoreduction efficiency. Background: WBC fragments may induce HLA-immunization and other adverse events in recipients. We investigated the occurrence and production of WBC fragments in platelet concentrates (PCs) and plasma units during storage and filtration using real-time (RT) PCR and flow cytometry. Methods: To study the occurrence of WBC fragments "male" WBCs were spiked into double filtered "female" PCs in a concentration series of 0.03-100 WBCs/mL (n = 4 per level). To study the production of WBC fragments "male" WBCs were spiked into "female" plasma units to 4 ¥ 10 9 WBCs/L and stored at room temperature prior to filtration (n = 4 per storage time; t = 0, 24 or 48 h). DNA was measured by both albumin-RT-PCR (for "female" and "male" DNA) and Y-RT-PCR (for "male" DNA). Intact WBCs were counted by flow cytometry. The amount of WBC fragments was calculated by subtracting cell-free DNA (RT-PCR on supernatant) and intact WBCs (flow cytometry) from total DNA amount (RT-PCR). Results: Spiking of "male" WBCs into "female" PCs showed that the Y-RT-PCR is linear and had a reproducible quantitative range down to 0.03 WBC/mL, but that the albumin-PCR in unspiked samples revealed a total amount of 6-10 WBC-equivalents (eq)/mL. After centrifugation, half of this (3-5 WBC-eq/mL) was observed as cell-free DNA in the supernatant, suggesting that the remaining DNA derived from WBC fragments. Results of plasma units after filtration are shown in the table.

Conclusion Our blood center would require five OrbiSac BC Systems. The platelet yield in OrbiSac produced PCs was 8% higher than our standard prepared PCs. The number of leukocytes was below the limit of specification (1 ¥ 10 6 ), but higher than in our standard PCs. The quality of the OrbiSac produced PCs was high judged by metabolism and activation parameters during storage for up to seven days.

Conclusion: WBC fragments occur in WBC-reduced PCs and increase when products are stored prior to filtration up to levels that might induce (secondary) HLA-immunization. Therefore, preparation methods for WBCreduced PCs should be optimized not only to minimize the number of WBCs but also the amount of WBC fragments.

Background: Autologous platelet gel (APG), which is platelet rich plasma (PRP) plus calcified thrombin (CT), is currently being advocated as a means of improving post-surgical healing in a variety of procedures. The release of platelet (Plt) and white cell growth factors into the surgical wound may accelerate bone growth and tissue repair. This study was designed to validate the performance of two devices for the production of APG. Methods: Units of freshly collected blood were manipulated to produce whole blood (WB) aliquots with a range of starting hemoglobins (Hb). WB was then processed into PRP in triplicate at each Hb level using the Magellan (Medtronics, Minneapolis, MN) and Symphony II (DePuy Spine, Raynham, MA) devices. Each PRP was gelled with CT and incubated in a water bath at 37 degrees C for 24 hrs. Plt, Hb, and fibrinogen concentrations as well as P-selectin expression were measured before and after PRP processing. Concentrations of transforming growth factor beta (TGFb), platelet-derived growth factor (PDGF-AB), and vascular endothelial growth factor (VEGF) were measured after processing and at 0.5, 4, and 24 hours from the onset of incubation. Results: (sGP1B) is indicative of poor quality in stored platelets. A relationship between these surface markers has not been previously described. Methods: Prestorage leukoreduced whole blood derived platelets were manufactured (day 0) using the leukotrap PL ® system (Pall Medical, Covina, CA) and stored with agitation in CLX containers at 20-24°C for 9 days. Samples were taken on days 1, 5, 7 and 9. These samples were labelled with FITC-conjugated Annexin V (for sPS detection) or PE-conjugated anti CD 42b (for sGP1b detection) and tested in a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA). The results of gated platelets were expressed as percent surface expression. Data were analyzed using Pearson's correlation. Results: Nine products were studied. sGP1b decreased and sPS increased with storage time, as expected. The overall correlation (r) was -0.77 (p < 0.01) for all data points. A subpopulation of donors exhibited a marked decrease in sGP1b and increase in sPS on day 1. This pattern became more pronounced throughout the storage period. Day 1 sPS correlated with day 5 sPS, r = 0.89 (p < 0.01); day 5 sPS correlated with day 7 sPS, r = 0.95 (p < 0.01) and day 7 sPS correlated with day 9 sPS, r = 0.73 (p = 0.02). Means ± 1SD are shown in the table for each study day. solution by either apheresis or buffy coat method. Multiple PCs were pooled and divided over 6 storage containers, to obtain 3 PCs with a "low" (groups A,B,C) and 3 with a "high" platelet concentration (groups D,E,F). Units were not agitated between day 1-3 (groups A,D), day 1-5 (groups B,E), or were continuously agitated (control groups C,F). Units were sampled on day 1, 5, and 7, and pH was measured and hypotonic shock response (HSR) was determined. Results: Day 7 results, see table (mean ± SD, n = 20). The composition of the PCs was not different. The controls had significantly higher pH values than the study groups (all p < 0.001, except D vs. C,F: not significant). HSR was significantly lower in group E (vs. A, B: p < 0.05; vs. F: p < 0.01; vs. C: p < 0.001). However, in the low concentration group, the differences were not significantly different from the control group. Conclusion: WBC-reduced PCs in PASIIIM can withstand up to 4 days without agitation with maintenance of acceptable pH and HSR responses, provided the platelet concentration is not too high.

Conclusions: sGP1b and sPS, as early as day 1, could be sensitive indicators of platelets from donors who store poorly in vitro and may possibly be predictive of poor in vivo recovery or survival. 

M O Barrera, C J Jennings, S West, P Schuman, S Inouye, P Washington, Gulf Coast Regional Blood Center, Houston, TX Background: As automated red blood cell (RBC) collections become more prevalent, sterile dock filtration will continue to be an important means of leukoreduction (LR). This study evaluated the performance of the Baxter Sepacell R500II sterile dock leukocyte reduction filter set with RBCs collected using the Haemonetics MCS+ 8150, which were refrigerated and filtered 68-72 hours post-collection. Methods: Sixty RBC units were collected using the MCS+ 8150 (Haemonetics Corp., Braintree, MA). These units were refrigerated and filtered with the Sepacell R500II (Baxter Healthcare, Deerfield, IL) 68-72 hours post-collection. Pre-and Post-filtration hematocrits were performed on the Cell-Dyn 3700 (Abbott Laboratories, Abbott Park, IL) to determine RBC percent recovery. Residual leukocytes in the final product were counted using the FacSCAN flow cytometer (BD Biosciences, San Jose, CA). Results: Relevant data is presented in the table below.

Conclusion: This study demonstrated that the R500II sterile dock filter yielded red cell units that met or exceeded the FDA and AABB standards for leukoreduction when refrigerated and filtered 68-72 hours post collection. All LR-RBC units had white blood cell (WBC) < 5 ¥ 10 6 . Pharma MTL1-WB system (Test), to the Baxter RS 2000 system (Control). Study Design & Methods: Healthy volunteers donated 500 ± 50 mL of Whole Blood in CPD. Each site evaluated 20 subjects with the Test products (in vitro and in vivo), and 8 subjects with the control products (in vitro only). Ten test units and 4 control units were filtered at room temperature (R.T) 5.5 ± 1.5 hours after collection, and 10 test units and 4 control units were filtered at 1-6°C on day 3 after collection at each site. AS-1 RBCs were prepared and stored at 1-6°C for 42 days. Sampling was performed on WB (pre-and post-filtration), and AS-1 RBCs (pre-storage and day 42). In vitro tests included: filtration time, residual WBC (Nageotte), % RBC recovery (by weight), plasma hemoglobin, sodium, potassium, glucose, blood gases, and ATP levels. Standard methods were used for all assays. In vivo 24-hour % recovery was measured using the 51 Cr/ 99 Tc double label method. Results: Relevant data (Mean ± SD) are shown in the table.

Conclusion: Transfer of AS1 RBC from RT to 1-6°C during filtration with the Leucolab system does not adversely affect filtration outcome and should be an acceptable option during production of leukocyte reduced RBC. It is probable that additional testing will support use of the Leucolab system with CP2D/AS3 RBC.

C J Jennings, M Barrera, S West, Gulf Coast Regional Blood Center, Houston, TX; M Charboneau, W Goforth, C Lusk, J Klein, H Sowdagar, K Riggins, V Mason, Oklahoma Blood Institute, Oklahoma City, OK Background: As automated red blood cell (RBC) collections become more prevalent, sterile dock filtration will continue to be an important means of leukoreduction (LR). Some currently marketed sterile dock filters require a hold time of 6-8 hours prior to filtration. This study evaluated the performance of the Baxter Sepacell RS-2000 ® sterile dock leukocyte reduction filter set with RBCs collected using the Haemonetics MCS+ 8150, which were filtered either immediately at ambient temperature or refrigerated and then filtered 6-8 or 68-72 hours post-collection. Methods: Three hundred thirty RBC units were collected using the MCS+ 8150 (Haemonetics Corp., Braintree, MA) and filtered with the Sepacell RS-2000 ® (Baxter Healthcare, Deerfield, IL). Units within Group A (n = 98) were filtered at ambient temperature immediately after collection. Units within Group B (n = 120) and C (n = 112) were refrigerated and filtered at 6-8 or 68-72 hours post-collection respectively. Pre-and post-filtration hematocrits were performed on the Cell-Dyn 3700 (Abbott Laboratories, Abbott Park, IL) to determine RBC percent recovery. Residual leukocytes in the final product were counted with the FacSCAN/FacSORT flow cytometer (BD Biosciences, San Jose, CA). Results:

Conclusions: Data from this study suggests that leukoreduction with the Leukoflex MTL1-WB system is highly efficient and meets FDA guidelines for 24-hour in vivo recovery (≥75%) and in vitro RBC Recovery (≥85%), residual WBC (£5 ¥ 10 6 /unit), and hemolysis (£1%).

G M George, L Kline, American Red Cross Holland Laboratory, Rockville, MD Background: A new leukoreduction system for red blood cells (MacoPharma Leucolab) was recently introduced in the US for use at room temperature (RT) or 1-6°C with red blood cells (RBC) with CPD anticoagulant and AS-1 additive solution (AS1 RBC). In many cases workflow may be optimized by beginning filtration at RT, then transferring the filtration in progress to 1-6°C for completion. However, the manufacturer's instructions for use (IFU) do not address use of the Leucolab under these conditions. This study was designed to demonstrate that filtrations with the Leucolab started at RT and completed at 1-6°C meet the FDA guidelines of less than or equal to 5 ¥ 10 6 residual leukocytes (WBC) per product and at least 85% red cell recovery post filtration. In addition, a small number of AS3 RBC (CP2D anticoagulant and AS3 additive solution) were evaluated for use with the Leucolab. Study Design: AS1 or AS3 RBC were prepared from 500 ± 50 mL whole blood (WB) and held at RT for less than 8 hours post-collection. Test units (N = 32 AS1 and 3 AS3 RBC) were filtered according to the IFU except that filtrations were initiated at RT and then transferred to 1-6°C when filtration was approximately half completed. Filtration of control units (N = 8 AS1 and 3 AS3 RBC) was performed entirely at RT. The head height was maintained at 60 inches throughout the filtering process regardless of temperature. Filtration time, plasma hemoglobin (Hgb), residual WBC and % RBC recovery measured by weight (wt) only and by wt + Hct were evaluated on the day of filtration. Results: Regardless of product type or filtration conditions, all filtered products had residual WBC below the sensitivity level of the Nageotte assay (<0.5 ¥ 10 6 WBC per unit) and plasma Hgb levels that correspond to <0.1% hemolysis. Filtration times (min) and RBC recovery (%) are presented as mean ± 1 SD (range) for CPD/AS-1 RBC; and only as a range for CP2D/AS3 RBC (due to the small number of units evaluated).

Conclusion: This study demonstrated that the Sepacell RS-2000 ® (Leuko-Connect) sterile dock filter yielded red cell units that met or exceeded the FDA and AABB standards for leukoreduction. In Groups A, B and C, 94 out of 98 (96%), 118 out of 120 (98%), and 111 out of 112 (99%) LR-RBC units had white blood cell (WBC) uniformities of < 1 ¥ 10 6 . In total, 327 out of 330 (99%) LR-RBC units met the AABB WBC standard of < 5 ¥ 10 6 . These results suggest that the RS-2000 ® (LeukoConnect) sterile dock filter may be used to LR automated RBCs immediately at ambient temperature and up to 68-72 hours refrigerated after collection.

D D Elliott, E Raymond, K Stauffer, Puget Sound Blood Center, Seattle, WA; R Juntunen, Puget Sound Blood Center, Renton, WA Background: Red blood cells (RBCs) with an additive solution that are prepared from units of whole blood (WB) are filtered to remove the residual leukocytes (WBCs). Units with visible clots are discarded prior to attempting filtration. Additional RBC units are discarded during the filtration process when clots clog the filter. Method: One factor that may have an impact on clot formation in a unit of blood is the timing of stripping the blood bag tubing to mix the blood in the tubing with the anticoagulant in the collection bag. After the desired amount of blood is collected, our current procedure is to collect the virology samples, remove the needle from the donor's arm, take the unit to a sealing station and strip the blood bag tubing. A trial was performed at our fixed collection sites where the blood bag tubing was stripped at the bedside immediately following the collection of the blood and prior to the collection of virology samples. Results: With the current procedure, the average time from the end of the blood collection to the stripping of the blood bag tubing was 2.14 minutes. During the trial, the average time was 25 seconds.

71A 2004-Vol. 44, Supplement Conclusion: While reducing the average time between the end of the blood collection to the stripping of the blood bag tubing resulted in a decrease in both the number of units that were clotted and the number of units that failed leukoreduction, the results are not significantly different between the control and trial groups (p = 0.6 and 0.5 respectively).

D D Elliott, E Raymond, Puget Sound Blood Center, Seattle, WA; R Juntunen, Puget Sound Blood Center, Renton, WA Background: Red blood cells (RBCs) with an additive solution that are prepared from units of whole blood (WB) are filtered to remove the residual leukocytes (WBCs). Units with visible clots are discarded prior to attempting filtration. Additional RBC units are discarded during the filtration process when clots clog the filter. Method: An important step in the collection process to prevent clotting is to ensure that the WB is adequately mixed with the anticoagulant in the blood bag. At our blood center the staff mix the WB with the anticoagulant at the beginning of the collection process and several times through out the donation. A trial was performed at our fixed collection sites using two different types of automated collection mixers where the blood would mix continuously throughout the donation process. Results: clots present or an extended collection time. One control unit was found to have a higher titer than both the failure samples and another unit had all three results with a titer of 1 : 1 at 4°C. The remaining 12 units had a range for the room temperature tube testing of 1 : 32 to 1 : 256 at 4°C, with the refrigerated segment testing range of 1 : 2 to 1 : 64 at 4°C and control tube testing range of 0 to 1 : 2 at 4°C. Conclusion: Naturally occurring cold agglutinin antibodies appear to be responsible for filter failures when performed at cold temperatures. Segment testing was not a good indicator of the presence of cold agglutinins due to the cold auto adsorption of the antibodies while the filtration was occurring in the refrigerator. Background: An increasing number of molecular methods for genetic disease diagnoses requires automated formats for high throughput DNA extraction and processing. Leukodepletion technology (Leukosorb TM ) was specifically developed for leukocyte depletion of blood for transfusion products. We evaluated the use of Leukosorb TM to entrap leukocytes from whole blood directly to provide a source of DNA for PCR analyses and diagnostic tests. Methods: Using proprietary compression seal technology we were able to seal Leukosorb TM into a 96-well Acroprep plate format. EDTA anticoagulated whole blood (200 ul), stored up to 7 days at 4°C, was added to each well and filtered within 1 minute using vacuum. Leukocytes were trapped on the matrix, lysed in situ, and, following several washes, DNA was eluted from the membrane. Depending on the downstream application, DNA was eluted directly into a collection plate or into an ultrafiltration plate (Acroprep UF) for concentration and desalting. After collection DNA quality was evaluated using PCR amplification with 3 different primer sets namely, VTNR, and the C282Y and H63D that represent mutations associated with hereditary hemochromatosis (HH). The fidelity of the amplification was validated using Restriction Fragment Length Polymorphism (RFLP; Bcl1 for H63D and Rsa I for C282Y) and single nucleotide polymorphism (SNP) assays using Pyrosequencing technology. DNA quality was compared between samples extracted from whole blood using silica gel or other binding matrices. Results: High molecular weight DNA suitable for PCR analysis was prepared routinely using the Leukosorb TM method. The yield of genomic DNA from 200 ml of whole blood was adequate for PCR and other downstream assays. We estimate that the processing time for DNA extraction using the 96-well format is approximately 90 minutes. PCR amplification from 3 primer sets produced the correct molecular weight DNA bands, hybridized as expected in RFLP assays, and gave the correct sequence in Pyrosequencing assays. Results obtained from the Leukosorb TM DNA were comparable in all respects with the DNA extracted using a silica gel binding kit or other binding matrices. Conclusions: The use of Leukodepletion technology, together with a multiwell filter plate, enhances automation of DNA sample prep directly from stored whole blood. The resulting samples can be used for PCR or SNP mapping when screening large populations of patients.

V A Collins, T E Reynolds, M F Bertholf, Florida-Georgia Blood Alliance, Jacksonville, FL Background: Due to increasing customer demand, the leukocyte-reduced red cell inventory at our blood center had grown to approximately 70% by the fourth quarter of 2003. All leukoreduction was being performed with dockable filters by Components Laboratory staff. A solution was sought to relieve the increasing pressure for space and personnel in the laboratory while maintaining a growing leukoreduced red cell inventory. Methods: The Terumo IMUFLEX ® WB-RP Blood Bag System with Integral Whole Blood Leukocyte Reduction Filter blood bag allows for the production of a 42-day red cell product which can be filtered as whole blood at room temperature between 30 minutes and 8 hours after collection. This whole blood filtration system was not designed to save platelets; therefore, we selected three outlying centers that, because of their distance from the Components Laboratory, did not routinely collect whole blood for platelet preparation. Terumo staff assisted in the initial group training of collection staff from all three centers, followed by individualized training, product validation, and implementation at each center as it was converted to the in-line filter bags. As A staff survey indicated that the alarm feature, which alerts when the blood flow is slow, was favorable and staff liked not having to agitate the bag during the donation. Conclusion: The results show that both collection mixers give clotting rates that are significantly less than the control method (p = 0.02). While we also realized a reduction in the number of units that failed leukoreduction, the findings were not significantly different from the control method.

C A Aronson, R Kakaiya, J Keene, M Volny, W Patterson, LifeSource Blood Services, Glenview, IL Background: Whole Blood (WB) units collected with an in-line filter are routinely refrigerated and then filtered in the cold refrigerator. Some donor units fail to filter for no apparent reason. Of these unknown failures, one observation was the appearance of grainy or clumped RBCs in the segments and/or whole blood unit. This phenomenon is often seen when high titer (£1 : 32) cold agglutinins are present. Methods: Units found to have failed filtration overnight, for no apparent reason, were tested for the presence of cold agglutinin antibodies. Units from donors of African American decent were tested for the presence of hemoglobin S. Units were evaluated for prolonged collection time or the presence of clots. Samples were tested at room temperature and then at 4°C. An O Positive adult cell was used for titration. Initial testing was performed on 4 units with plasma from the segment line of the failed unit and the plasma from 2 other units that successfully filtered on the same day for controls. The second group consisted of 16 units with plasma from the segment line of the failed unit and a red top tube that had remained at room temperature from the same unit. A red top tube from a successfully filtered unit was run as a control. Results: Initial samples showed 2 units with no cold agglutinin titer at all and 2 units with cold agglutinin titers at 4°C of 1 : 4 and negative at room temperature. One unit was from an African American donor who tested negative for hemoglobin S. Samples appeared to have undergone a cold autologous adsorption while filtration was attempted in the refrigerator. Therefore, follow up testing compared refrigerated samples to room temperature samples. The second test group consisted of 16 failures out of 5643 WB inline filter products. None of the failed units were from African American donors therefore hemoglobin S testing was not performed. All samples tested negative at room temperature. Three units that tested negative for cold agglutinins were found to have TRANSFUSION 2004-Vol. 44, Supplement other staff rotated through these select centers, they, too, were trained on site. Results: The collection staff quickly adapted to the procedure, and integrated it seamlessly into the daily workload. This included "batching" the filtration process to maximize efficiency. Because the collection staff at these centers saw this additional step in the process as a new learning opportunity and a chance to add value to the operation, morale improved significantly. We estimate that approximately 20% of the total leukoreduced red cell inventory we produce are now filtered by the collection staff, which translates to an average of 800 fewer units per month that the Components Laboratory staff must filter using the sterile docking method. Conclusions: With thorough training and support, collection staff are fully capable of filtering whole blood units using the IMUFLEX ® WB-RP system without compromising the quality of the red cell component produced. Although the filtering of a whole blood unit immediately after collection can be more costly due to a positive test result or other reason for laboratory discard, these numbers are small in comparison with the space and labor savings realized in the laboratory.

P F Van der Meer, L A Liefting, R N Pietersz, Sanquin Blood Bank North West Region, Amsterdam, Netherlands Background: Flow cytometric white blood cell (WBC)-counting methods are used for quality control of WBC-reduced blood components. However, due to the 20 mL "neat" sample volume, the limit of accurate detection is 3 WBC/mL, equivalent to 1 ¥ 10^6 WBC/unit. Because the results can thus only be used as pass/fail criterion, we investigated a novel flow cytometric method that counts up to 70 mL "neat" sample. Methods: Double-filtered red cell concentrates (RCCs), platelet concentates (PCs) and plasma were spiked with whole blood to 1, 3 and 10 WBC/mL. The samples were lysed and subsequently stained with propidium iodide, and measured with the Personal Cell Analyzer (PCA) flow cytometer (Guava Technologies). Each sample was measured in nine-fold. We required that the accuracy (Acc.) was >80% and the CV < 20%. Results: See table (in WBC/mL; mean ± SD, n = 9). Apart from a high CV for RCCs at the level of 1 WBC/mL induced by one outlier, all levels conformed to our requirements. Conclusion: This counting method is potentially able to accurately and precisely count WBCs at or below 3 WBC/mL. A formal comparison with current counting methods must be performed.

Conclusions: The MacoPharma Leucolab LCG2 and Leucoflex MTL1 meet FDA requirements for in vitro RBC Recovery (≥85%) whether or not the calculation includes hematocrit values. Elimination of the hematocrit offers very significant savings in time and cost, and avoids variables that reduce the accuracy of the calculation. The "weight only" approach also permits use of a "look-up" table of pre-filtration weights to determine easily whether the ≥85% recovery criterion will be met, and avoids filtration of units predicted to fail due to a small starting volume. Background: Expanded volume additive solutions (AS) provide better and longer red blood cell (RBC) storage. However, patients may not tolerate the additional volume. Optimizing conventional volume AS provides another path to improving RBC storage. Study Design: Using bicarbonate buffering, phosphate supplementation, mannitol and hypotonicity for membrane protection, and alkaline pH, an optimized AS was formulated and compared to conventional and expanded volume additive solutions using10-week storage of aliquots of pooled RBC. RBC ATP concentrations, hemolysis, and membrane loss were measured during storage. Membrane loss was measured as osmotic fragility and microvesicle protein. In a second study, 12 volunteers donated units of whole blood-derived RBC which were stored for 6 or 8 weeks in the optimized additive solution followed by measures of 51 Cr in vivo recovery, hemolysis, and membrane loss. Results: RBC stored in the optimized additive solution had higher RBC ATP concentrations and reduced hemolysis compared to storage in AS-3. RBC stored for 8 weeks (n = 6) in this solution exhibited 87 ± 2% 24-hr 51 Cr in vivo recovery and 0.4 ± 0.2% hemolysis. Membrane loss after 8-week storage was less than half that observed with 6-week storage in AS-1. Conclusions: It is possible to store RBC for 8 weeks in a conventional volume additive solution with excellent recovery, reduced membrane loss, and low hemolysis. Combining buffering, metabolic support, and membrane protective strategies will allow the development of a family of better RBC storage solutions. Both are dry filters with identical materials; they differ in that LCG2 has a rigid housing and MTL1 has a soft PVC cover that permits centrifugation. These MP filters have no significant effect on the concentration of RBC, and MP petitioned the FDA to simplify the calculations by eliminating the pre-and post-filtration hematocrits. Studies demonstrated that sampling errors and imprecise hematocrits result in less accurate recovery values vs. "weight (wt) only" calculations. Methods: Data from several whole blood and RBC filtration studies indicate that the MP filters consistently have a ratio of pre-to post-filtration RBC concentration approaching 1, with approximately half of the values clustered above and half below the "zero difference" point. The RBC recovery calculations done by both methods are compared below. Results: Both RBC and Whole Blood leukoreduction filtration with MP filters show insignificant alteration of the RBC concentration in studies at 13 centers processing 715 units.

73A 2004-Vol. 44, Supplement This study investigated the nature and possible causes of WPM formation. Study design: Whole blood units were collected from 18 healthy subjects using 3 different types of collection sets. Six units were collected into sets from lots that had been implicated in observations of WPM (Baxter), six were collected into sets from the same manufacturer but from lots that had not been implicated, and six were collected with sets from a different manufacturer (Medsep). Units were divided into 4 equal parts and stored for 4 hours; two parts were stored at room temperature (RT) and two at 4°C. RBC components were then prepared from each quarter unit; two by heavy spin (5000 ¥ g) and two by light spin (2000 ¥ g). One heavy and one light spin component were prepared from the RT-stored as well as the 4°C -stored units. The divided whole blood units were inspected for WPM every 15 minutes for the first hour and every 30 minutes for the remaining 3 hours. The prepared RBC components were inspected for WPM every 15 minutes for 1 hour. Donor cholesterol levels, triglyceride levels, and complete blood counts were measured and blood counts in the whole blood and RBC components were analyzed. Results: No WPM was detected in whole blood units during the four hours of storage, but WPM was detected in at least one RBC component from 9 of the 18 donations. Among the 72 RBC components prepared from 18 donations, WPM was detected in 25 of the 72 or 35%. The 36 components prepared by heavy spin were more likely to contain WPM than the 36 prepared by light spin (50% versus 19%, p < 0.02). Among the 36 components stored at room temperature, 31% had WPM compared to 39% of those stored at 4°C (p > 0.6). WPM was found in 4 of 24 or 17% of components prepared in MedSep bags, in 9 of 24 or 37% of those prepared in non-implicated Baxter bags and in 12 of 24 or 50% of those prepared in implicated Baxter bags. Donors of RBC components with WPM had higher total cholesterol levels than donors of components without WPM (191 ± 20 vs. 163 ± 32 mg/dL, p < 0.04), but there was no significant difference in triglyceride levels, platelet, or WBC counts among the two groups. Conclusions: WPM was found in greater than one third of RBC components studied and was more likely to be present in units prepared by heavy spin and from donors with higher cholesterol levels. WPM appears to be an expected consequence of standard RBC manufacturing methods. Background. During preparation and storage at 4°C of RCC, 2,3-DPG levels fall rapidly. We developed an additive solution based on the "chloride-shift" principle (Meryman, Vox Sang. 1991), allowing maintenance of high 2,3-DPG levels, without concurrent ATP decline. Methods. Leukoreduced RCC were washed with PAGGS-M or experimental solution (PAGGS-M containing gluconate instead of saline; pH 8.2), using a Haemonetics ACP215. RCC were stored at 2-6°C and weekly sampled. Results. 2,3-DPG content of RCC washed with experimental solution increased during storage, up to 350% of initial value (7 to 25 mmol/g Hb) and started decrease after 42 days. ATP content at day 1 was 15% lower than in PAGGS-M, but at day 42 ATP was still at starting value (4.2 mmol/g Hb). In experimental solution, the intracellular pH starts at 7.4 vs. 6.4 in PAGGS-M and gradually decreased to 6.2 at day 42. At day 42, hemolysis was 0.4% vs. 0.5% in PAGGS-M and 1% of cells exposed PS (Annexin-V binding), in contrast to 3% for PAGGS-M. Glucose consumption and lactate production was increased in our experimental solution, but this was not reflected in a lower pH. Plasma stability (hemolysis induced upon dilution in plasma) was found similar for both control and experimental solution. If directly added to RCC (no wash), the experimental solution showed similar, but less pronounced and less durable effects (not shown). In experiments using PAGGS-M with a pH of 8.2 it was found that both the replacement of gluconate by saline and the increased pH were necessary to reach these results, the increased pH alone resulted in limited effects (not shown). Conclusions. Using a modified PAGGS-M, we were able to store RCC for at least 42 days with a 2,3-DPG content similar to that in fresh whole blood, without compromising other quality parameters such as hemolysis, plasma stability and ATP content.

M P Sindaco, J Matok, D Triulzi, D Rodgers, J Kiss, Insitute for Transfusion Medicine, Pittsburgh, PA Background Our blood center received a request to supply CPDA-1 Red Blood Cells (RBCs) for neonatal transfusion due to the perception that they have a lower potassium (K+) content. Since CPDA-1 units were not being collected, this would have resulted in increased cost and manufacturing complexity. The extracellular K+ content of concentrated RBCs increases during storage, as has been noted in CPDA-1 RBCs. However, data is scant describing K+ changes in CPD RBCs. We examined the levels (mMol/L) and loads (mMol) of supernatant K+ in CPDA-1 and CPD RBCs during the first week of storage. Study Design Twelve units each of CPD (500 ml) and CPDA-1 (450 ml) whole blood were collected using standard procedure. Each CPDA-1 RBC was manufactured by centrifugation at 3800 rpm for 4 minutes, followed by expression of 200-225 ml of plasma. Each CPD RBC was manufactured by centrifugation at 4000 rpm for 7 minutes, followed by removal of approximately 225 ml plasma. On the day of component preparation (Day 0), the hematocrit (HCT) of each RBC was measured and heparinized plasma from each RBC was sent for measurement of K+. Additional samples of heparinized plasma from each RBC were sent for K+ measurement on days 2, 4, 6, and 7. Results Mean potassium levels (mMol/L) and loads (mMol) were as follows (SD in parentheses):

Mean HCT on day 0 was 65.2 (SD = 8) for CPD RBCs and 65.7 (SD = 5.8) for CPDA-1 RBCs. Mean plasma volume on the packed RBCs was 150.5 ml (SD = 35.4) for CPD RBCs and 130.9 ml (SD = 22.4) for CPDA-1 RBCs. The CPD RBCs exhibited slightly higher potassium levels than did the CPDA-1 RBCs (p = NS). With respect to potassium loads, the CPD RBCs were significantly higher than the CPDA-1 RBCs on days 2, 4, 6, and 7 (all p < 0.002). There were positive correlations (ranging from 0.69-0.93) between whole blood, CPD RBC, and CPDA-1 RBC HCTs and K+ level and load. However, mean RBC HCT were not significantly different. Conclusions K+ content increased rapidly in both RBC types, 2-to 3-fold during the first two days of storage. Due to the increased plasma volume, the CPD units have higher K+ loads than CPDA-1 units but the differences are not clinically significant. In this study, an automated closed system (ACP 215, Haemonetics, Braintree MA) was evaluated for its efficiency in washing RBCs suspended in a variety of preservative solutions. RBC quality was also evaluated 24 hours post-wash. Study Design: Whole blood and apheresis derived RBC units (+/-Leukocyte reduction-LR) were collected, maintained at 4°C in standard conditions (all units held for 28 days except CPDA1 units which were held for 21 days), then washed and resuspended in saline-dextrose (0.2%). In vitro testing was performed pre-wash, post-wash (Day 0) and 24 hours later (Day 1). Parameters evaluated included Hct, Hgb, pH, ATP, lactate and, supernatant Hgb, potassium, and IgA. Results: IgA ELISA demonstrated plasma protein washout greater than 99% for all units. Post-wash (Day 1) RBC units demonstrated a mean of 0.55% hemolysis and mean ATP of 3.7 mmol/gHgb. (Table 1 ) Mean supernatant potassium decreased from 45 mEq/L to 3.3 mEq/L on Day 1. Lactate and pH were acceptable and as expected. Mean Hgb recovery for all units was 91% (±5%). ATP 24 hours post-wash correlated strongly with pre-wash ATP levels (r 2 = 0.80) Conclusion: Using this automated closed system, RBC units can be washed and stored for 24 hours with >99% plasma protein removal and satisfactory RBC quality 24 hours post-wash with £1% hemolysis and ATP ≥ 2.5 mmol/gHgb. Background: Hemoglobin (Hb) content is usually measured on an automated cell analyzer. However, these machines are large, expensive, require maintenance and need trained personnel to operate. Recently, the CompoLab Hb machine (Fresenius HemoCare) became available, that can rapidly measure (Hb) using a small device, and is operated by filling a cuvette with the blood sample. It was our aim to investigate this machine. Methods: A concentration series of samples with Hb of 0-16 mmol/l was prepared to determine the linearity, accuracy and precision of the method. These series were prepared in triplicate and measurements were performed with the Hb meter (Hemocue) and the CompoLab Hb (Fresenius Hemocare). Also, the Compolab cuvettes were used in the HemoCue machine. The requirements were: CV < 5%; accuracy ≥ 80% and r 2 ≥ 0.99. Results: See table (mean ± SD (CV). n = 3). All requirements were met at all levels, except a slightly higher CV for the HemoCue at the level of 16 mmol/L. Conclusion: The Com-poLab Hb meter seems suited for determination of the Hb in the whole range between 4 and 16 mmol/l. The cuvettes can also be used on other Hb meters.

All welds were inspected visually, and the Compodock welds showed a reduced diameter at the site of the weld. With the Terumo SCD, the tubing diameter was almost unchanged. There were no significant differences in the composition of the blood products. At the end of storage, there were no significant differences in percentage hemolysis in the RCCs, and CD62 expression in the PCs. Conclusion: Although Compodock welds have a somewhat reduced diameter at the site of the weld, this does not influence the quality of either stored RCCs or PCs.

O C Illoh, B L Greb, J M Davis, University of Virginia Medical Center, Charlottesville, VA; K Illoh, Stroke Branch, NINDS/NIH, Bethesda, MD Backround: Red Blood Cells (RBCs) for transfusion can be stored for up to 42 days. As these units get older, cytokine accumulation occurs and this has been shown to be associated with the occurrence of febrile non-hemolytic transfusion reactions. Chemokines and chemokine receptors play a key role in many inflammatory events. The expression of these receptors following the storage of blood products is unknown. The aim of this study was to investigate the pattern of chemokine receptor expression on the white blood cells in RBCs over storage time. Materials and Methods: Randomly selected segments from RBCs were obtained from the blood bank. The white blood cells were evaluated by staining with a panel of fluorescent labeled antibodies to CD3, CD4, CD8 and chemokine receptors CXCR3 and CCR4. After incubation, the cells were analyzed on a flow cytometer. Statistical analysis was performed by regression analysis. A p value of <0.05 was considered statistically significant. Results: Twenty samples were analyzed. They ranged in age from 5 to 38 days. There was a statistically significant increase in the population of CD4+ T cells expressing CXCR3 over storage time, and a progressive decrease in the CD4+ T cells expressing CCR4. Conclusion: Increased storage time increases the expression of the inflammatory chemokine receptor CXCR3 on T cells suggesting a Th-1 bias. This may be a contributory factor in the pathogenesis of transfusion reactions.

Background: A flow cytometric assay was developed to detect and enumerate residual red blood cells (RBCs) in platelet (plt) products. Residual RBCs may be of special concern in different clinical situations. One important situation to note is that of avoiding rhesus (Rh) immunization in Rhnegative females. It was previously reported that the risk of alloimmunization after Rh-D incompatible plt transfusions using plt concentrates prepared by modern technical methods appears to be small in patients with haematological disease, but significant in non-immunocompromised patients. Methods: We validated the assay by preparing 6 sets of serial dilutions (3 PBS spiked with whole blood and 3 plt products spiked with whole blood) with known RBC concentrations ranging from zero cells per uL to 10E4 cells per uL. Each tube in the dilution series was created in triplicate, stained and incubated appropriately and then analyzed via this flow cytometric assay. The data was plotted in Excel to verify the linear range of the assay. The assay utilizes the Anti-glycophorin-A (conjugated to phycoerythrin (PE) fluorescent dye) and CD61 (conjugated to fluoroisothiocyanate (FITC) fluorescent dye) antibodies to detect RBCs and plts, respectively. We evaluated one hundred twenty-one plt samples collected from Trima Accel Version 5.1 (within 24 hours of collection) at room temperature with this assay. Results: Validation data (in plt products) determined that the assay has a linear range SP140

Background: During an evaluation of the Compodock (Fresenius Hemo-Care) sterile connection device (SCD), we observed sometimes irregularities on the inside of the tubing at the site of the weld. It was our aim to investigate the effect of these observations on the quality of blood products. Methods: Three leukoreduced red cell concentrates (RCCs) were pooled and divided over 3 bag systems: one without weld in the connecting tubing, one with a Compodock-weld, and one with a weld made with the Terumo SCD. The RCC was transferred 10 times over this tubing to have maximum result if the weld had deleterious effects. The RCCs were stored in PVC containers, and sampled on day 1, 28, 35 and 42, and free hemoglobin (Hb) was measured. The same procedure was also performed using platelet concentrates (PCs), but these were stored in polyolefin containers, sampled on day 1, 6 and 8, and CD62 expression was measured. Ten experiments were performed per blood component. Results: See table (mean ± SD, n = 10).

2004-Vol. 44, Supplement of 10 to 1000 RBCs per uL (R 2 = 0.9934). A normal probability plot was used for the 121 data points generated by the samples collected by Trima Accel. Trima Accel plt products had a mean value of 17 RBCs per uL, with 99% containing less than 100 RBCs per uL. Conclusion: The two-color flow cytometric assay can be used to detect RBCs in a linear range of 10 to 1000 RBCs per uL. Ninety-nine percent of Trima Accel plt products had less than 100 RBCs per uL, or less than ~3E10 7 RBCs for a 300 mL plt product. The clinical effects of these products need to be further investigated.

N C Thomas, D K Gozelanczyk, S Adolph, J Gibble, C Conry-Cantilena, American Red Cross, Baltimore, MD Background: pH testing of apheresis platelets is performed on at least 1% of collections. There were no pH failures (pH < 6.2) among 1300 Amicus blood cell separator (ABCS)-derived products tested from Jan 2001-Jan 2004. From Jan-March 2004, 3 platelet apheresis products from a new fixed collection site (Site A) had measured pH values less than 6.2. Objective: To determine and correct the cause for pH failures. Methods: All pH quality control (QC) was performed within 24 hours of consignment or at product outdate using an automated pH test system. ABCS lot numbers and transportation times from collection to processing site were examined. Site A platelet collection staff were queried about procedures. Amicus pH QC data from Site A was compared with fixed sites B, C and D. Ambient temperatures were measured at fixed sites in those areas where platelets are collected. Results: No differences were noted in platelet volume, concentration, white blood cell count or color among sites. There were no differences in lot numbers of apheresis sets or transportation times to blood center. Site A staff noted that the three products with pH < 6.2 were produced by two ABCSs located along sun-filled windows. A third always shaded ABCS had no pH failures. When handling platelet products they were occasionally noted to be warm to the touch. On review of floor plans, no ABSCs were near windows at sites B, C, and D. Site A ambient temperatures were noted to exceed 30°C while all other fixed sites maintained temperatures in the 20-24°C range despite outside weather conditions.

(unless the bag was not returned to the laboratory). This testing system was phased in over a one-month period. Results Over a three-month period, 672 pH tests were done. Duplicate testing was done on 9 SDPs, so a total of 663 SDPs were tested. Of these, 538 were transfused and 125 were either forwarded to another hospital for transfusion or discarded. Cultures were done on 55 SDPs with pH < 7.0. (Three units were not cultured because they were transfused emergently before the pH results were reported.) One of these cultures grew coagulase negative staphylococcus; the unit was not transfused. We observed lower pH values in imported SDPs even though these units were often fresher than the locally collected SDPs. A total of 16 transfusion reactions due to SDPs were reported during this time; five of the SDPs associated with reactions did not have pH testing done; one was not returned for culture. The table compares results for SDPs collected locally versus those imported from other regions.

ABCSs at Site A were relocated away from sunny windows. The average pH of Site A platelets prior to relocation was 6.84 (range 5.53-7.42) and 7.06 (range 6.80-7.32) post-relocation. There have been no additional pH failures. Conclusion: There are small inter-site differences in platelet apheresis product mean pH and ranges. Discussion with collection staff regarding site conditions was helpful in focusing on temperature differences. Despite maintaining platelets under standard storage conditions, temperature elevation of apheresis platelet products during and shortly after collection may be an important factor to consider in pH failures.

Background Due to a sentinel event, our medical executive committee (MEC) requested that the transfusion service initiate testing for bacterial contamination in platelet units. Even with the understanding that the local blood supplier would begin culturing platelet units, the MEC recommended that some form of testing be done in-house. The decision was made to test the pH of each unit as a screening test and then do culture and Gram stain on those units that were below the acceptable limit. Methods Only single donor apheresis platelets (SDPs) are transfused at this institution. Before a pH meter was validated, sterilely acquired segments from SDPs were sent to the blood gas laboratory for pH testing. The results reported are only those that were evaluated using the blood gas analyzer. A Gram stain and culture were done on any SDP with a pH less than 7.0. Also, any transfused SDP that resulted in a transfusion reaction was sent for Gram stain and culture * p < .0001 by Mann-Whitney U test ** The SDP bags from one allergic reaction and one febrile reaction each had a positive culture; both were considered false positives.

Conclusion Although SDPs imported into this region are fresher at the time of pH testing, they have a lower mean pH than locally collected SDPs. This might reflect differences in the collection system or storage bags in different regions. It might also reflect the variable conditions of long distance transport.

Background: No test exists to routinely measure platelet quality. Therefore, platelet concentrates are released without quality testing. Annually, about 15% of platelet concentrates have to be discarded by blood operators because the platelets reach their 5-day expiry date before they can be transfused. With the implementation of bacterial testing and pathogen inactivation, platelet quality remains the major determinant for the expiration of platelet concentrates. Study Design: To reach the ultimate goal of a zero out-date discard rate and to provide patients with the best products possible, blood operators require an automated test for platelet quality. In this study we compared a novel, single-step platelet quality test utilizing dynamic light scattering to current methods for platelet testing. Methods: The dynamic light scattering parameters characterizing platelet quality and viability are (1) the platelet size distribution, (2) the number of microparticles and (3) the response of platelets to temperature changes. These three independent parameters define a computational matrix where each sample is represented by a unique combination of results. Results: Fresh platelets from platelet concentrates are characterized by large size (hydrodynamic size above 1600 nm), few microparticles and a low response to temperature change. With increasing age, platelets bud off microparticles resulting in reduced platelet size, increased number of microparticles and significant temperature-dependent shape change. All three parameters were obtained during one measurement in a 100 microL sample. The dynamic light scattering parameters show excellent correlation with platelet morphology scoring, extent of shape change and hypotonic shock response. Conclusions: Dynamic light scattering is a new automated method to determine a quality-based rather than a fixed platelet out-date. Quality control is required in order to provide patients with the best product.

Background. Evaluation of the quality of stored platelets is only possible by the use of in vitro tests; there is no information about the in vivo behaviour of transfused platelets. The TEG is used in situations were point of care testing of hemostasis is desired. The main parameters of the TEG are (a) the reaction time ® , the time until the initial fibrin formation and comparable with the coagulation times PT and APTT; (b) clotting time (K), the time until a fixed level of clot firmness is reached; (c) the angle (a) is closely related to K and measures the rapidity of fibrin build up and gives information about the clot strength; R, K and a are prolonged by anticoagulants and factor deficiencies; (d) maximum amplitude (MA) is a measurement of maximum strength or stiffness of the developed clot; it is especially influenced by platelets and fibrin. Methods. In 17 patients treated with chemotherapy for hematologic malignancies, TEG's were performed during the second and third chemotherapy course, when blood count has reached normal values after the first chemotherapy. During the subsequent decrease of platelet count TEG parameters were measured in native whole blood samples. We compared the parameters obtained before the start of the first platelet transfusion with those obtained in the aplastic period in which the patient was transfused with platelet concentrates, we assumed that in that period only transfused platelets were circulating. Results. TEG parameters obtained with platelet counts < 20 ¥ 10 9 /l, 20-50 and 50-100 in the period without platelet transfusion were compared with those obtained in periods with platelet count < 20, 20-50 and 50-100, during platelet transfusions. Platelets < 20, R resp 68 ± 25 min (n = 9) and 51 ± 25 (n = 16), p = 0.1; K resp. 46 ± 11 min and 51 ± 31, p = 0.7; a resp. 5 ± 1 deg. and 8 ± 9, p = 0.4; MA resp. 30 ± 5 mm and 35 ± 8, p = 0.1. Platelets 20-50. R resp 53 ± 26 min (n = 14) and 69 ± 36 (n = 72), p = 0.1; K resp. 30 ± 13 min and 41 ± 22, p = 0.1; a resp. 10 ± 6 deg. and 8 ± 7, p = 0.5; MA resp. 44 ± 10 mm and 41 ± 12, p = 0.4. Platelets 50-100. R resp 46 ± 32 min (n = 21) and 52 ± 32 (n = 16), p = 0.6; K resp. 19 ± 11 min and 24 ± 15, p = 0.3; a resp. 17 ± 9 deg. and 14 ± 9, p = 0.3; MA resp. 54 ± 10 mm and 50 ± 10, p = 0.2. Conclusion. TEG parameters demonstrate no statistically significant differences in function between platelets produced in the patient and transfused, allogeneic platelets; the TEG can be used to study the in vivo behaviour of transfused, stored platelets.

Background: Thromboelastography (TEG) is an accepted method in the evaluation of platelet function in whole blood, but can also be applied to platelet rich plasma (PRP) or platelet poor plasma (PPP). The maximum amplitude (MA) measured in mm is dependent on both fibrinogen concentration, platelet count and platelet function. Subtraction of the MA in PRP from the MA in PPP (delta MA) appears a useful measure of the contribution of platelets to the in vitro clot. Methods: Prestorage leukoreduced whole blood derived platelets were manufactured (day 0) using the leukotrap ® system (Pall Medical, Covina, CA) and stored with agitation in CLX containers at 20-24°C for 9 days. Samples were taken on days 1, 5, 7 and 9 and the platelet count adjusted to 300 ¥ 10 9 /L using PPP. A small volume (330 mL) of either PRP or PPP were used in the TEG cups and clotting intitiated in the TEG (Haemoscope, IL) using Russel Viper Venom (10 mL) and Ca Cl 2 (20 mL, 0.2 M). The MA of the PRP and PPP was recorded and the delta calculated by subtraction. Data were analyzed by paired t tests. Statistical significance was p < 0.05. Results: Eight products were studied. Results are shown in the Table. Data are the mean ± 1SD. approximate 7% decline in function per day after day 5. Conclusion: The decline in MA between day 5 and day 7 is similar to the decline reported in other in vitro measures of platelet function, such as hypotonic shock response and in vivo survival and indicates that the delta MA may be a useful in vitro test of platelet functionality. Background: There is no agreed upon protocol for radiolabeling small aliquots of fresh blood platelets (PC), in vitro. Such protocols will be needed, for example, to provide controls in support of in vivo radiolabeled platelet recovery and survival studies for new extended (7 day) platelet storage studies submitted to the FDA for licensure. Study Design & Methods: After IRB approval, we radiolabeled autologous PRP from normal donors (N = 5) with 51-Cr (chromium) and 111-In (indium), in plastic tubes. Fresh blood (100 mL) was drawn via venipuncture and divided into two conical tubes. For each label, approximately 50 mL of blood was anticoagulated with 7 mL of ACD-A. PRP was prepared and the platelets labeled with either 200 mCi of 51-Cr or 100 mCi of 111-In. Appropriate elution samples and standards were prepared. Approximately 20-40 mCi of both isotopically labeled autologous platelets were sequentially (simultaneously) injected immediately after preparation. Post-injection samples (20 mL) were taken at 1, 2, and 3 hours, 24 hours, and daily on days 2-7 and day 10. Counting was for 5 minutes in a Wallac gamma well counter with a 2" NaI crystal. Data was analyzed with the COST program. Results: Mean donor platelet counts were 290 ± 83 ¥ 10E3/mL. Mean labeling efficiency was 15 ± 6% for 51-Cr and 76 ± 8% for 111-In. For 51-Cr, injectates were 5.5 ± 0.3 mL and 20 ± 9 mCi; for 111-In we injected 3.4 ± 1 mL and 35 ± 3 mCi. In vivo Recovery and Survival results (p > 0.05, NS) are seen in the Table. Conclusion: We conclude that despite low 51-Cr labeling efficiency, labeling a 50 mL aliquot of fresh blood platelets in a tube with 111-In or 51-Cr is feasible. Our pilot study also demonstrated equivalence between the two radionuclides for in vivo assessment of platelet storage.

No difference is observed between day 1 and day 5 (p = 0.05), but a 15% deterioration is evident between day 5 and day 7 (p < 0.01) and an approximate 30% decline between day 5 and day 9 (p < 0.01). This indicates an

B Newman, American Red Cross Blood Services, Detroit, MI; A J Roth, Pfizer Pharmaceutical Corporation, Ann Arbor, MI BACKGROUND: Estimating the probability of a vasovagal reaction or fatigue is useful for blood donor education, blood donor medical care, and to identify where to apply solutions to mitigate these adverse events. In a previously reported study, 1,000 general whole-blood donors were observed during a 500-ml phlebotomy and interviewed three weeks after the phlebotomy for adverse events. The interview enhanced the comprehensiveness and accuracy of the collected data. The vasovagal reaction and fatigue rates were 7.0% and 7.8%, respectively, and these two events were clinically significant because they decreased blood donor return presentations by approximately 45% and 38%, respectively, based on a one-year follow-up. METHOD: The probability of an adverse event (vasovagal reaction or fatigue) was represented via a logistic regression model in the form e R / (1 + e R ), where e is about 2.71. R was allowed to depend on age, weight, sex, race, and first-time donor status, as well as interactions between them.

Stepwise techniques using data from the 1,000 blood donors were used to decide which of these factors belonged in the final model, resulting in a formula to estimate the probability of a vasovagal reaction or fatigue. The contribution of each factor in the model was quantified, and a p value was generated to assess its statistical significance. RESULTS: Weight, age and first-time donor status affected vasovagal reaction rates at p values of <0.0001, 0.015, 77A 2004-Vol. 44, Supplement and 0.054, respectively. No other factors (including interactions) entered into the model. The R value is: -1.415 -0.0162 (weight-110 lb) -0.0208 (age -17 years) + .499 (if a first-time donor). Hence a 17-year-old, 110-lb, firsttime donor has a 28.6% probability of having a vasovagal reaction, while a 50-year-old, 175-lb, repeat donor has a 4.1% probability. Sex and first-time donor status were statistically significant contributors to rates of fatigue at p values of 0.005 and 0.023, respectively. The model also included effects due to weight (p = 0.065) and sex by weight interaction (p = 0.11), whereby increasing weight mitigated against fatigue for females but not for males. The R value for fatigue is: -3.311 + 1.626 (if female) +0.571 (if first-time donor) -0.0131 (weight-110 lb) (if female). A first-time, 110-lb female donor has a 24.7% probability for fatigue, while a repeat, 170-lb male donor has a 3.6% probability. CONCLUSION: We were able to statistically identify and quantify the important factors that contribute to a vasovagal reaction or fatigue and provide a formula to estimate its rate of occurrence. Background: Up to 36% of blood donors may experience donation-related complications. While most reactions are minor and short-lived, some require additional medical care. We sought to evaluate donor reactions necessitating care from outside medical facilities. Methods: A retrospective review of donor complications from 1/1/03 to 12/31/03 was undertaken from a single blood center. Reactions requiring additional medical care were entered into an electronic relational database created in Microsoft Access. Reactions were classified into reaction type and to whether donors were treated as outpatients or required hospital admission. Descriptive statistics were calculated for the above classifications based on age, gender, and type of donation. Results: Overall, 62 donor reactions necessitated care from an outside medical facility (0.02% of 251,379 donations). Twenty (32%) were referred by the blood center nurse, blood center physician, or taken by ambulance for further care. The remaining donors (68%) sought additional medical care on their own. The data for reaction type and gender are shown in the table below.

B Newman, N Crooks, L Zhou, American Red Cross Blood Services, Detroit, MI; L A Chambers, B Dy, J Kennedy, American Red Cross, Biomedical Headquarters, Washington DC, DC BACKGROUND: A multicenter blood system implemented a national donor complication monitoring program in 2003. Donor complications were classified into 7 categories and subdivided according to whether or not outside medical care was accessed by the donor. Three of the categories were hypotensive, probably vasovagal events: syncope, syncope with injury and prolonged recovery. Three categories were related to arm injury: hematoma, nerve irritation, and potential arterial puncture. The remainder of events were listed as "other complications". METHODS: The rate of complications with outside medical care for the entire system, and for one contributing center, were compared. The individual cases at the center were characterized in detail. RESULTS: Forty-six of 222,500 whole blood donors at the center (2.11 per 10,000) and 2,065 of 7.03 million whole blood donors nationally (2.94 per 10,000) had a complication that was treated by an outside medical care provider. An additional 21.8 cases/10,000 on a national level did not obtain outside medical care. A review of the local cases showed the following: The 13 syncopal reactions with injury were mainly to the head; all 13 occurred at the blood collection site and more often in younger persons. Syncope without injury and prolonged recovery events often occurred after the donor had left the blood collection site (8 of 14) and occurred more often in females (13 of 14); only 1 of 14 led to an admission. CONCLUSIONS: The incidence of blood donor complications leading to outside medical care was defined. Syncope with injury that needs outside medical care tends to occur on-site. In contrast, half of the cases of syncope without injury that sought or were sent for medical care occurred after the donor left the site.

B Newman, B Siegfried, L Buchanan, American Red Cross Blood Services, Detroit, MI BACKGROUND: Vasovagal reactions have been reported to be decreased in African-American high-school blood donors (n = 226) and in African-American blood donors (n = 62) who donated in a hospital-based program. These are relatively small numbers, so the present study evaluated several thousand first-time (FT) African-American, Hispanic, and Caucasian blood donations for the incidence of vasovagal reactions. METHODS: First-time donations were considered to be the most unbiased donations because the donors did not self-select themselves to not return based on a previous vasovagal reaction. 4,473 consecutive FT African-American blood donations in 2003; 3,659 consecutive FT Hispanic blood donations between 1999 and 2003; and 4,011 consecutive FT Caucasian blood donations during the first five days of each month in 2003 were entered into the study. The groups were evaluated for three variables: sex, age, and the presence of a vasovagal reaction. The three ethnic groups' vasovagal reaction rates were compared, and then compared again with the additional risk factors of being 17 years old or being 17 years old and female (lower weight). RESULTS: The vasovagal reaction rates were 3.4%, 6.9%, and 8.3% in African-Americans, Hispanics, and Caucasians, respectively. In 17-year-old donors, the reaction rates were 6.3%, 11.8%, and 14.5% in African-Americans, Hispanics, and Caucasians, respectively, and in 17-year-old, female donors, the reactions rates were 9.0%, 15.5%, and 16.0% in African-Americans, Hispanics, and Caucasians, respectively. African-Americans had a lower vasovagal reaction rate than Caucasian or Hispanics for all comparisons with an exact Fischer test at p < 0.001. Hispanics had a slightly lower vasovagal reaction rate than Caucasians (p = 0.02) and Caucasian 17-year-olds (p = 0.04) but a similar vasovagal reaction rate to Caucasian17-year-old females (p = 0.87). CONCLUSIONS: African-Americans had a lower incidence of vasovagal reactions than Caucasians or Hispanics. Ethnicity does influence vasovagal reaction rates.

Of the 62 reactions, 54 (87%) occurred in whole blood donors (0.02% of all whole blood donations), 5 (8%) occurred in apheresis donors (0.03% of all apheresis donors), and 3 (5%) occurred in autologous donors (0.15% of all autologous donors). Of the 62 reactions, 59 (95%) donor reactions were treated as outpatients while only 3 (5%) required hospitalization. All 3 donations requiring hospitalization were for vasovagal reactions; and all were female donors; all 3 hospitalizations were for one night. The average age of blood donors experiencing reactions requiring additional medical care was 35.4 (range 17-77 years). The majority of the reactions occurred in the 17-21 age group (n = 22, or 35%). Donors greater than 50 years old accounted for an additional 19 (31%) of the reactions. Conclusions: Approximately 1 in 5000 blood donations resulted in donors obtaining additional medical consultation. The rate was highest in autologous donors (1 in 667), then in apheresis donors (1 in 3333), and lowest in whole blood donors (1 in 5000). Donors who were female, or donors who were less than 21 years old or greater than 50 years old were most likely to experience a donor reaction requiring additional care at an outside medical facility. 

B Newman, S Satz, N M Janowicz, B Siegfried, American Red Cross Blood Services, Detroit, MI; D Frattarelli, Wayne State University School of Medicine, Detroit, MI BACKGROUND: The Japanese collect 200-and 400-mL whole-blood collection volumes and have a vasovagal reaction rate of 0.6-0.7%. Decreasing the collection volume can decrease donor vasovagal reaction rates and would be most beneficial to young, low-weight, first-time, Caucasian blood donors, who have the highest risk. METHOD: The present study evaluated 8,135 consecutive 17-year-old, first-time, Caucasian high-school (HS) students, each of whom donated 525 mL (one 481-mL unit) of whole blood between October 1, 2003, and March 23, 2004. Each donor's gender and weight were recorded. A best-fit curve was developed for each sex relating the reaction rate to the ratio of collection volume to donor weight. The asymptote was calculated and was considered to be the percent of reactions that occurred at every weight. Utilizing the curve for each gender, we calculated the expected vasovagal reaction rate for 450-, 400-, 350-, 300-, and 250-mL collection volumes for this donor population. We also performed the same calculations on donors who weighed 130 lb or less since lowweight donors have the highest risk. RESULTS: There were 4,031 females and 4,104 males. The vasovagal reaction rate for a 481-mL unit was 16.3% and 7.1% for females and males, respectively. The asymptote was estimated to be 6.8% for females and 1.2% for males. The table shows vasovagal reaction rates in females and males as collection volumes decrease from 481 mL to 450 mL and then by 50-mL increments down to 250 mL. At the extreme (250-mL unit), males' reactions would decrease by 70% and females' by 49%. An evaluation of donors who weighed 130 lb or less showed that at the extreme, males' reactions would decrease by 69% and females' by 53%, which are similar to the relative decreases for the total population. CON-CLUSION: Decreasing the collection volume would significantly decrease vasovagal reactions in both females and males. directly related to the severity of donor reaction and is prolonged in several types of donors who have experienced reactions.

Background: Blood donor reactions are associated with reduced donor return rates. For a given population of donations, the distribution of intervals from donation to subsequent presentation determines the return rate and also may give more information on donor characteristics. Methods: Presentations for allogeneic whole-blood donation from September 9, 1998, through July 8, 2003, were studied. These consisted of 197,859 first-time donations; 800,573 repeat donations; and 142,835 deferrals. For each index donation from September 9, 1998, through May 5, 2001, the next presentation was identified; and the interval from the index donation to that presentation was determined. The donor's age at the time of the index donation, whether the donation was from a first-time donor or a repeater, and the severity of any donor reaction were noted. During the period studied, the definition of a light reaction included no loss of consciousness, whereas a moderate or severe reaction included loss of consciousness. Results: There were 1,141,267 presentations, consisting of 197,859 first-time donations; 800,573 repeat donations; and 142,835 deferrals. Presentation frequencies peaked at multiple seven-day intervals after the index donation and around the anniversary of the donation. The intervals from index donation to subsequent presentation increased as reactions progressed from absent to light to moderate/severe. These prolongations could be due to youth and/or firsttime-donor status rather than to reactions. However, while reactors with those traits contributed to the prolongation, so did other reactors, including older, repeat donors. Conclusions: The interval from blood donation to subsequent presentation is a useful indicator of donor behavior. The interval is SP155

B Newman, S Satz, N M Janowicz, American Red Cross Blood Services, Detroit, MI BACKGROUND: Two recent studies suggested that female gender does not increase vasovagal reactions and that different vasovagal reaction rates in females and males were due to differences in weight between the two sexes. One study evaluated 1,076 Caucasian high-school (HS) students, including 855 that were first-time blood donors; and the other study evaluated 1,000 random donors who were also interviewed. Neither study found that vasovagal reaction rates between male and females of equal weight were significantly different. METHOD: The present study evaluated 8,135 consecutive 17-year-old, first-time, Caucasian HS blood donors, each of whom donated 500 mL of whole blood between October 1, 2003, and March 23, 2004. The study was limited to Caucasian HS blood donors because a previous study showed that Caucasian HS blood donors had a much higher vasovagal reaction rate than African-American HS blood donors. Each donor's gender and weight were recorded. Female and male donors' reaction rates were compared by a Fischer Exact Test. RESULTS: Females had higher vasovagal reaction rates than males in each 10-lb weight interval from 120-129 lb through 190-199 lb. No statistical significant differences were detected at and above 200 lb, even when intervals were combined to increase the numbers of blood donors compared. The table shows the vasovagal reaction rates, relative rates (RR), and p values (two-tailed test) for selected weight groups. CONCLUSION: There is a gender difference with regard to vasovagal reaction rates in 17-year-old, first-time, Caucasian blood donors. Females had a higher vasovagal reaction rate than males, when similar weight groups were compared. This phenomenon was observed only for blood donor weights less than 200 lb. * Less than 20 donors. ** Less than 10 donors.

Syncope is defined as transient loss of consciousness. These episodes may be as mild as dizziness and progress to prolonged loss of consciousness and seizure, requiring more urgent intervention. This study examines the effect clinic environment, age and frequency of donations contribute to the frequency of syncope. Many syncopal episodes result in an 79A 2004-Vol. 44, Supplement incomplete blood donation and a loss of a unit of blood. METHODS AND OBJECTIVES: All adverse reactions were recorded over a period of 12 months. The clinic environments were categorized into Residential, Business/Factory and Educational facilities. The age group and whether the donors were new or repeat donors was also recorded The study excluded autologous donors and people who experienced syncope without donating blood e.g. potential donors having their haemoglobin tested and bystanders. RESULTS:

T Hoekstra, J Ten Velde, W De Kort, P Van Noord, Sanquin Blood Bank Southeast Region, Nijmegen, Netherlands Background It has been observed that during the summer, compared to other seasons, a higher percentage of donors is deferred due to a low haemoglobin (Hb). Biological explanations might be heat acclimatization (haemodilution) or differences in nutrition and activity patterns of donors. Furthermore, it is possible that a difference in population characteristics might be the underlying reason (selection bias). Plasma aphaeresis is not likely to affect Hb and plasma donors generally have a high donation frequency. A population of plasma donors is therefore ideal to study a possible seasonal effect on Hb levels. Methods The study population consisted of plasma donors who visited the blood bank at least twice for a plasma donation during the study period (January 2002 until December 2003 . From all visits physical examination data were collected (e.g. Hb and blood pressure). A substantial part of the donors also had a number of whole blood donations during the study period, but these were excluded from the analyses. Furthermore, we excluded the first plasma donation following a whole blood donation. Hb values were correlated with the mean temperature per day (provided by the Royal Netherlands Meteorological Institute). Results Data were available for 2,357 plasma donors. The mean number of examinations in the two-year study period was 8.70 (±5.3), resulting in 20,515 examinations. In men Hb values were higher than in women (9.5 vs 8.6 mmol/L). In men, Hb values were negatively associated with age (b = -1.06, P < 0.001) while in women Hb values increased with age (b = 0.16, P < 0.001). In women 4,1% of the examinations resulted in deferral due to a low Hb (<7.8 mmol/L), compared to 1,2% in men (<8.4 mmol/L). In the summer lower Hb values were observed, both in men and women. In both men and women Hb levels were significantly associated with environmental temperatures (r = -0.05 and -0.04 respectively, P < 0.001). Per donor a mean winter-Hb (mean of Hb measurements November through February) and a mean summer-Hb (June through September) was calculated. The differences between these means were used as measurement of a seasonal effect within the donor. In men, a mean difference of 0.05 mmol/L and in women of 0.04 mmol/L (both P < 0.0001) were observed. Conclusions Hb levels in plasma donors vary with the season, resulting in somewhat higher deferral rates for Hb during the summer. The seasonal effect was also observed within subjects and thus cannot be explained by differences in donor characteristics. The data will be further explored by using longitudinal data analysis techniques.

B Newman, N Crooks, J Roberts, N Trela, B Siegfried, American Red Cross Blood Services, Detroit, MI BACKGROUND: Recent medical literature suggests that drinking 16 oz of water 25-35 minutes prior to blood donation might decrease the incidence of vasovagal reactions. However, drinking water shortly before blood donation might influence the blood donor's hemoglobin concentration. METHOD: Twenty-five female adult subjects were enlisted to study the effects of 16 oz of water on hemoglobin concentration. A baseline venous sample to test hemoglobin concentration was drawn. It was tested twice on a Hematron 3700 (Abbott, Abbott Park, Illinois), CV £ 1.2%, and the mean hemoglobin concentration was calculated. The subject then drank 16 oz of water over 5-10 minutes and had a second venous sample drawn 30 minutes after the baseline sample. It was also tested twice, and the mean hemoglobin concentration was calculated. RESULTS: The mean hemoglobin concentration at baseline and 30 minutes was 13.22 and 13.08 g/dl, respectively, for a net decrease of 0.14 g/dl. Seventeen of 25 subjects (68%) had a decrease in hemoglobin concentration ranging from 0.05 to 0.7 g/dl; 7 of them had a decrease of 0.3 g/dl or more. One subject (4%) would have been deferred based on a hemoglobin concentration change from 12.6 to 12.4 g/dl. Seven subjects had a slight increase in hemoglobin concentration ranging from 0.05 to 0.3 g/dl and 1 subject did not change. The results before and after drinking water were statistically different by Fisher exact test (p = 0.02) and by analysis of variance (p = 0.02). The medical literature reports that drinking 32 oz of water over 2 minutes causes a biphasic change in hemoglobin concentration; the hemoglobin is concentrated during the first 15 minutes and diluted after 15 minutes, with maximal dilution at 30-35 minutes. CON-CLUSION: Drinking 16 oz of water 30 minutes prior to a venous sample leads to a slight decrease in hemoglobin concentration. It potentially could Of a total of 127,139 blood donations 2.4% or 2,323 of donors experienced syncope. Reactions were more common amongst younger donors in education facilities at 6.55%. Residential and business clinics showed a lower rate of adverse reactions at under 2%. Young, new donors in the 17-30 year age group were found to have more episodes of syncope than repeat donors.

CONCLUSION: Psychogenic syncope-anxiety, fear of needles and hysterical fainting as well as chain reaction fainting were contributors. Environment, immaturity and the slightly charged atmosphere at educational institution clinics appeared to be a factor. It is anticipated that the number of adverse reactions can be reduced by educating and reassuring potential donors and increasing the number of experienced personnel in attendance at high incidence clinics. A reduction in the number of reactions would also reduce the cost of uncompleted donations and wasted blood bags. Introduction: Low hemoglobin/hematocrit (Hgb/Hct) values account for the highest number of donor deferrals. An iron supplement therapy program was created to help women of childbearing age donate blood regularly. Blood donors supplemented with iron were evaluated to see if they donated more frequently with fewer Hgb/Hct deferrals than in the year prior to signing up for the program. Materials and Methods: A carbonyl iron supplement (45 mg of elemental iron per caplet) was provided for 60 days. Menstruating women over the age of 18 were eligible to participate at fixed drawing sites. Ineligible donors included women who: 1.) do not menstruate; 2.) have a family history of hereditary Hemochromatosis; 3.) have severe chronic gastrointestinal disorders; 4.) have a first-degree family history of colon cancer or polyps. Results: There were 872 donors who filled out applications/consents to participate in the program, and 648 (74%) were accepted. These participants averaged 1.9 donations per year in the year prior to joining the program. 621 (71%) of the participants have had at least one 60-day cycle elapse since entering the program, and averaged 1.6 donations per year. If we consider only those donors who have successfully returned (n = 294), their average is 3.1 donations per year. Fifty-nine (9.5%) donors returned 79 times but were not able to donate. Of those unsuccessful attempts, 10% had taken the iron for less than 40 days, and 55% attempted more than 30 days beyond their iron supply. Donors who started in the first month have completed a full 6 cycles of iron. In the year prior to the program, they were deferred on average 2.2 times; but during the program's first year, these donors who returned were only deferred an average of 0.6 times_a decrease of almost 400%. Conclusion: A low dose iron program significantly increased the number of annual donations by women. Women actively participating in this program are donating at a higher annual rate with fewer deferrals. We will continue to monitor these women to see if this trend continues. cause a blood donor deferral. It would be prudent to do the pre-donation hemoglobin test prior to drinking the water to prevent a possible deferral.

S M Randolph, E Waltman, United Blood Services of New Mexico, Albuquerque, NM; D J Costello, Cell Robotics International Inc., Albuquerque, NM BACKGROUND: Blood donors are routinely screened for anemia with an initial screening test using copper sulfate solution (CuSO 4 ). If the donor fails the initial screening a secondary screening using microhematocrit (m Hct) is performed. The quality of capillary blood has been reported to make a significant difference in the primary screening failure rate. A capillary blood sampling device which produces a fingertip wound by laser ablation may be expected to change the result of pre-donation screening tests. This is because the physiological mechanism of wound production is significantly different from the conventional lancing procedure using steel blades or needles. The purpose of this study is to compare screening results obtained from use of a laser-based lancing instrument to those obtained using conventional lancets. METHODS: Seventy-six donors reporting to both mobile and fixed donation centers and who had failed the initial (CuSO 4 ) screening test were enrolled as subjects. The subject population included 3 men and 73 women. The standard lancing device was a safety lancet with a springloaded blade (Surgilance). Following screening by Hct in accordance with standard procedures, both screening tests were repeated using capillary blood drawn from the opposite hand with the Lasette laser lancing device. RESULTS: A 38% improvement in deferral rate due to anemia screening was observed when the laser-lancing device was used. Matched pair analysis indicated no significant difference (p < 0.001) between the Hct values obtained from samples taken with the two different instruments. The rate of laser samples passing the initial screening but failing secondary screening was 5%. Table 1 . Anemia Screening Data accomplish it (132 ± 13 g/L vs. 143 ± 11 g/L, respectively; p < 0.0001). In the group of patients who underwent a knee replacement, the mean hemoglobin value at the entry of the study was lower in the group of transfused patients than in the group of non transfused patients (137 ± 10 g/L vs. 143 ± 13 g/L, respectively; p = 0.014). Conclusions: The use of vitamin supplements does not enhance the accomplishment of our PABCP. A high hemoglobin value at the entry of the study is a predictive variable to accomplish our PABCP. In the group of patients who underwent a knee replacement, a high hemoglobin value at the entry of the study prevents the need of blood transfusion after the surgery.

Introduction of the Self-Administered Donor History Questionnaire: One Blood Center's Experience R R Gammon, D G Richards, J B Ward, South Florida Blood Banks, Lake Park, FL Background: In July 2003 the US Food and Drug Administration (FDA) permitted blood centers the opportunity to allow the donor to answer all of the pre-donation screening questions without direct oral questioning by donor service personnel. All answers required review by donor service personnel before phlebotomy occurred. We attempted to determine if implementation of the self-administered donor history questionnaire (SADHQ) resulted in a change in the number of errors among our donor service personnel as well as a change in the number of FDA reportable post-donation information (PDI) events. Study Design: At our blood center donor questions were presented on printed forms. We reviewed the first three months of implementation of the SADHQ and compared it to a three-month period one year earlier. We compared pre-and post-implementation error rates (included: donor screening and determination of donor suitability) among donor service personnel as well as the number PDI events that required submission of biological product deviation (BPD) reports to the FDA. Sixty-five events were reviewed for each donation. Results: Total rates of donor screening errors were 797/1,780,025 events (0.045%) with 42 BPD reports issued in 1/1/03-3/31/03 vs. 408/1,627,925 events (0.025%) with 29 BPD reports issued in 1/1/04-3/31/04. Top PDI categories requiring issuance of a BPD did not change between the study periods (European residence/military service, medication/medical conditions, tattoos/piercings). Conclusions: Introduction of the SADHQ at our blood center resulted in a decrease in the rate and number of errors in the donor screening and the determination of donor suitability process by our donor service personnel. There was also a reduction in the number of PDI events requiring submission of a BPD report to the FDA as well as product recall and consignee notification.

P D Cumming, Talisman Ltd., Vienna, VA; L M Katz, Mississippi Valley Regional Blood Center, Davenport, IA; E L Wallace, Center for Management Systems, Naples, FL Background. Previously reported results from studies at one Midwestern regional blood center (RBC) on implementation of audio video (finger) touch screen computer assisted donor self interviewing (AVT-CASI) have shown the technology to be superior to face-to-face (FTF) interviewing. Donor and staff satisfaction increased by a minimum of a factor of three and often the preference strength was much greater. Donor time increased by four minutes but staff time decreased by five minutes. Reportable errors and omissions declined more than 60%. Reporting of socially sensitive and stigmatizing behaviors increased by a factor of nine. Other advantages such as increased ability to manage donor surges were also reported. Studies to replicate and expand these results are in process at four RBCs. Methods. Pre & post AVT-CASI implementation studies at four RBCs use surveys to obtain donor and staff satisfaction measures across seven or more variables, recording of donor and staff time required to complete the processes and data from Blood Establishment Computer Systems (BECSs) on donor truthfulness and retention, as well as staff errors and omissions. The design provides current data from the original RBC that has been using the technology for three years plus two RBCs with one year of experience and one center just implementing AVT-CASI. It includes over sampling to fill in gaps in previous research such as new and mobile donor satisfaction and donor retention and expanded sample sizes. RBC collection volumes range from 30,000 to 150,000 units per annum. The new RBCs are located in Tennessee and Louisiana. Results. Survey and time study work at three centers is scheduled for completion in May 2004 and one year pre & post AVT-CASI CONCLUSION: The use of a laser-based lancing device for anemia screening decreases the donor deferral rate. The result is similar to that of previous investigations of the use of venous blood or the use of a second conventional fingerstick for Hct screening. The laser lancing procedure appears to provide this improvement during initial screening.

Background: It is a common practice to administer vitamin supplements to blood donors who are candidates for a preoperative autologous blood collection program (PABCP). Our main objective was to investigate the role of iron and folic acid as vitamin supplements to enhance the accomplishment of our PABCP. Our secondary objectives were to detect some predictive variables for the accomplishment of our PAPCP and for the risk of receiving blood transfusion after the surgery. Study design: We prospectively, randomly assigned 180 autologous blood donors who underwent to elective orthopaedic surgery to receive no vitamin supplement, iron (105 mg of elemental iron/day), or iron (105 mg) and folic acid (5 mg). All of the donors had a hemoglobin value ≥115 g/L at the start of the study. Supplement vitamins were started one month before the donation of the first autologous unit. We used a weekly schedule for blood collection and we drew two (for knee surgery) or three (for hip and scoliosis surgery) autologous units. Blood collection was performed if hemoglobin value was ≥10 /L. Results: There were 161 donors who were eligible to study. They were 56 (35%) men and 105 (65%) women with a median age of 69 years (range: 15-82). The PABCP was accomplished in 44 (86%) out of 51 donors who receive no vitamin supplement, in 49 (86%) out of 57 donors who received iron supplement and in 46 (87%) out of 53 who received iron+folic acid supplement. The mean hemoglobin value at the entry of the study was lower in the group of patients who do not accomplish our PABCP than in the group who do BECS data is scheduled to be available in August for two RBCs. Satisfaction results from 1000 donors at each of three RBCs will be presented by computer skill level, gender, ethnicity, age, race, education, type of donation and number of prior donations. Based on the literature and early RBC feedback, new and mobile donors are expected to prefer AVT-CASI even more than the repeat fixed site donors of past studies. Staff satisfaction surveys from all AVT-CASI experienced staff at each RBC will be similarly analyzed and presented. Sample sizes are near 100 at one RBC. BECS data sample sizes cover a year of RBC collection volumes, i.e., 30,000 to 75,000 units where studies are expected to be complete this summer. Results of all studies will be compared to previously published work. Conclusion. The research is new to blood but consistent with other literature. Earlier AVT-CASI findings are reinforced and extend to new, mobile and demographic subgroups of donors.

D Kessler, L Milan-Benson, R Harkin, J E Valinsky, New York Blood Center, New York, NY Background: Recent literature has alluded to the limited value of CUE in identifying individuals at risk for transmission of infectious diseases (ID) via transfusion, who are not otherwise detected using the sensitive EIA and NAT screening tests. We compared the current efficacy of CUE in our center relative to evaluations made at this center 10 years ago (Transfusion, 33:35S). Methods: In our current procedure for the use of CUE, donors can choose to exclude the current donation from use for transfusion by the application of a bar coded label to the donor registration form. These donations are removed from inventory, but the donors are not deferred. We reviewed viral marker data for donors who used CUE to exclude units during the period 1/1/01-10/1/03. We compared these results to those donors who did not selfexclude during this same period as well as to data obtained in a 6-month period in 1993. Additionally, the records of the recent group of CUE donors were also reviewed to see if viral markers were positive on previous, the same or subsequent donations. Results:

Referral Program (HARP) was developed to address this issue. Methods-Donors were provided with written materials about hypertension and HARP. BP was checked as part of donor intake. Donors with BP ≥140/≥90 mm Hg were informed of the BP reading and the need for medical follow-up. Donors were referred to their personal health provider or an area hospital outpatient facility. The hospital, in partnership with the blood center, provided expedited appointments to blood center referrals, physician services pro bono to uninsured donors and assistance in enrolling in public insurance programs. The referred donors were followed up by mail survey at 6-8 weeks post-referral. Non-responders were given the opportunity to complete the survey by phone. Results-A preliminary study of all donors presenting in the blood center's regional catchment area during one month (n = 969) revealed that 24% were hypertensive. From December 2003-April 2004 we evaluated 2,961 donors at a fixed site and mobile drives in the region and identified 224 (8%) with elevated BPs. Of these 71% had a BP of 140-159/90-99 mm Hg, 24% had a BP of 160-179/100-109 mm Hg, and 5% had a BP of >180/>110 mm Hg. Demographically, 60% were male, approximately 30% African American, 35% Caucasian, and 16% Hispanic. More donors at mobile drives (9%) had elevated BP than at the fixed site (6%), although severity did not differ. We contacted 68 donors with high BP in the first 3 months of the study; 40 responded. Of these, 72% saw a health care provider in response to the blood center referral and 20% intended to within two months. Only one donor sought care at the referral hospital. Conclusions-Blood centers can reach a largely healthy community and identify significant health risks. In a 4-month period, 8% of donors tested had elevated BP and were referred for medical follow-up, although preliminary studies showed that more (24%) might have high BP readings. Initial data indicate that some potentially at-risk donors are seeking further medical assessment as a result of HARP. Most donors chose to see their personal provider. We plan to continue and expand HARP as part of our efforts to position the blood center as a community health resource. Background To study autoantibodies (IgM rheumatoid factor (IgM-RF) and anti-cyclic citrullinated peptide antibodies (anti-CCP)) and C-reactive protein (CRP) in serial blood samples of blood donors, who developed rheumatoid arthritis (RA) later, compared with control donors. Methods Since 1984, our blood bank has stored one mL aliquots of serum from each blood donation. From 5000 patients with RA, we identified 79 patients who had been blood donor at this blood bank before the onset of RA. A computer program was developed to trace all blood donations from a particular RA patient in the period 1984 to 1999. The deep frozen blood samples were retrieved from each donation together with one control donor sample, matched for age, gender and time of donation. All samples were measured with ELISAs for IgM-RF and anti-CCP on an ES300 analyser. C-reactive protein (CRP) was measured using a latex enhanced highly sensitive assay (Roche Diagnostics). The development over time of the CRP concentration in the patient and control group was estimated with random coefficient analysis. Results A median of 13 samples (range 1-51) per patient was available; the earliest donation was at a median of 7.5 years (range 0.4-14.5) before the onset of symptoms. 1078 patient and 1070 control samples were tested. 39 patients (49%) were positive for IgM-RF and/or anti-CCP at least once before the start of the symptoms. There was a steady increase in time of the percentage of patients positive for one or both of the antibodies in a period of 14.5 years before disease onset. Of the control samples, 1.2% was positive for IgM-RF and 0.7% for anti-CCP. The development of the mean CRP concentration of the patient group shows a statistically significant increase in time, with the highest value at the start of the symptoms, whereas the mean CRP value over time of the control group remains stable. Conclusions Half of patients with RA have specific serological abnormalities years before the development of symptoms. Furthermore, CRP values increased in time, with the highest value at the start of the symptoms.

There were significantly higher viral marker rates in donors who excluded units using CUE than those who did not (chi sq p < 0.001) both for data obtained in 1993 and in 2001-2003. In addition, the overall marker rates were higher in 1993 than 10 years later. We evaluated results on 53 donors in 2001-2003 who were screening reactive (RR) for HCV, HBsAg or HIV and who used CUE to exclude donations. Of these 32 were RR for a viral marker on the same donation as CUE was used, 17 were RR on previous donations, 2 on the next donation (of these 2 donors 1 was confirmed HIV positive-HIV WB and NAT positive, the other donor was HCV RR/RIBA and NAT neg), 1 on the 4 th subsequent donation and 1 on the 5 th . Conclusions: The use of CUE declined by 50% in the last 10 years. In the period 01/01-10/03, we identified 1 case (1/~700,000) in which CUE may have prevented HIV transmission. This evaluation supports the view that there is a declining value of the use of CUE. The incremental improvement in risk reduction (~1/700,000 in our study) is small, especially in light of the introduction of more sensitive donor screening tests. Nonetheless, CUE may be of value in very rare situations. Background: Blood donation may have cardiovascular (CV) health benefits. Some blood centers have provided donors with cholesterol screening and other measures of CV health. It is important to understand the interest of current and potential donors in these potential health benefits. Methods: While conducting a study of the effects of blood donation on CV health, we noted a difference in the participation rate among different donor segments. Donors 50-75 years old were randomly recruited based on donation patterns. High (H) frequency donors donated at least 8 donations in the last two years. Low (L) frequency donors donated at least two lifetime donations but no more than once a year in the past two years. 650 donors (150 H and 500 L) received an invitation for the study from the blood center Medical Director and the academic hospital PI. The letter said the purpose of this study was to determine how blood donation affects blood vessel function. If the donor participated in this study, they attended a study visit at the academic hospital; their medical history, blood pressure, and pulse were recorded; blood vessel function in their arm measured with ultrasound; and blood assayed for glucose, cholesterol, and genes important in iron metabolism and blood vessel function. The subjects were told expenses would be compensated and they would receive a record of blood pressure, glucose and cholesterol. The rate of donor participation as a proportion of letters mailed was analyzed by donation status (H or L), sex, and age. Results: 23% of H blood donors participated in the study compared to 5% of L donors (p < 0.001, Chi square). Although, as expected, H donors were more likely to be male, female donors were equally likely to participate in the study (H donor participation: 26% of female and 22% of male donors. L donor participation: 3% of female and 6% of male donors). 19% of donors over 65 years old participated compared to 7% of donors 50-64 years old (p < 0.001, Chi square). This age difference was found in both H and L donors. Conclusion: Frequent donors were 4 times more likely than infrequent donors to respond to a study of the health benefits of blood donation which also provided the participant with important personal health information. Donors 65 years and older were more than twice as likely to participate. Older and more frequent donors appear to be more interested in these health benefits from blood donation, although other factors such as available time to participate in the study or increased desire to be helpful to the blood program could also explain the results.

S Rummler, U Werner, D Barz, Institut für Transfusionsmedizin, Jena, Germany; F Hofman, Gambro BCT NV/SA, Zaventem, Belgium Background To ensure donor safety, regulatory guidelines for apheresis donations define limits for annual blood component losses, which require blood centers worldwide to carefully track and manage individual donor losses on an annual basis. A European blood center examined its ability to consistently meet BA*P4K (Bundesärtzekammer) guidelines for red cell losses; specifically, annual volume limits of 1000 ml for women and 1500 ml for men. The center relied on manual tracking methods until implementation of the Vista TM Information System, a donor management system that automatically tracks plasma and red cell volume losses on a 365 rolling calendar basis. This study reconstructed procedure histories from a sample of male and female donors in order to compare the results of manually tracked red cell losses with results obtained through Vista's automated tracking capabilities. Study Design Procedure histories from 40 platelet donors were obtained from manual records and entered into Vista to calculate cumulative blood volume loss. Total donation volumes, including RBC losses, were calculated using the value inputs in Table 1 .

Results Vista calculations of blood loss associated with reconstructed donation histories indicated that 50% of female and 5% of male donors exceeded the annual RBC volume loss limits specified by BA*P4K guidelines. Conclusions These results underscore the need for careful and accurate management of donor volume losses on a cumulative basis, and highlight the difficulty associated with maintaining accurate records manually. Vista tracks blood loss volumes and identifies donors at risk for over-collection. This capability could improve donor safety by ensuring compliance with component blood loss guidelines.

J Ringwald, I Mertz, V Weisbach, R Zimmermann, E Strasser, S Achenbach, University Hospital Erlangen, Germany, Erlangen, Germany; S Seyboth, E Richter, Red Cross Blood Donation Service Baden-Wuerttemberg-Hessen, Baden-Baden, Germany; R Eckstein, University Hospital Erlangen, Germany, Erlangen, Germany Background: Despite serological testing and donor selection the risk of transmitting hepatitis B virus (HBV) via blood transfusion still exists. Consequently the introduction of HBV NAT testing is a matter of discussion today. Beside more costly testing another possibility of protection against HBV infection is possible: the active vaccination of blood donors (BDs). So far no data are available displaying the percentage of already HBV vaccinated BDs. We compared the HBV vaccination status and demographics of two BD populations (BDPs), apheresis donors (ADs) of a university based (UBBDS) and whole blood donors (WBDs) of a Red Cross blood donation service (RDBDS). Methods: Over 6 months all incoming ADs of the UBBDS were requested to answer a standardized questionnaire. Therein it was asked whether they ever had been vaccinated against HBV and for what reason. Additionally demographic data like gender, age and occupation were registered. In a second step the questionnaire was given to randomly appearing WBDs of the RCBDS. Results: The questionnaires of 260 ADs and 6812 WBDs were included for analysis. Whereas the distribution of gender was not different between the two BDPs (p = 0.372), ADs had a more academic social background, donated blood more frequent, were younger and more often HBV vaccinated (table 1) . Overall women had a higher vaccination rate than men (ADs: 54% vs. 31%, p < 0.001; WBDs: 26% vs. 19%, p < 0.001). The percentage of HBV vaccinated BDs declined with increasing age. The main reason for the HBV vaccination was work-related although foreign travel was named about as much from WBDs. Conclusions: ADs are significantly more often HBV vaccinated than WBDs. This might be mainly due to different social background and age. As ADx typically donate blood quite often a high proportion of HBV vaccinated ADs might already reduce the risk of transmitting HBV via blood transfusion. Therefore, one should rise the question whether HBV vaccination of non-immunised ADx could be as effective as introducing HBV NAT testing. standing donor demographics and donor characteristics is crucial for any blood center in understanding how to mobilize potential donors to donate blood. Study Design and Methods: The study population included all whole blood donors from January 1-December 31, 2003 who donated blood either at the Urumqi City Blood Center or at one of its mobile blood collection buses located throughout Urumqi city. There were a total of 29,784 whole blood donors who donated within this time period. Results: The majority of donors, 71%, were first time donors with only 19.4% having donated twice and 9.6% having donated more than twice. Students had the highest proportion of firsttime donors, 80.5%, and this group tended to have fewer repeat donors with only 19.5% donating 2 or more times. Laborers, and professionals tended to have proportionately more repeat donors with 32.3% and 33.9%, respectively, donating 2 or more times. Most donors, 73%, donated only 200 ml of blood which undoubtedly contributes to blood shortages, particularly because most of these donors do not return for subsequent donations. Among the 946 donors who were positive by screening tests for on or more transfusion transmissible infections (TTI) including HBV, HCV, HIV or syphilis, 77.1% were among first-time donors. Among all donors, the positive rate for TTI at screening test were 3.5% for first-time donors, 2.7% for donors who donated twice and 2.1% for donors who had donated more than 2 times or more than 4 times. Discussion: Sustained evaluation of blood donor demographic and donation characteristics is essential for maintaining an adequate and safe blood supply. The typical blood donor in Urumqi based on the data collected in this study can be characterized as male, less than 36 years of age, Han Chinese, having a high school education or better, being a first-time donor and donating at a mobile blood collection bus. Higher rates of TTIs were observed among first-time donors which is consistent with research in other blood donor settings. By better understanding the demographics and donation characteristics of its study population, the blood center can create and implement specific recruitment and retention campaigns and can identify high-risk donor groups.

Background. Each third dose of whole blood in Russia is donated by firsttime blood donor. There are two reasons for attention to this kind of donors: 1) possible risk of infectious disease in seronegative study; 2) possible risk of donation for person with contraindication. We investigated role of regional deferred donors registry (RDDR) in by first-time donor selection. Methods. Moscow RDDR includes parts: HIV, viral hepatitis, syphilis, tuberculoses, malaria, drug users, psychiatry, 60 days after blood donation. RDDR was complete and our center began actively work with it since last year. Each donor has to be registered in RDDR and automatically checked for deferral reason. Effectiveness of RDDR was investigated. Results. First-time donors donate less than 10% blood in our center. About a quarter of them are deferred before possible donation (table 1). Part of donors deferred by RDDR has been significantly increase in 2003 (X2 = 19,6; p < 0,001) at the expense of seropositive people.

M Schmidt, L Walch, M K Hourfar, E Seifried, W K Roth, German Red Cross, Frankfurt, Germany BACKGROUND: After the introduction of PCR testing of blood donations the expected residual transfusion risk for hepatitis B virus infection is estimated to be 1 : 500,000 and about 40 times higher than for HIV or HCV. The main reason for the elevated transfusion risk are chronic anti-HBc positive, HBsAg and PCR negative donors with low virus load. MATERIAL/METHODS: 13,136 blood donors at our blood donation centre were tested for Anti-HBc (Abbott PRISM® HBc, Abbott PRISM® HB core) to evaluate the diagnostic sensitivity and specificity. Initially reactive samples were confirmed by Anti-HBc AxSym CORETM assay or Anti-HBs (AxSym AUSAB) assay. Furthermore assessment of sensitivity and specificity for three Anti-HBc assays (PRISM® HBc, PRISM® HB core and AxSym CORE) was made by testing 20 highly positive samples which were diluted 1 : 5 in five consecutive steps. RESULTS: 270 and 255 out of 13,136 donors were initially reactive for HBc (2.1%) and HBcore (1.9%), respectively. For Prism HBc assay 249 could be confirmed by AxSym CORETM. 157 (58.2%) out of 270 Prism HBc reactive samples were also positive for Anti-HBs and 115 (73.3%) out of 157 Anti-HBs positive results had an Anti-HBs titer over 100 IU/l. For Prism HBcore assay 251 could be confirmed by AxSym CORETM. 152 (59.6%) out of 255 Prism HBcore reactive samples were also positive for Anti-HBs and 111 (73.0%) out of 152 Anti-HBs positive results had an Anti-HBs titer over 100 IU/l. Diagnostic specificity was 99.87% and 99.97% for Prism HBc and Prism HBcore, respectively. Only in 0.2% and 0.03% out of 13,136 donors for HBc and HBcore assays, respectively, initial reactive Anti-HBc results could not be confirmed. The analytical sensitivity revealed a 95% hit rate at a dilution of 1 : 12.5 for HBc and HBcore, whereas the CORETM assay still detected a dilution of 1 : 27.7. CONCLUSION: Both Prism Anti-HBc assays revealed numerous initially reactive donors. However most of them could be confirmed by a second Anti-HBc (AxSym CORETM) assay or Anti-HBs test. The new Prism HBcore assay had a better diagnostic sensitivity and specificity than the Prism HBc assay, whereas the analytical sensitivity was best for the AxSym CORETM test. Further examinations are necessary to estimate the infection risk by initially anti-HBc positive blood donors.

J L Blejer, Fundacion Favaloro, Buenos Aires, Argentina; A O Chiera, Instituto de Hemoterapia de la Provincia de Buenos Aires, La Plata, Argentina; M C Saguier, H J Salamone, Fundacion Favaloro, Buenos Aires, Argentina Background: Anti-HBcore (aHBc) testing in blood donors (BD) is questioned because of its low value as surrogate test and an important percentage of false positive results. The aim of this work was to evaluate the performance of a new kit available in our country comparing it with other two routine tests. Methods: Routine tests: A) Hepanostika anti-HBc Uniform, Biomerieux, B) Core Abbott Axsym System. Evaluated test: C) Wellcozyme anti-HBc, Abbott-Murex. Repeatedly reactive (RR) samples were tested with other markers for HBV: a-HBs, a-HBe and a-HBc IgM Abbott Axsym. In order to evaluate the relative sensibility we selected 140 BD samples RR for test A with other markers and a ratio cut off/donor sample absorbance (CO/DA) >2 (A1) or £2 (A2) and without other markers and CO/DA >2 (A3) or £2 (A4); and 89 BD samples RR for test B with other makers and CO/DA >3 (B1) or £3 (B2) and without other markers and CO/DA >3 (B3) or £3 (B4). In order to evaluate the relative specificity, there were studied simultaneously 293 BD samples for tests A and C and 445 for tests B and C. Results: The evaluated test showed a good relative sensitivity because all the samples except one (Group A2) from the groups with other markers (A1, A2, B1, B2) were reactive; reactivity in groups without other markers were lower (Table) .

Conclusion. RDDR is effective for blood donor selection and decreases necessity in laboratory screening. First-time blood donors have to be examined before blood donation. If they have not contraindications, donation can be performed up to 10 days before examination and screening.

These samples were selected with the assays employed in the laboratory routine, which implies to make a bias in favor with these techniques. As regards specificity, from the samples assayed with tests A and C, 2 were reactive for both, 288/293 were negative for test A, and 291 for test C, and from the samples assayed with tests B and C, 4 were reactive for both, 438/445 were negative for test A, and 441 for test C. The 6 samples negative for test C and reactive for A or B presented values of CO/DA between 1.1 and 1.4 and were negative for a-HBs, a-HBe and a-HBcIgM. Conclusion: the relative sensibility and specificity of the evaluated test is very suitable to be used in BD screening. There is an agreement for the 3 a-HBc tests when the samples are frankly positive, but for weak samples (low CO/DA) without other markers, the evaluated test was useful in order to avoid the unnecessary discarding of blood units beause of its higher specificity.

S Caglioti, S Cyrus, R Spizman, F Stallone, G Robertson, Blood Systems Laboratories, Tempe, AZ Background: A large donor testing laboratory implemented a new HBsAg EIA for screening and confirmation in late July 2003. This assay showed improved sensitivity and equivalent specificity in clinical trials. Several facilities implementing this assay identified a dramatic increase in repeatedly reactive (RR) donations. The confirmed positive (POS) rate did not increase but a large number of POS donations were nonreactive (NR) for anti-HBc. Data presented at a Blood Products Advisory Committee (BPAC) meeting led to recommendation that blood centers submit a variance request to FDA to allow evaluation of these HBsAg POS, anti-HBc NR donors for re-entry. Analysis: Test data for the new test was analyzed to identify the number of donors eligible for re-entry evaluation using a modified re-entry protocol. A total of 1,649,000 donations were screened for HBsAg using the new assay through March 2004 with 2079 (0.126%) RR donations. A total of 1995 RR donations were evaluated by neutralization with 702 (35%) found POS. Of the 702 confirmed positive donations, 175 (25%) were also NR for anti-HBc. These 175 donors would be eligible for evaluation using the modified reentry protocol described at the BPAC meeting. A small study was performed using an alternate, licensed 3.0 HBsAg EIA and 32 of 53 index donations were NR. Conclusion: With the current RR rate for the new HBsAg assay, approximately 9% of RR donors could be evaluated for re-entry. Each facility with an FDA-approved variance would be allowed to test these index donations using an alternate, licensed 3.0 HBsAg EIA. Donors with an RR alternate EIA would remain permanently deferred. However, donors with a NR alternate EIA could be contacted to submit am 8-week follow-up sample. Based on our study, 60% of HBsAg POS, HBc NR donors would be eligible to submit a follow-up sample. This follow-up sample would then be tested for HBsAg, anti-HBc and anti-HBs. Donors with NR results for these tests could be re-entered as regular donors. This modified HBsAg re-entry protocol would allow blood centers an opportunity to re-enter donors that might have been deferred inappropriately using this new assay.

F F Stallone, E Robertson, S Caglioti, R Spizman, G Kulkarni, Blood Systems Inc., Tempe, AZ Background: A large regional blood screening laboratory implemented HBsAg System 3 (System3) on 07/28/03. Within 60 days of implementation the initial reactive (IR) and the repeat reactive (RR) rates increased four-fold over HBsAg System 2. A formal investigation was initiated with the vendor for failure to meet package insert reactive rates: IR (0.28%) and RR (0.05%). Study: A Work Flow Analysis was performed to identify areas of potential process improvements. An analysis was performed between OSP and manual processing of this assay. Lower rates utilizing the manual process was demonstrated. The manual wash process seemed more effective in reducing the potential of non-specific binding. As a result, the following improvement steps were initiated: In early September, IR samples were held for 24 hours and then recentrifuged prior to retesting. This change addressed the potential for "fresh serum effect" and particulates suspended in the serum. In October the frequency of routine maintenance required on the Ortho Summit Processors was increased. The weekly wash manifold antimicrobial flush was now being performed daily. In October the working substrate was limited to two hours. The vendor increased the System 3 stability claim of the working substrate from 30 minutes to 8 hours. Plates processed with >2 hours old substrate seemed to have more reactive samples per plate. In October the usage of OSP disposable substrate reagent containers were reduced from 14 days to 7 days and the dispos-able syringes from daily to per shift. In December and January increased IR rates were noted. System 3 processing was switched to a different OSP. Routinely, dedicated OSPs are used for each assay.

Conclusion: Using Flow Charting and Process Flow analysis, potential areas of focus were identified. By implementing multiple process improvements, reactive rates were reduced to nearly HBsAg System 2 levels. The Confirmed to RR rate, although not the same as System 2 has improved significantly. The vendor is working to identify sub-components of their system that may have contributed to the December/January IR increase. Further investigation is necessary in this area. The results of this study indicate the potential to improve the performance of any assay performed on the Ortho Summit Processor. We are currently in the process of expanding these process changes to all our assays.

J Lowther, M Reutz, R Plumlee, S Mirpour, Gulf Coast Regional Blood Center, Houston, TX Background The implementation of Ortho ® HBsAg Confirmatory Test (System 3) caused an increase in the number of HBc non-reactive samples submitted for confirmatory testing. With the "rare occurrence" algorithm, almost 50% of the samples tested required additional dilutions. The increase in false reactives caused an immediate impact on the donor's future eligibility. The amount of time required for test resolution placed a strain on staffing, resources and our ability to complete testing in a timely manner. A complete investigation was performed to determine if these issues could be reduced by improving laboratory operations. Methods Ortho ® System 2 and System 3 pre-and post-implementation data, HBsAg initial and repeat reactive rates, dilution data and confirmation rates were analyzed. Confirmatory algorithm calculation changes, recommended by Ortho, as well as kit lot numbers, frozen vs. liquid samples, and HBsAg screens performed by the OSP (automated) method vs. the semi-automated method were also evaluated to isolate problematic areas. Data review indicates that neither the variables or the calculation change made a significant difference in the rate of confirmed reactive samples that were anti-HBc non-reactive, or influenced the number of dilutions required to obtain a final result. The following table represents a comparison of system 3 confirmation rates for HBsAG System 3 EIA repeat reactives performed using the OSP and semi-automated methods from January 14, 2004-May 12, 2004. Confirmation of Ortho ® System 3 Repeat Reactives CR = Confirmed Reactive CN = Confirmed Non-reactive.

The confirmed reactive rate for OSP reactives was 13.3% and for the semi-automated method was 23.7%. Even though the OSP data seems to indicate that there are more false positives associated with the automation, the semi-automated method still does not meet expectations. The review of confirmation data when Ortho ® System 2 was utilized as the screening test shows that the confirmed rate was higher at 35.9%. With the scope of our investigation and subsequent information from other sources, the deficiencies were determined to be test system derived. The 10-20% increase in the false positive reactive rate demonstrates the challenge of improving the test performance of an assay when increasing the sensitivity while trying to maintain the specificity. 

D Chien, A Tabrizi-Wright, Y Fong, G Ching, B Phelps, Chiron Corporation, Emeryville, CA Background: Due to its mode of replication by reverse transcription of its pre-genomic RNA, HBV has a high rate of mutation relative to other DNA viruses. Amino acid substitutions have been described in all HBV DNA encoded viral proteins such as polymerase, nucleocapsid protein (HBcAg) and surface antigen (HBsAg). There are increasing concerns about the contribution of variant HBsAg to vaccine escape, immune prophylaxis failure and false negatives in serological HBV diagnosis or blood screening assays. The group-specific "a" determinant region (amino acids 124-147) has attracted most attention, not only because mutations in the "a" determinant were found in 10-20% vaccine escapes, but also because of the fact that high-affinity antibodies in response to this region have been used in most HBV diagnoses. Method: To address the needs of an improved immunoassay capable of detecting the major mutations in the "a" determinant regions, four potential rabbit monoclonal antibodies, which exhibit a broad immunoreactivity and higher affinity to HBsAg mutant antigens, were selected for further characterizations. The goal of the study is to identify a single or a few rabbit monoclonal antibodies that could replace the current panel of mouse monoclonal antibodies for the detection of the various HBsAg mutants. Results: The results from the Biacore analysis and the sandwich ELISA analysis demonstrated that the rabbit monoclonal antibodies, 99S9, 96S1, 99S1, and 99S6 have superior sensitivities against the various HBsAg mutants compared to the mouse monoclonal antibodies. Conclusion: The study results also indicated that a single rabbit monoclonal antibody could substitute the multiple mouse monoclonal antibodies used in current assay development. Therefore, rabbit monoclonal antibodies may provide a tool for better detection of HBsAg variants. Background: Defining the conditions that preserve sample integrity is of paramount importance, taking into account the trend toward centralized testing. Therefore, we have investigated the stability of HBsAg, HBcAb, HBcAb-IgM, HBsAb, HBeAg, HBeAb, HIV I/II Ab & HCVAb-positive samples, for storing duration and conditions, beyond the usually tolerable limits. Method: During a two-month period, samples positive for each one of the above mentioned parameters, were collected from 3-6 patients and/or donors, and were frozen down to -20°C. After the collection procedure was completed, the samples were pooled according to each parameter and the pooled sera were distributed among three groups. The first group was frozen down to -20°C, the second group was stored in a conventional refrigerator (4-8°C), and the third group was kept in ambient temperature (20-25°C). The relative values of Signal to Cut-off (S/CO) were being determined for all the 8 parameters every 15 days and for an overall time-period of 4 months. The frozen samples were thawed each time, before the measurement. Results: All except one of the parameters surveyed, remained positive, under all storing conditions, during the four-month testing period. The HbsAb-positive sample unfolds under ambient temperature (20-25°C), a weak fading trend, during the last two measurements. Only the frozen (-20°C) HbeAg-positive sample appeared to behave as negative, during the last two measurements. The comparative (in %) variation, between initial and final relative values measured, is following: provides information that might be useful, under eventually extraordinary, and non-reversible conditions of sampling and storing, as for example in the case of an accident, a natural disaster, or another sort of emergency.

Objective Sanquin is screening for HIV-and HCV-RNA in pools of 48 donations with semi-automated extraction (NucliSens-BioMerieux) and cDNA PCR amplification (AmpliScreen-Roche). Sanquin is examining the introduction of HBV DNA testing. Decrease of pools size is necessary because HBV-DNA load is low during the pre-seroconversion phase of the HBV infection. HBV-DNA testing in pools of 6 with automated extraction (AmpliPrep, Roche) and PCR (AmpliScreen, Roche) is examined. Testing in pools of 48 for HCV-and HIV-RNA with the current system is still feasible by combining 8 pools of 6, making it possible to limit the number of samples to be amplified. The objective of the evaluation is to investigate the sensitivity and to determine the robustness in pool screening in the setting of blood bank routine. Materials Tested panels: Dilution series of the International Standard (WHO-IS) for HBV-DNA; and commercially available reference panel for HBV-DNA (PeliCheck, Sanquin-VQC). Members of these panels were tested in 24 replicates. Robustness was evaluated by screening of 600 donations in pools of 6 and 2400 donations in pools of 24. Cross contamination sensitivity was examined by alternate testing of high HBV DNA positive (10 8 geq/ml) and negative samples. Results Analytical sensitivity : HBV-DNA concentration for 95% success rate of detection (95% CI) as determined by Probit analysis with log conversion of loads.

Conclusions: The present project obviously does not aim to encourage any kind of violation of sera storing guidelines. However, the achieved evidence No cross contamination was found and >99.5% of the results were valid. Conclusions Based on the 95% limit of detection, the probability of detecting HBV DNA for the individual donation when present in a pool of 6 or 24 is respectively 30 IU/ml and 120 IU/ml. The assay proved to be robust during the validation period and could easily be implemented in the current operational setting with a modification of the pooling protocol by preparing intermediate small pools. Here we describe the development of a Target-Capture PCR technology based NAT assay for HBV. Methods: A magnetic bead-based protocol was used for isolation of HBV DNA target along with a related internal control in a single tube from 0.5 mL plasma. The Region X on the isolated target on the beads was amplified by TaqMan technology. The qualitative assay was validated for Analytical and Clinical Sensitivity, Specificity, and Reproducibility. The quantitative assay utilizes a 12 member panel of standards tested in triplicates in the range of 50-10 5 IU/mL. A blinded panel of HBV samples by QCMD were tested for detection and quantitation. Results: The analytical sensitivity of the qualitative Target-Capture PCR HBV Assay determined with WHO HBV standard indicated >95% positivity at 15 IU/ml. The assay detects all known genotypes, has very high specificity and tolerates a variety of anti-coagulants, interfering substances, and plasma from pathological conditions. The quantitative assay estimates HBV in the range of 10 2 -10 9 IU/mL for unknowns. The QCMD BBV results indicate the detection of the lowest copy member and also very close estimation of IU/mL for the A genotype. Conclusion: Using a combination of magnetic bead based target capture and TaqMan amplification-detection a rapid, sensitive, userfriendly, and accurate assay for HBV DNA detection and quantitation has been developed. Background: Previous Brazilian legislation required that donors were screened by ALT and deferred from future donations if they had elevated ALT level in two successive donations. Recently, a new Brazilian legislation has withdrawal the obligation of donor ALT screening. Since ALT elevation occurs a few days before antibody detection by third generation EIA we retrospectively tested all ALT elevated donations by NAT in order to evaluate the impact of this recent decision on the Brazilian blood supply. Methods: From January 1996 to December 2002, all donors who were 3 rd generation anti-HCV EIA negative and had elevated ALT level were selected to be included in the study. Donors were judged not to be HCV infected if: a) a repository sample from the ALT elevated donation was negative by NAT or; b) a successive donation/sample was negative for HCV-EIA and NAT. Results: During the study period 62,254 blood donations were screened for ALT and elevated levels of this enzyme were found in 1125 donations (1.8%) from 1064 HCV antibody negative donors. We have analyzed by NAT the HCV-RNA of 722 donors (67,9%) and all were found not infected by HCV. Interestingly, among the 722 donors investigated, 85 (out of the total 95-89.5%) donors had super-elevated ALT levels (values greater that 2 times the normal value). Conclusions: At least for 67,9% of the donors with elevated ALT levels investigated, ALT screening did not contribute to any additional protection to the blood supply. Although not definitive, this result supports the recent decision of the Ministry of Health to withdrawal the obligation of donor ALT screening in Brazil. The Vitros® Immunodiagnostic Anti-HCV Assay (ECi), unlike traditional enzyme immunoassays (EIA) used in blood bank screening, has a very broad dynamic range due in part to enhanced chemiluminescence as the positive signal generator. The assay combines good specificity and excellent sensitivity with the advantages of an automated, continuous random-access immunoassay system. We used the Anti-HCV ECi assay to better understand the dynamics of the humoral immune response versus viremia during Hepatitis C (HCV) infection. METHODS: Two populations of specimens were evaluated: 1) 363 Chiron TM RIBA TM HCV 3.0 SIA confirmed positive specimens from volunteer blood donors, i.e., a cross-sectional analysis of whole blood donors; and 2) 31 index and 512 follow-up samples from 31 acutely infected plasma donors detected by NAT screening. Of the 363 RIBA positive whole blood donor specimens, 92 (25%) were non-viremic using discriminatory HCV Transcription Mediated Amplification (dHCV TMA) (Chiron/Gen Probe); while 271 (74%) were viremic. Anti-HCV ECi results were calculated as signal to cut-off (S/CO) ratios. RESULTS: For the cross-sectional seropositive donor population, the mean ECi S/CO ratio for non-viremic specimens was 15.45 (standard deviation, 1.01) vs. a mean of 31.79 for viremic specimens (SD, 5.57); the difference was highly significant (p value = 0.0001). Use of receiver operating characteristic (ROC) curve analysis indicated that using an ECi S/CO of 18.0, it was possible to discriminate between HCV RNA positive and negative specimens with a sensitivity of 97.8%, a negative predictive value (NPV) of 90.8%, and a positive predictive value (PPV) of 88.3%. Furthermore, comparison of ECi S/CO ratios from HCV acutely infected plasma donors resulted in the following observations: 1) window period sensitivity of HCV 3.0 EIA and the anti-HCV ECi assay were identical, 2) on the average window period sensitivity of the anti-HCV ECi assay preceded HCV 2.0 EIA detection by 29 days (p = 0.0001), and 3) the anti-HCV ECi assay has the ability to measure evolution of the humoral immune response when compared to changes in number and intensity of RIBA bands. CONCLUSIONS:

These studies suggest that, the anti-HCV ECi assay demonstrates sensitivity equal to licensed 3 rd generation assays currently used for screening blood donations. Moreover, due to it's broad dynamic range, it is possible to identify changing anti-HCV titers as well as to discriminate between viremic and non-viremic individuals regardless of time of infection.

Immunoreactive HCV NS3/4a Protein Without NS3 Serine Protease Activity S Lin, A Medina-Selby, D Coit, P Arcangel, C McCoin, P Ng, A Gyenes, C Hu, B Phelps, D Chien, Chiron Corporation, Emeryville, CA Background: Hepatitis C virus nonstructural protein 3 (NS3) is 630 amino acid protein containing three functional domains. The serine protease domain is located in the amino terminus, whereas helicase and NTPase are in carboxy terminus. The serine protease of NS3 is responsible for the cleavage at the junction of NS3/4a, NS4a/b, NS4b/5a and NS5a/b. Previously we have expressed NS3/4a in yeast and purified the protein in non-denaturing conditions. The purified NS3/4a is a conformational protein, and is more sensitive than c200 or c33c antigen in early seroconversion antibody detection. In antibody assays using NS3/4a and MEFA 7.1 (multiple epitope fusion antigen) as antigens, we found the new antigens achieved 2-14 days of earlier detection of seroconversion antibodies in comparison with currently marketed assays. However, the NS3/4 protein did undergo self-hydrolysis and cleave MEFA 7.1 due to the protease activity. This study investigated if the immunoreactivity of NS3/4a might be preserved while the protease activity eliminated. Methods: The mutant proteins, NS3(S1165A) and NS3/4a(S1165A), were generated where Ser1165, one of the catalytic triad amino acids in the protease domain of NS3, is mutated to Ala. The protease activity and immunoreactivity of the mutant proteins were studied and compared with normal NS3/4a. Results: The mutants, NS3(S1165A) and NS3/4a(S1165A), had expression levels in yeast similar to that of NS3/4a, and was purified to similar purity as in NS3/4a (>90% purity). In antibody assays, NS3/4a(S1165A) showed ~90% immunoreactivity as compared with NS3/4a, whereas NS3(S1165A) had ~80% immunoreactivity. When cocoated with MEFA 7.1 on an ELISA plate, both mutant proteins achieved immunoreactivity very similar to that of NS3/4a in early seroconversion antibody detection. It has been confirmed on SDS-PAGE and Western blot, that the mutant NS3/4a(S1165A) showed the full length of the protein without cleavage of NS4a, and neither mutant protein cleaved MEFA 7.1. Conclusions: This is the first evidence demonstrating that the enhanced immunoreactivity of the NS3/4a protein is not affected by elimination of the serine protease activity of the protein. To further improve assay sensitivity and ease of assay development, we hope to adopt this conformational sensitive protein to well characterized streptavidin plate. To achieve that, studies are underway to derivatize the mutant with a short amino acid tag that can be recognized by a high affinity biotinylated antibody.

C Liu, F Ren, Q Lv, J Zhu, X Gao, J Li, Beijing Red Cross Blood Center, China, Beijing, China Background: To investigate whether processing and storage conditions might influence the stability and could alter the concentration of the HCV RNA in whole blood or in plasma; To determine the stability of HCV RNA during transport and storage of blood samples from donors in transfusion practice, prior to NAT testing. Methods: We studied the samples obtained from five patients known to be positive for HCV RNA. The samples were kept in different storage conditions with different anticoagulants. And at the end of processing the plasma samples were frozen at -80 degree C until fluorescent quantitative PCR testing. Results: There was no significant loss of HCV RNA titers in whole blood anticoagulated with CPDA or EDTA or ACD or none stored at 4 degree C after 48 h (P > 0.05); There was no significant difference in these four anticoagulants storage conditions at 4 degree C after 48 h (P > 0.05). No decline in HCV RNA level was observed after 48 h of storage of whole blood anticoagulated with ACD at 4 degree C or at 25 degree C, while significant declines occurred at 37 degree C after 48 h (P < 0.05). There was no decline in HCV RNA level of plasma samples at 4 degree C after 7 days (P > 0.05); While a significant decline was observed in HCV RNA level of plasma samples at 25 degree C (room temperature) after 7 days (P < 0.05). After four freeze-thaw cycles there still was no loss of HCV RNA in plasma samples (P > 0.05). Conclusions: The HCV RNA is stable relatively. We conclude that HCV RNA is resistant to degra-dation under routine laboratory handling and storage conditions or blood collection transport and processing conditions. The influence of different anticoagulants on the stability of HCV RNA is of no consequence. Blood samples would better be stored at 4 degree C after collection and plasma separated within 48 h. We believe that HCV RNA remains stable at 4 degree C for at least 7 days or at room temperature for 48 h, allowing greater flexibility in samples collection and transport in transfusion practice nowadays. HCV RNA in plasma samples subject to up to four short-term freeze-thaw cycles is still stable. Conclusions: In 24-26% of the HCV cDNA-PCR -ve, anti-HCV ELISA rr blood donors with a RIBA -ve or ind test result risk factors for HCV infection were present. In RIBA +ve donors a significant higher percentage risk factors were present. Case report In July 2002, a 56-year-old female unremunerated repeat blood donor tested positive in HCV-NAT. HCV antibody assays (ELISA (PRISM) and immunoblot (RIBA®)) were negative (the first in >2 ¥ 10 6 donations). Five days after the donation, the donor was invited to the blood bank to collect fresh blood samples (for anti-HCV ELISA, RIBA and cDNA-PCR test) and to obtain a standardized interview. Again, HCV antibody tests were negative and the donor tested positive in the HCV cDNA-PCR (viral load 23 ¥ 10 6 copies/mL in bDNA). Thoroughly counselling revealed no risk factors for HCV infection, except for a new short heterosexual relationship 6-7 weeks prior to the donation. There were no indications for co-existing venereal diseases. Eight days after the donation, the donor seroconverted (ELISA positive, RIBA indeterminate). That day a therapy with Peg-Interferon and Ribavirin was started for 24 weeks. We did a weekly follow-up for HCV antibody and cDNA-PCR testing. Follow-up After 5 weeks the anti-viral therapy, the donor remained negative in consecutive cDNA-PCR testing. Two weeks later the RIBA changed from positive to indeterminate. One year after the donation, the donor was discharged for follow-up from her HCV infection. The male sexual partner could be traced for testing and counselling. The interview revealed risk factors for HCV infection (IVDU and tattoo 1977) . Tests revealed a HCV infection (anti-HCV ELISA, RIBA and cDNA-PCR positive, bDNA: 15 ¥ 10 6 copies/mL, anti-HIV negative). Genotyping showed a HCV type 3a in both donor and her sexual partner. Sequencing of the NS5B part of the HCV genome demonstrated that the virus in donor and partner were identical. Conclusion We report a case of acute hepatitis C in a repeat donor, most likely caused by sexual transmission. Immediate anti-viral therapy probably contributed to clearance of HCV. Background: Serological based donor screening reactive results are supplemented by alternate EIA or confirmatory assays. Although NAT for HIV-1/HCV RNA tests are highly sensitive and specific there is currently no standard for supplemental testing of reactive samples. There are two manufacturers licensed in blood donor screening. Therefore using initial reactive NAT specimens, we evaluated the Procleix assays with the COBAS AmpliScreen PCR HIV-1 and HCV assays as cross-supplemental NAT confirmatory tests. Methods: Procleix HIV-1/HCV (Mx) assay initially reactive samples (N = 240) were tested for HIV-1 and HCV RNA with one or both of the COBAS AmpliScreen PCR assays: AmpliScreen Standard Specimen Processing (STDPrep) and/or MultiPrep Specimen Process (MLTPrep). A repeat test was performed using the Procleix Mx and Procleix discriminatory assays (dHIV-1, dHCV). Samples were also evaluated by National Genetics Institute (NGI) PCR HIV-1, HCV assays. Samples were considered reactive only if two or more of the methods generated a reactive result. STDPrep tested 179 HIV and 178 HCV samples. MLTPrep tested 213 HIV and 217 HCV samples. Procleix dHIV and dHCV tested 239 HIV and HCV samples. NGI PCR assay tested 153 HIV and 234 HCV samples. Results: Of the 240 specimens evaluated for HIV, 3 specimens (1.3%) were true positive (TP), 180 were true negative (TN) (75%) and 57 (23.7%) had discordant results. Of the 57 discordant specimens, there were 55 false positive (FP) Procleix Mx, 1 FN STDPrep and 1 discordant sample: Procleix Mx Rx, MLTPrep negative. Of the 240 specimens evaluated for HCV, 54 specimens (22.5%) were TP, 176 TN (73.3%), and 10 (4.2%) had discordant results. Of the 10 discordant specimens, there were 6 FP samples: 4 Procleix Mx, 1 STDPrep and 1 NGI; the 2 FN were: 1 dHCV and 1 NGI. The remaining 2 discordant samples were positive by Ampliscreen MLTPrep and STDPrep but negative by Procleix and NGI. The majority of FP samples were found with the Procleix Mx assay for both HIV-1 and HCV RNA and the NGI assay had some FN samples. Evaluation of MLTPrep assay results with Procleix discriminatory tests only, found 100% concordance for HIV-1 and 98.8% concordance for HCV. Conclusions: The data show there is 100% sensitivity and high concordance between the Procleix dHIV-1, dHCV and the Ampliscreen MLTPrep PCR HIV-1 and HCV assays, 100% and 98.8% respectively. Therefore, the COBAS MLTPrep and Procleix discriminatory assays are appropriate for use as cross-supplemental NAT confirmatory tests for blood donor screening. The few discrepancies between the COBAS Ampliscreen results and the Procleix discriminatory results will be investigated further.

Background: Post donation counseling of blood donors with positive microbiological markers is not a routine practice in India and universal guidelines are not available, unlike the developed world. This effort aimed to counsel Hepatitis B (HbsAg) positive blood donors to reduce the risk of disease transmission in the community. Methods: Questionnaires were used to determine donors' preferred mode to communicate their test results. A protocol for post donation counseling was developed using the NBS, Colindale, UK guidelines. Donors were referred to the specialist clinic after evaluation of test results of investigations like HbsAg, ALT levels and HbeAg. Results: In spite of 16% deferral due to history of jaundice, the prevalence of Hepatitis B among the blood donors is 3.2% at our center. Most donors (96%) were keen on knowing the test results, 58% preferred information by mail, 30% by direct personal contact and 12% wanted to be informed on the telephone. On call 89% (57/64) donors turned up for post donation discussion. On testing the donors for HbeAg, 5.2% (3 of 57) were reactive by ELISA method. One of the donors, his 3 siblings and mother, excluding father were positive for HbsAg, HbeAg and HBV DNA with high ALT levels. The transmission seemed vertical through mother who has developed hepatocellular carcinoma. The family is being managed at the specialist clinic. Conclusion: Uninformed, blood donors positive for microbiological markers remain potential source of onward disease transmission. The issue merits global attention and universal guidelines need to be developed and mandated. Understanding the origin of infection can identify areas to improve blood safety. 

Objective: In Sanquin, blood donations are currently screened by NAT for HIV and HCV in pools of 8. Sanquin is now assessing the introduction of HBV DNA testing. For this purpose the evaluation of the Procleix Ultrio Assay was performed. Materials and methods: Dilution series of WHO International Standards for HBV, HCV and HIV and commercial reference panels for HBV-DNA genotype A, HCV-RNA genotype1 and HIV-1 RNA genotype B (PeliCheck, Sanquin-VQC) were tested in 24 replicates. For HBV genotype detection, dilution of genotypes A to G samples were used. Robustness was evaluated by screening of 8800 donations in pools of 8. Results: Analytical sensitivity: concentration of the target for 95% success rate of detection (95% CI) as determined by Probit analysis with log conversion of concentrations.

All genotypes were detected at 100 geq/ml or better. No false positive result was found and 99.9% of the results were valid. Conclusions: Based on the 95% limit of detection, the probabilities of detecting HBV DNA, HCV-RNA and HIV-RNA for the individual donation when present in a pool of 8 are respectively 72 IU/ml, 24 IU/ml and 256 IU/ml. HBV genotypes (A-G) are detected with comparable sensitivity. The assay proved to be robust. Conversion of the routine testing from pools of 48 to pools of 8 might be done with a limited expansion of workload.

M A Gonzalez Jr., V Régine, V Piccinini, F Vulcano, A Giampaolo, H J Hassan, Blood Transfusion Methodology Section, Department of Hematology, Oncology and Molecular Medicine, Istituto Superiore di Sanità, Rome, Italy Estimating risk of transfusion-transmitted infections is essential to monitor blood safety. Risk infection toward transfusion is evaluated by the probability to develop an infection in spite of the accuracy of the transfusion practice and analytical testing. The magnitude of this residual risk depends on the sensitivity of the assays used for screening and on the incidence of the infection among the population. In this study the residual risk of transfusiontransmitted infections was estimated according to the "Incidence Rate/Window Period" model: the incidence rate of viral infections among repeat blood donors (expressed per 100,000 person-years) was multiplied by the mean length of the window period before seroconversion (expressed as a year-fraction). The results are based on the Surveillance System of Transfusion-transmitted Infections (SSTI) and, on a sample of 5850 repeat donors from 39 Italian Transfusion Structures, in order to have an estimate of the number of person-years. 1,079,281 Italian repeat-donors were surveyed by the SSTI during the study period 1999-2001. In the three years study period 302 new infections were observed: 77 donors seroconverted to anti-HIV (ELISA-3, confirmatory western blot) and 225 to anti-HCV (ELISA-3, confirmatory RIBA-3). Sampled donors gave blood a median of 5.2 times during the follow-up; the median interval between donations of the repeat donors population was 224 days (95% CI: 212-236); the median follow-up of sampled donors was 2.25 years (27/36 months); the estimated person-years of the Italian repeat donor population was 2,430,567. The estimated residual risks per 1,000,000 donations were 1.91 (probable range 0.52-3.32) for HIV and 16.74 (probable range 9.57-24.01) for HCV. Combined residual risk of HIV and HCV seroconversion was 18.65 per 1 million units donated, it means that annually 37 donations from repeat donors might probably be infectious for these diseases in Italy. These data show that the risk of acquiring a transfusion-transmitted infection is in the same order of magnitude of European countries and can be considered low. Background: A new triplexed Nucleic Acid Test, Procleix Ultrio Assay (Gen-Probe/Chiron) using Transcription Mediated Amplification to detect HCV, HIV-1 and HBV was evaluated in an Italian multi-center trial. The study was segmented into: Analytical sensitivity, specificity and robustness Clinical sensitivity and HBV DNA yield analysis compared to third generation HBsAg assays Evaluation of Ultrio feasibility in routine compared to Procleix HIV1/HCV Assay Method: For analytical sensitivity, dilution series of the ISS reference standards (ISS HCV RNA 0102; ISS HIV-RNA 0103 and HBV DNA 0501) were prepared and assayed in 24 replicates. For clinical sensitivity, 44 samples from HBV acute and chronic patients, 100 anti-HBc positive alone and 212 anti-HBc/anti-HBs positive blood donors samples, and ten HBV seroconversion panels (Impath/Bioclinical partners) have been tested. For routine testing analysis, 9.732 volunteer blood donors from 14 different blood centers have been tested in pool of 8 or in individual donor testing in parallel with Procleix HIV-1/HCV Assay. A total of 21,582 tests was performed in 143 runs. Results: Based on probit analysis the analytical sensitivity, (95% detection probability) for HCV was 2.4 IU/ml (IC: 1,6-4,9 IU/ml), for HIV 5,2 IU/ml (IC: 4,1-8,0 IU/ml) and for HBV 13,9 IU/ml (IC: 7,0-74 IU/ml). Window period closure results with HBV seroconversion panels showed a mean closure of 16 days neat and 12 days in 1: 8 pools and in all cases exceeded the HBsAg screening assay. All samples from patients were correctly detected. In routine donor testing, one yield case (HBsAg negative and alternative NAT positive) was identified. Feasability analysis showed 0,1% of tests as invalid. Specificity was 100%. Test turnaround time was equivalent to the Procleix HIV-1/HCV Assay. Conclusion: These results demonstrate sensitivity greater than EIA testing, HBV window period closure of 12-16 days, and test performance adequate for routine HBV NAT implementation. The yield of 1 : 9732 donations shows that implementation of this test would increase blood safety compared to HIV/HCV testing alone.

J Lowther, D Vyas, R Plumlee, S Mirpour, C Cousins, Gulf Coast Regional Blood Center, Houston, TX Background Sensitive and specific screening tests for the detection of transmissible, infectious diseases are essential for maintaining the purity, potency and safety of the blood supply. NAT testing has recently received licensure and supplemented traditional methods of screening assays for HIV and HCV. A retrospective study was performed to evaluate post-licensed reagents' performance for the detection of HIV and HCV. Study Design Data from April 9, 2003-April 30, 2004 was analyzed. The total number of HIV/HCV tests performed during this time period was determined. ELISA RR results were evaluated against NAT and confirmatory results to assess the sensitivity and specificity of the assays post licensure. Results A total of 215,316 samples were tested for HCV by the NAT method. Of this total, 406 were Repeat Reactive (RR) by ELISA testing. RIBA was used as the confirmatory method. During the same time period, a total of 215,312 samples were tested for HIV by NAT. Of these samples, 313 were RR by ELISA testing. Immunofluorescence Assay (IFA) was used as the confirmatory method. The following tables demonstrate the relationship between NAT results, initial ELISA results and the confirmatory testing. ELISA HCV RR = 406

Conclusions Both HIV and HCV data analysis demonstrated correlation rates comparable to those seen during the clinical trial phase. NAT negatives demonstrating RIBA positive results are explained through virus resolution with persistent antibody titer. NAT HCV testing during this time period detected 2 additional HCV window cases (NAT+ / ELISA =). HIV data analysis demonstrates 100% correlation between NAT positive samples and IFA confirmation. There were no NAT negative samples that confirmed by IFA. Additionally, 1 HIV window case (NAT+ / ELISA =) was detected during this time period. This data analysis indicates that NAT testing continues to provide the same level of safety and effectiveness that was demonstrated during the clinical trial phase of testing.

S A Herman, Roche Molecular Systems, Inc., Pleasanton, CA; N Newton, Roche Molecular Systems, Inc., Alameda, CA; S Dugenny, H Huang, Roche Molecular Systems, Inc., Pleasanton, CA; S Will, Roche Molecular Systems, Inc., Alameda, CA Background: A four color, real-time RT-PCR test that simultaneously detects and differentiates HIV, HBV, HCV and an internal control (IC) was demonstrated. This multi-target assay was first developed as a two color test, in which the probes for all viral targets were labeled with the same fluorescent dyes, and the IC probe was labeled with a different reporter dye. In the present work different reporter dyes were used for the HIV, HBV, HCV and IC probes so that the viral targets and IC could be detected in separate channels on the COBAS TaqMan instrument. Methods: RT-PCR reactions contained primers for HIV-1, HIV-2, HBV and HCV and probes for the viral targets and IC. The IC was an in vitro RNA transcript containing primer binding sites from one of the viral targets and a unique probe binding sequence. Each probe was labeled with two dyes, a reporter and a quencher, such that the reporter fluorescence was quenched by the quencher dye. In RT-PCR reactions containing target nucleic acid, the corresponding probe is digested by the nuclease activity of DNA polymerase, separating the dyes from each other and unquenching the reporter dye so its fluorescence can be observed. We employed a non-fluorescent quencher dye on all probes and different reporter dyes on the HIV, HBV, HCV and IC probes. The probes for HIV-1 Groups M and O, and HIV-2 used the same reporter dye and were detected in the same channel. The prototype assay was evaluated on a COBAS TaqMan instrument with appropriate excitation and emission filters for fluorescent measurement of all four reporter dyes. Results: The specificity of the fluorescent signal due to each target was evaluated by testing ~1,000 copies of HIV RNA, HBV DNA, HCV RNA and IC RNA in separate reactions, and measuring the fluorescence in all four color channels. Each target yielded a strong fluorescent signal in the expected channel at ~3, 25, 100 and 20 times the baseline fluorescence, respectively. Three of the targets generated a small signal in one other channel (<1.2X baseline), which can be easily corrected with data analysis algorithms for fluorescence cross-talk compensation. The performance of the four color assay was compared to that of the two color test by analysis of serial dilutions of positive controls for HIV-2, HBV and HCV. The four-color and two color assays had similar performance. Conclusions: The feasibility of a four color, real-time RT-PCR assay for simultaneous detection and identification of multiple pathogens was demonstrated. Initial results show that the fluorescent signals in each channel are specific, and that the sensitivity of the four color assay is similar to that of the two color screening assay. 

E N Padilla, P V Holland, BloodSource, Sacramento, CA; J Akers, E Kyger, Y Yang, C Parkhouse, Roche Molecular Systems, Inc., Pleasanton, CA Background The objective of this study was to assess performance of a new automated NAT method for the COBAS s200 system. Various samples were tested to obtain pre-clinical data on the limit of detection and specificity of the test. Additionally, seroconversion panels were tested for comparison to more traditional serologic test results. Study Design/Methods This test uses the Roche COBAS AmpliPrep instrument for sample extraction, and the COBAS TaqMan analyzer for real-time PCR amplification and detection. The test is a multiplex test designed for simultaneous detection of HBV, HCV, HIV-1 Group M, HIV-1 Group O, and HIV-2. It is also designed concurrently to detect an internal control to ensure proper test performance. Three studies were performed on the system. All test runs contained negative and positive kit controls; runs with invalid kit controls were repeated, if sufficient sample remained. Specificity study: 500 plasma samples from normal blood donors, screened and found negative for HCVAb, HBsAg, HBcAb, HIV 1/2Ab, HCV NAT, HIV NAT, and HBV NAT, were tested. Initial positives were repeated in duplicate and the 2/3 rule applied. Limit of detection study: WHO referenced standards for HBV, HCV and HIV-1 (group M) were serially diluted and tested as 22 replicates. Seroconversion panels: Three commercially available panels were diluted 1 : 6 in normal human plasma and tested singly. Results Specificity study: 1/500 initial reactive, 0/500 repeat reactive = 100% specificity Limit of detection study: HBV 86%(19/22) detection at 1.5 IU/mL HCV 84% (16/19, 3 replicates invalid) detection at 3.3 IU/mL HIV 75% (12/16, 6 replicates invalid) detection at 10.0 IU/mL Discussion/Conclusions This test system was easy to use and operator training was very quick, particularly since it uses very similar equipment to what is already in use for WNV screening. During the study, the system performed well mechanically. Throughput for the system would necessitate one instrument set per approximately 500 donor samples screened per shift, if tested in pools of 6. The results of the study warrant further investigation of this method as a possible NAT multiplex blood-screening system. ated, & the assay is compatible with a variety of anti-coagulants. Selected concentrations for all internal & external controls generated consistent positive results under normal conditions. Conclusions Data obtained with the 5-parameter multiplex assay show excellent performance. Moreover, expansion capability to concurrent 6-plex target screening without compromise relative to single target sensitivities was demonstrated. The advantageous combination of sensitive multiplex PCR & parallel multi-dye detection of targets & IC will allow for a more comprehensive & efficient screening of blood & thus, may significantly contribute to making the blood supply safer. Background: The Procleix Ultrio Assay simultaneously detects HIV-1, HCV RNA and HBV DNA in plasma. The study objective was to generate specificity data as part of a clinical study to evaluate the Procleix ® Ultrio TM Assay (ULT) and the associated Ultrio Discriminatory Assays (dHIV-1, dHCV, dHBV). Data from these assays were compared to licensed screening assays: Procleix® HIV-1/HCV (Mx) and Discriminatory dHIV-1, dHCV as well as serologic tests for HBsAg and anti-HBc. Methods: Specificity was evaluated using 52,016 donor specimens in paired master pool tubes (MPTs) of 16 specimens and 6,012 ULT individual donor specimens (IDS). MPTs and IDS were tested across 3 conformance reagent lots. Reactive (Rx) Procleix ULT or Mx IDS were tested with the respective Discriminatory Assay(s). Discordant specimens were further tested with an alternate NAT PCR. IDS were tested for the ULT discriminatory specificity analysis. Results: Within 3,251 MPTs, the nonreactive (NR) ULT mean S/CO was 0.08 ± 0. 06 

J Dockter, C Militar, A Umali, C Giachetti, Gen-Probe Incorporated, San Diego, CA Background: The Procleix® Ultrio TM Assay (Ultrio) is a blood-screening assay based on Transcription-Mediated Amplification that simultaneously detects HIV-1, HCV and HBV. The HIV-1, HCV and HBV Discriminatory Assays contain specific probe reagents and are designed to discriminate the three viruses in Ultrio reactive specimens. In this study, the analytical and subtype sensitivities of the four assays were determined in the Procleix System (Procleix) and in the fully-automated TIGRIS® System (TIGRIS). Methods: Serial dilution panels of an HIV-1 virus stock as well as the HIV-1, HCV and HBV WHO International Standards were prepared and tested in the Ultrio and appropriate discriminatory assays using both the Procleix and TIGRIS Systems. Dilutions of HIV-1, HCV and HBV genetic variants of known concentrations were prepared and tested in the four assays on each With a stringent thermo-cycling program, amplification & kinetic detection is reduced to 2 h. An optimized signal threshold detection algorithm is being used for accurate +/-result differentiation. Full process controls are constructed using armored RNA (3 HIV targets/HCV) & protected DNA (HBV) to monitor assay performance across the entire automated process, from initial sample preparation to amplification & concurrent fluorescence signal acquisition. Results Analytical sensitivity (determined via PROBIT 95% CI) was found to be approximately 15 cp/mL (26 IU/mL) for HIV-1M, 18 cp/mL (6 IU/mL) for HCV, 12 cp/mL (2 IU/ml) for HBV, & <50 cp/ml for HIV-1-O & HIV-2. Inclusivity studies indicate comprehensive subtype/genotype recognition (HIV-1M/O, gt A -H, O tested, HCV gt 1-6 tested, HBV gt A-G & precore mutants tested). Specificity was significantly improved (~0.1% false positives). Potentially interfering chemical substances (e.g. peak pathological levels of h-BG-DNA, HSA, heparin, bilirubin, lipid) are fairly well toler-system using three reagent lots. Results: Probit analyses of the analytical sensitivity results indicate that the predicted 95% detection level for the HIV-1 virus stock was approximately 38 copies/mL and for the HIV-1, HCV and HBV WHO International Standards were approximately 18, 4 and 8 IU/mL, respectively, in the Procleix System. In the TIGRIS System, the predicted 95% detection level for the HIV-1 virus stock was 29 copies/mL and for the HIV-1, HCV and HBV WHO standards approximately 20, 3, and 10 IU/mL for each respective virus. In the Ultrio assay, 60 unique specimens infected with HCV genetic variants (types 1-6) were detected at 300 copies/mL in both systems. Forty-eight unique HIV-1 specimens [groups M (A-G), N and O] and 57 unique HBV specimens (types A-G) were detected at £300 copies/mL in each system. Comparable performance was observed in the discriminatory assays with similar results obtained across reagent lots. Conclusion: These results indicate that the Ultrio and HIV-1, HCV and HBV Discriminatory Assays are very sensitive for the detection of HIV-1, HCV and HBV genetic variants in the Procleix System and in the TIGRIS System.

P Estes, K Brookes, S Wang, M K McCormick, C Giachetti, Gen-Probe Incorporated, San Diego, CA Background: The Procleix ® Ultrio TM (Ultrio) Assay and associated Procleix Ultrio HIV-1, HCV, and HBV Discriminatory (dHIV-1, dHCV, and dHBV) Assays are amplified nucleic acid tests (NATs) designed for the simultaneous detection of human immunodeficiency virus type 1 (HIV-1) and hepatitis C virus (HCV) RNA, and hepatitis B virus (HBV) DNA in routine blood screening. As part of an on-going clinical trial, the current study was conducted to determine the Ultrio, dHIV-1, dHCV, and dHBV assays' early detection capability on the fully-automated TIGRIS ® System (TIGRIS) in commercially-available seroconversion panels containing HIV-1, HCV, and HBV. Methods: Thirty seroconversion panels (10 panels of each virus) collected from plasmapheresis donors were tested in the Ultrio Assay (neat and 1 : 16 diluted) and in the dHIV-1, dHCV, and dHBV Assays (neat only). The test results were compared to antigen or antibody tests as follows: Abbott HIVAG TM -1 Monoclonal Blocking Antibody and Abbott PRISM ® Anti-HIV1/2 for HIV seroconversion panels, Ortho Anti-HCV 3.0 Antibody or Ortho ELISA Anti-HCV 3.0 Antibody for HCV seroconversion panels, and Abbott PRISM HBsAg for HBV seroconversion panels. Median days to detection versus antibody or antigen tests were calculated for each of the four assays.

The results show that the Ultrio Assay on TIGRIS detected HIV-1, HCV, and HBV earlier than antibody/antigen tests in panels tested neat and diluted. Similar results were obtained from panels tested neat in the dHIV-1, dHCV, and dHBV Assays.

Target-Capture PCR assays were developed at Chiron for HBV, HCV and WNV with respective internal controls using oligonucleotide based magnetic bead capture, followed by TaqMan assay for quantitation based on WHO standards. The HIV-1 RNA was tested by ARUP laboratories. The Quantitative Target-Capture PCR assays have also been qualified through blinded studies for HBV (QCMD panel) and WNV (BCP panel). Results: The BBI samples for all four viruses, CBER sample for HCV, VQA sample for HIV-1, AcroMetrix samples for HCV and HBV were in the anticipated range. The VQC samples for all viruses including HCV and HBV are overquantitated by the vendor. The BCP/Impath WNV sample was underquantitated 36 fold by the vendor. Conclusion: The above study illustrates significant variations in vendor claims versus WHO Standards. Therefore we have created vendor and virus specific conversion factors to WHO units. Objective: Applying nucleic acid tests (NAT) for organ and tissue donor screening may reduce risk of the transmission of HBV, HCV and HIV-1 viruses to recipients. The COBAS ® AmpliScreen (CAS) HBV, HCV and HIV-1 Tests, which are PCR-based assays for blood screening, were further evaluated for use in cadaveric screening. Materials and Methods: Due to a relatively high inhibition rate, a test algorithm was designed specifically for use of CAS assays for cadaveric specimens. Specimens were diluted fivefold and tested using the Multiprep Specimen Diluent (MP DIL) and using the Multiprep Specimen Processing Procedure (MP). If inhibition occurs, the inhibited specimen is retested in duplicate. 30 HBV, HCV and HIV-1 seronegative post-mortem specimens were tested, with and without spiked HBV, HCV and HIV-1 targets at a concentration near the LOD level of each CAS Test, to evaluate the CAS assays' sensitivity and specificity, respectively. For this study, all specimens, inhibited or not, were retested in duplicate following the initial test. Results: The initial inhibition rate of the test is 3.3%, 10.0%, and 6.7% in the sensitivity study and 0.0%, 5.0% and 6.7% in the specificity study for HBV, HCV and HIV-1 assays, respectively. After retesting in duplicate, the final resolved inhibition rate is 0.0%, 3.4% and 3.4% in the sensitivity study and 0.0%, 5.0% and 1.8% in the specificity study for HBV, HCV and HIV-1 assays, respectively. Based on the preliminary data, sensitivity is 100%, 89% and 96.4% for HBV, HCV and HIV-1 assays, respectively, while the specificity is 100% for all three assays. Conclusions: With the proposed testing algorithm, the CAS Tests may provide sensitive and specific NAT assays to screen HBV, HCV and HIV-1 infections in organ and tissue donors.

S Anderson, M J Bankowski, S Belzer, B Lembke, ViroMed, Minnetonka, MN Background: The use of HIV-1 and HCV nucleic acid testing (NAT) by gene amplification has resulted in the detection of viral RNA in seronegative plasma from transfusion-associated whole blood and blood components. The "window" period between infection and antibody production varies from about 22 days for HIV-1 to as long as 50 days or more for HCV. In the blood donor community, NAT has contributed successfully to the reduction of this window and the prevention of HIV-1 and HCV transfusion-associated infections. The tissue donor community has a similar concern over insuring that the transplanted tissue is both safe and free of both HIV-1 and HCV. Methods: We have performed a verification of the COBAS AmpliScreen TM HIV-1 and HCV test kits for the screening of both infectious agents in plasma from cadaveric blood specimens. Performance characteristics for both assays were very agreeable. Quantitative standards for both HIV-1 and HCV were spiked into cadaveric specimens and multiple replicates of serial dilutions were tested for the limit of detection. Results: The analytical sensitivity for HIV-1 and HCV were 20-50 copies/ml and 20-50 IU/ml respectively. The analytical specificity was 100% using such agents as HTLV-I, HTLV-II, CMV, EBV and HSV. The clinical sensitivity and specificity consisted of testing previously positive and negative HIV-1 and HCV cadaveric specimens. Both clinical sensitivity and specificity were 100%. Accuracy was also 100% for both HIV-1 and HCV. Clinical precision for HIV-1 and HCV were 99.5% and 98.6% respectively. Prospectively, over 1000 specimens have N/A = not applicable.

Conclusion: These results suggest that the Ultrio, dHIV-1, dHCV, and dHBV Assays on TIGRIS are highly sensitive NATs that can aid in preventing the transmission of HIV-1, HCV, and HBV infection through blood components. The implementation of fully-automated NATs such as the Ultrio Assay on TIGRIS will further reduce the window for HIV-1, HCV, and HBV detection in blood screening.

V Shyamala, J Cottrell, D Madriaga, P Arcangel, M Hartman, K Gasuad, D Chien, B Phelps, Chiron Corporation, Emeryville, CA Background: Since 1997, WHO has had international standards for nucleic acid technology (NAT) assays. These are primarily used to develop secondary reagents for routine use. Several vendors market such secondary reagents. Claims of analytical sensitivity by NAT assay developers are based on the vendor provided information. Here we evaluate the claims of the vendors using the quantitative Target-Capture PCR assay. Methods: Commercially available materials were obtained from VQC, BBI, AcroMetrix, CBER, VQA and BCP/Impath, for HCV, HBV, WNV and HIV-1. Quantitative been tested with a positive rate for HCV at 1.2% and 0% for HIV-1. Conclusions: These results are very encouraging for the use of the COBAS AmpliScreen TM HIV-1 and HCV test kits for screening cadaveric blood from tissue donors.

R Garcia, L Walashek, T Dettori, C Giachetti, J Dockter, Gen-Probe Incorporated, San Diego, CA Background: The Procleix® Ultrio TM (Ultrio) Assay simultaneously detects HIV-1, HCV and HBV in a Transcription-Mediated Amplification based assay. The HIV-1, HCV and HBV discriminatory assays, which contain specific probes to each virus, are used to discriminate Ultrio reactive specimens. The assays are designed to be run in the semi-automated Procleix System and the fully-automated TIGRIS® System. In this study, the specificities and sensitivities of the Ultrio and discriminatory assays were evaluated in cadaveric specimens from tissue donors tested in both systems. Methods: Cadaveric serum specimens were collected up to 24 hours post-mortem from HIV, HCV and HBV serological negative donors. Control specimens from normal blood donors were included in the study. A set of specimens (cadaveric and controls) was tested in the Ultrio Assay and in each discriminatory assay to determine specificity. To evaluate assay sensitivity, three sets of specimens were spiked with low levels of HIV-1, HCV or HBV and tested in the Ultrio Assay and in the appropriate discriminatory assay. .4%] were observed with cadaveric and control HBV spiked specimens, respectively. Conclusion: No significant differences were observed in assay specificities and sensitivities between the cadaveric and control specimens tested on both instrument platforms. These results show that the performance of the Ultrio and discriminatory assays is not compromised in cadaveric specimens tested in the TIGRIS and Procleix systems.

Francisco, CA; C Nass, American Red Cross Blood Services, Baltimore, MD; B Wang, K Loughlin, D Smith, Westat, Rockville, MD Background. High HTLV-I proviral load (VL) has been associated with an increased risk of HTLV associated myelopathy (HAM), but little is known about variation in HTLV-I or -II VL by demographic characteristics and risk behaviors. Methods. We measured HTLV-I and HTLV-II VL in a large cohort of 127 HTLV-I and 328 HTLV-II seropositive former blood donors using real time PCR using tax primers. Multivariable linear regression was used to control for confounding by relevant covariates. Results. Mean log 10 VL was 3.28 per 10 6 PBMC (range 0.5-5.3) for HTLV-I 2.60 per 10 6 PBMC (range 0.05-5.95) for HTLV-II (p < 0.0001). HTLV-II VL was higher in those with subtype IIA (mean = 2.82 log 10 copies per 10 6 PBMC) compared to subtype IIB (mean = 2.29 log 10 copies per 10 6 PBMC)(p = 0.005). Higher HTLV-I VL was associated with previous receipt of a blood transfusion (p = 0.04), and lower HTLV-II VL was associated with female gender (p = 0.007). These associations persisted in virus-specific multivariate linear regression models controlling for potential confounding variables. Conclusion. VL was significantly higher in HTLV-I than in HTLV-II, and in HTLV subtype IIA than in subtype IIB infection. Chronic HTLV VL may be related to infectious dose acquired at the time of infection, with higher VL following blood transfusion and lower VL following sexual acquisition.

K Land, P Davenport, Carter Bloodcare, Bedford, TX; F Nizzi, L Sutor, University of Texas Southwestern Medical Center, Dallas, TX; M Sayers, Carter Bloodcare, Bedford, TX Background: In order to understand risk factors of HTLV I/II infection in blood donors, we reviewed our experience for the six year period since implementation of HTLV I/II screening in 1997. Methods: Enzyme Linked Immunoassay (EIA) was used for screening. Western Blot, Immuno-Fluorescent Assay (IFA) and Recombinant Immuno-Precipitate Assay (RIPA) were used for supplemental testing for donor counseling purposes (one or more were used during various periods of the 6 year time). When possible, donors with positive supplemental tests were interviewed and counseled in person. Results: Eighty-four donors out of 1,557,955 total donations between 1997 and 2003 (0.0054%) were reactive in screening tests and at least one or more supplemental tests. Thirteen donors were excluded from the study because of inconclusive results (i.e., RIPA non-reactive, or repeat antibody testing non-reactive), leaving 71 donors in the study, 20 of these were autologous donors. Thirty-five of the 71 (49%) consented to a personal interview and counseling, with only 4 of the 20 (20%) autologous donors consenting. Two of the 35 interviewed donors conceded a risk factor (IV drug use) not revealed at time of donation. Both of these donors were also anti-HCV and HCV NAT reactive. Twelve of the 35 (34%) had a history of STD, multiple sex partners (self assessed), or former sex partner with history of IV drug use, but were eligible to donate under our current eligibility criteria. Five (14%) had a history of transfusion prior to 1990. Additional positive infectious disease markers were present in 20 (28%), either HBV, HCV, syphilis, or combinations of these. Fifty-five of the 71 (77.5%) were female. Thirtyseven of the 71 (52%) were interpreted as HTLV type II, 24 of 71 (34%) as type I, and the remainder were non-specific. Conclusion: Half of the group of donors found to be positive for HTLV I/II do not agree to a personal interview and counseling. The majority of counseled donors do not reveal a risk factor during the interview. The majority of positive donors are women. Background. In sub-Saharan Africa, the percentage of screened blood is limited to approximately 75 for anti-HIV, 50 for HBsAg and 19 for anti-HCV, mainly due to the cost of tests and to the impracticality of microtiter plate enzyme immunoassays (EIA). Screening with high performance rapid tests is better adapted to small daily number of units collected. The cumulative prevalence of HIV, HCV and HBV infection ranges between 10 and 25% and pre-donation screening of candidate blood donors may reduce costs by saving blood bags and permit subsequent donor care where communication with donors is difficult. Determining the efficacy of this novel approach was the purpose of this study. Design and methods. In a 12-month period of 2002-2003, candidate volunteer and replacement blood donors in a regional teaching hospital in Ghana, West Africa were screened pre-donation for viral markers with rapid tests. The efficacy of screening was assessed by simultaneous amplification of HBV, HIV and HCV genomes (NAT) in a Triplex assay applied to pools of 10 seronegative plasmas. Positive and some negative pools were resolved. Deferred candidate donors were referred to a donor care program. Results. 9,372 people were screened and 1,563 (16.7%) were deferred. In collected units, no HIV or HCV RNA-containing samples remained undetected by rapid tests unless a human testing error (1 HIV, 11 HCV, 8 HBV) was involved. In contrast, 1.3 and 3.0% of HBV DNA-containing blood units were negative with rapid tests but were detected in individual donations with EIA and/or NAT, respectively. Half of these RNA+ units were detected in pools of 10 (median 77 IU/ml) or in individual samples (median 35 IU/ml). One third of deferred candidate donors attended the donor care program and received information, clinical examination and ALT testing, counseling and, when appropriate, were referred to hospital physicians. Conclusions. In high endemic areas, pre-donation screening for HIV and HCV but not for HBV markers is effective when properly performed, limits the cost of blood and contributes to public health. HBsAg rapid tests or EIA as well as NAT in pools of 10 samples failed to detect some HBV infectious units. 

BACKGROUND: South Africa has one of the highest HIV prevalence rates in the world. Our transfusion service is situated in the Western Cape Province, which has the lowest prevalence rate in South Africa. However, in the past 3 years the rate of HIV infection has risen from 8.6% to 15%. In contrast the HIV rate for blood donors is significantly lower at 0.03% of all donations. This study examines the role of the Confidential Exclusion Questionnaire (CEQ) and personal interviews in maintaining this low HIV seropositivity rate amongst donors. METHOD: In 2001 the CEQ was revised to make the questions more direct, user friendly and understandable to our multilingual society. High risk clinics were identified, and personal interviews introduced. Tertiary Education Institutions, Secondary Schools and our walk-in clinic were identified as higher risk than stable residential or community clinics. We analyzed our HIV positivity rates prior to the introduction of donor interviews and compared them with the post interview rates. RESULTS:

HIV p24 antigen positive (VL = 599,000 copies/ml) but still negative for antibody. The day 10 sample was again HIV p24 antigen positive (VL = 231,000 copies/ml) but now it was antibody positive by EIA. The donor was oriented to seek medical advice upon confirmation of the acute HIV infection. Conclusions: We have documented the first Brazilian case of NAT interdiction of a blood donation at the time when donor was HIV infected during the antigen and antibody window period. Donor was followed until seroconversion and therefore proved to be HIV infected. Until a definite algorithm is developed to confirm NAT positive donations diagnosis of HIV infection should only be demonstrated by donor seroconversion. only antigenemia had no other marker, that means that the surrogate tests are not enough to improve safe transfusions. Although a larger population needs to be studied to validate these results, we consider p24Ag testing as a valuable alternative until the implementation of Nucleic Acid Testing and the development of an altruistic BD program could be possible.

Background: Heterophilic antibodies found in normal donors are known to result in non-specific reactivity in screening tests to detect transmissible disease markers that are based on ELISA method. These antibodies may cause a spurious elevation of the measured value that is independent of the true analyte concentration, resulting in biological false positive interpretations of the testing. Method: blood samples with non-specific reactivity for anti-HIV 1,2 (i.e. non-reactive by IFA and NAT) or anti-HTLV-I/II (i.e. nonreactive by alternate ELISA and LIA) were treated using the Non-Specific Antibody Tube (NABT) to remove heterophilic antibodies, and retested using the ELISA reagents from the same manufacturer that gave the biological false positive results. An untreated sample was tested in the same run as a control. Results: Of 39 samples that showed false reactivity for anti-HIV 1,2, the non-specific results were eliminated in 10 (25%) samples. For the 42 samples with false positive reactivity for anti-HTLV-I/II, heterophile antibody blocking did remove non-specific reactivity in only 5 (12%). Conclusion: While heterophile antibodies are responsible for one fourth of the samples that have non-specific reactivity for anti-HIV 1,2, non-specificity due to heterophile antibodies is observed in a significantly lower proportion of the samples tested for anti-HTLV-I/II by ELISA method.

Despite the HIV rate increasing rapidly in the region, the HIV seropositivity rate of blood donation dropped. At the same time the pre-donation deferral rate of potentially at risk donors rose rapidly. CONCLUSION: Personal interviews and a carefully designed CEQ have proved to be very effective in lowering the number of HIV positive donations and at the same time effectively screening out high-risk donors. It is recommended that personal interviews should be used in all high prevalence areas. Background: Nucleic acid testing (NAT) for HIV/HCV has not become mandatory for blood donor screening in Brazil. However, because NAT significantly decrease the HIV and HCV window period we have implemented NAT testing in our blood banks. We report a case of HIV infection in a donor with a p24 antigen and antibody negative results but positive on HIV-NAT. Methods: Since May 2003 every blood donor has been routinely screened by HIV/HCV NAT by the transcription mediated amplification (TMA) technology (Procleix, Chiron). Positive samples were submitted to a discriminatory NAT assay to unravel the infectious agent (HIV or HCV). In the present case, the donor was invited to return for counseling and to collect follow up samples until complete seroconversion was documented, as judged by western blot. Viral load (VL) was retrospectively evaluated by Amplicor HIV-1 Monitor V 1.5. Results: The donation sample (day zero) was positive on NAT with a VL of 24,100 copies/ml but negative on HIV p24 antigen and antibody (3 different EIA manufacturers). Five days later the donor has returned for counseling and admitted a single episode of an active and unprotected anal intercourse with another man (who later was found to be HIV positive), 24 days before the blood donation. The day 5 sample was TRANSFUSION 2004-Vol. 44, Supplement 

J Hu, S Lee, S Kerby, O Wood, A Machuca, S Daniel, M Rios, CBER, Rockville, MD; A Bih, L Zekeng, Cameroon Ministry of Health, Cameroon, MD; I K Hewlett, CBER, Rockville, MD Background: Several serologic assays have been licensed for detection of antibodies against HIV in blood donors. Two NAT assays have also recently been licensed for early detection of HIV in blood donor samples. However, the sensitivity of these assays for detection of HIV variants has not been fully established. We evaluated a number of commercially available assays for their ability to detect HIV in samples from Cameroon where a number of diverse HIV strains are known to exist. Methods: 239 plasma samples from two blood donation sites in Cameroon were tested in this study. Of these 149 were determined to be positive and 90 negative based on a rapid antibody test in routine use in the country. In our study we used 2 qualitative NAT assays, 1 quantitative NAT assay, an HIV-1 p24 antigen assay, 3 licensed antibody assays and an IFA. All reactive samples were further tested by a Western blot assay. Results: Of the 149 samples previously determined to be positive,131 were confirmed for their HIV status by all assays. 4 samples were negative by all methods and determined to be negative. 14 were discordant among the various EIA and NAT assays. 5/14 were confirmed as positive based on Western blot, 4 had indeterminate band patterns, and 5 were negative by WBlot and NAT. 2/90 negative samples were positive by 2 EIAs and 1 NAT assay and confirmed by Western blot. 6 additional samples were reactive on at least one of the EIA tests but only one had an indeterminate pattern, all others being negative by WBlot. All samples except one were negative by the HIV-1 p24 antigen assay. Conclusions: Our results indicate that although HIV assays currently in use in the US are highly sensitive for detection of viral variants, some samples may not be detected by all assays. The results suggest that surveillance studies to determine test performance with viral variants in areas where they are prevalent are a useful means to the impact of viral diversity and evolution on assay performance. Genetic sequencing is being performed on plasma viral RNA isolated from the discordant samples to determine if they could represent new variants.

J A Rios, R J Benjamin, New England Region, American Red Cross, Dedham, MA BACKGROUND: Blood center and hospital staff may be exposed to HIV and hepatitis via needlesticks and splashes from donor blood. The management of employees with blood exposures should be based on local data regarding the likelihood of disease transmission from the blood of volunteer donors. METHODS: The risks for the transmission of HIV, Hepatitis B, and Hepatitis C for employees with blood exposures was estimated by adding the estimated number of window period units plus the number of units presumed to be viremic as evidenced positive by nucleic acid testing (NAT) using data from blood collections during calendar year 2002. It was assumed that all NAT positive units could possibly transmit a virus after blood exposure. Since blood donors are not tested with NAT for Hepatitis B it was assumed that all Hepatitis B Surface antigen positive units were viremic. Autologous donors were excluded. RESULTS: In 2002 our blood center collected 334,042 whole blood and apheresis donations (primarily single donor platelets). Based on antibody or antigen testing, 3 donors were HIV confirmed positive (all RNA positive); 141 were Hepatitis C confirmed positive (105 RNA positive); and 39 were Hepatitis B Surface antigen confirmed positive. No donor was antibody negative and NAT positive.

CONCLUSION: Healthcare workers exposed to volunteer blood have a very low risk of becoming infected with HIV and hepatitis. The low risk of disease transmission from volunteer donor blood vs patient blood is attributable to the screening process required for blood donations. All healthcare workers with potential exposure to volunteer blood should be offered the Hepatitis B vaccine. Because the risk of infection after an exposure with volunteer blood is extremely low, antiretroviral chemoprophylaxis against HIV is not recommended by the CDC. Hepatitis B infection remains the highest risk, however prophylaxis with the Hepatitis B immunoglobulin is not needed in employees with immunity through vaccination.

J Hu, S Lee, O Wood, R Taffs, P Akolkar, A Machuca, I K Hewlett, CBER/FDA, Rockville, MD Background: HIV is a highly diverse virus consisting of many different subtypes and circulating recombinant forms. Current NAT assays are capable of detecting most subtypes although with varying degrees of accuracy. Therefore well-characterized NAT standards are needed to ensure sensitivity of tests for subtype detection. Methods: Virus isolates of subypes A-G of group M, group O and group N were cultured in PBMCs to high titers. HIV subtype B was tested in-house using assays based on 3 different NAT methodologies. A dilution series of the isolate based on consensus values was sent to 15 participating laboratories to verify copy number assignment. Virus stocks of non B subtypes were sent to 5 different laboratories for copy number determination. Results: The results of the collaborative study indicated that most participating labs were able to detect and quantitate the panel members of the HIV subtype B panel with a high degree of accuracy. There was good agreement between the various laboratories for copy number determinations of non B subtypes. Conclusions: The HIV subtype panel will be useful for standardization of NAT assays and serve as reference materials to harmonize NAT in the international setting Background: Screening for HIV-1 in Source Plasma donations by NAT or p24 antigen testing is required in the US. Studies were performed to support a plasma indication for the Roche COBAS AmpliScreen TM HIV-1 Test v1.5 in 96-donation minipools (previously licensed for whole blood donations in pools of 24). Materials and Methods: The following studies were performed: Seroconversion study (10 seroconversion panels), yield study (20 Agpositive, Ab-negative yield samples), positive pool deconstruction (ten 96sample minipools, 1-10 positives each), and a clinical specificity study (>100,000 plasma donations). Results: For all seroconversion panels, HIV-1 RNA was detected at 1 : 96 concurrent with or prior to HIV-1 p24 Ag and prior to HIV-1 Ab. All 20 yield samples were positive when diluted 1 : 96. Fifty of 51 positives in the deconstruction study were correctly identified. The undetected, spiked sample was derived from a labile clinical specimen containing significant precipitate, which may have limited the ability to extract sufficient RNA for detection. The clinical study comprised 104,448 donations representing 35,905 applicant and qualified donors. Three PCR positive donations from 3 unique donors were identified. Two were p24 Ag-/Ab-and one was p24 Ag-positive only. HIV-1 positive status was confirmed by licensed HIV-1 assays. Conclusions: The COBAS AmpliScreen TM HIV-1 test was highly sensitive and specific to HIV-1 in 96-sample minipools, while meeting the throughput and performance needs for HIV-1 screening of source plasma donors.

Background: The new PRISM HIV Combination assay is a chemiluminescent immunoassay for simultaneous detection of anti-HIV-1/HIV-2 IgG/IgM and HIV p24 antigen, which was designed to further close the seroconversion window compared to assays detecting HIV antibodies exclusively. Sensitivity and specificity of PRISM HIV Combination assay were determined and compared to other licensed HIV antibody or HIV Ag/Ab Combination assays. Method: The PRISM HIV Combination assay was evaluated in a multi-center study. Specificity was investigated on blood donor specimens. Antigen quantitative sensitivity was determined by evaluation of the French Antigen-AFSSAPS panel. Enhanced antibody sensitivity was evaluated on selected HIV-1 and HIV-2 positive specimens, HIV-1 group O specimens, and commercially available seroconversion panels. Results: Apparent specificity of 99.95% was observed on blood donor specimens, respectively, with IR and RR rates of 0.19% and 0.05%. Antigen quantitative sensitivity against the AFSSAPS subtype B p24 quantitative antigen panel averaged 33.44 pg/ml. Compared to the competitor HIV Ag/Ab Combination assay, the PRISM HIV Combination assay showed superior seroconversion sensitivity. The PRISM HIV Combination assay detected all HIV-1 group M and O and HIV-2 antibody positive specimens. Conclusion: The newly developed PRISM HIV Combination assay showed excellent sensitivity, thereby reducing the risk of transfusion transmitted HIV compared to assays detecting HIV antibodies exclusively. Specificity was comparable to HIV assays designed for antibody detection only.

K Osada, Jikei University, School of Medicine, Tokyo, Japan Background: Although leukocyte enzymatic activities are important for nonspecific protection against bacterial infections in HIV-infected patients, traditional methods for the detection of intracellular enzymatic activities require complex assays. Very little information is available on the intracellular enzymatic activities on the intracellular proteasic activities of phagocytes in HIVinfected patients. We used flow cytometriy to investigate the intracellular enzymatic activities of proteases: elastase, collagenase and cathepsin D, using CellProbeTM fluorogenic substrates in unfractionated viable peripheral blood leukocytes from HIV patients and healthy controls. Methods: Blood samples were collected from ten HIV-infected patients and ten healthy controls. Three CellProbeTM reagents, elastase, cathepsin D and collagenase (Beckman Coulter) were used in this study. Each of these reagents comprises synthetic, non-fluorescent enzyme substrates of two leaving groups conjugated to a dye molecule (Rhodamine 110). Conjugation of the leaving groups to the dye quenches the dye's fluorescence (CellProbeTM). When the enzyme cleaves the bonds between the leaving groups and dye, the fluorescence is released. The samples were analyzed by flow cytometry. Results: All three enzymatic activities studied were significantly higher in monocytes than in polymorphonuclear cells (PMNs) and lymphocytes independent of the group considered. Collagenase avtivity yielded the strongest fluorescence intensity. Comparison of HIV-infected patients and controls showed that the elastase activity of PMNs and lymphocytes was higher in HIV-infected patients than in controls, but was not different in monocytes. Collagenase activity was significantly lower in monocytes of HIV-infected patients but it was unchanged in lymphocytes and PMNs. Cathepsin D activity was significantly higher in PMNs and lower in monocytes from HIV-infected patients compared to controls. In HIV-infected patients, Cathepsin D and elastase activities were correlated in lymphocytes, PMNs and monocytes. In monocytes, there was also a strong correlation between collagenase and Cathepsin D activities. In PMN, collagenase and Cathepsin D activities were also correlated. There was no correlation between the enzymatic activities and CD4 positive cell count or viral load in HIV-patients. Conclusion: Significant variations were observed in the peripheral blood of HIV-infected patients and different patterns were especially evident in monocytes and PMNs. The differences noted in HIV patients com-pared to controls may indicate previously unreported alterations in the innate immunity of such patients.

N Boschetti, I Niederhauser, C Kempf, ZLB Bioplasma AG, Bern, Switzerland; A Stühler, J Löwer, J Blümel, Paul-Ehrlich-Institute, Langen, Germany Background: Parvoviridae are highly resistant, small, non-enveloped viruses. Since low pH is frequently applied to manufacturing process intermediates or final products, we assessed the impact of such conditions on the human erythrovirus B19 (B19V) and the mouse parvovirus (MVM) which is frequently used as a model for B19V. Methods: Virus inactivation was monitored by decrease of infectivity and loss of capsid integrity. Infectious B19V was quantified by detection of virus-specific mRNA from Ku812Ep6 cells. To measure capsid integrity, endonucleases were added after exposure to low pH and the encapsidated virus DNA was quantified by real-time PCR. Results: B19V was inactivated >5 log after 2 h at pH 4, whereas MVM was resistant over 9 hours. Infectivity data strongly correlated with data obtained by the endonuclease assay. Capsid disintegration was observed in IgG-as well as in different albumin-solutions. Temperature and pH showed concerted impact on B19V capsid disintegration. Conclusion: Our data show, that B19V is much more vulnerable towards low pH conditions than MVM. Together with the previously reported susceptibility of B19V towards wet heat conditions this is the second report showing, that the erythrovirus B19V is less resistant to physico-chemical treatment than viruses from the parvovirus genus. Since virus-validation studies have to cover a broad range of viruses, the use of animal parvoviruses as model for small robust viruses is still appropriate. Background: The plasma industry and the US Food and Drug Administration have recognized that Parvovirus B19 and Hepatitis A virus (HAV) are important targets for the plasma industry as in-process controls. To meet the increasing requirements for high quality testing, novel NAT TaqMan assays have been developed for qualitative detection of Parvovirus B19 DNA and Hepatitis A virus RNA in human plasma. The assays are designed to run on the COBAS ® AmpliPrep and COBAS ® TaqMan instruments, which fully automate generic nucleic acid isolation and kinetic PCR. Methods: Total nucleic acids were extracted from human plasma using a generic extraction technology on the COBAS ® AmpliPrep instrument. The qualitative detection of Parvovirus B19 DNA and Hepatitis A Virus RNA occurs via kinetic PCR using rationally designed primers and probes along with appropriate amplification mix on the COBAS ® TaqMan analyzer, either as single target or duplex assays. The workflow of these assays will also apply to an integrated next generation blood screening analyzer with increased throughput. Results: The assay performance for Parvovirus B19 or HAV was similar for both the single or duplex target format. Sensitivity for Parvovirus B19 was approximately 20 IU/mL (traceable to the WHO International Standard NIBSC Code 99/800). Precision was shown to be <1% at 1E4 IU/mL (based on the threshold cycle value) for Parvovirus B19. The HAV assay demonstrated a sensitivity of approximately 3 geq/mL (traceable to the HAV RNA Working Reagent NIBSC Code 01/488). Precision was shown to be <1.6% at 1E5 geq/mL (based on the threshold cycle value) for the HAV assay. Specificity was 100% for both parameters using routine blood donations provided as EDTA plasma. Conclusions: The feasibility of generic nucleic acid isolation and newly designed TaqMan assays for Parvovirus B19 DNA and Hepatitis A virus RNA has been demonstrated and may prove to be a valuable tool for in-process testing. 

L Qu, Institute for Transfusion Medicine, Pittsburgh, PA; D Rowe, University of Pittsburgh Graduate School of Public Health, Pittsburgh, PA; D Triulzi, Institute for Transfusion Medicine, Pittsburgh, PA BACKGROUND: EBV infects and establishes latency in B lymphocytes of the majority of adult population, including blood donors. The effect of RBC storage on the stability of EBV viral genome is not known. METHODS: Four units of randomly selected AS-5 RBC were divided into 7 equal aliquots using sterile docking devices and stored refrigerated at 1-6C until use. The EBV genomic copy numbers in CD19+ B lymphocytes were quantified at following time points: Fresh (48 to 96-hour), Day 7, Day 14, Day 21, Day 28, Day 35, and Day 42. Each sample was initially processed by density gradient centrifugation to isolate lymphocytes. B cells were further purified by CD-19 magnetic bead selection of the lymphocyte fraction. DNA extracted from B cells was assayed in duplicate for EBV genomic copy number by quantitative TaqMan® PCR. RESULTS: Three of the four units selected for the study were EBV positive by sensitive PCR. The number of EBV genomes detected in 10 5 CD19+ B lymphocytes ranged from 1.98 to 482.29 in the three units and remained fairly constant during the storage each of RBC. (Table) ble adverse effects in immuno-compromised individuals, children and the elderly. Though there is presently no documented case of transmission of vaccinia virus through blood transfusion from vaccinated donors, potential mass vaccination has raised some concerns on the safety of the Nation's blood supply due to the possibility of vaccinia viremia in donor blood that could be transmitted to susceptible individuals during blood transfusion. Hence this study was undertaken to explore the presence and duration of vaccinia viremia in healthy vaccinees as well as the potential interference of vaccination against smallpox on the performance of FDA approved test kits used to ensure blood safety. The results presented here could impact decisions regarding blood donor deferral of recent smallpox vaccinees. Methods: Attempts were made to culture vaccinia virus from the plasma of donor vaccinees enrolled in smallpox vaccine clinical trials. Our plasma viremia assay had a sensitivity of approximately 66 viruses/ml using a Vero cell culture assay. Virus detection is also being studied by real-time PCR. The possibility of false-positives due to cross reactivity resulting from vaccinations were evaluated with FDA-licensed serologic test kits used to screen blood for HIV, HCV, and HBV. Results: Six hundred and sixty five plasma samples were obtained from ninety-five subjects. Vaccinia viremia was not detected by from day 3 to day 56 after vaccination by virus culture. Very little cross-reactivity was observed with plasma samples from vaccinees with most of the test kits. Conclusions: This study establishes that in the case of mass vaccination, there would be negligible risk of transmission of vaccinia virus by transfusion of blood from healthy donors. No virus could be detected in the healthy donor blood under our assay conditions. Vaccination against smallpox did not interfere with tests used to screen blood donations in this study.

M Schmidt, M K Hourfar, E Seifried, W K Roth, German Red Cross, Frankfurt, Germany BACKGROUND: Biological weapons are very attractive for terrorism. As a result of international collaboration under the WHO eradication program, smallpox was declared eradicated in 1980. Therefore, the immunization programs were discontinued worldwide. Because most people are immunologically naive, variola virus is considered to be a potential threat agent or bioterrorist weapon. Real-time PCR for orthopoxviruses was developed for fast and safe analysis. Melting analysis following PCR enables the differentiation beween variola and other genotypes like vaccinia or camelpox virus. STUDY DESIGN AND METHODS: RealArtTM Orthopox LC PCR Kit was used to amplify Orthopox virus (OPV) sequences from blood donor samples. We tested 31,500 blood donor samples in mini pools of up to 96 samples. To validate the sensitivity of the assay, routine donor mini pools (90 ± 6 samples per pool) were spiked with vaccinia virus and used as positive controls. RESULTS: Specificity was 100% because none of 31,500 blood donors were positive for OPV. The detection limit of the assay was 1.41 input TCID50/PCR. Therefore we calculated a sensitivity of 211.5 TCID50/mL. Overall, 0.28% of test results had to be considered invalid due to negative internal controls. CONCLUSION: The real-time Orthopox PCR kit enables detection of OPV in viremic blood donor samples even in the beginning of the disease when patients present minor clinical symptoms and could be implemented in routine screening procedure immediately. Thus the assay could potentially help to prevent dispersion of viral agents by blood transfusion in case of bioterrorism. The initial discovery of TT virus (TTV) was followed by the identification of several TTV-like viruses (TTLVs) including TTV-like mini virus (TLMV), SEN virus (SENV), and PM virus (PMV). Initially, some of the viruses were reported to be associated with hepatitis of non A-G origin. To date, the link between these viral infections and liver disease remains uncertain. This uncertainty is partially due to sequence heterogeneity that results in incomplete detection of some genotypes by PCR. To establish a standardized detection system, multiple sets of primers were validated by testing large CONCLUSIONS: EBV genomic copy number per B cell remains fairly constant during RBC storage. This indicates that EBV latently infected quiescent B cell can survive the normal storage conditions for the RBC. The infectivity of EBV PCR positive RBC units must be assessed through clinical studies.

K E Nelson, Johns Hopkins University, Baltimore, MD; S C Dollard, Centers for Disease Control, Atlanta, GA; M J Cannon, Center for Disease Control, Atlanta, GA; P M Ness, V Stambolis, Johns Hopkins University, Baltimore, MD; P E Pellett, Cleveland Clinic Foundation/Department of Virology, Cleveland, OH Background: There is evidence that about 3.5% of blood donors in the U.S. are seropositive for HHV-8, the causative agent of Kaposi's sarcoma (KS) and that the virus may be transmitted by illicit injection drug use. However, the transmission of HHV-8 by blood transfusion in the U.S. has not been demonstrated. Methods: We examined HHV-8 antibody status in 406 cardiac surgery patients who were enrolled in 1986-1992 before and 6 months after surgery. Overall 82% of patients were transfused during surgery, 66% were men, 90% were white and their median age was 65. Results: Using a lytic immunofluorescence assay and a specimen dilution of 1 : 80, we identified 3 seroconverters with post surgery titers of 1:160-1:1280. The seroconverters were transfused with 12, 13 and 22 units of blood products, compared to a median of 3 units for all study patients. We found no seroconversions among 71 cardiac surgery patients who were not transfused. We do not have data from linked donors; however the rate of seroconversion is higher than would be expected in this population from community-acquired infection. If seroconversion was caused by transfusion, the risk was 0.7% per patient and 0.08% per unit. Conclusion: This is the first report of probable transfusion-transmitted HHV-8 in the United States. Because of the association between HHV-8 and KS, further evaluation of the risk is needed.

K Srinivasan, O Wood, S Lee, D Sylvester, S Kerby, K Francis, P Akolkar, R E Taffs, I K Hewlett, CBER/FDA, Rockville, MD Background: Vaccination with live vaccinia virus vaccine could result in several vaccine adverse events known earlier. The recently reported association of myopericarditis is also of particular concern, besides other possi-numbers (424) of thalassemia patients and healthy individuals for the known viruses. The gene sequences of over 400 cloned PCR products, amplified by the different primers, were compared with reference genotype sequences available in GenBank. The results indicate that 22 of 23 known genotypes of TTV, including SENV and PMV could be detected by two sets of primers (i.e. NG and TT primers). An additional set of primers, specific for genotype 21, is needed to detect all known genotypes of TTV and TTLVs including a new group of TLMV. Two sets of TLMV-specific primers plus the NG primers are needed for detection of all three known groups and a new group of TLMV. Thus, five sets of primers are required for efficient detection of all known genotypes and groups of TTV and TTLVs. We described a detection system for all TTV and TTLVs, thereby fulfilling an essential prerequisite for the proper assessment of the pathogenesis of these viruses.

Y Nie, Guangzhou Blood Center, Guangzhou, China; J X Wang, Institute of Blood Transfusion, Chinese Academy of Science, Chengdu, China; Y S Fu, C F Jiang, Guangzhou Blood Center, Guangzhou, China Background: Severe acute respiratory syndrome (SARS) is an acute infectious disease, which has been found to spread mainly via respiration in 30 countries and areas of the world. The first case was found in Guangdong, P.R.China, which was also referred to as acute infectious atypical pneumonia (AIAP). This disease was a severe and acute disease with strong infection. The high transmission rate from direct contact with infected individuals has lead to thousands of cases worldwide. The mortality rate is greater than 7%. SARS is caused by a novel cornonavirus, referred to as SARS-CoV. Genomic RNA from this virus has been detected in the serum of patients following the acute symptomatic phase of infection. There is still no effective medicine to cure it up to now. The main therapy is corticosteroids, antibiotics and mechanical ventilation. In China, in mainland and Hongkong district, it is proved to be effective for some severe SARS patients to transfuse plasma with high titer antibody to coronavirus donated by recovery patients from SARS. In Guangzhou, 26 voluntary recovery patients from SARS were involved to donate their plasma, The plasma were screened in HBsAg, anti-HCV, anti-HIV, syphilis test, ALT and antibody to coronavirus as well. In order to assess the safety of the donated plasma by the voluntary recovery patients from SARS, we tested the coronavirus RNA in the plasma by real time PCR. Methods: The samples were collected and treated in 1 hour immediately after the plasma was donated. Tow different reagents by two different reagent companies were involved to detect the coronavirus RNA in the donated plasma respectively. The coronavirus RNA were tested by fluorescence real time polymerase chain reaction. Genomic RNA in the plasma was extracted, adverse transcript and polymerase chain reaction followed. Polymerase chain reaction system included primer, TaqE, dNTP and buffer. The total reaction volume is 50 ml. Three steps were involved in the reaction parameters. It is 93°C for 2 minutes, 1 cycle in the first step; 93°C for 15 seconds, 55°C for 15 seconds and 72°C for 20 seconds, repeating 10 cycles in the second step; 93°C for 15 seconds, 58°C for 30 seconds, repeating 35 cycles in the third step. The internal control and the negative control were treated simultaneously. Result: No coronavirus RNA was detected in the 26 plasma samples from 26 voluntary, recovery patients from SARS, while simultaneously detecting the internal control and without detecting the negative control. Conclusion: The detection of the coronavirus RNA in the plasma from 26 voluntery, recovery patients from SARS showed that the donated plasma by the recovery patients of SARS show no coronavirus RNA. Chlamydia pneumoniae is an occasional cause of pneumonia, bronchitis, sinusitis, and pharyngitis. DNA of Chlamydia pneumoniae has been found in peripheral blood of 20-59% of heart disease patients in Sweden (1) and Germany (2), as well as in atherosclerotic plaques. In small studies, it has been reported in peripheral blood of 9-26% of U.S. blood donors (3,4), 13% of a German normal population (2), and 46% of Swedish blood donors (1). The present study was conducted to determine whether the reported prevalence of Chlamydia pneumoniae DNA in U.S. blood donors could be confirmed. Methods: Buffy coat DNA from 211 normal blood donors from the Washington, DC metropolitan area, Los Angeles, CA, San Francisco, CA, and New York, NY was assayed by touchdown PCR targeting the 16S ribosomal RNA of C. pneumoniae and real-time quantitative PCR targeting the C. pneumoniae pmp4 gene. Results: None of the 211 normal blood donors had detectable C. pneumoniae DNA in their buffy coat DNA when assayed with a limit of detection of 100 gEq/pg DNA. Validation tests on donor buffy coat DNA preparations, performed by spiking each of 10 samples with 1000 gEq/pg of C. pneumoniae DNA, were all positive. Conclusions: C. pneumoniae DNA could not be detected in any of 211 buffy coat samples from normal blood donors in the US. Since the concentration of the C. pneumoniae DNA in buffy coats may be lower than that in concentrated preparations of peripheral blood mononuclear cells, additional testing is in progress on purified monocytes from normal blood donors. 

Background: Blood transfusion is the second most common way of Chagas' disease in endemic countries. With migrations of people from affected areas to other countries, blood transmission may become a worldwide problem. The aim of this work was to study the reactivity with different techniques and to compare the sensitivity and specificity of a recombinant ELISA with three conventional screening assays. Materials and Methods: Samples: 78,342 consecutive blood donor samples (BDS) for seroprevalence, 775 negative consecutive BDS for specificity and 138 selected BDS for sensitivity. Conventional tests: Chagatek ELISA, Biomerieux, Argentina, Indirect Hemagglutination, Polychaco, Argentina, Particle Agglutination, Fujirebio, Japan. Due to the absence of a gold standard confirmatory technique, BDS that reacted in at least two tests were considered positive. The BDS were considered strong positive if they were reactive in ELISA with a donor sample optical desity (OD)/cut-off OD) >2 and one or two other reactive tests, weak positive if they were reactive in ELISA with a OD/cut-off OD <2 and other reactive test and false positive if they were reactive in only one test. As regards seroprevalence percentages we calculated the confidence interval at 95%. We considered p < 0.05 significant in all cases. Results: The seroprevalence in 78,342 BDS was 1.88% (SE: 0.048; CI:1.781-1.972, p < 0.05) and 1.34% (SE: 0.041; CI: 1.260-1.423, P < 0.05) were confirmed positive. It is shown in the table, how many BDS were strong positive, weak positive and false positive.

We selected 60 strong positive, 18 weak positive and 60 false positive samples. All positive and weak positive and 21 from the false positive samples were reactive with the evaluated test. The difference between the mean OD from ELISA Wiener recombinant (1.71, CI:1.57-1.84) and ELISA Biomerieux (0.71,CI:0.63-0.78) was significant:(1.0, p < 0.05). Specificity was 99.87%. Conclusion: The data observed when a low risk population such as blood donors is studied are sometimes inconclusive. We have obtained an important number of false positive results with ELISA reactive without other test. Our results show that the recombinant test has a high specificity and sensitivity, with higher OD values than the ELISA test we use. This test could be useful to confirm discordant results as well as a routine screening test. 

L H Tobler, L Pitina, M P Busch, Bloodsystems Research Institute, San Francisco, CA; P Contestable, H Groth, S R Lee, Ortho Clinical Diagnostics, Rochester, NY; G Blackburn, S Shaffer, Blood Systems Laboratory-Bedford, Bedford, TX BACKGROUND: Chagas disease imported into the US blood supply by blood donors born in South and Central America represents a serious blood safety problem. Chronically infected blood donors, can transmit the parasite Trypanosoma cruzi via blood transfusion. The FDA Blood Products Advisory Committee recommended implementation of Chagas antibody screening of U.S. donors as soon as a suitable assay is licensed. We conducted an anonymized pre-clinical study of a prototype T. cruzi lysate-based enzymelinked immuno-adsorbent assay (ELISA) developed by Ortho Clinical-Diagnostics. METHODS: Two populations of specimens were evaluated: 1) 9835 sequential donations from blood donors residing in the El Paso, Texas area; and 2) 180 specimens from South America which were presumptively positive for Chagas antibody and purchased from commercial vendors. All testing was performed using the Ortho Summit TM Sample Handling System and in accordance with the manufacturer's specifications and donor screening algorithms. RESULTS: 9832 (99.97%) out of the 9835-screened donor specimens were non-reactive, while 3 (0.03%) tested reactive. The 3 initially reactive specimens tested repeat reactive (3/3 reps) by ELISA, and confirmed by radio-immunoprecipitation analysis (RIPA, D. Leiby, American Red Cross). Based on Antibody Profile Analysis (California Department of Health Services, CA), 2 of the 3 Chagas positive specimens were from the same donor. Observed specificity of the test was 100%. The mean S/CO ratio of the non-reactive specimens was 0.050; while the mean S/CO ratio of the reactives was 4.60. Eight (0.08%) grey zone specimens (S/CO ratio ≥0.5 and <1.0) were identified among the 9832 non-reactive specimens; there was adequate volume in 6 of 8 specimens for RIPA confirmation, and all were RIPA negative. Of the specimens from S. America, 177 out of 180 (98.3%) were repeatedly reactive by the Chagas ELISA. RIPA testing of the three non-reactive specimens is pending. CONCLUSIONS: These studies indicate that the prototype Chagas ELISA has excellent sensitivity and specificity for detection of Chagas antibody in donors. Moreover, among donations from a geographically selected collection region of the US, observed seroprevalence was three in 9835 donations (0.03%).

P Contestable, H Groth, P Hosimer, E Grogan, S Hess, E Woodward, H Warren, S Lee, Ortho-Clinical Diagnostics, Rochester, NY Background: Indirect immunofluorescence (IFA) and enzyme-linked immunosorbent assays (ELISA) are currently two of the most frequently used serologic tests for T.cruzi. We have recently developed an ELISA to detect antibody to T. cruzi, suitable for screening blood donations in a low prevalence setting. We evaluated the performance of this ELISA in comparison to an IFA method carried out a commercial laboratory. Radioimmunoprecipitation assay (RIPA) was used to resolve all discrepant results between the ELISA and IFA methods. Methods: The Ortho screening assay for detection of antibody to T. cruzi uses a standard ELISA procedure carried out in a microtiter plate coated with T. cruzi antigens prepared from lysed epimastigotes. The cutoff value for the assay is calculated by the mean absorbance of the Positive Calibrator at 492 nm multiplied by a constant. In the IFA procedure, the appropriate sample dilution is overlaid on with the cell substrate and incubated at room temperature for the specified time. After the excess sample is removed, the cell substrate is incubated with the FITClabeled antibody to human immunoglobulin (anti-IgG) and after washing off excess FITC-labeled antibody, fluorescence is assessed under UV light. Observation of the expected fluorescent pattern at sample dilutions of 1 : 16 or greater indicates a positive antibody result. Clinical sensitivity and specificity was assessed in both methods using 175 blinded samples including 76 specimens from South America which were consensus positive in serological assays as well as 99 previously unscreened low risk US donor samples (presumed negative). Specimens positive for only one of the two methods underwent supplemental testing by a radioimmunoprecipitation assay (RIPA) to confirm antibody status. Dilutional sensitivity was assessed using dilutional series from 5 known positive samples tested by both methods. Results: Of the 76 presumed positive samples, all 76 were posi-tive by ELISA and 71 were positive by IFA for sensitivity of 100% and 93.4% respectively. Of the 99 presumed negative samples, all 99 were non-reactive by ELISA and 97 were negative by IFA for specificity of 100% and 98.0% respectively. ELISA results correlated with those by RIPA. The ELISA demonstrated greater dilutional sensitivity than IFA with all 5 dilutional series tested with reactive results at dilutions ranging from 10x to 50x greater than those detected by IFA Conclusion: The newly developed ELISA appears to be able to detect antibody to T. cruzi with sensitivity and specificity that is superior to currently established methodologies such as IFA and shows high correlation with RIPA. Introduction: Statistical spatial analysis has been used in medicine for spatial location of data, by use of Geographical Information Systems (GIS), a computer-based set of tools that capture, edit, store, integrate and analyze data that are spatially referenced. Spatial analyses are divided into three main groups: point, area and continuous data. This work relates to preliminary continuous data analysis derived from the rate of rejected blood bags only due to a reactive test for Trypanosoma cruzi (Chagas disease) from Blood Transfusion Services (BTS). Methods: Data were derived from a national questionnaire sent to more than 600 BTS, of which 218 (coming from 175 municipalities) replied with simultaneous rates for blood collection and rejected donations for Chagas disease, as performed by several screening methods (EIA, indirect haemagglutination or indirect immunofluorescence). Thus, we cannot define the actual national prevalence rate for this marker. By means of a combination of several GIS (commercial or public), data were analyzed for exploratory spatial descriptive analysis (ESDA-mean, median, outliers), proportional maps (equal, quintiles and trimmed intervals). From the original point data, inference and interpolation methods were applied to the whole territory: tessellation and Delaunay triangulation, inverse and normal distance weighting, point and block anisotropic ordinary kriging (the latter performed after defining no surface trend-as evidenced by median polishing-and semivariogram generation, including definition of anisotropy axis for maximum continuity). These methods allowed some different models that were cross-validated in order to check their accuracy. Results: A total of 218 BTS reported 1,577,607 whole blood donations in 2002, where 17,235 (1.09%) were discarded by any combination of reactive screening test for Chagas disease. The models produced displayed a considerable spatial heterogeneity distribution all over the territory, where spatial clusters and hot spots could be predicted and estimated (maps not presented here). Conclusion: These results indicate a high spatial distribution for rejected blood donations for Chagas disease, providing valuable tools for epidemiological mapping of other transfusion-transmitted diseases, as well as future management of BTS, including identification of areas that deserve more attention from health authorities in order to provide safe blood to the whole national territory.

D O Shah, C D Chang, K Y Cheng, L Jiang, V A Salbilla, A S Haller, G Schochetman, Abbott Laboratories, Abbott Park, IL BACKGROUND: The Abbott PRISM ® is a fully automated, chemiluminescence based immunoassay analyzer (ChLIA). It has been used to screen blood specimens for 5 different viral markers. A prototype assay is being developed on the PRISM analyzer for the detection of Chagas' disease (American trypanosomiasis). The disease, caused by the protozoan T. cruzi, is endemic to most regions of Latin America. Of the estimated 18 million infected people, approximately 50,000 die from this disease yearly. The estimated seroprevalence of the disease in the US blood donor populations was as high as 0.48% and the trend is increasing with the increase in the size of the Hispanic population. METHODS: The assay uses (a) microparticles coated with rAgs of T. cruzi to capture antibodies to T. cruzi, (b) an acridinium derivative-labeled mouse anti-human IgG to tag the captured human Ab on the microparticles, and (c) alkaline peroxide to trigger the chemiluminescence. Samples with anti-T. cruzi antibodies gave significantly higher signal compared to normal control. RESULTS: The sensitivity and specificity of the assay were evaluated by testing a number of T. cruzi positive specimens, random donor negative serum and plasma donations to set a provisional cutoff to differentiate positive and negative responses. Preliminary data showed 100% detection of confirmed T. cruzi positive specimens (n = 222). Specificity was estimated to be 99.83% (n = 3,008; RR = 5). Further optimization of the assay is in progress. CONCLUSIONS: A prototype Abbott PRISM ® Chagas assay based on rAg and chemiluminescence detection was demonstrated. The assay has the potential to be developed into a product for blood screening for the prevention of transmitting Chagas' disease through blood transfusion.

T T Gonçalez, C Almeida-Neto, N A Salles, Fundação Pró-Sangue/Hemocentro de São Paulo, São Paulo, Brazil; D A F Chamone, Departament of Hematology and Hemotherapy, Medical School, University of Sao Paulo, São Paulo, Brazil; E C Sabino, Fundação Pró-Sangue/Hemocentro de São Paulo, São Paulo, Brazil BACKGROUND: Domiciliary transmission of Chagas' disease as a result of exposure to the Triatoma infestans has been controlled in Sao Paulo State since the 1980s and throughout Brazil since the 1990s. Because the parasite also infects other mammals, it is excepted that the sylvatic cycle of the parasite will continue. With good current control of transfusion-related cases, other forms of human transmission should be evaluated. OBJECTIVE: To review the trend in the prevalence of Chagas' disease and the risk factors among young blood donors in São Paulo, Brazil. STUDY DESIGN AND METHODS: The data was obtained at Fundação Pró-Sangue/Hemocentro de São Paulo (FPS/HSP) in the years of 1995 to 2003 for 1st time blood donors under the age of 20 at the time of blood donation. Samples were considered positive if they were reactive to the three serological tests used at screening (indirect immunofluorescence-IIF, indirect hemagglutination-IHA and enzyme-linked immunosorbent assay-EIA). Reactive donors were asked to return and were interviewed for risk factor for Chagas' disease. RESULTS: The prevalence of Chagas' disease has decreased from 13.23/10,000 in 1995 to 0.97/10,000 in 2003. The overall prevalence was more than two times higher among replacement blood donors than among altruistic donors (9.05 versus 3.34 cases per 10,000). This difference was reduced over the course of the years analyzed, and by 2000, the prevalence were very similar among altruistic and replacement blood donors. Of 89 seropositive blood donors detected in the period, 23 returned for an interview. Only one case could not be associated to domiciliary transmission. This case was a considered vertical transmission because the mother and two of the seven sibling were infected and they have always lived in urban areas of São Paulo. The remained 22 vectorial cases had lived in rural area of the followed regions: Bahia (8), Minas Gerais (8) states, 05 were from other northeast states (Pernambuco, Piaui, Ceara), and 01 from another country, Bolivia. CONCLUSION: The prevalence of Chagas' disease is decreasing rapidly in the youngest population (92% in 9 years). Vectorial transmission was still the major cause for acquiring the infection in this groups and were restricted to specific areas of Brazil.

C R Seed, A Cheng, A J Keller, Australian Red Cross Blood Service, Perth, Australia; T M Davis, School of Medicine and Pharmacology, Fremantle, Australia Background Current "risk exposure" strategies to prevent transfusion transmitted malaria are effective but lead to significant wastage in the form of discarded components or deferred donors. This study investigated the feasibility of combining malarial antibody testing with the current cellular component restriction strategy. The two key objectives were; to evaluate the sensitivity and specificity of two malarial antibody EIAs and, to estimate the risk associated with implementing either test with a shortened (six month) cellular component restriction period for blood donors with an identified malarial risk exposure. Materials and Methods Blood donors were recruited into four distinct groups (Non-exposed [control], Malarial area "visitors", "residents" and "previous infection") and screened using two commercial malarial antibody EIAs. Assay sensitivity was evaluated in serologically confirmed positive blood donor samples as well as acute clinical samples. A mathematical model was derived to estimate the risk associated with implement-ing the combined testing/questioning strategy. Results The incidence of asymptomatic parasitaemia in malarial "visitors" returned for less than six months was 1 in 338 while no parasitaemic donors were detected amongst 402 "visitors" or "residents" returned for more than six months. The incidence of malarial antibodies within the exposed blood donor groups was 1.33% (10/751). EIA 1 detected 106/108 (98.1; 93.5-99.5%) acute (smear positive) P. falciparum cases compared to 58/108 (53.7; 44.3-62.8%) for EIA 2 (p value < 0.0001). For P. vivax, EIA 1 detected 12/12 (100, 75.8-100%) acute cases compared with 6/13 (46.2; 20.4-73.9%) for EIA 2(p value < 0.0001). The estimated additional risk exposure of the proposed new strategy was one infectious P. falciparum donation per 166 or 1 per 7 years for the EIA 1 and EIA 2 respectively. Conclusion The study findings support the efficacy and safety of a targeted screening strategy combining antibody screening with a six month cellular component restriction period for donors with a declared malarial risk. Background Tick-borne infections are increasingly recognized as threats to blood safety in the United States. Over 40 cases of transfusion-transmitted Babesia microti infections have been reported in the U.S., and the presence of enduring, silent Babesia parasitemia in healthy blood donors represents a challenge to blood safety. Human granulocytic ehrlichiosis (HGE) is caused by the bacterium, Anaplasma phagocytophilum, also reported as transmitted by transfusion. Both agents are endemic to the northeastern U.S. This study sought to determine the prevalence of exposure to both pathogens among healthy blood donors on the island of Martha's Vineyard. Methods From October, 2002 to April, 2004 , donors presenting at quarterly blood drives on Martha's Vineyard were invited to donate an extra tube of blood for testing for antibodies to B. microti and other tick-borne infections by immunofluorescent antibody (IFA) testing. Prevalence and temporal occurrence of seroconversion were recorded for both agents, and duration of antibody response was noted for HGE. Babesia-seropositive donors were indefinitely deferred and thus made no further donations. Results Thirty-one of 590 (5.3%) donations had antibodies to a tick-borne infection. Eleven (1.9%) donations were positive for antibodies to B. microti alone, while 19 (3.2%) were positive for antibodies only to A. phagocytophilum. One donation (0.2%) was seropositive for both agents. Twenty-seven of 362 (7.5%) donors tested positive to one (26) or both (1) agents. The age range of antibody-positive donors was 19 to 67, and 18 of the 27 were females. Positive test results were recorded at each blood drive. Altogether, 12 seroconversions were confirmed over the 21-month period. Interestingly, three seroconversions to HGE were confirmed in January for donors who tested negative three months before, suggesting infections acquired during coldweather months not known for tick biting activity. Each of 4 HGE-seropositve donors tested at the next quarterly drive reverted to seronegative status, while 3 tested six months later were no longer seropositive. The one donor who tested positive to both agents remained seropositive for the agent of HGE 9 months later. Conclusions The seroprevalence of 5.3% of donations and 7.5% of donors for the agents of babesiosis or HGE indicates that many whole blood donors on the island have been exposed to tick-borne pathogens in the absence of diagnosed infection. With intensified surveillance confirming the presence of both agents in a wider geographic area than once thought, this study highlights the attendant need for enhanced vigilance to blood safety in endemic regions. For a 12-county metropolitan area that includes some endemic areas, cases of babesiosis linked to transfusion reported to the regional blood center, mandatory state health department incident reporting system, and mandatory communicable disease reporting system from 2000 to 2003 were studied. Cases that proved to be confirmed or probable transfusion-associated cases were identified. The number of transfusions in the study area during the study period was determined from mandatory transfusion activity reports. The number of community-acquired cases of babesiosis was determined from communicable disease case reports. RESULTS: experience any signs or symptoms of infection, 17 days after inoculation. A 5 log reduction was achieved.

S Wagner, A Skripchenko, D Thompson-Montgomery, H Awatefe, American Red Cross, Rockville, MD Background: The use of photosensitizers for pathogen reduction in RBC has been extensively studied but progress in developing methods with >6 log reduction of agents has been hampered by photoinduced hemolysis. A flexible, nucleic acid-intercalating thiopyrylium (TP) dye was utilized that is only photochemically active in the bound state in order to reduce RBC membrane photodamage arising from unbound dye. In addition, the band 3 ligand, dipyridamole (DP), was used to compete for some of the TP binding sites on the RBC membrane in order to reduce photoinduced hemolysis stemming from bound dye. Study design and methods: Oxygenated leukodepleted 20% hct RBC suspensions in Erythrosol were deliberately inoculated with extracellular virus, cells containing provirus, virus infected cells, or bacteria, incubated with 200 mM DP and various concentrations of TP, and subsequently illuminated with 1.1 J/cm 2 of red light. Controls and treated samples were assayed for virus or bacteria. At the most stringent phototreatment conditions studied for pathogen reduction (200 mM DP, 160 mM TP, 1.1 J/cm 2 red light), identically prepared, but uncontaminated RBC suspensions were phototreated, concentrated to 45% hct, and assayed for in vitro RBC storage properties. Results: In the presence of DP, phototreatment of RBC suspensions resulted in >8.4 log of extracelluar VSV at 100 mM TP, >6.3 log extracellular PRV at 15 mM TP, >5.7 log extracellular DHBV at 10 mM TP, >6 log extracellular BVDV at 4 mM TP, >7.5 log extracellular HIV at 80 mM TP, 6.2 ± 0.1 log intracellular HIV at 80 mM TP, and >7.6 log S. epidermidis, >6.3 log S. aureus, 5.9 ± 0.6 log Y. enterocolitica, 5.4 ± 1.0 log P. fluorescens, >7.1 log S. liquefaciens, 6.2 ± 1.0 log S. marcescens, and 7.1 ± 0.5 log E. coli at 100 mM TP. DP prevented binding of approximately 36% of the TP that would otherwise bind to RBCs. RBCs suspended in Erythrosol and treated with 160 mM TP, 200 mM DP and 1.1 J/cm 2 light and subsequently stored for 42 days exhibited acceptable yet statistically different levels of hemolysis (0.32 ± 0.09 vs. 0.19 ± 0.06%), morphology scores (94.5 ± 1.5 vs. 98.3 ± 0.7%), extracellular pH (6.48 ± 0.07 vs. 6.38 ± 0.06), ATP (2.63 ± 0.07 vs. 3.61 ± 0.22 mmole/g Hb), similar glucose utilization rates (6.20 ± 0.97 vs. 7.08 ± 0.31 mM/day), and similar lactate production (26.7 ± 1.9 vs. 28.2 ± 1.0 mM on day 42) compared to controls. Treated samples exhibited substantially increased potassium efflux (16.6 ± 4.2 vs. 4.4 ± 0.4 mM on day 1) compared to controls. Conclusions: Use of TP photosensitizer and DP enables significant levels of pathogen reduction for a number of organisms while retaining most, but not all RBC properties during 42 day storage.

H E Purdum, CryoFacets, Inc., Raleigh, NC Background: Transfusing red blood cells (rbc's) exposes the patient to the risk of deadly infections by a wide variety of pathogens. Screening and testing can reduce, but not eliminate, this risk. It is therefore necessary to develop effective means of decontaminating rbc's. Conventional decontamination approaches employ various chemical additives to bind or destroy the pathogen DNA and/or RNA. Unfortunately, these additives, and their activating processes, can severely damage the rbc's. Furthermore, residual additives can induce antibody responses. An ideal decontamination process would therefore require no chemical additives. Also, an ideal process would combine multiple, independent treatments so that any pathogen that might escape one treatment would likely not escape the second. Two such techniques are Ultraviolet-C (UVC) and ozone, but these techniques are also known to have several problems that have thus far prevented their use for rbc's. Methods: Packed human rbc concentrates were first diluted 2 : 1 with saline and exposed to ozone. Because conventional ozone systems damage the cells by inducing shear around the gas bubbles and by producing excessive local concentrations, a proprietary system was developed to introduce the gas smoothly and uniformly throughout the cell suspension. Also, the ultraviolet absorption of the ozone in solution was monitored to maintain the ozone concentration within ± 2%, thus ensuring uniform exposure. After ozone treatment, the cells were then irradiated with UVC. Because conventional UVC systems generate highly reactive species from dissolved oxygen, the cells were first treated in a proprietary system to reduce the Seven cases were reported in which an implicated donor could be identified, with an observed frequency of 1/400,000 RBC units. All patients were elderly and/or immunocompromised. No cases were found in recipients of co-components. Transfusion was the attributed source of infection in 2.5% of reported babesiosis cases. Of the 7 implicated donors, 6 resided in or had traveled to babesiosis endemic areas, consistent with the 88% of reported community-acquired cases who resided in or reported travel to endemic areas. One implicated donor lives and works in a county where only 2 cases have been reported. This donor reported a history of Lyme disease, but a risk factor for babesiosis was not identified. CONCLUSIONS: Transmission of Babesia organisms continues to be a transfusion risk, more frequent than viral transmissions in this high-risk area. However, at-risk donors are not limited to known high-risk areas. BACKGROUND: Treatment of red blood cell (RBC) suspensions with riboflavin and light has been used to reduce the number of certain pathogens. Orientia (formerly Rickettsia) tsutsugamushi, the causative agent of scrub typhus, is an obligate intracellular bacterium that grows free in the cytoplasm of infected cells. Clinical manifestations range from small skin lesions to organ failure. The organism is sensitively and specifically identified in a well-characterized mouse bioassay, and as an intracellular bacterium it provides a robust system well suited for evaluating the effectiveness of a pathogen reduction technology. This study evaluated the capability of the Riboflavin and Light in RBC suspensions to inactivate Orientia tsutsugamushi as measured by mouse infectivity. METHODS: RBCs were inoculated with human mononuclear cells infected with Orientia tsutsugamushi. Groups of three mice were injected intraperitoneally with untreated RBCs infected with 10 0 to 10 5 organisms (n = 9). The treatment group was inoculated with 10 5 organisms suspended in RBCs treated with 500 mM Riboflavin and 60 J/cm 2 of light. Mice were monitored daily for up to 17 days for signs and symptoms of infection (e.g. lethargy, labored breathing, rough coat) and euthanized upon appearance of symptoms or at day 17 post infection. Real-time PCR on blood and Giemsa stains from peritoneal exudates were performed. RESULTS: Animals from all untreated groups exhibited enlarged spleen, liver, and bloody peritoneal exudates in addition to the above signs. Orientia tsutsugamushi was inactivated to below the limit of detection by treatment with 500 mM Riboflavin and 60 J/cm 2 of light. In addition, no significant changes were detected in RBC protein structure when treated and untreated RBCs were evaluated with infrared spectroscopy. CONCLUSIONS: Riboflavin and light is effective in reducing Orientia tsutsugamushi, as assayed in an animal model. Mice infected with 10 5 organisms in RBCs treated with 500 mM Riboflavin and 60 J/cm 2 of light did not oxygen content of the solution. Next, the cells were passed through a proprietary thin film UVC exposure chamber. To determine the decontamination effectiveness of the two systems, the cells were spiked with Bovine Viral Diarrhea Virus (BVDV). Cell integrity was monitored by crystal violet staining for Heinz body detection and by measuring the hemolysis and potassium leakage. Measurements were taken at days 0, 1, and 7. Results: UVC and ozone both reduced BVDV by Log 5, the limit of detection. No Heinz bodies were observed. Hemolysis was less than 1.0% for controls and all treated samples, while potassium levels were within 12% of controls for all samples. Conclusions: UVC and ozone are both highly effective at decontaminating BVDV in rbc solutions, and it is thus expected that their combination would be even more effective. Neither technique causes excessive cell damage. Additional work is necessary to determine the system effectiveness against other pathogens, and to assess the functionality of the treated cells in vivo and over extended time periods. Background: Viral clearance validation is a required process to be carried out in the production of biologics including blood-derived products to minimize the risk of hazadous virus contamination. One of possible viral contaminants in plasma-derived products is HCV(Hepatitis C Virus). Determination of HCV contamination not only in donated bloods but in pooled-plasmas using NAT(Neucleic acid amplification techniques) assay to identify the presence of HCV RNA genome is recommended internationally. Thus, in this study we aimed to establish HCV removal validation in a production process of coagulation factor IX by quatitative PCR method. Method: HCV RNA plasmas obtained from Blood Transfusion Research Ins. Korea were tested for enzyme immunoassay for anti-HIV, HbsAg, anti-HCV and by PCR for HBV DNA, HIV RNA, HCV RNA. The genotypes of HCV were also determined. Compared to WHO international standard, HCV RNA units were determined and the plasmas were processed and used as a quantitative reference. The nested RT-PCR was performed after extraction of viral RNA from samples by QIAamp viral RNA isolation kit. The first round of PCR was carried out by GeneAmp PCR System 9700 TM and for amplicons were amplified by LightCycler TM to quantify HCV RNA. Results: The quantitative reference was established and assigned an activity of 5.9 ¥ 10 6 IU/mL of HCV RNA. The linearity was manifested on the diluted series of this reference ranging from 5.9 ¥ 10 4 IU/mL to 5.9 ¥ 10 2 IU/mL. The coefficients of variation values of the inter-and intra-assay reproducibility were 6.61% and 0.79%, respectively. Coagulation factor IX spiked with HCV RNA positive plasma was filtered with viresolve NFP filter to simulate virus filtration process and the reduction factor was 4.63. Conclusion: We established a quatitative reference for HCV NAT assay and applied a sensitive and quantitative PCR method for removal validation of HCV in plasma-derived products. Using this HCV reference and assay system, we have demonstrated that we can provide a system as a tool for HCV removal validation in the plasma-derived products with a high level of confidence in sensitivity and reproducibility.

H Mohr, A Redecker-Klein, U Gravemann, T Müller, Blood Center of the German Red Cross Chapters of NSTOB, Springe, Germany Background: In plasma containing 1 mmol/L methylene blue (MB), West Nile virus (WNV) is rapidly inactivated by illumination with visible light. Because of the limited titre of the virus stocks used to spike the samples, inactivation factors higher than 5-6 log10 cannot be determined. Using an approach in which plasma units were repeatedly challenged with WNV and re-illuminated, it was investigated if there is additional capacity to inactivate WNV by MB/light treatment. Methods: Plasma units were isolated from blood donations by centrifugation and stored in standard PVC bags. The volume of plasma units was approx. 300 ml. They were spiked with WNV. MB was added at a concentration of 1 mmol/L. The plasma units were illuminated with fluorescent tubes emitting white light. Light intensity was 45,000 lux. Samples were assayed for viral infectivity after 1, 2 and 5 min, respectively. New virus was added and the plasma was re-illuminated. This sequence (addition of virus-illumination for 1, 2 and 5 min) was repeated three more times. Results: In each of the four treatment cycles the WNV titres were reduced by 2-3 log10 within 1 min light exposure; after 2 min titre reduction was by 4 and after 5 min by > 5.7 log10. There was no indication that the inactivation kinetics between cycles 1 and 4 were slowed down. Conclusions: The results of the rechallenge experiments demonstrate that photodynamic treatment with MB/light is highly effective against WNV. The rechallenge approach may be useful to prove the true capacity of this and other decontamination procedures to inactivate pathogens. After treatment we take a sample of 10 ml from each bag in order to carry out several T. cruzi viability assays. Viability assays:1) Microscope direct inspection.2) In vitro cultures of parasites: 7,5 ml aliquot was taken from each bag, centrifuged and re-suspended in 600 ml of LIT parasite differentiation medium, and seeded in triplicate in 96 well tissue culture plate. On day 0 and then once a week, cultures were observed microscopically, and parasite concentration determined 3) Inoculation to a highly sensitive mouse strain deficient in inducible nitric oxide syntethase (iNOS -/-): 7,5 ml aliquot was taken from each bag, centrifuged and re-suspended in 500 ml of uninfected and untreated serum. Each aliquot was used for the inoculation of iNOS -/mice. On day 0 and then twice a week, a blood sample from each mouse was taken and the number of parasites was counted. The experiment was followed up until mice survival or death could be determined. Results:

Conclusions: Photochemical treatment with MB plus light reduces T. cruzi in FFP as an extent that avoids mice infection. Filtration reduces the number of parasites transfused, but not avoids mice infection. Freezing itself appears as being as effective as MBT.

L Bellio, S Rossini, M A Introini, P Ronchi, SIMT H S Raffaele, Milano, Italy

To improve the quality and safety of platelet transfusions, we evaluated a treatment to inactivate pathogenic agents in platelet concentrates. We chose a nucleic acid-targeted photochemical treatment using amotosalen HCL (S-59) and ultraviolet A (UVA) light (Intercept, Cerus and Baxter Healthcare Corporations). The pathogen inactivation process was performed in platelet components prepared with the buffy coat method from a pool of 5 whole blood collections. Pooled platelet concentrates were leukoreduced by filtration followed by photochemical treatment with 150 mM S-59 and 3 J/cm 2 UVA treatment for pathogen and leukocyte inactivation. This study was performed to evaluate the feasibility of introducing this technology into the routine daily preparation of hemo-components. It was conducted using a pool of 20 platelet concentrated buffy coats stored and tested for 7 days preand post-treatment and evaluated for pH (37°C), pCO2, pO2, glucose, lactose, swirling phenomenon, and platelet count. The unit received the UVA treatment if the platelet number was 2,5-5,0 ¥ 10 11 and the red cell number was < 4 ¥ 10 6 /ml. No differences were observed between platelet concentrates with or without treatment. In a clinical study, a total of 35 thrombocytopenic patients with haematological disease received 53 leukoreduced platelet components. Twenty patients received 29 treated platelet concentrates and 15 patients received 24 non treated platelet concentrates. There were no selection patients neither stratification in these study. More than 70% of the platelet components were stored for only 2 days before transfusion. The mean 1 hour corrected count increment for the test group was 7058 ± 4400 vs 9837 ± 5000 for the control group. The mean 24-hour corrected counts were 4018 ± 2000 vs 5560 ± 3000 respectively. No adverse events were recorded in the group using treated platelets. Conclusion: the inactivation process proved to be easy and fast to set-up; there were no differences in clinical haemostasis or hemorrhagic adverse events between patients that received treated and non treated platelet concentrates. Background INTERCEPT Blood System (Baxter, Deerfield), treatment allows pathogen inactivation in platelet products, both isolated from whole blood or obtained through thrombapheresis, addressing the residual risk of viral or bacterial infection. The procedure also blocks DNA transcription, which eliminates cytokine production by contaminant white cells, and might therefore lower the rate of transfusion reactions. The procedure was routinely introduced in the BTC of Mont-Godinne in October 2003 . Study Between October 2003 and March 2004 , 1502 INTERCEPT Blood System procedures have been performed on thrombapheresis products from 1097 procedures on Amicus Crescendo (Baxter, Deerfield). All products were treated in the hospital blood bank, and transfused to patients in the institution. Transfusion reactions were signaled through the routine institutional procedure and compared with the frequency of transfusion reactions during 9 months preceding the introduction of INTERCEPT Blood System. Results In terms of technical feasibility, despite the use of Integrator, 9% of collected doses needed a diminution of the platelet content before being processed in order to meet the process requirements (the limits for the process are set to 2.5-6.0x 10e11 platelets/treated dose). In 32 cases out of 1502 (2%), the product could not be processed, either because the platelet content was too high (29/1502), or too low (3/1502). Only 2 technical failures occurred, 1 of which was disposable-linked and 1 user linked. Interestingly, there was a tendency towards the reduction of febrile non-hemolytic transfusion reactions (0,82% vs 1,02%). Although this difference is not statistically significant, it should be noticed that almost half of the transfusion reactions in the test group were observed within the first weeks after the introduction of INTERCEPT Blood System, possibly reflecting a bias. Conclusion Routine introduction of INTERCEPT is feasible even in a small transfusion center. Technical problems are rarely seen. INTERCEPT Blood System probably leads to a diminution in the rate of febrile non-hemolytic transfusion reactions.

L Sawyer, D Hanson, K Dupuis, Cerus Corporation, Concord, CA Background: Current assays to detect infectious Parvovirus B19 (B19) are cumbersome and yield variable results. We have developed and previously reported on an ELISpot assay to detect infectious B19, and have used this assay to demonstrate inactivation of B19 by Helinx technology, as used in the INTERCEPT Blood System. A limiting factor in inactivation of nonenveloped viruses may be the time necessary for amotosalen (S-59) to penetrate the virus capsid. We investigated the impact of pre-incubation with S-59 prior to UVA treatment on inactivation of B19 in infected platelet units.

Two separate studies were performed using B19 infected plasma obtained from Alpha Therapeutics. Platelet concentrates (PC) were prepared by centrifugation and resuspension of single donor apheresis platelets in 35% B19-infected plasma and 65% platelet additive solution. Each PC unit was divided into 30 mL mini-units, each containing approximately 2.5-6.0 ¥ 10 10 platelets. Amotosalen was added to each unit to a final concentration of 150 mM, and the units were kept at room temperature until UVA treatment. After the appropriate incubation (Table) , each mini-unit was treated with 3.0 J/cm 2 UVA illumination. Samples were taken prior to illumination to determine B19 input titer and each mini-unit was sampled immediately after illumination to detect residual virus. Viable virus titer was determined by infection of CD34+ cells and detection of progeny by ELISpot assay. Results: See Table. The B19 input titer averaged 7.7 logs spot forming units/mL in the two replicates and the genome titer was approximately 11.2 logs genome equivalents/mL. Conclusions: Up to 5.8 logs of Parvovirus B-19 in platelet concentrates were inactivated by Helinx technology, as used in the INTERCEPT Blood System. have also demonstrated the conservation of the FVIII procoagulant activity, the von Willebrand factor antigen concentration, the ristocetin cofactor activity and the multimeric composition of the von Willebrand complex. Phase 3 clinical trials have been conducted in patients with 1) congenital coagulopathies, 2) acquired coagulopathies and 3) thrombotic thrombocytopenic purpura (TTP). Results from the first 2 trials demonstrated that INTERCEPT Plasma is comparable in efficacy to conventional FFP and results of the TTP trial have not yet been unblinded. As decreased levels of vWF cleaving protease (vWF:CP, ADAMTS-13) have been shown to be involved in the pathogenesis of TTP, we measured the impact of photochemical treatment on vWF:CP activity in INTERCEPT Plasma. Methods: Fresh apheresis plasma collected in 600 mL jumbo units using the Haemonetics PCS apheresis platform was treated with amotosalen HCl and UVA. All of the major blood groups were represented. Aliquots of baseline and post-treatment plasma were frozen prior to analysis. VWF:CP activity was determined by assessing the degradation of exogenous, purified vWF by electrophoretic analysis of vWF multimers in sodium dodecyl sulphate-agarose gels and also by differential binding of residual vWF in a collagen binding assay. Results: Results were expressed as a ratio relative to a normal control (N = 12).

Multimeric analysis showed essentially no change in vWF:CP activity after photochemical treatment. Collagen binding after treatment was 95% of the pre-treatment binding. All results were within the reference ranges. Conclusion: INTERCEPT Plasma retains von Willebrand factor cleaving protease activity at levels comparable to conventional FFP. 

B Donnelly, L Pinkoski, L Corash, Cerus Corporation, Concord, CA BACKGROUND: INTERCEPT Blood System for Plasma, using Helinx ® technology (amotosalen HCl and UVA light), inactivates pathogens and leukocytes while retaining acceptable coagulation function. Prior clinical studies for congenital and acquired factor deficiencies and plasma exchange demonstrated that INTERCEPT Plasma supported hemostasis. Practical pathogen inactivation processes should be sufficiently robust to conserve coagulation factor activity over a range of operating conditions. This study evaluated a range of treatment conditions believed to be challenging to coagulation factor activity. METHODS: Three experiments were performed, each with six replicates. For each experiment, apheresis plasma units (600 mL) were pooled and separated into 2 Test and 1 Nominal arm before INTERCEPT treatment including reduction of amotosalen levels using a flow-through compound adsorption device. Processing variables included amotosalen concentration (150 mM Nominal vs. 160 mM Test), UVA dose (3.0 J/cm 2 Nominal vs. 3.4 J/cm 2 Test), plasma volume (600 mL Nominal vs. 400 and 665 mL Test), and processing temperature (22°C Nominal vs. 15°C and 30°C Test). Coagulation factors were assayed using Clauss assay (fibrinogen), one-stage PT-based clotting assays (V and VII) or one-stage aPTT-based clotting assays (VIII, IX, XI). RESULTS: Coagulation factor concentrations and activities after treatment were within reference ranges.

Samples taken after INTERCEPT treatment were used to determine viable titer after inactivation. Samples for determination of viable K. pneumoniae titer were cultured on Luria Bertani agar plates and colonies were enumerated after overnight incubation at 37°C. HIV infectivity was quantified by microplaque assay on MT-2 cells in 96-well plates. Log inactivation was calculated from the number of viable organisms in treated and untreated samples. RESULTS:

CONCLUSION: The INTERCEPT Blood System for Plasma retains in vitro coagulation function over a range of processing conditions believed to be challenging to coagulation factor activity. (model for HCV and  Dengue) , WNV, HTLV-I/II, K. pneumoniae, and P. falciparum (malaria) was evaluated in jumbo plasma units. These studies were conducted to evaluate the inactivation characteristics of the INTERCEPT Blood System for plasma using treatment components and conditions suitable for high throughput commercial use. Methods: ABO-matched, single-donor jumbo apheresis plasma units were pooled as necessary to make units of 585 mL, inoculated to a final concentration of ~10 6 organisms/mL and treated with 150 mM amotosalen and 3.0 J/cm 2 UVA treatment. Samples were taken before illumination to determine input titer and after illumination to detect residual viable pathogens. Titer was determined by plaque assay in MT-2 cells (HIV), BT cells (BVDV), vero cells (WNV), or BHK21pA18G cells (HTLV-I and II), growth in RBC culture (P. falciparum), or colony formation on LB agar (K. pneumoniae). Results: See Table. N of 2 or less indicates that studies are in progress and results are preliminary. Conclusions: High titers of five viruses, P. falciparum and K. pneumoniae were inactivated in jumbo apheresis plasma units to or below the limit of detection by treatment with the INTERCEPT Blood System for plasma. BACKGROUND: Orientia (formerly Rickettsia) tsutsugamushi, the causative agent of scrub typhus, is an obligate intracellular bacterium that grows free in the cytoplasm of infected cells. It occurs over a wide area of eastern Asia and the western Pacific region where it is a serious health problem for both military personnel and local residents. Clinical manifestations range from small skin lesions to organ failure. Because O. tsutsugamushi is a bloodborne organism, the potential exists for blood transfusion-associated transmission. The organism is sensitively and specifically identified in a well-characterized mouse bioassay, and as an intracellular bacterium it provides a robust system well suited for evaluating the effectiveness of a pathogen inactivation technology. This study evaluated the capability of the INTERCEPT Blood System for plasma to inactivate O. tsutsugamushi as measured by mouse infectivity. METHODS: In two replicate experiments human plasma was inoculated with human mononuclear cells infected with O. tsutsugamushi. In each experiment, groups of three mice were injected intraperitoneally with untreated plasma serially diluted to contain 10 0 to 10 5 organisms/inoculation (0.5 mL). The treatment group (n = 3) was inoculated with plasma which had been contaminated with 10 5 organisms/0.5 mL, then treated with 150 mM amotosalen HCl and 3.0 J/cm 2 UVA illumination. Mice were monitored daily for up to 17 days for signs and symptoms of infection (e.g. lethargy, labored breathing, rough coat) and euthanized upon appearance of symptoms or at day 17 post infection. Real-time PCR on blood and Giemsa stains from peritoneal exudates were performed. RESULTS: Animals from all untreated groups exhibited enlarged spleen, liver, and bloody peritoneal exudates in addition to the above signs. Mice inoculated with treated, O. tsutsugamushi-infected plasma showed no clinical or microbiological evidence of infection 17 days after inoculation, indicating a reduction of >5.5 logs. In addition, no significant changes were detected when several clotting factor assays were performed on treated and untreated plasma.CONCLUSIONS: The INTERCEPT Blood System for plasma is effective in inactivating Orientia tsutsugamushi, as assayed in an animal model. Orientia tsutsugamushi was inactivated to below the limit of detection, >5.5 logs, by treatment with amotosalen HCl and UVA illumination.

V Chandrasekaran, S Lakshminarayanan, Long Island Jewish Medical Center, New Hyde Park, NY Introduction: Severe hemolytic disease of newborn (HDN) requiring transfusion is caused by antibodies in Rh or Kell blood group system, mostly from anti c, Kell, and anti-k, since the advent of Rh immune globulin. HDN resulting from anti E or from maternal IgG autoantibody is often mild. We present a case of severe HDN with Hydrops Fetalis resulting from combined allo anti E and maternal IgG autoantibody. Case Study: Thirty-four year old Caucasian female with prior history of massive blood transfusion in 1994 for trauma and a spontaneous abortion in 2002 was admitted for elective Caesarian Section for poor fetal tracing. The blood bank history from 2002 revealed a positive direct coombs with anti IgG and a panagluttinating warm auto antibody in the eluate. No alloantibodies were identified. Prenatally the patient's sample was tested at an outside lab where only warm autoantibody was identified. On admission patient had Hgb of 12.9 gram/dl and normal indirect bilirubin. Blood bank sample typed as A,Rh positive. Direct coombs was 2+ positive with polyspecific and anti IgG Coombs sera,with panagglutinating warm autoantibody reacting equally in the eluate panel. The plasma panel had a pan agglutinating IgG autoantibody reactive in the MTS Gel column (Ortho) and a stronger allo anti E. The titer of anti E at indirect coombs phase was 64 and the titer of warm autoantibody with E negative cells 1. The neonate was pale and jaundiced at birth. Cord blood typed O,Rh positive with 2+ direct coombs with anti IgG. Elution revealed a warm panagluttinating IgG antibody and a stronger anti E. Phenotype for E antigen was positive. Hemoglobin was 6.8 grams with hematocrit of 22%. Total bilirubin 3 hours after birth was 20 mg/dl with predominantly indirect bilirubin. The peripheral smear revealed polychromasia, spherocytes, and nucleated RBC's. The baby had hepatosplenomegaly and generalized edema. Immediate phototherapy and two volume red cell exchange transfusion with E negative,O,Rh positive PRBC was initiated. Patient responded to the treatment with drop in the indirect bilirubin and increase in hemoglobin to 14.5 grams/dl. Two weeks later neonate was discharged in good physical condition. Conclusion: This is a rare and unusual case of severe HDN with Hydrops Fetalis, resulting from anti E and a warm maternal autoantibody. It is likely that the severity of the HDN was related to the combined effect of anti E and warm autoantibody with apparent anti Rh specificity. Other causes for hemolysis need to be ruled out also. It is interesting that the mother did not have active hemolysis during the pregnancy.

E Lee, M De Silva, National Blood Service, London, United Kingdom UK Guidelines recommend that pregnant women with anti-K/-c should be tested 4 weekly to 28 weeks then 2 weekly to delivery, irrespective of the partner's phenotype. We analysed retrospectively the data to identify women with anti-K and anti-c whose partners were K-and c-respectively. In this group of women it is very likely that the anti-K and anti-c was transfusion induced. Method: The records of 203 pregnant women with anti-K and 199 with anti-c, referred between 2001-3 were scrutinised to obtain the partner's phenotype and woman's transfusion history. Results: The partners of 91 out of the 199 women with anti-c were tested and only 3 (3.3%) of them were c-. The partners of 168 out of 203 women with anti-K were tested and 138 (82.1%) were K-. Out of these 138 women, the transfusion history was available for 76 women, all of them were transfused. Conclusion: 90% UK donors are K negative and all donations are K typed. Implementing a policy of selecting K negative blood for all women with potential for child bearing should be easy and could have prevented 82.1% women making anti-K, thus avoiding up to 8 blood tests each during pregnancy. 96.7% of women with anti-c had a c positive partner and selecting c negative blood would have less impact on c sensitisation. Furthermore, only 20% of donors are c-and all Rh positive women would need to be c typed to identify those eligible for c negative blood. Reference: 1. BCSH Guidelines. Transfusion Medicine, 1996;6:71-74.

G E Carnahan, University of South Alabama, Mobile, AL Background Alloantibody formation during the first four months of life is a notably rare event. Anti-D formation upon transfusion of platelets, phersis is reported despite concurrent administration of RHIG. Case Report A 23-yearold African-American (AA) woman, G2P1, spontaneously delivered a 635 gm female after a 23-week-long gestation. Prenatal course had been uneventful before the mother presented with a severe urinary tract infection, developed preterm labor and failed to respond to terbutaline. The newborn was suspected to have congenital sepsis-in addition to sequelae of extreme prematurity (respiratory insufficiency; patent ductus arteriosus). Blood cultures yielded no bacteria. However, the infant was treated with cephalothin and gentamicin for a seven day course immediately after birth. However, episodes of suspected sepsis, anemia, and thrombocytopenia continued and recurred, prompting addition of ceftazidime and amphotericin B to the antimicrobial regimen. On the eight day of life, a diagnosis of necrotizing enterocolitis lead to a jejunoileal resection and ileostomy. With and following this procedure, multiple transfusions of Rh negative red cells and platelets were necessary. A Grade I intraventricular hemorrhage occurred. Recurrent suspicion of sepsis caused additional rounds of antimicrobial treatment without any positive cultures except for Staphylococcus epidermidis associated with catheter colonization. Despite multiple complications, the patient made slow progress. However, persistent patency of ductus arteriosus (PDA) despite indomethacin therapy lead to a demonstrable left-toright shunt, left atrial and ventricular enlargement, and intermittent arrhythmias. Surgical intervention was needed. A pre-operative platelet count of 17,000 resulted in an order to transfuse platelets. Since only Rh positive platelets, phersis were available, platelet transfusion was immediately followed by an intravenous infusion of RHIG (WinRho 270 micrograms) on the 45 th day of life (post-conceptual age 206 days). The PDA ligation pro-ceeded uneventfully. Additional complications of short bowel syndrome, bacterial overgrowth, hyperalimentation associated cholestasis, and RSV infection occurred. When elective closure of the ileostomy was planned, the patient was found to have a positive antibody screen. Anti-D was identified at 13 weeks of age (132 nd day of life; post-conceptual age 293 days). Conclusion A four-month-old, A Rh negative, AA female developed an anti-D alloantibody after a single exposure to the D antigen in a single aliquot of apheresis platelets despite administration of intravenous RHIG.

Severe Hemolytic Anemia Following Rh o (D) Immune Globulin Intravenous Administration in ITP Patient Associated with Anti-C J Schwartz, T Hamilton, K Kapoor, K Grima, New York Blood Center, New York, NY Background: Immune thrombocytopenic purpura (ITP) is an autoimmune disease mediated by antiplatelet antibodies which bind to circulating platelets and result in accelerated platelet destruction. Rh o (D) immune globulin intravenous (WinRho®) is approved for treatment of ITP in D-positive, nonsplenectomized patients. The primary mechanism of action involves competitive binding of anti-D-sensitized red blood cells (RBCs) to macrophage Fc receptors within the spleen, thereby reducing the extent of sequestration and destruction of the patient's antibody-sensitized platelets. Hemoglobin decreases, but it is usually mild. We had a patient who developed severe hemolytic anemia after treatment with WinRho® which was found to have anti-C in addition to the anti-D. Case report: 67 year old woman with ITP for 5 months was admitted for GI bleeding. Her platelet count was 16K and her Hgb was 11.6. When her platelet count dropped to 9K the next day, she received 3500 mg of WinRho®. Two days later her Hgb was 6.9; laboratory data was cosistent with hemolytic episode (Table) . patient received 1 unit of red cells. Pretransfusion workup showed + direct antiglobulin test (DAT), 3 + IgG and weakly + for complement. Patient phenotype was positive for the C antigen. Antibody screen was positive with both screening cells. Antibody work-up identified an allo anti-D in the patient's undiluted serum sample. The eluate revealed anti-D and anti-C consistent with hemolysis secondary to WinRho®. Patient was further treated with pulse steroids and anti -CD20 monoclonal antibody (Rituxan®). Conclusion: The anti-D in a 300-mg vial of WinRho effects the sensitization and sequestration of approximately 17 ml of D+ RBCs, which are presumably destroyed primarily via extravascular hemolysis. In most studies a mean decrease in Hgb of 1.7 g/dl is seen. Only 3% of cases show a decrease in Hgb > 4 g/dl and the need for transfusion is rare. WinRho® contains more than 90% polyclonal IgG anti-D as well as low titer anti-A, anti-B, anti-C, and anti-E. It might be postulated that the presence of anti-C in our patient contributed to the severity of the hemolysis. Transfusion services should be aware that WinRho® can cause severe hemolysis and other alloantibodies in addition to anti-D may be transferred with the administration of this medication.

for hemolysis. We report the use of a chemiluminescent assay (CLA) and IgG subclass determination in predicting fetal outcomes for anti-D and -K sensitized pregnancies. Methods: Serial antenatal serum samples on 18 women sensitized with either anti-D or -K were evaluated using a standard CLA assay. A standard curve using dilutions of RhIg and single donor R 2 R 2 cells was prepared for each assay. For each sample RLU units were expressed as a percentage of the maximum chemiluminescence of the control. IgG subclass determinations were performed by gel card using WHO reference antibodies. Fetal outcome determined by chart review. Results: Fetal outcome was stratified into four categories; Category 1: no treatment; Category 2: phototherapy only (PT); Category 3: exchange transfusion (ET); Category 4: IUT. The CLA correctly predicted the outcome of 14/18 cases (see Table) . The CLA incorrectly predicted a category 1 outcome in an infant who required PT and vice versa, and incorrectly predicted a category 2 outcome in a fetus who required ET. One infant with a category 4 outcome sensitized to anti-K was predicted to require PT by the CLA. In category 1, the outcome of both fetuses in pregnancies sensitized with anti-K was correctly predicted by our CLA. IgG 1 and 3 were both present in the maternal sera of many of those with the most severe HDN. Conclusion: We found the CLA to be an accurate predictor of fetal outcome. The assay correlated well with the titer method of predicting fetal outcome, required minimal training of staff and had less inter-observer error. Although the number of patients studied is small, the CLA is capable of accurately predicting mild outcomes in anti-K sensitized pregnancies. The two most serious incorrect predictions featured both IgG1 and 3. Perhaps the CLA assay is insensitive to the IgG3 component when present in high titre. While IgG subtypes 1 and 3 were both present in most maternal sera with severe disease, further study of the significance of IgG subclasses on fetal outcome is required. Background: Hemolytic disease of the newborn (HDN) can result in significant fetal mortality and morbidity. The standard procedure in the clinical laboratory to evaluate whether invasive therapy is necessary for an at-risk pregnancy is maternal antibody titers. This is a laborious technique prone to significant inter-observer variability and may not reflect the true potential Hemolytic Disease of the newborn has decreased secondary to Anti-D prophylaxis. However, it remains common enough to warrant investigation into improved management. If an exchange transfusion can be avoided, benefits would include decreased blood exposure for the infant and the possibility of avoiding transfer to another health care institution. The use of IVIG to avoid an exchange transfusion is not a new idea, but recently it has been combined with treating the mother with Phenobarbital prior to delivery and initiating it in the newborn. We present a case study combining Phenobarbital and IVIG. The mother was a 30-year-old G3 P1 T1 LC1 woman who was blood type A negative. She had become Rh isoimmunized during pregnancy, and had an Anti-D titer of 1 : 64. She was treated with Phenobarbital prior to delivery. The baby boy was born at an estimated gestational age of 36 weeks. His DAT was positive; Hematocrit was 39, Total bilirubin was 3.2 mg/dL; and blood type was A positive. The baby was treated with IVIG, Phenobarbital and phototherapy. His bilirubin rose to 5.5 mg/dL at 5 hours of age and 7 mg/dL at 24 hours of age. His total bilrubin peaked at 11.1 mg/dL by day 4, then started to drop. His hematocrit slowly dropped until he was discharged asymptomatic at 24%. He remined steady and asymptomatic for the next two weeks. It is very likely that an exchange transfusion was avoided in this case. With the very low morbidity that this treatment entails, further research is warranted.

The Krever Commission (1997) recommended an informed consent policy on transfusion in Canada. In November 2002 the Canadian Standards Association included this policy as a licensing requirement for hospital transfusion services. The current compliance is not known. Study The Transfusion Ontario Program in Ottawa conducted a survey on Pediatric Transfusion practices (Nov 02-Apr 03) in pediatric and general hospitals with pediatric services across Canada. There were questions on practices and compliance to a transfusion policy. Eight children and 21 general hospitals with pediatric services completed the survey. This sample data is summarized below. Results p < 0.001); fibrinogen from 1.46 ± 0.75 to 1.66 ± 0.59 (t = 1.25; p > 0.05). Platelets decreased from 145.4 ± 107.4 to 124.7 ± 75.4 (t = 1.26, p > 0.05). There were 2 fatalities (16.7%). One child died from concomitant severe congenital pathology. There was one haemorrhagic death in a 5 years old following liver biopsy. PT pre-transfusion was 16.6 (upper level 15.5); APTT 34.3 (upper level 40.3); fibrinogen 2.14. Following Octaplas infusion (16 ml/kg) there was no improvement in clotting time. Conclusion: (SD) plasma transfusion results in amelioration of PT and APTT in paediatric patients with liver disease; however, as with FFP, this may not be universal or complete. The prevention of haemorrhage following liver biopsy requires further studies.

V Chekrizova, W G Murphy, Irish Blood Transfusion Service, National Blood Centre, Dublin 8, Ireland

Background: Since March 2002 the Irish Blood Transfusion Service has replaced fresh frozen plasma (FFP) from Irish donors with pooled solventdetergent (SD) plasma from US voluntary donors, as a response to the threat posed by variant Creutzfeldt-Jakob disease (vCJD). Approximately 25,000 units of (SD) treated plasma are used annually for patients requiring plasma in place of FFP. AB plasma is provided as Uniplas, a blend of A, B and AB plasma to provide a low titre of anti-A and anti-B in the final product. Considerable clinical experience in adults with (SD) plasma has been reported, however its use in paediatric patients has not been systematically reported. Methods: To address this problem we studied the efficacy and safety of Uniplas in children requiring plasma transfusions. Results: 18 children got a total 67 transfusions of Uniplas over 18 months (16-8770 ml per patient, mean 66.3 ± 134.0 ml/kg); maximum dose received was 515.9 ml/kg. In one patient 4 total plasma exchanges (for TTP) were performed with Uniplas. Underlying conditions comprised: sepsis, cardiac surgery, neonatal pulmonary haemorrhage, haemolytic-uraemic syndrome in leukaemia, thrombotic thrombocytopenic purpura (TTP). Age at presentation was from 1 day to 4 years (9 neonates, 6 infants, 3 older children). The distribution according blood groups was following: O-4, A-2, B-4, AB-8. No adverse reactions or complications were observed for Uniplas. There were 3 fatalities from underlying conditions (16.7%): one neonate (group AB) with meningitis, one neonate (group O) with severe congenital heart defect and one child of 18 months (group B) with TTP and encephalitis associated with influenza. 17/18 patients also received red cells (mean 68.9 ± 49.4 ml/kg), 13/18 got platelets (mean 75.2 ± 74.1 ml/kg) and albumin (mean 97.0 ± 176.2 ml/kg). Pre-and post-transfusion APTT and PT were measured in 19/67 transfusion episodes, fibrinogen in 15/67, platelets in 24/67. Mean APTT improved from 63.2 ± 15.6 to 46.1 ± 9.8 (t = 7.43; p < 0.001); PT from 27.2 ± 12.3 to 21.1 ± 8.9 (t = 3.05; p < 0.01); fibrinogen from 1.55 ± 0.68 to 2.13 ± 0.73 (t = 3.05; p < 0.01), platelets from 87.6 ± 67.7 to 93.1 ± 62.7 (t = 0.4; p > 0.05). Conclusion: Uniplas in therapeutic doses has good overall clinical tolerance in children and is associated with correction of coagulopathy.

Background: Hemostasis is a concern in the operating room (OR) as it affects duration of surgery and amount of blood products transfused. This study assessed non-invasive parameters related to intra-operative postcardiopulmonary bypass (CPB) hemostasis in pediatric cardiac surgery, to provide baseline data for future studies assessing efficacy of blood conservation strategies. Methods: Intra-operative blood loss is difficult to assess during pediatric cardiac surgery, thus the length of time needed to achieve hemostasis was measured in this study. The hemostatic time was time to sternal closure (time from protamine administration at completion of cardiac repair to sternal closure after satisfactory hemostasis was achieved). The time from protamine administration to the time that patient exited OR was also measured. Potential hemostasis-related variables recorded included the patient's age, weight, history of cardiac surgery, pre-operative diagnosis, length of CPB, lowest core temperature, laboratory parameters pre-and post-CPB. Results: 83 consecutive cases of open-heart surgery were monitored over a two-month period. 3 patients who required post-surgical extracorporeal membrane oxygenation were excluded from analysis. 39 cases of cyanotic congenital heart disease (CCHD), mean age 17 months, mean weight 8 Kg; and 41 cases of non-cyanotic congenital heart disease Conclusion This sampling identified 69% of hospitals have an informed consent policy and 50% have an informed consent form. 86% report having discussions prior to transfusion but documented this in only 2/3 of cases. The children hospitals report similar data except fewer have a separate informed consent form. Seven years after the Krever recommendations 30% of the surveyed hospitals have yet to comply with the informed consent policy. The next phase in studying the implementation of the Canadian Standards Association guidelines may be to explore the issues for noncompliance.

V Chekrizova, W G Murphy, Irish Blood Transfusion Service, National Blood Centre, Dublin 8, Ireland

Background: Since March 2002 the Irish Blood Transfusion Service has replaced fresh frozen plasma (FFP) from Irish donors with pooled solventdetergent (SD) plasma from US voluntary donors, as a response to the threat posed by variant Creutzfeldt-Jakob disease (vCJD). Approximately 25,000 units of (SD) treated plasma are used annually for patients requiring plasma in place of FFP. Considerable clinical experience in adults with (SD) plasma has been reported, however its use in paediatric patients has not been systematically reported. Methods: To address this problem the safety and efficacy of Octaplas in correcting coagulopathy in children with liver disease was assessed. Results: 15 patients (age from 12 days to 16 years) received 33 transfusions of Octaplas over 18 months (35 -2200 ml per patient, mean 38.0 ± 41.5 ml/kg). Liver pathology comprised: autoimmune hepatitis, cystic fibrosis, hepatoblastoma, Still's disease, hepatic venooclusive disease, sclerosing cholangitis, ischaemic liver injury in cardiac surgery, liver trauma, neonatal liver failure, extrahepatic biliary atresia. 3 transfusions were given during surgery (2 liver resections, 1 liver trauma); 6 were performed for liver biopsies. 10/15 patients received also red cells (mean 68.7 ± 71.9 ml/kg), 8/15 got platelets (mean 87.2 ± 210.9 ml/kg), 7/15 received albumin (mean 532.9 ± 708.9 ml/kg). Pre-and posttransfusion APTT, PT and platelets were measured in 22/33 transfusion episodes, fibrinogen in 13/33. Mean APTT improved from 61.5 ± 33.0 to 47.8 ± 12.5 (t = 5.15; p < 0.001); PT from 24.4 ± 10.0 to 19.9 ± 4.2 (t = 5.05; (NCCHD), mean age 36 months, mean weight 14 Kg, were compared. The cardiopulmonary bypass time was marginally longer in CCHD than in NCCHD (122 min vs 100 min, P = 0.049), but the hemostatic time was significantly longer (CCHD 74 ± 36 min, NCCHD 47 ± 30 min, p = 0.0005). The time from protamine administration to patient exiting OR was also significantly longer (CCHD 108 ± 39 min, NCCHD 80 ± 35 min, p = 0.001). There was no difference in the baseline activated thrombin time prior to surgery, suggesting that the differences in hemostastic times were related to the type of surgery rather than pre-existing coagulopathy. Hemostatic times did not appear to correlate with age, weight or history of previous cardiac surgery. The blood use by donor exposures in the 2 groups was not significantly different (8.1 vs 5.9, p = 0.2) but the volume of blood component transfused by Kg body weight was significantly higher in CCHD (258 ml vs 143 ml, p = 0.0005). There was good correlation between the hemostatic times and the volume of blood transfused. Conclusions: Hemostatic times can be used as surrogate markers of blood loss and transfusion needs in pediatric cardiac surgery. Further study is planned to show that interventions which shorten hemostatic times will also decrease blood transfusion requirements.

A D Slonim, J G Joseph, W M Turenne, N L C Luban, Children's National Medical Center, Washington, DC Background: For over 50 years, the United States has had an organized blood collection system. Currently, nearly 27 million units of blood and blood products are transfused each year in the United States. These transfusions are provided to patients across all age groups, multiple diagnoses, and in a variety of settings. While investigations of blood and blood product transfusion practices in specific subsets of pediatric patients have been performed, broad epidemiologic approaches regarding practice patterns have not been published for hospitalized children. Study Design: A non-concurrent cohort study of pediatric patients < 21 years of age hospitalized at one of 32 academic Children's Hospitals in calendar years 2001-2003. These hospitals voluntarily contribute data to a performance improvement collaborative entitled the Pediatric Health Information System (PHIS). Once submitted, data must meet a threshold criterion to assure validity before inclusion in the dataset. A severity level was assigned based upon a validated algorithm (APR-DRG) that encodes age, diagnoses, and procedures. Each sampled discharge was surveyed for an ICD-9 procedure code indicative of a blood product transfusion. The patient and hospital characteristics associated with transfusions and their complications were identified for each discharge. Bivariable testing of patient and hospital characteristics with the occurrence of a transfusion or complication was performed using a p value of < 0.05 as the significance level. Results: From a cohort of 1,009,096 discharges, 48,195 (4.8%) received a blood product transfusion during their hospitalization. The most frequent blood product transfused was packed red blood cells (n = 41,577, 86.3%). Important patient characteristics associated with blood product transfusions included race, payer status, severity level, and outcome. Asians were 2 times more likely than whites to receive a transfusion and privately insured patients were less likely to be transfused than publicly insured patient (p < 0.001). Transfused patients had a higher severity level and mortality rate than non-transfused patients. Transfused patients also had a hospital LOS, which was more than 3 times that of non-transfused patients (17 vs. 5 days, respectively). Hospitals with more personnel, more beds, a higher census and serving a larger population were more likely to provide transfusions and have more transfusion reactions. Conclusions: Blood product transfusions in academic pediatric hospitals are a fairly common procedure. Important patient and hospital level associations exist and may provide insight into the variability in transfusion practices and their association with disparities in care and outcome.

K Nicol, T Hudson, P Moder, Columbus Children's Hospital, Columbus, OH Background: Platelet Pheresis (PP) demand has grown since leukoreduction and the implementation of bacterial detection. Efficient utilization of these products is necessary with the supply somewhat limited. However, concerns involving CO 2 retention have been raised when low volume transfusions are desired, primarily in children. We sought to demonstrate the safety and adequacy of multiple aliquot usages throughout the life of the PP. Method: The pH of PP product "residuals" (stored in the original bags @ 22°C with constant agitation) was measured over 34 days using a Mini Lab pH Meter TM as platelet aliquots were made for transfusion using a sterile connection device. The "age" of the PP, and the remaining platelet volume were then noted. Patient's post transfusion platelet increments and the mean platelet volume (MPV) were documented when spontaneously ordered. We arbitrarily divided patient results into 3 groups (for comparison) based on platelet residual volumes from which their aliquots were made (£150cc, 151-300cc, and ≥301cc). Results: Seventy aliquots from 42 PP in Gambro ELP bags were evaluated. Residual platelet volumes (remaining after aliquot production) ranged from 60-548cc; 6 aliquots were made from residual volumes of < 100cc and 14 aliquots were made from residuals of 101-150cc. Twenty aliquots were made from residual units with < 24 hours remaining until expiration. All had a pH > 7.0 except 3 aliquots (pHs of 6.8, 6.7, and 6.9) (~96%); 2 of the 3 were imported (out of region) PPs, and the other was from a residual of 118cc having 7.5 hours until expiration. Using the unpaired Student t-Test there were no significant differences between unsolicited post-transfusion platelet increments or the patient's MPV (as a measure of platelet function) between any patient groups. Deviation from typical platelet ordering practices was not noticed. Conclusion: Several institutions may be considering pediatric use of PP aliquots for small volume transfusions. While this practice has not been heavily studied, our facility has used pheresis aliquots for nearly five years and has achieved expected therapeutic effect with minimal wastage of product. From these and previous observations, there has been no significant degradation of quality of multiple small volume aliquots as seen in pH or post-transfusion indices over time.

R C Framil, J Doherty, M Beaton, S Sloan, Childrens Hospital Boston, Boston, MA Background: In January 2003 a universal irradiation policy was established for all red cell containing units released from the Blood Bank. The policy was intended to prevent Transfusion-Associated Graft versus Host Disease (TA-GVHD) in patients whose immunologic status was unknown. While there have been no known cases at Children's Hospital Boston, when manifested, TA-GVHD is approximately 90 percent fatal. The most significant challenge associated with a universal irradiation policy concerns inventory management. This policy leads to an increased inventory of irradiated units that have a shortened shelf life and an increased concentration of free potassium that may cause complications for some infants. This analysis was performed to study the implications of a universal irradiation policy on inventory management at a large pediatric hospital. Design: All irradiated red blood cell units dispensed from February to December of 2003 were analyzed. The criteria for analysis included patients under one year of age, red cell units utilized on patients under one, divided units and red cells released after having been irradiated greater than three days. Also included were locations throughout the hospital that were ordering and returning irradiated blood. Results: During the analysis period, the monthly return of irradiated blood averaged approximately 400 units. Of these 400 units, 250 were returned from the Main OR, 80 units were returned from the Cardiac Catherization Lab, and the remainder was returned from the various medical surgical wards with no particular trends noted. Of the 400 units returned, 90 units were from 70 patients less than 1 year of age and 40 units were divided or split units. To prevent transfusion induced hyperkalemia in infants less than one year old, a policy was adopted that prohibited transfusion of any units that had been irradiated for more than 3 days. 84 units that were crossmatched or reserved for these patients were subsequently released to general inventory because the time since irradiation had exceeded 3 days. Of these, 48 were whole units 26 divided units were from patients ≠4 months old and 7 whole units and 3 divided units were from patients between 5 months and 1 year old. Conclusion: The majority of irradiated units returned, are from the Main OR and Cardiac Catherization Lab. Patients in the ICU, hematology/oncology, and NICU, whose patients are more often immunocompromised, rarely return blood. However, patients that are 4 months old or less often have units ordered that are either not used or are released from that patient. Based on these findings it has been recommended to reevaluate the blood product ordering practices of the Main OR and Cardiac Catherization Lab. We report the occurrence of a TRALI-type reaction during an exchange transfusion in a one day old infant. Case Report: The patient was a stable premature 36 week gestation 2.6 kg group O, Rh negative male with hemolytic disease of the newborn (maternal anti-D titer = 128). During the first of two double volume exchange transfusions, he experienced cyanosis, acute oxygen desaturation, temperature rise (37.2 to 37.8C) and tachypnea.

His chest x-ray, previously normal, showed newly acquired mild pulmonary vascular congestion and a small amount of fluid in the right minor fissure, consistent with early pulmonary edema. Echocardiogram displayed persistent pulmonary hypertension with right to left shunt at patent ductus arteriosus and foramen ovale, and normal left ventricular function. The baby was intubated and placed on mechanical ventilation with 100% oxygen. Inhaled nitric oxide was given. Dopamine was administered for pressor support. There was a transient episode of renal failure treated with furosemide and fluid restriction. The baby was extubated to room air three days after the onset of his respiratory distress. He was discharged with apparent full recovery. The implicated exchange unit was composed of pooled red blood cells and plasma, from two different donors. The red blood cell donor tested negative for HLA and neutrophil antibodies. The unit plasma demonstrated HLA Class II anti-DR 52, non-specific HLA Class I antibodies and weak IgM antineutrophil antibodies, identified by flow cytometry. Molecular genotyping of the baby's buccal smear two months later was positive for the HLA Class II DRB 3 gene, corresponding to the anti-HLA DR 52 serologic specificity. Conclusions: Transfusion of a reconstituted whole blood-equivalent unit, containing anti-HLA DR 52 in its plasma, was associated with a TRALI-type episode in a neonate with a corresponding HLA genotype. There is at least one report of a similar occurrence of acute respiratory insufficiency in a neonate, following the transfusion of anti-neutrophil antibodies in donor plasma. These apparently reacted in vivo with a subsequently transfused granulocyte concentrate and they demonstrated in vitro cytotoxic reactivity with the neutrophils of the granulocyte donor (Transfusion 1988; 28:173-76) . For appropriate management of patients, discussions with parents and donor deferral, it is important to be alert to the possibility of TRALI-type reactions in neonates as well as in older transfusion recipients, secondary to passively acquired anti-leukocyte antibodies.

E C Wong, C A Colvin, V R Criss, M Gallagher, S J Soldin, N L C Luban, Children's National Medical Center, Washington, DC Background: Lead exposure adversely affects intellectual development at blood lead levels (BLLs) as low as 10 mcg/dL in vulnerable populations including infants and children. Transfusion of blood with elevated BLL to infants assigned to a single packed RBC unit for the duration of their hospitalization or who receive multiple transfusions would lead to a BLL resulting in adverse health risk. Changes in water treatment with chloramine in November 2000 resulted in increased leaching of lead from homes, schools and businesses with lead piping. Longitudinal studies of BLL in children, cross-sectional study of lead levels in homes and schools drinking water (MMWR 53:2004) and water sampling in area schools in the Maryland and Virginia suburbs demonstrated a significant public health issue. We therefore initiated a study to determine the prevalence of elevated blood lead levels in blood donors presenting to our donor center. Study Design and Methods: We describe a prospective study of blood donors over a 3 month period at a tertiary care pediatric hospital donor center. Samples collected at the time of blood donation were tested for lead levels using a Perkin Elmer Atomic Absorption Spectrophotometer (model 600 A Analyst) in a Centers for Disease Control approved lead testing site. State of origin and zipcode were extracted from donor history cards. Results: Overall, there were a total of 565 donations (both whole blood and apheresis platelets) from 499 donors. 444 donors donated once, 44 donors donated twice, 10 donors donate three times and 1 donor donated four times. Of the 565 donations, 97 were from the District of Columbia, 350 from Maryland, 99 from Virginia, and 19 from outside of the DC metropolitan area. The majority of donations (n = 516, 91%) had BLL less than 3.0 mcg/dL, with 49 donations (9%) having elevated levels between 3.0 and 10.0 mcg/dL. The highest BLL measured was 9.2 mcg/dL. Of the donors that donated more than once (n = 55), 5 had lead levels > 3.0 mcg/dL. Two donors had a level of > 3.0 mcg/dL for only one of their donations; while the other four had levels > 3.0 mcg/dL for all of their donations.

Conclusions: Elevated BLL (greater than 3.0 mcg/dL) can be seen in a small proportion of blood donors. Further study is needed to evaluate the significance of transfusing RBC units with elevated lead levels to selected pediatric populations. 362:332, 2003) suggested a daily permissible intravenous lead exposure of 0.36 mcg/kg derived from WHO recommendation of oral intake. Because elevated lead levels have been found in packed RBC units, we undertook a study to predict the amount of the lead that would be transfused to patients based on our donor population BLL and compared these predicted exposures to post-transfusion BLL. Study Design and Methods: Prospective study of pre-and post-transfusion BLL over a 3 month period at a tertiary care pediatric hospital donor center. Diagnoses, age, gender, patient weight, number and volume of transfused packed RBC units were extracted from the medical chart or blood bank records. Samples were tested for lead levels using a Perkin Elmer Atomic Absorption Spectrophotometer (model 600 A Analyst) in a Centers for Disease Control approved lead testing site. Predicted daily intravenous lead exposure was determined from the 25th, 50th and 75th percentile of BLL of blood donors and the volume of transfused RBCs to recipients. Results: A total of 6 patients (5 males and 1 female) ranging in age from <1 month to 34 months were evaluated. Diagnoses included extracorporeal membrane oxygenation because of left congenital diaphragmatic hernia repair or as a bridge to heart transplant (n = 2), cardiac surgery for congenital heart disease (n = 2) and extreme prematurity with necrotizing enterocolitis or respiratory distress syndrome (n = 2). Median (range) pre-and post-transfusion BLL was 0.4 mcg/dL (range 0.0 to 2.5), and 1.2 mcg/dL (range 0.5 to 2.8), respectively. Although median post-transfusion BLL demonstrated an increase relative to pre-transfusion blood levels in 5/6 patients, this increase, however, was not significant (p = 0.17, two tailed Wilcoxon Signed Rank Test). However, predicted daily lead exposure based on the median (50 th percentile, 1.0 mcg/dL) of BLL in donors who presented at our donor center demonstrated a significant correlation (Pearson r = 0.87 and p = 0.0254). Furthermore, lead exposure theoretically exceeded 0.36 mcg/kg/day in 1/6, 2/6 and 4/6 patients based on 25 th , 50 th and 75 th percentile of blood donor BLLs. Conclusions: Predicted daily intravenous lead exposure based on RBC transfusion demonstrates significant correlation with post-transfusion BLL. RBCs from donors with relatively normal BLL can theoretically contribute to intravenous exposure to lead that exceed permissible limits in neonates and small children. Background: Cord blood (CB) from a sibling donor is an increasingly important hematopoietic stem cell (HSC) source for transplantation. In certain clinical circumstances, it is critical to ascertain whether an unborn fetal donor's CB may be histocompatible with an existing sibling recipient. To address this problem non-invasively (without amniocentesis), we have conducted a pilot study for analysis of paternally inherited fetal HLA sequences using mater-nal peripheral blood. Materials and Method: We collected blood prospectively from 13 pregnant women during the first and second trimester; blood from the father was also obtained. The mother and father were typed for HLA-DR and DQ using a real-time quantitative PCR typing panel which included 12 DRB types (DR 1, 3, 4, 7, 8, 9, 10, 11, 12, 13, 15, 16) and 9 DQB types (DQB 2, 3, 4, 5, 6.01, 6.02, 6.03, 7, 8) . The mother's plasma and whole blood were then assayed for the presence of informative paternal DR and DQ sequences; when sufficient DNA was available, these samples were also probed for the alternative paternal allele as an added control. Four replicates were studied for plasma and 8 for whole blood. Accuracy of the non-invasive prenatal typing was determined by comparison with reference DR and DQ typing which was performed directly on CB collected after delivery. Results: The Table depicts the informative paternally inherited DR allele which was amplified in maternal blood or plasma, the trimester the sample was drawn, and the number of aliquots giving a positive result. Control amplifications of the non-inherited paternal DR sequence were negative. Background: Isohemagglutinins (anti-A, anti-B) are naturally occurring alloantibodies made by immunocompetent individuals who lack the corresponding A or B antigen on their red cells. Isohemagglutinin titers are used to assess immune function in patients with primary or secondary immunodeficiencies. In pediatric patients, isohemagglutinin titers in potential heart transplant recipients are used to predict rejection of ABO mismatched allografts. However, there is no standardized method for performing isohemagglutinin titers. The majority of studies examining isohemagglutinin titers in blood donors or blood products employ a 30 minute incubation, whereas our laboratory have always used a one hour incubation method for testing patients. Thus, a study was conducted to evaluate isohemagglutinin titers, comparing 30 minute and 60 minute incubation. Methods: 20 patient samples were tested in parallel. Each serum sample was serially diluted with normal saline, incubated with 3-5% cell suspension of appropriate A or B cells, and O cells as control. After incubation for 30 or 60 minutes, the tubes were centrifuged and read macroscopically. The end point was the last 1+ reaction. Results: 16 of 20 samples showed total concordance in the titers. 4 samples showed a one-tube difference. The titers were higher after 60 minutes in 3 samples (0, 8, 32 at 30 minutes, and 1, 16, 64 at 60 minutes). The titer was lower at 60 minutes in one sample (4 at 30 minutes, 2 at 60 minutes). We also examined isohemagglutinin titers performed on a day when samples were left to incubate for 2 hours because of unforeseen circumstances. These samples were not run in parallel, so we searched for other isohemagglutinin titers performed on the same patients. For 2 patients, there were titers performed within 4 months of the index date, one was tested before, the other was tested both before and after. There was no difference in the titers in these 2 patients, after 2 hour incubation on the index date or 60 minute incubation on the other dates. Conclusions: Isohemagglutinin titers read after 30 or 60 minute incubation showed no significant difference in results. A two-hour incubation may also provide acceptable results. We conclude that the length of incubation is not critical in measuring isohemagglutinin titers, and laboratories may choose a length of incubation suitable for the workflow in their laboratory. Background: Patients who are refractory to platelet therapy frequently have high percent reactive antibodies (PRAs). In order to identify a suitable platelet for transfusion two different crossmatching approaches are frequently employed. The single donor (SD) crossmatch approach utilized by our center is independent of HLA, ABO, or the presence of platelet specific antibodies; however, it is inventory dependent and therefore lends itself to a relatively rapid turn-around-time. The HLA crossmatch approach, while more selective, may result in a delay in product dispensation despite the fact that our donors are routinely typed. The purpose of this study was to examine the success rates of these two approaches and the relative contribution of anti ABO and anti platelet antibodies. Materials and Methods: 24 archival plasmas from patients with PRA's >80% were evaluated for ABO titer and antiplatelet antibodies. Crossmatch data was available for SD only (n = 7), HLA only (n = 4), both HLA and SD (n = 11). A total of 931 SD and 220 HLA crossmatches were performed. All were performed by solid phase red cell adherence (SPRCA) with the Immucor Capture-P kit (Norcross, GA). Platelet antibodies were determined by ELISA with the PAK 12 Platelet Anti-Conclusion: Using the methods we have developed, it is possible to determine accurately the paternally inherited HLA DR sequences during the 2 nd and/or 3 rd trimester in most cases. This information has potential applications in sibling cord blood banking and in the planning of alternative donor sources for selected patients. The results should be confirmed and extended in larger subject populations. Background. During induction of remission therapy (IRT), children with acute lymphoblastic leukemia (ALL) are exposed to diverse chemotherapeutic agents, including the enzyme L-asparaginase, a major factor for thrombotic complications during ALL treatment. Objective: to prospectively evaluate changes in the activity of natural anticoagulants in ALL children under 13 years of age, and document thromboembolic episodes. Methods: The functional activity of natural anticoagulants was measured in 15 ALL patients before administration of L-aspar and at the end of IRT. Antithrombin (AT) and protein C were determined through a colorimetric assay. Protein S activity was measured employing a microlatex particle-mediated immunoassay. Lymphoblast immunophenotype was determined with a FACSCalibur cytofluorometer (Becton Dickinson, USA), including CD10, CD19, CD20, and CD22. The analysis included average values, range, and standard deviation. Student's "t" test was employed to test for differences between average values, with significant differences recognized at p < 0.05. Results: The basal percentage of proteins C, S, and AT activity was within the normal range, but a consistent decrease was found on the day of the last L-asparaginase dose. The mean reduction on the activity of protein C was 26%, 18% for protein S, and 29% for AT (table 1) . This level of reduction was significant (p < 0.05), when compared with the initial values and normal controls. One ALL child developed central nervous system thrombosis immediately following administration of the third dose of L-asparaginase, with hemipharesia persisting as a sequel. An additional child suffered a superficial venous thrombosis, which resolved with no further complications. Although in both cases a decrease in natural anticoagulants was found, the decrements were not statistically different from those observed in the remaining children who did not sustained thrombotic complications. Immunophenotype distribution was the expected, with early pre-B and pre-B phenotypes accounting for all the cases. Conclusion. A consistent, statistically significant decrement in the percentage of activity of all three natural anticoagulants was documented. In two cases this decrement was accompanied by a thrombotic complication.

body Screening Kit (GTI, Waukesha, WI). ABO IgG Titer >1 : 128 were selected as potentially significant. Logistic regression model was used for SD and HLA match rates. Paired t-test was used to test the difference between the SD match rate and the HLA match rate. Results: Successful crossmatch rates for the HLA match matched platelets were significantly higher than the SD match rate (P HLA = 0.42, P SD = 0.12, p-value = 0.01). Antiplatelet antibodies were identified in 9 cases; 2 cases with multiple antibodies. The distribution was anti HPA 1b (3), HPA 5b (3), HPA 4a (3), and one each for HPA 1a, 5a, and Ib/IX. ABO IgG antibodies >1 : 128 were detected in 9 patients with blood groups: group O(5), groupB(3), and group A(1). The presence of antiplatelet or anti ABO titers did not significantly affect the rate of successful HLA crossmatches. In contrast, the presence of antiplatelet antibodies, but not anti ABO IgG, was associated with a lower crossmatch success rate with SD platelets (p-value = 0.0126). Conclusions: The provision of HLA crossmatched platelets for patients with high PRAs is considerably more efficient and cost effective. In centers that routinely HLA type the donor population, the SD crossmatch rates may be optimized if existing inventory is checked against the center's database for HLA type. Background: With the increasing use of high-dose granulocyte transfusions, better post-transfusion absolute neutrophil counts are being observed in the recipients. These circulating neutrophils could either be of donor or recipient origin, and not infrequently, it is difficult to assess the continued need of subsequent transfusions. Moreover, a sustained elevation in the recipient's post-transfusion absolute neutrophil count might suggest an "ineffective transfusion" due to the presence of circulating donor neutrophils that are incapable of tissue margination. This study evaluated the use of flow cytometry in distinguishing mixed neutrophil populations, in an attempt to explore its possible utility in identifying donor from recipient neutrophils. Methods: Granulocyte suspensions were prepared from freshly collected Na 2 EDTA whole blood of volunteer donors using a Ficoll-Hypaque discontinuous double density gradient (density 1.080 and 1.120 g/mL). Samples for analysis were prepared using a mixture of seven different ratios of HNA-1a or HNA-2a positive and negative cells (2.5, 5,10,50,90,95 and 97 .5% antigen positive). Two well-characterized antisera specimens containing HNA-1a and HNA-2a, respectively, were incubated with the mixture of target neutrophils. Unbound material was removed by washing the cell suspension with Hanks Balance salt Solution + 20 mM HEPES + 1% NCS. A second incubation with a FITC-labeled F(ab') 2 fragment of anti-human total immunoglobulin followed. The measurement of cell-bound fluorescence by flow cytometric analysis was carried out with a Beckman Coulter EPICS® XL-MCL (Miami, FL). Results: Baseline separation of HNA-2a positive and negative cell populations was observed in all neutrophil mixtures containing at least 2.5% of HNA-2a cells. In contrast, a distinct HNA-1a positive cell population was not observed with anti-HNA-1a. Conclusion: Flow cytometry using polyclonal reagents can distinguish distinct populations of neutrophils from different individuals who are positive or negative for the neutrophil antigen HNA-2a, but not HNA-1a. Additional evaluation of monoclonal reagents or different antigen/antibody systems may permit separation of other mixed populations of neutrophils. Therefore, identifying donor neutrophils from recipient's neutrophils using flow cytometry seems conceivable, and further studies are needed to explore this avenue.

M F Leach, J P Aubuchon, Dartmouth Hitchcock Medical Center, Lebanon, NH Background: Solid phase red cell adherence assays (SPRCA) are utilized by many transfusion service laboratories for the detection of HLA and/or platelet-specific antibodies. During routine testing, inconsistent results were obtained when EDTA samples were used. We investigated the cause of these unexpected positive reactions. Study Design: Reactivity with 100% of the reagent platelets was noted when certain EDTA samples were tested. Evaluation of this reactivity included performance of an ELISA (GTI PAT12, Waukesha, WI) as well as repeat SPRCA screens using a serum sample; lymphocytotoxicity (LCT) testing was performed on selected samples. Chloroquine diphosphate (CDP) treatment of the reagent platelets was performed; the SPRCA screen was then repeated using EDTA samples. Serum samples from patients found to have unexpected reactivity with EDTA samples were added to EDTA vacutainer tubes and incubated at room temperature for 15 minutes; the SPRCA screen was repeated using the treated serum. Control testing was performed using EDTA samples from 24 normal donors; reactive samples were tested by ELISA. Results: Between January 2002 and May 2004, a total of 401 HLA screens were performed using SPRCA assays; 152/401 (38%) were known to be performed using serum samples. Of the remaining 249 samples, almost all of which were EDTA plasma, 15/249 (6%) showed the presence of unexpected reactivity. These 15 samples were further evaluated using serum, CDP-treated platelets, and ELISA methods; 4 samples tested by LCT were negative. The only reactivity observed was seen when EDTA samples were tested using SPRCA. Reactivity was also observed when serum samples from patients exhibiting the positive reactions with EDTA were added to tubes containing EGTA, which also anticoagulates by binding calcium. There was no effect of tube type (glass vs. plastic) on the reactivity of these antibodies. Control testing showed 4/24 (17%) samples reactive by SPRCA; two were also positive by ELISA techniques for detection of HLA Class I antibodies (8% of total samples). The 2 samples positive only by SPRCA (8%) exhibited the reported reactivity. Conclusion: The use of EDTA samples for performance of HLA screens by SPRCA technology was found to result in the detection of EDTA-dependent antibodies reactive with an HLA epitope in approximately 6% of the samples tested. The presence of these antibodies should be considered when EDTA samples are utilized in testing and all of the reagent platelets are found to be reactive in routine testing. The problem can be avoided by not using EDTA samples.

Background: Drugs such as tirofiban (T), abciximab (A) and eptifibatide (E) bind to platelet glycoprotein (GP) IIb/IIIa, which may also be the reactive site for some drug-associated platelet antibodies (DAPA). We investigated the pre-treatment of platelets with these drugs to determine if DAPAs to metoprolol (M), famotidine (F) or omeprazole(O) (designated below as group A drugs), reactive by solid phase red cell adherence (SPRCA) methods, were blocked from reacting when T, A or E treated platelets were tested. Study Design: Serum containing anti HPA-1a was used to determine the optimal drug concentrations, media and incubation time that inhibited reactivity with treated platelets.Testing was performed with samples containing DAPAs known to react at GP IIb/IIIa (sulfamethoxazole, quinine, and quinidine). Samples containing DAPAs against group A drugs were tested to determine if GP IIb/IIIa was the reactive site. Controls were run with all samples to detect false positive reactivity. Results: Blocking of anti HPA-1a reactivity was optimal when drug concentrations ranging from 0.977 ug/mL (A) to 375 ug/mL (E) were utilized; T and E had optimal blocking when no LISS was added, while A required the addition of LISS. Reaction times of 10 minutes were optimal with strongly reactive antibodies, while 35 minutes were required for weaker antibodies. The reactivity of DAPAs known to react with GP IIb/IIIa was blocked when T, A and E treated platelets were tested. Testing of DAPAs against M and O reacted with treated platelets while F DAPAs were inhibited. False positive reactions were seen in 26% of samples tested, almost always due to reactivity with platelets pretreated with A, E or T. These false positives have been reported in 2-74% of healthy donors (Aster,Blood, 2002) . Conclusion: The use of platelets treated with drugs that bind to GP IIb/IIIa inhibited reactivity of anti HPA-1a and DAPAs known to bind to GP IIb/IIIa.Testing suggests that F DAPAs have a GP IIb/IIIa reactive site, while M and O antibodies react at other sites. Testing using SPRCA and platelets pre-treated with IIb/IIIa binding drugs may be useful in determining the reactive sites for DAPAs. BACKGROUND Activation of platelet by anti-platelet antibodies (Abs) may define or modify the severity of transfusion reaction. Although some antiplatelet Abs produced by transfusion or pregnancy has been shown to induce platelet activation, little is known about anti-CD36 Abs. We experienced a case of nonhemolytic transfusion reaction followed by thrombocytopenia after transfusion of anti-CD36 Abs-containing fresh frozen plasma (FFP). In this study, we examined the in vitro response of platelets derived from normal healthy subjects to this anti-CD36 Abs-containing plasma. METHODS Platelet-rich plasma (PRP) from CD36-positive and -negative healthy subjects was incubated with the anti-CD36 Abs-containing plasma that appeared to be produced by pregnancy in a CD36 type I deficient individual. The platelet aggregation and RANTES release were measured. The synergistic effect of epinephrine and the role of FcgRIIa in the plasmainduced platelet activation were studied. At the same time, the platelet surface levels of CD36 and FcgRIIa and the polymorphism of FcgRIIa were analyzed. Furthermore, for the platelet responses, anti-CD36 Abs-containing sera obtained from other individuals were tested. RESULTS Platelets from 20 CD36-positive individuals showed variable platelet responses to the plasma. Platelet activation was induced by the plasma alone in 4 subjects. In 7 subjects, platelet activation was synergistically induced by the plasma and a subthreshold concentration of epinephrine. Platelets of 9 subjects failed to respond to the plasma. The platelet activation induced by either the plasma alone or by synergy with epinephrine required the involvement of FcgRIIa. The different responsiveness of platelets was partially associated with the surface levels of CD36 and FcgRIIA. In addition to this plasma, one serum among 16 anti-CD36 Abs-containing sera tested also caused activation of platelets from healthy subjects. CONCLUSION This is the first demonstration that anti-CD36 Abs cause heterogenous platelet activation among healthy human subjects. The heterogeneity of platelet response to the anti-CD36 Abs might determine the occurrence and severity of anti-CD36 Abs-related transfusion reaction.

L Barea, R González, J L Bueno, P Rodríguez, E Castro, Centro de Donación de Sangre de Cruz Roja Española, Madrid, Spain BACKGROUND: A fast and simple method for large-scale platelet phenotyping is desirable for identifying PL A1 negative donors when transfusions of this phenotype are needed. AIM: To evaluate the usefulness of a solid phase microplate immunoassay (Capture P, Immucor), for platelet immunophenotyping and the feasibility of automation of this assay with the Galileo (Immucor) device. MATERIALS AND METHODS: The study was divided in three phases. In the first one, 104 samples of platelets rich plasma (PRP) from platelet-apheresis were tested manually with the Capture-P. These samples were incubated with the Capture-P positive control serum (human origin) included in the kit, that has an anti-Pl A1 specificity. The test was carried out following the manufacturer's instructions. In the second phase the same assay was automated with the Galileo auto-analyser, and a validation of the automation was performed. The most critical point was the adaptation of the washing steps. During the third phase, 238 PRP (obtained after maintaining the blood collected in EDTA for 8-12 hours at 4°C) were tested automatically in the Galileo device. In this case the anti-Pl A1 employed was a commercial antiserum of human origin (Immucor). All specimens that rendered a negative result with the antiserum were confirmed using NAT techniques. RESULTS: The manual method requires minimal additional equipment, but needs more time of technician dedication than the automatic one (60 vs 15 minutes in processing a total of 144 samples). The automatic method allows to test a total of 144 samples per run in 80 minutes. If a second run (94 samples) is tested it takes only 20 additional minutes. In both visual or automatic versions the interpretation of the reactions was very clear.

CONCLUSIONS: The rate of confirmed Pl A2 donors correlates well with the reported percentage in Caucasians (2.25%). The technique is easy to automate and to perform for large-scale screening. This immuno-phenotyping method let restrict the use of NAT techniques only to confirm negative or unclear results.

Y-J Chen, J Y Lyou, P S Chen, H M Liu, H Y Hu, C H Tzeng, Taipei Veterans General Hospital, Taipei, Taiwan Republic of China

The ABO blood group system is the most important blood group system in transfusion medicine. Besides the major A, B, and O alleles, many rare alleles have been reported, including Cis-AB alleles. In our institution, we have defined a novel Cis-AB Taipei allele. PCR-RFLP and PCR-SSP are useful for genotyping but unable to detect unknown substitution(s) in amplified DNA fragments. PCR-SSCP on the other hand, can be useful to identify the common ABO alleles and rare, unknown ABO alleles as well. Methods: A total of 300 AB phenotype healthy random blood donors were enrolled in this study. Four pairs of primers were designed to make up two sets of multiplex PCR which amplified four fragments that spanned the whole exon 6 and its immediate flanking regions (fragment 1, F1), and nt 431 to nt 1065, and 3¢ untranslated region of exon 7 (F2, F3 and F4) of the ABO gene. Representative DNA samples with common and rare ABO alleles, including two different Cis-AB alleles and one B(A) allele were used as controls in each SSCP run. The electrophoresis was processed with a commercial 12.5% polyacrylamide gel and run by a GenePhor electrophoresis unit at 600 V, 25 mA, 15 W for 4 hours with a running temperature of 15°C/5°C for multiplex-1/2. DNA fragments were then visualized after silver staining. For those with diverse banding patterns as compared to normal ones of A 1 B/A 1v B, SSCP analyses were performed in duplicate and samples cloned and sequenced for exon 6 and exon 7. Results: six samples were found to have different banding patterns at either F2 or F4. As shown in the Table, five novel alleles including four B alleles and one A allele were identified. Conclusions: In the present study, we have uncovered five new ABO alleles which have not been reported in the literatures. It is estimated to have a frequency of 0.0033/0.0166 for the new alleles of A/B genes for Chinese population in Taiwan. We have demonstrated that PCR-SSCP analysis is capable of detecting point mutations, including substitusion or deletion at various positions in the amplified fragments. Optimal control of the temperature during electrophoresis is essential in securing a reliable reproducibility. In conclusion, we have established a cost-effective and very informative PCR-SSCP ABO genotyping technique. Table Background: The cisAB first discovered in 1964 is a rare phenotype. Unlike the general AB blood groups (transAB), people with cisAB phenotype inherited both A and B antigens from a single parent. Serologically, cisAB phenotype is characterized by the presence of A, weakened B, and high level of H antigens on RBCs. In addition, most cisAB sera contain weak anti-B antibody. Previous publications indicated that the gene sequence of the cisAB had three types of alleles. In this study, we delineated the serological and molecular analysis of cisAB phenotype found in Taiwan. Methods: Samples were collected from two families, six members of two generations in Dai's family and five members of three generations in Li's family. Blood and saliva samples from those subjects were analyzed serologically and assayed for the activity of glycosyltransferases. Genomic DNA was extracted with a commercial kit (QIAmp DNA mini kit). PCR was performed on exon 6 and exon 7 using primers mo-46/mo-57 and mo-101/mo29a, respectively. PCR-RFLP was applied to detect the 261G deletion of O allele. The sequences of exon 7 were analyzed by a DNA autosequencer (ABI 3100). Results: Three members in the Dai's family and two members in the Li's family were defined as cisAB phenotypes. The serological results of cisAB individuals in both families showed that the RBCs reacted with anti-A strongly, but not reacted with anti-A 1 . However, the cells showed mix-field agglutination with anti-B only in the Dai's family but showed strong reaction in the Li's family. In addition, both families'sera contained weak irregular anti-B. In saliva samples of cisAB individuals in both families, A, H substances and weak B substance expressions were observed. In addition, the a-N-Acetyl-D-galactosaminyl transferase in cisAB sera failed to convert O cell to A. The exon 7 DNA sequencing data of cisAB individuals in both families were identical to A101 allele except for two nucleotide substitutions: C to T substitution at nucleotide 467 and G to C substitution at nucleotide 803. These nucleotide substitutions were previously reported by Yamamoto. Conclusion: This is the first cisAB phenotype report in Taiwan. The serological characteristics were not identical between the two families but were consistent within the same family. The same mutations were found in the cisAB phenotypes of both families.

G Wehrli, N Kaufman, K Snow, Tufts-New England Medical Center, Boston, MA Background: It is well known group O people form naturally occurring anti-A, anti-B, and anti-A,B in both IgG and IgM forms. ABO incompatibility between a mother and fetus may cause positive reactivity in the newborn's DAT, although only rarely does hemolysis requiring transfusion occur. We report a case of maternal hemolytic anti-A in a woman with a history of recurrent abortions. Case Report: A 36 year old, pregnant female presented at 15 weeks estimated gestational age for a spontaneous abortion. This was her sixth pregnancy with the same man. Her previous 5 gestations included two live children, a full term miscarriage, a 15-week spontaneous abortion and an abortion of unknown gestational age. She had no additional past medical history. In our Blood Bank and Transfusion Services (BBTS) the patient typed as Group O, Rh positive with a negative antibody screen. Interestingly, the reverse typing showed hemolysis of the A 1 reagent red blood cells (RBCs). The ABO forward and reverse typing was repeated to yield the same results, hemolysis of the A 1 reagent RBCs. The anti-A was titered to 1 : 128 at the immediate spin phase and 1 : 1024 at the antihuman globulin (anti-IgG) phase. At this point the mother's physician was notified of the hemolytic anti-A in the context of whether it might have complicated the pregnancy. The additional pregnancy history was elucidated and a sample from the father of the fetus was requested. The follow-up samples were collected 8 weeks after the abortion. The father typed as Group A, Rh positive. The mother's anti-A titer was unchanged, 1 : 128 at the immediate spin phase and 1 : 1024 at the antihuman globulin (anti-IgG) phase. The mother's serum reacted 2+ with the father's RBCs after immediate spin and then hemolyzed after a 1 minute, 37°C incubation. Follow-up samples from the two live children have been requested and are currently pending. We believe the 2 children will be Group O, explaining their live births. Conclusions: Perhaps this is not the first case of anti-A in a pregnant woman leading to fetal demise, but the interaction of the BBTS with the patient's obstetrician brought light and possible explanation to the patient's history of recurrent abortions. The obstetrician might now provide counseling for future pregnancies and avoid costly patient work-up. Again, this reminds us of the importance in leaving the comfort of our service to discuss patients with their primary physicians. Background: Several RHC/c genotyping assays have been developed on the basis of DNA sequences of individuals of Caucasian ancestry. These assays may not be applicable to populations of highly diverse ancestry because of the genetic variation in RH between different ethnic groups. Objective: We determined the RHC/c alleles in 350 unrelated blood donors with a highly diverse background and compared the results with those obtained by serological testing of RBCs. Methods: DNA samples from these blood donors were tested for C and c alleles of RHCE by two PCR assays: a multiplex PCR that includes primers recognizing a C-specific sequence in RHCE intron 2 and primers recognizing a c-specific sequence in RHCE exon 2 (Singleton et al, 2000) and a promoter region-based PCR-RFLP system (Tanaka et al, 2001) . Sequencing of RHCE exon 1, intron 2 and exon 2 was performed on discrepant samples results. Results: Phenotype and genotype results of the multiplex PCR agreed in 347/350 (99.1%) of the cases. Genotyping for RHC/c using the promoter region-based PCR-RFLP provided 76 (21.7%) discrepant results compared with those obtained by phenotyping. Sequencing of RHCE exon 1, intron 2 and exon 2 of the discrepant samples agreed with the phenotyping results in all samples except one that were phenotyped as Cc, genotyped as cc and presented the dCce allele. Conclusion: The multiplex PCR using primers recognizing a C-specific sequence in RHCE intron 2 was more reliable for RHC and RHc genotyping than a promoter-region-based PCR-RFLP assay in this population. However, caution is necessary for interpretation of RHCc genotyping in populations with genetic mixture. The frequencies of hybrid alleles as dCce s should be considered. Background: The immunogenicity of RBCs with a weak expression of D is presumed to be lower than that of RBCs with normal expression of D. The weak D phenotype is caused by many different RHD alleles encoding altered RhD. Among the common weak D types in whites, only weak D type 2 has showed to induce alloanti-D immunization (Flegel et al, 2000) . It has been postulated that sufficiently potent monoclonal anti-D would detect all immunogenic weak D samples, even without the antiglobulin test (AGT). Case report: A donor phenotyped as RhD-negative at his first donation with two commercial IgM monoclonal anti-D reagents, was found to be weak D with an IgG monoclonal anti-D in conjunction with the antiglobulin test at his second donation. The weak D type was determined by a PCR system using sequence specific primers that detect the common weak D types (Wagner and Gassner, 2001) as weak D type 1. This result was confirmed by sequence performed on PCR products amplified from genomic DNA using RHD-specific primers for exon 6, which covers a missense mutation (809 T > G) diagnostic for the weak D type 1 allele. A look-back was conducted for the transfusion recipient who showed alloanti-D in her serum. This recipient was exposed to other 4 allogeneic blood units that were rechecked for Dtype and confirmed to be D-negative by phenotyping and genotyping. The weak D type 1 RBC unit, initially assumed to be D-negative had by flow cytometry, a mean fluorescence intensity comparable to that of the weak D type 2 used as control. Conclusion: Our results indicate that weak phenotypes may cause alloanti-D immunization even with reduced D antigen densities. IgM monoclonal anti-D reagents cannot detect weak or very weak expression of D. The sensitivity of detecting weak D depends on monoclonal anti-D reagent and on the exact conditions of the methods.

T J Legler, E Binder, Department of Transfusion Medicine, Goettingen, Germany; E Smart, South African National Blood Service-East Coast Region, Pinetown, South Africa; M Prager, BAG, Lich, Germany; J-H Maas, Department of Transfusion Medicine, Goettingen, Germany Background: In individuals from African descent the RHD and RHCE locus is highly polymorphic. In order to evaluate the reliability of newly developed PCR-SSP assays, Rh phenotype and genotyping results were compared. Study Design: EDTA anticoagulated blood was obtained from South African blood donors of which 53 were Blacks and 45 of mixed race, referred to as "Coloured" in South Africa. After serological testing for ABO and D, samples were shipped to Germany and serological testing of the Rh factors was completed within two months after blood sampling using gel test cards (Diamed, Cressier, Switzerland) . DNA was automatically extracted with magnetic beads (Agowa, Berlin, Germany). Samples were tested with SSP kits ( The prediction of the D, c, E and e phenotype from DNA with SSP seems to be reliable in Black and Coloured South African blood donors. Since two independent SSP methods indicated the presence of the Cde s allele in a donor who tested Cnegative with stored red cells, a repeat phenotype test is required in order to interpret this single discrepancy in this study.

L Castilho, A Rodrigues, J Pellegrino Jr., F F Costa, Hemocentro, Unicamp, Campinas, Brazil Background: The D-phenotype may be caused by the lack of functional RhD protein or by the presence of aberrant forms of RhD that do not express D antigen. RHD positive antigen D negative alleles can be frequent in populations of African admixture and in these populations several RHD genotyping systems are associated with a high rate of false positive results. The specificity of RHD genotyping can be improved by a systematic characterization of RHD positive antigen D negative alleles. Such systematic knowledge could have considerable impact for typing and transfusion strategy in a mixed population as the Brazilian population. Objective: We examined the RHD alleles of 395 Brazilian blood donors of heterogeneous ethnic origin phenotyped as D-negative in order to determine the molecular cause of such alleles and their frequencies. Methods: RHD genotyping was performed on DNA samples from 395 D-blood donors and tested for intron 4, exon 10, dCce s and common weak D types by PCR-SSP, RHDc and hybrid genes by multiplex PCR, dispersed missense mutations by PCR-RFLP and sequencing to identify RHD alleles. Results: RHD was deleted in 364 (92.2%) of the 395 D-blood donors studied. We detected 31 (7.8%) RHD positive samples. Four samples (1%) were characterized as weak D (3 type 2 and 1 type 5), 3 (0.76%) as partial D (1 DAR, 2 DVI), 11 (2.8%) as RHDc and 8 (2%) as dCce s . Two samples (0.5%) presented the RHD(48 G>A) in exon 1. The molecular basis of 3 samples is still pending. Conclusions: RHD deletion is the most frequent cause for the absence of a functional RhD protein in Brazilians, followed by RHDc, dCce s and RHD(48G>A) alleles. The high frequency of weak D and partial D found in these donors classified as D negative shows the importance of re-evaluate the reactivity of commercial anti-D reagents used in the routine. Our findings are important to establish reliable genotyping and phenotyping in this population. L Castilho, Hemocentro, Unicamp, Campinas, Brazil; M Rios, DETTD, CBER, FDA, Rockville, MD; J Pellegrino Jr., F F Costa, Hemocentro, Unicamp, Campinas, Brazil Background: People with partial D antigens such as D category IIIa and DAR can produce anti-D if exposed to D-positive red blood cells. DIIIa is characterized by three point mutations in RHD: 455 A>C (exon 3); 602 C>G (exon 4) and 667 T>G (exon 5) and it is usually recognized only after the patient makes anti-D since anti-D monoclonal antibodies do not readily distinguish the D category IIIa from D+. The partial D allele DAR presents common polymorphic sites in the exon 4 and exon 5 of the RHD and presents an additional polymorphism (1025 T>C) in exon 7. DAR shows weaker reactions with monoclonal and polyclonal antisera against D but any current serological typing does not recognize it. This study aimed to investigate the occurrence of RHDIIIa and DAR alleles in a cohort of SCD patients phenotyped as D+ and weak D. Methods: DNA samples from 130 SCD patients were tested for 455A>C, 602 C>G, 667T>G and 1025T>C by PCR-RFLP and sequence analysis. Results: The PCR-RFLP showed that 8 (6.2%) of the SCD patients had D variants. Sequencing results showed that 2 (1.5%) samples were heterozygous RHDIIIa, 4 (3.1%) were heterozygous DAR and 2 (1.5%) samples carried a partial D with 4 mutations: 455A>C (heterozygous), 602C>G and 667T>G (homozygous) and 1025T>C (heterozygous). These results suggest that these patients carry one RHDIIIa allele and one DAR allele. Anti-D was detected in the serum of three of these 8 patients. Conclusion: These findings suggest that SCD patients who are candidates for chronic transfusion, may benefit from genotyping for RHDIIIa and DAR to prevent alloimmunization.

M L Virata, M-Y W Yu, US Food and Drug Administration, Center for Biologics Evaluation and Research, Rockville, MD Background: Some adverse reactions to intravenous immunoglobulin (IGIV) therapy are hemolytic events due to anti-D. Few manufacturers of USlicensed IGIV products have set specifications for the maximum permissible concentration of anti-D. In the UK, Thorpe et al (2003, Vox Sang 85:80) found that two lots associated with hemolytic events had high levels of anti-D (endpoint dilution titers ≥64). Using a simple microtiter plate hemagglutination assay, they established a positive reference, i.e., a titer of 8, suitable for defining the upper limit of anti-D in IGIV products distributed in Europe. Methods: A total of 125 lots of 7 different IGIV products made by US-licensed manufacturers in 2000-2003, plus 14 lots of 3 newly licensed IGIV products, were tested for anti-D according to the hemagglutination assay of Thorpe et al. All lots were diluted to a starting concentration of 5% IGIV, after which 2fold serial dilutions were prepared prior to testing. Reference preparations NIBSC 02/228 and 02/226 described in Thorpe et al were used as positive and negative controls, respectively. Results: The majority of the IGIV lots (102/139, 73%) had no detectable anti-D. Of the 37 lots that were anti-Dpositive, 32 had titers £4. The highest observed titer was 16, which occurred in one lot. That one lot was a product made from recovered plasma not screened by the manufacturer for anti-D. That particular product also had the highest prevalence of anti-D (17/19, 90%), with titers ranging from 2 to 16. In general, manufacturers who tested their starting plasma material or final-container products produced IGIV with little or no anti-D. Among the 4 products with low levels of anti-D, two of the manufacturers tested each lot for anti-D before release, whereas another manufacturer screened its firsttime donors for circulating anti-D. Conclusions: IGIV products currently made by US-licensed manufacturers do not have appreciable levels of anti-D, although there are differences in the prevalence of anti-D-positive lots. Because IGIV doses vary depending on clinical indication, it would be difficult to set a limit of anti-D in IGIV that would circumvent all hemolytic events. Nevertheless, to ensure product safety, steps, such as plasma and finalcontainer screening, may need to be taken to minimize anti-D levels.

L R Dake, W J Judd, R D Davenport, University of Michigan Hospitals and Health Centers, Ann Arbor, MI BACKGROUND: Using low-ionic-strength tube (LIS-tube) indirect antiglobulin tests (IATs), Shirey et al. found the incidence of anti-c in R 1 R 1 (DCe/DCe) patients with anti-E to be 32% (Transfusion 1994; 34:756-8) . We report a much higher incidence of anti-c in such patients when using gel technology in conjunction with ficin-pretreated RBCs. METHODS: Results of antibody identification studies on R 1 R 1 patients with anti-E (n = 82) were reviewed. Ficin-treated RBCs and low-ionic-strength saline (LISS, Löw and Messeter formulation) were from Immucor/Gamma, Norcross, GA. ID-MTS gel technology was from Ortho, Raritan, NJ. IATs with untreated RBCs were performed by LISS-tube technique (10¢ at 37 C, anti-IgG) or on anti-IgG gel cards. IATs with ficin-treated RBCs were performed on anti-IgG gel cards; in some cases they were also done by tube technique (30¢ at 37 C, anti-IgG + C3). RESULTS: LISS-tube or gel IATs with untreated RBCs demonstrated anti-c in 32 patients with anti-E. A further 21 patients with anti-E were found to have anti-c demonstrable in ficin-gel tests, among these, weak anti-c was also seen by gel with untreated RBCs in 2 cases, and in 5 cases by ficintube IAT. No anti-c was found by ficin-gel in 29 additional R 1 R 1 patients with anti-E. CONCLUSION: The incidence of anti-c among R 1 R 1 patients with anti-E in our study was 32/82 (39%) by routine methods, and 53/82 (65%) when the ficin-gel data are included. The latter is significantly higher than that reported by Shirey et al., (p = 0.0001). Accordingly, we now test all R 1 R 1 patients with anti-E for anti-c by ficin-gel technique. Our data strongly support the selection of R 1 R 1 blood for all c-negative patients with anti-E.

L Castilho, I Machado, J Pellegrino Jr., R Barini, Hemocentro, Unicamp, Campinas, Brazil Background: Noninvasive prenatal identification of fetal RHD status is a goal of obstetrical practice, in order to prevent maternal immunization and to help in the management of alloimmunized pregnant women, without inherent risks of the invasive procedures. We assessed the accuracy of fetal RHD genotyping by analysis of distinct RHD regions from maternal plasma in a mixed ethnic population. Methods: We analyzed plasmas from 81 RhD-negative pregnant women between 4 and 36 weeks of gestation. Peripheral or umbilical cord bloods from respective neonates were used as control. We tested for intron 4, exon 10 by PCR-SSP and for RHDc and dCce s alleles by multiplex PCR Commercially available kit were used to extract DNA from maternal plasma (Easy-DNA Invitrogen ® ). Results: Two samples presented the RHDc and 1 sample presented the dCce s allele. These three samples showed to be RhD-negative by hemmaglutination at birth. Discrepancies were observed between intron 4 and exon 10 RHD regions results in 8 samples. According to neonatal typing better results were obtained by using intron 4 gene region. General accuracy of fetal RHD genotyping in this mixed population was 97.3%, sensitivity of 98.3% and specificity of 93.8%. Conclusion: Our findings indicate that fetal RHD genotyping of maternal plasma should take account for RHDc and dCce s alleles and that the accuracy of isolated RHD gene regions is insufficient for clinical application in a mixed population. Exhaustive analysis of polymorphisms, population frequency and distribution of RHD alleles are required. Background The low incidence Rh antigen Crawford is present in less than 1% of blacks and is assigned to ISBT number RH43. Since the first report in 1980 (Transfusion 1980:20; 631) , the antibody has been described in many manufacturer examples of monoclonal anti-D. We describe the characterization and management of a donor with RH43. Case Report A regional blood center had an African-American donor with a history of typing as Gp A, D Negative. Upon a subsequent donation tested at a centralized donor testing lab, the donor's red cells and serum typed as Gp A, with a questionable D typing. The donor's D testing at immediate spin (IS) was weakly reactive with Gamma Biologicals, Inc anti-D; however, the weak D typing was non-reactive at AHG. The donor's red cells were then tested with second source Gamma Biological, Inc. anti-D and Immucor anti-D. The donor's D testing at IS was reactive with the second source Gamma anti-D antisera and was negative at AHG, similar to the above stated results. The donor's D testing at immediate spin and at AHG was non-reactive at both phases using the second source Immucor anti-D antisera. The donor's red cells were then Ficin-treated and testing was repeated with all three anti-D sources. The donor's Ficin-treated red cells were reactive at immediate spin and at AHG using both Gamma sources of anti-D; however, the donor's Ficin-treated red cells were non-reactive at AHG using the Immucor anti-D antisera. After consultation with Immucor/Gamma, the donor D typing was consistent with the Crawford antigen. The donor was labeled as Gp. A, D Positive. The submitting regional blood center phenotyped the donor as positive for RH43, re-confirmed the D typing pattern obtained by the regional donor testing lab with different sources of anti-D and determined that the donor cells were unable to adsorb and elute anti-D. The regional blood center notified the donor of the discrepancy in the donation typing (A, D Positive) and the patient typing (A, D Negative). The donor was informed that if blood transfusion was necessary, only D-Negative blood should be transfused. Conclusions The serology completed by the centralized donor testing lab and the regional blood center were similar to previously reported examples of RH43. The coordination and communication of the donor typing results from the centralized donor testing lab and the regional blood center were critical for the correct labeling of the donated products and the management of the donor.

R M Kakaiya, J Cseri, B Jochum, LifeSource Blood Services, Glenview, IL; L Gillard, S Silberman, Loyola University Medical Center, Maywood, IL Introduction: The e antigen is present in 98% of the Caucasians. Individuals with a variant form of the e antigen that lacks hr S could make allo-anti-hr S due to pregnancy or transfusion. We report a case of maternal allo-anti-hr S that did not cause HDN. Case report, methods and results: A 24-year old Gravida III and Para III African American full term gestational woman was admitted for C-section delivery. Her past history included the presence of anti-e during her second pregnancy. The second child was delivered by C-section and the newborn did not show any signs of HDN. There was no information available to determine if the first child was affected by the HDN. On current admission of the patient, her serum showed RBC antibody (1 to 2+ strength) with apparent anti-e specificity on routine LISS and PEG panels. Reactions with the PEG panel were stronger than with the LISS panel. Her RBCs were e+. The presence of anti-e in a patient who was e+ alerted us to the possibility that our patient could be e variant. Therefore, we performed additional studies. The maternal serum gave negative reactions with two cells that were e+ and hr S -. Also, the maternal serum gave positive reactions with two cells that were e+ and hr B -, thus ruling out the presence of anti-hr B . A healthy full term child was delivered by C-section. The newborn's hemoglobin level at birth was 18.8 gm/dL. The DAT on cord blood was negative. Total circulating bilirubin levels on the first day of life ranged from 2.5 to 4.0 mg/dL. No clinical evidence for hemolysis was detected. At 2-months of age the infant's RBCs were typed and they were e+ and hr S +. Discussion: Serological and clinical aspects of our case indicate that our patient had allo-anti-hr S in the serum and that this antibody did not cause HDN. This was our patient's third pregnancy and delivery. At least the last two children were born without any evidence for HDN. It has previously been shown that the individuals who produce anti-hrS appear to have inherited certain mutational alleles that allow them to recognize certain antigenic epitopes. Therefore, it is possible that the clinical signficance of the allo-anti-hr S may be dependent upon the specificity of the antibody directed against one or more of the antigenic epitopes. Our case confirms the lack of clinical signficance of allo-anti-hr S and this may be due to the variability in the antibody specificity.

A P Cozac Jr., A Zanelli, T Rodrigo, D Covas, Hemocento de Ribeirão Preto, Ribeirão Preto, Brazil; L Castilho, Hemocento deda UNICAMP, Campinas, Brazil Background: Up to the present time, little is known about D partial individual susceptibility to D antigen alloimmunization and about the hemolytic power of the anti-D developed by them. We reviewed six D antigen alloimmunization developed by D partial patients. Study design: 6 D partial patients at polytransfusion therapy were analyzed after the anti-D development. Each patient was diagnosed as D partial after genotyping by AS-PCR and RFLP-PCR techniques. Three out of 6 patients received RhD+ RBCs by mistake after the diagnosis. Each case was reviewed in order to verify the length of time the antibody could still be detected at the serum after its development. The hemolytic power of the anti-D was indirectly analyzed by the increment count after the RhD incompatible transfusions. Results: The average time it took between the fist transfusion event and the D antigen alloimmunization development was 10,17 years (1-17). Five patients (83%) developed other alloantibodies before the anti-D. The other one (17%) developed only anti-D. The average of RhD+ RBC transfusion events before the alloimmunization was 120 (12-219). After the establishment of the diagnosis three patients out of 6 were transfused with RhD+ RBCs by mistake. The first one received 2 units in a single transfusion event. Despite the incompatible crossmatch due to the presence of anti-D at the serum, the patient presented an expected increment in hemoglobin. The second one received 126 RhD+ RBCs in a total of 32 transfusion events presenting the expected increment in hemoglobin after each episode and in spite of repeated antigenic stimulus, the anti-D never reappeared at the serum. The last one received 17 RhD+ RBCs but it was impossible to analyze the increment because of a colon angiodisplasy bleeding, however the anti-D wasn't identified at the serum. Anti-D was detected just once in two patients and in the other 4 the average of it's detection was 11,7 months (3-32).

The anti-D development susceptibility was inferior when compared with the others antibodies. Anti-D didn't have any clinical significance in all cases presented.

C M Jenkins, S T Johnson, J L Gottschall, Blood Center of Southeastern Wisconsin, Inc., Milwuakee, WI Background: The incidence of weak D has been reported to be between 0.23 and 0.5 percent in Europe and 0.4 percent in the United States. All studies have been done before the introduction of potent monoclonal anti-D reagents. Using today's commercial reagents our laboratory evaluated Rh positive samples for the presence of weak D. Methods: Rh positive donors, typed by the Olympus PK7200 using diluted monoclonal blend anti-D (Immu-corGamma, Norcross, GA) and diluted human polyclonal anti-D (Immu-corGamma, Norcross, GA), were selected by sampling batches of 100-200 samples from the previous day's collection. Anti-D reagents used on the Olympus PK 7200 are required to detect weak D positive red cells which do not agglutinate at immediate spin (IS) when tested with human polyclonal anti-D by manual tube methods. Greater than 95% of donors tested were Caucasian. Using tube methods with two different monoclonal blend anti-D reagents (ImmucorGamma, Norcross, GA) and one human anti-D typing reagent (ImmucorGamma, Norcross, GA), the presence or absence of the D antigen was evaluated after the IS reading. Donors found negative or weakly positive (<2+) at IS were further typed for weak D by the indirect antiglobulin test. The weak D samples were RHD genotyped by allelespecific PCR. Results: Of 1,005 Rh positive donors tested, 4 (0.4%) were classified as weak D positive by one or more anti-D reagents. The table below describes the varied reactions with the different reagents.

Human Anti-D reagent demonstrated weaker reactions at IS followed by the Immucor Anti-D (Series 4) Monoclonal Blend. Gamma-Clone ® Monoclonal Blend had the strongest reactions. All weak D positive samples were found positive for exon 4, intron 4, and exon 10, a finding consistent with most Rh D-positive samples. Conclusion: The incidence of weak D (0.4%) has not changed significantly from earlier studies using human polyclonal reagents. System antigens tested, however, no expression of Kx antigen was observed. A family study was performed by: serological techniques, flow cytometry, Kell system genotyping, Western immunoblotting for Kell/XK protein complex, Southern Blot, PCR and sequencing of XK gene in order to investigate a possible deletion. Serum enzymes and peripheral blood smears were also analyzed. Results: One apparently healthy blood donor demonstrated low expression of K:2, K:4, K:5; K:7; K:14; K:22 and no Kx antigen in his RBCs. His brother and mother showed the same results and his father showed normal expression of these antigens. Both siblings presented elevated serum creatine kinase and rare acanthocytic RBCs. Flow cytometry studies confirmed mother's status as a McLeod carrier female. Genotyping determined the presence of KEL2 and KEL4 alleles in mother and siblings. In the Western blot, the father was the only one presenting a normal Kell/XK complex. Southern blot with an exon 1 probe showed fragments shorter than predicted for the siblings and the mother, suggesting a deletion. PCR with primers spanning exon 1 and flanking regions displayed a similar pattern. DNA sequence allowed the precise characterization of a deletion of 392 nucleotides, beginning 5¢ of the coding region up to nucleotide 201 of exon 1, which putatively abrogates the production of XK protein. Conclusion: We identified two McLeod phenotype in a Brazilian blood donor population. The molecular basis for this phenotype is a 392 bp deletion of the XK gene, spanning from upstream of the coding region to exon 1, never described before.

B J Bryant, S L Weber, A J Indrikovs, University of Texas Medical Branch, Galveston, TX Background: Few cases of antibodies to the Cromer (a) antigen have been described during pregnancy. In these, the IgG Anti-Cr a titers decreased during pregnancy, and although the newborns were Cr a positive, the direct antiglobulin tests (DAT) were negative and hemolytic disease of the newborn (HDN) was not observed. The Cromer blood group system antigens reside on the short consensus repeats of the decay accelerating factor (DAF), which is preferentially expressed on the apical aspect of the fetal derived syncytiotrophoblast layer of the placenta. It has been postulated that Cromer antibodies are not transported to the fetus, but are bound to placental DAF thereby protecting the fetus from HDN and causing the observed disappearance of Cromer antibody in maternal plasma during pregnancy. This case report is the first to demonstrate Cromer antibody sequestration by the placenta. Case Report: A G4P1 woman with an anti-Cr a presented for prenatal care during her 4 th pregnancy. Anti-Cr a had been identified during her 3 rd pregnancy with a titer of 128 at 11 and 15 weeks and a titer of 8 at 20 weeks gestation. The 3 rd pregnancy ended at 22 weeks gestation with the delivery of a non-viable male infant. Her 4 th pregnancy was conceived 4 months after the delivery of the 3 rd infant. The anti-Cr a titer decreased during her fourth pregnancy from 64 at 7 weeks gestation to undetectable after 25 weeks. This pregnancy was complicated by multiple episodes of preterm labor, but resulted in the delivery of a healthy female infant at 37 weeks gestation with no evidence of HDN. At delivery, the infant had a negative DAT, and the maternal plasma, cord plasma, and the cord blood eluate were negative with screening cells, the infant's cord cells, and the mother's cells. To demonstrate sequestration of the maternal anti-Cr a by the placenta, placental eluates were prepared using EluKit TM II (Gamma Biologicals, Inc., Houston, TX) and revealed the presence of anti-Cr a . Conclusions: This is the 4 th case report of a Cromer (a-) woman producing anti-Cr a during pregnancy. As in the previously reported cases, the anti-Cr a titer decreased during the pregnancy, and the viable infants delivered had no signs of HDN. This is the first report demonstrating anti-Cr a sequestration in the placenta. The presence of anti-Cr a in the placental eluate, but not in the cord plasma, maternal plasma, or cord blood eluate strongly supports the hypothesis that DAF at the fetomaternal interface absorbs anti-Cr a from the maternal circulation blocking its passage to the fetus and protecting the fetus from HDN.

L Cooling, University of Michigan Hospitals, Ann Arbor, MI Background: LKE antigen is a globo-ganglioside related to the P blood group family and is the reported physiologic receptor for P-fimbriated, uropathogenic E. coli. In adults, LKE is present on 98-99% donor RBC, with 80-90% typing as LKE-strong (LKE-S), 10-20% LKE-weak (LKE-W) and 0.7-2% LKE-negative (LKE-N). We previously found a correlation between LKE-W and Lewis null, due to an allelic variant in _1,3 galactosyltransferase III (B3GalT3), which synthesizes both type I chain and galactosylgloboside precursors. Because Lewis antigens are differentially expressed during early childhood, we screened cord RBC for possible LKE expression. Methods: Adult (n = 449) and cord RBCs (n = 122) were typed for LKE, ABO and Lewis antigens by hemagglutination. LKE typing was performed with washed, ficin-treated RBC and monoclonal antibody MC813-70. Donors were typed as LKE-S, LKE-W and LKE-N based on the presence/absence of agglutination at immediate spin and IAT. All LKE-W and LKE-N samples were typed twice. Results were compared by chi-square (P < 0.05) and odds ratio (OR, 95% CI). Results: LKE is expressed on cord RBCs. There was a relative increase in LKE-W phenotype among newborns. In adults, Le (ab-) was increased among LKE-W donors (P = 0.006; OR 3.85, [1.26-11.57, 95% CI]). There was no association between Le (a-b-) and LKE in cord samples (P > 0.50). There was also no association between LKE type and ABO.

Conclusions: Like P k and P antigens, LKE antigen is synthesized and expressed on cord RBC. There appears to be an increase in LKE-W phenotype among cord samples. Unlike adults, there is no correlation between Le (a-b-) phenotype and LKE-W.

Serological Evaluation of Two New Monoclonal Anti-Ok a G R Halverson, R W Velliquette, A Schawalder, New York Blood Center, New York, NY Background: Anti-Ok a was first described in 1979 (Morel et al. Vox.Sang. 1979 . Since that time only 8 probands have been identified, all of Japanese extraction. The Ok a antigen (ISBT 024.001) is located on CD147 glycoprotein. The purpose of this study was to characterize two new monoclonal antibodies (Mabs), 4C5 and 8A4, and to compare their reactivity with the original human polyclonal anti-Ok a . Methods: Mice were immunized with transfected cells to produce monoclonal antibodies to various red blood cell (RBC) antigens. Mouse spleen cells were fused with mouse myeloma X63-Ag8.653 cells using PEG 3350 to form antibody-secreting hybridomas. The Mabs were tested by indirect antiglobulin tests (IAT) in v-well trays, in test tubes, and in gel cards with anti-mouse IgG as the secondary antibody. Normal RBCs and those pre-treated with papain, trypsin, a-chymotrypsin, pronase and neuraminidase as well as 0.2M DTT were tested with 4C5, 8A4 and the original anti-Ok a . Results: Both Mabs reacted with normal adult (n = 12) and cord (n = 1) RBCs by tube IAT with 4C5 giving the stronger reactions (3+), and 8A4 giving slightly weaker reactions (2+). The Mabs did not react with one example of an Ok(a-) RBC. In addition, they were tested with rare cells that lack antigens of high prevelence (n = 18) and were found to be equally reactive. We tested RBCs of the following phenotypes: K 0 , Rh null , O h , Lu(a-b-), Jk(a-b-), Co(a-b-), Di(b-), Kx-, Vel-, Lan-, Jr(a-), Cs(a-), Wr(b-), Dr(a-), Gy(a-) and one example of PNH-III RBCs. Both 4C5 and 8A4 were resistant to enzyme treatment with papain, trypsin, a-chymotrypsin and neuraminidase. Interestingly, 4C5 was non-reactive with pronase-treated RBCs while 8A4 demonstrated a radical reduction in strength (from 2+ to microscopic) with the pronase-treated RBC sample. Similarly, a reduced reactivity was observed with 4C5 when tested against RBCs treated with papain. The strongest reactivity with 4C5 was observed with trypsin-treated RBCs, and 8A4 was strongest with neuraminidase-treated RBCs. One example of M k M k RBCs reacted stronger with 8A4 which correlated with the observed reaction of 8A4 with neuraminidase-treated RBCs as both have markedly reduced levels of sialic acid. The reactivity of the human polyclonal anti-Ok a showed the strongest reactions with trypsin-treated RBCs, and unlike Mabs 4C5 and 8A4, the polyclonal anti-Ok a reactivity was not reduced in testing with pronase-treated RBCs. Conclusion: Mabs 4C5 and 8A4 are specific for the Ok a antigen. Both reacted in gel cards and by tube agglutination testing and produced comparable results. Both Mabs should prove to be useful for antigen typing blood to find additional Ok(a-) blood donors.

Background: Two isotypes of C4 exist and carry different blood group antigens. C4A usually expresses Rg while C4B expresses Ch. A previous study conducted among an elderly Hungarian population reported a decreased frequency of C4B null alleles (C4B*Q0, Ch-) and suggested that C4B*Q0 could be a genetic risk factor for early death. The purpose of our study was to determine if the frequency of C4B*Q0 among a mixed population of elderly Caucasian Americans was similarly decreased. Materials: 287 individuals participated in this study, including 119 above the age of 65 (elderly group), and 168 under the age of 65 (young group). The elderly cohort included both healthy individuals and those confined to nursing homes. Methods: C4 allotyping was performed using neuraminidase and carboxypeptidase treated EDTA plasma followed by high-voltage agarose gel electrophoresis. The gels were immunofixed, stained and allotype assignments made using accepted nomenclature. In addition, total C4 levels were quantified using radial imunodiffusion. Results: The frequencies of either a C4A*Q0 or C4B*QO allele did not differ between young and old (C4A*Q0 = 22.6% vs 27.8% and C4B*Q0 = 17.3% vs 16.0%). When analyzed by a Chi Square test, the differences between the young and elderly groups were not statistically significant. However, the Rg-phenotype (homozygous C4A*Q0) was markedly increased (19.5%) in a healthy elderly group while the Ch-(homozygous C4B*Q0) was absent. The average total C4 levels in the elderly group were 36.9 mg/dl for the individuals not having C4 null genes and 27 mg/dl for the individuals having C4 null alleles. These data were similar to the young group where the average total C4 levels were 33.4 mg/dl and 28.6 mg/dl, respectively. Conclusion: We conclude that there is no significant difference between young and elderly individuals with regard to the frequency of C4 null genes or in total C4 plasma levels. However, the increased frequency of the Rg-phenotype in the healthy elderly cohort is intriguing as this type previously has been linked to autoimmune disease. Background: Typing red cell samples for antigens in the Dombrock blood group system is notoriously difficult due to a dearth of reliable anti-Do b . For this reason, DNA-based assays were used to type donor and patient samples. Our purpose was to make a monoclonal (Mab) anti-Do b for use in phenotyping. Methods: Balb/c mice were immunized with transfected 293T cells expressing the Do b antigen. Hybridomas were produced in fusions with X63-Ag8.653 mouse myeloma cells using PEG 3350. A resulting Mab (Clone 6B5) was made by standard cell culture methods. Hemagglutination was done in gel cards using anti-mouse IgG. We compared the results of RBC phenotyping using the Mab anti-Do b in125 patient and donor samples to DNA assays performed by PCR-RFLP for single nucleotide polymorphisms of the DO. Results: The isotype of Mab 6B5 is IgG2a k. In initial testing, undiluted 6B5 agglutinated all RBCs except three examples of Gy(a-) RBCs. It also agglutinated RBCs of common phenotype treated with papain or achymotrypsin, but did not agglutinate RBCs treated with trypsin or pronase. At a dilution of 1 : 20 in gel cards, 6B5 is anti-Do b , and this was validated by testing 28 samples of known type [11 Do(a-b+), 7 Do(a+b+), 10 Do(a+b-)]. As expected, RBCs from people with the HY/HY or HY/JO genotype reacted weakly or not at all with the antibody. The result of testing the 125 samples is given in the table. Three DOB/DOB samples (indicated with *) were weakly-reactive or non-reactive with 6B5. Conclusion: Mab 6B5 a useful tool for phenotyping RBCs. However, like human anti-Do b , some RBCs that are expected, based on DNA analysis, to be Do(b+) are non-reactive or weakly reactive with Mab 6B5. The cause for this weaked expression is under further investigation. Introduction: IgA autoantibodies are not uncommon in patients with warm autoimmune hemolytic anemia. Usually, the specificities of IgA autoantibodies are associated with the Rh system; autoantibodies associated with the Gerbich and Vel systems have been described. We report an unusual case of an IgA autoantibody having anti-"N" specificity. Case report and results: A post splenectomy sample of a 32 year old female, presenting with warm autoimmune hemolytic anemia (AIHA), was referred to our Blood Centre, because of a positive antibody screen found using a solid phase technique at the referring hospital. The patient was B+, R 1 R 2 , K-, M+ N-S+. The DAT results were IgG 3+, IgA 4+. Strong pan-reactive antibodies against papainized cells and microscopic IAT reactions using polyspecific antiglobulin reagent were seen. Autoantibodies reacting strongly by IAT (using anti-IgA) were detected in the serum and eluate. Titration of an eluted antibody against selected panel cells showed a titre of 32. IBGRL confirmed only Ss-U-M+ N-cells were compatible with the patient, but not S-s-U-M+ N+ blood. The specificity was subsequently confirmed to be auto anti-"N". Due to the rarity of S-s-U-N-blood, N-K-blood was used for transfusion. A total of 12 units of blood were transfused without any signs of overt haemolysis or other evidence of a hemolytic transfusion reaction (HTR). Conclusion: Anti-"N" (MN30) is a rare antibody than can be produced in an individual who lacks red cell membrane glycophorin B, phenotyping as S-s-N-. We believe that this case is the first example of an IgA auto anti-"N". There are insufficient data to allow clinical judgement as to the significance of anti-"N", no HTRs or cases of haemolytic disease of the newborn caused by the antibody have been reported to date. The evidence presented here indicates that patients with IgA auto anti-"N" may be safely transfused with N-S+/s+ blood. Background: Autoimmune hemolytic anemia can be caused by a variety of underlying illnesses. We report a case of an auto anti-N which caused a severe hemolytic anemia. Case Report: A 6 year-old female (28.9 kg) with history of vomiting, fever and abdominal pain presented to a local hospital with initial Hct of 8% and abnormal liver function tests. She was transferred to our hospital and her admission lab values were: Hgb = 2.7 g/dL, Hct = 7.4%, WBC = 22.5 ¥ 10 9 /L, lactate dehydrogenase (LD) = 7856 IU/L, and total bilirubin = 12.3 mg/dL. Peripheral smear showed lymphocytosis, polychromasia and varying shapes and sizes of RBCs. Anti-nuclear antibody was positive (speckled pattern), blood cultures and other immunology tests were negative. Two O negative emergency release RBCs were transfused upon admission. Antibody screening showed a 1+ reaction with 2 cells of a 3 cell screen at 37 C using the ID-MTS gel test. Antibody panel results showed reactivity with N positive cells at RT and 37 C polyethylene glycol (PEG) incubation. The auto-control was positive at RT and 37C PEG incubation. The direct antiglobulin test (DAT) was 4+ with polyspecific antihuman globulin (AHG), 2+ with anti-IgG, and 3+ with anti-C3d. The patient's RBCs phenotype was N positive. ABO/Rh typing was initially discrepant with forward typing and the Rh control was positive. A one-hour chloroquine treatment, resolved the forward typing and Rh control problem and patient was typed as A positive. The antibody was interpreted as a Warm Autoimmune Antibody with N specificity. An AHG crossmatch of the two O negative emergency release units showed one of the units incompatible with the patient's serum. Further testing determined this unit was N positive; the other unit was N negative and compatible. The patient received one additional N negative AHG crossmatch compatible RBC unit, corticosteroid therapy was administered and the Hgb increased to 12.9 g/dL. Eight days following admission, a DAT was negative with polyspecific AHG. Patient was discharged home on corticosteroids with lab values of: Hgb = 13.2 g/dL, Hct = 36.7%, WBC = 9.2 ¥ 10 9 /L, LD = 1773 IU/L, and total bilirubin = 1.4 mg/dL. Follow up transfusion testing 2 months later showed a negative antibody screen, negative DAT and no ABO/Rh typing discrepancy. Conclusion: Following corticosteroid therapy, the autoimmune hemolytic anemia resolved with disappearance of the auto anti-N. Etiology of the autoimmune hemolytic anemia was assumed to be related to the acute illness which was suspected to be of viral origin. There have been at least 3 other case reports of auto anti-N hemolytic anemia published. Two of 3 were in pediatric patients (<18 years old) and as in our patient, all had antibody resolution following steroid therapy. The Post serum sample antibody screen was positive only in PEG-AHG; LISS-AHG & Gel were both negative. The Post serum was incompatible with unit #4 only, which was Fy(a+b-). Units #1 and #2 transfused on 3/30 were both Fy(a+); yet, compatible with the Post serum sample. Unit #2 was Fy(a+b-) the same as #4. The follow-up sample antibody screen was positive by all testing methods. Only unit #4 was positive in the LISS-AHG testing phase. All units were incompatible in PEG-AHG crossmatch. Fy(a+) and S+ transfused cells remained in the patient's circulation 10 days posttransfusion. This case demonstrates that no single testing technique will identify all antibodies & the importance of follow-up samples for FNTR's.

T J Greenwalt, U J Dumaswala, L C Tse, M M O'Leary, Hoxworth Blood Center, Cincinnati, OH Background: No previous studies measured the transfer of cholesterol between RBCs. Transfer to small unilamellar vesicles and plasma has been reported. The aim was to investigate the possibility of transfer of cholesterol between RBCs suspended in an ionic medium. Methods: Group A RBCs were labeled by incubating with [4-14C]-cholesterol for six hours at 37°C with constant agitation (n = 4). They were then washed repeatedly and resuspended in Basal Eagles Medium. Mixtures of equal volumes 10% suspensions of the labeled group A and unlabeled group O cells were incubated for 32 hours at 37°C in a shaking-water bath. The A cells were agglutinated with a monoclonal anti-A and trapped in glass bead columns that allowed the O cells to pass through freely. Flow-cytometry established that the recovered O cells had no contaminating A cells. Liquid scintillation spectrometry was used to measure the amount of cholesterol transferred to the lipid extracts of the O cells (n = 4). Results: 48 ± 1.6% SD of the labeled cholesterol was transferred from the labeled A cells to the unlabeled O cells. Conclusion: Cholesterol was demonstrated to transfer from one intact population of RBCs to another in an ionic medium by contact alone.

T J Greenwalt, U J Dumaswala, L C Tse, Hoxworth Blood Center, Cincinnati, OH Background: Transfer between the erythrocyte and the free cholesterol of serum has long been known (Hagerman J.S., Gould R.G., Proc. Soc. Exper. Biol. Med. 1951; 78:329) . The aims were to measure the transfer to plasma and to determine if adenosine-5¢-triphosphate (ATP) is needed. Method: Fresh RBCs (ATP-4.08 mmol/gHb and aged RBCs (ATP-0.37 mmol/gHb) were labeled with [4-14C]-cholesterol by incubation in a shaking water-bath for six hours at 37°C (n = 4). They were washed repeatedly and 10% suspensions prepared in Basal Eagles Medium. Two mL of the 10% suspensions were centrifuged and the supernatants removed. The RBC buttons were resuspended in their autologous plasma. Each sample was incubated at 37°C in a shaking water-bath for 32 hours. The radioactivity in the supernatant plasma was measured by liquid scintillation spectrometry (n = 4). Results: The transfer of cholesterol radioactivity from fresh cells to plasma was 49.55 ± 0.15% and from the ATP depleted cells 49.6 ± 0.15%. Conclusions: Cholesterol in the RBC membrane exchanges with plasma cholesterol. ATP is not required for the process. The 4/16 sample had a weakly positive DAT (mixed field); however, the Post sample was negative microscopically. An acid eluate performed on the Post cell sample was negative. An eluate performed on the follow-up cell sample identified anti-Fy a & anti-S.

R M Leger, G Garratty, American Red Cross Southern California Region, Los Angeles, CA; L Calhoun, A Ziman, C Wheeler, P I Figueroa, UCLA Medical Center, Los Angeles, CA Background: Autoagglutinins in cold agglutinin syndrome (CAS) usually have a high titer (>1000) at 4C; the most common specificity is anti-I. Some CAS agglutinins are low titer, but high thermal amplitude (react at or greater than 30C). Cryoglobulinemia may be associated with CAS. Case Report: A 60 yo male with hemolytic anemia (≠ bili, ≠ LDH, ≠ retic, Ø Hct) had a positive DAT (2+ with anti-C3 only, after 37C washes) and a 2+ room temp (RT) autoagglutinin that also reacted 2+ in gel-IgG. Prewarm indirect antiglobulin test was negative. The agglutinin was not detected in a sample collected (maintained at 37C prior to separation) the next day. When albumin (alb) was added, the agglutinin reacted 2-4+ at 37C. The specimens were referred to a reference lab for classification of the hemolytic anemia. Reference lab results: The EDTA RBCs from admission were spontaneously agglutinated (37C washed RBCs = 2+ with 6% alb); DAT on dithiothreitol-treated RBCs: IgG = 0, C3 = 3+, alb = 0. In contrast to the hospital's findings, an agglutinin was detected in the plasma from the next day which led to further investigation of this discrepancy. Titer of agglutinin: 37C = 0, 30C = 8, 22C = 16, 4C = 8. The serum agglutinated untreated RBCs at 22C and agglutinated and hemolyzed enzyme-treated RBCs at 22C and 37C. The specificity was not anti-I. Further testing showed weaker agglutination (1-2+) at RT with RBCs from one manufacturer (used by the hospital) compared to 3-4+ with 3 other sources of RBCs; this difference did not appear to be RBC agerelated. Three other CAS samples did not show reactivity difference between sources of RBCs. The agglutinin was weaker (1-1 1 / 2 + vs. 2 1 / 2 -3+) when other RBCs were tested in the presence of the diluent or hydrocortisone (a known constituent) from the one manufacturer. Investigation was confounded by a gel that had formed in all separated plasma samples (but not in the sole serum tube). This gel was not observed until 2 weeks after storage at 4C. The gel was not completely resolved after warming to 37C. The agglutinin in supernate from gelled plasma reacted 2+ vs. 3+ with the re-mixed plasma. Conclusion: The activity of the low titer, high thermal amplitude agglutinin in this unusual CAS appears to be depressed in the presence of one manufacturer's diluent (possibly due to the presence of hydrocortisone); this has not been previously reported. This depressed plasma activity led to false negative results at the hospital. The weaker reaction with the supernate from the gelled plasma suggests cryoglobulinemia or cryofibrinogenemia, though incomplete clotting was not definitively excluded; additional samples were not available.

C C Sanz, D Abello, A Pereira, Service of Hemotherapy-Hospital Clinico, Barcelona, Spain Background and purpose: RBC alloimmunization is a major problem in MDS, though both its incidence and correlation with the number of transfused RBC units are not fully established. Methods: Clinical records of all MDS patients who underwent transfusion at a tertiary care hospital were reviewed retrospectively. Patients were included in the study if they had received at least two RBC transfusions after June 1992 and no RBC antibody was detected at the time of the first transfusion. Immunohematology assays were performed at the time of each transfusion. For the purpose of this study, complex alloimmunization was defined as that resulting in 3% or less probability of finding compatible RBC. The actuarial probability of RBC alloimmunization was calculated by the method of Kaplan and Meyer. Comparisons of follow-up, incidence of alloimmunization or number of transfused RBC across groups of patients were done by the log-rank and Mann-Whitney tests. Results: Fifty-two patients met the eligibility criteria. Median age was 71 (16-89) years and 32 were males. Median follow-up from the first to the last RBC transfusion was 19 (3-125) months, and the median number of RBC units transfused per patient was 47.5 (4-401). Sixteen patients developed 40 alloantibodies and 5 autoantibodies, and 5 evolved to a complex RBC immunization during follow-up. Six patients formed one alloantibody, 2 patients formed two, and 8 formed three or more alloantibodies. Anti-E and anti-K1 were seen most frequently (9 instances each), followed by anti-c (5), antiJk a (3), and anti-Wr a (3). All the examples of anti-Wr a appeared in patients who had autoantibodies. The actuarial probability of presenting at least one antibody increased with the number of transfused RBC and reached 64% (SE: ±16,5%) after 130 RBC units. By contrast, the probability of complex immunization grew up to 11.3% (SE: ±5.4) at 38 RBC units, but then it remained at this ceiling despite continuos transfusions. Conclusions: In patients with MDS, the probability of RBC alloimmunization increases continuously with the number of transfused RBC units. By contrary, complex alloimmunization reaches its ceiling early during the course of disease.

Fatal Delayed Hemolytic Transfusion Reaction Due to Anti-Fy b J Kosanke, B Dickstein, L A Chambers, A Ng, M E Wissel, American Red Cross, Columbus, OH; C Nicely, OhioHealth, Columbus, OH Background: Hemolytic transfusion reactions may be difficult to recognize in patients with multiple alloantibodies and complex health problems, especially when medical care is provided at more than one facility. The management of an unusually severe delayed hemolytic transfusion reaction due to an uncommon antibody, anti-Fy b , was complicated in the presence of all of these factors. The reaction was ultimately fatal. Case Report: A 77 year old Caucasian woman was referred to a large urban teaching hospital (facility #1) for knee replacement (day 0). A single dose of Keflex was given prior to surgery. She had multiple red cell antibodies (to E, K and S) and received two units of antigen-negative RBCs on day 1 before being transferred to a hospital near her residence (facility #2) for physical therapy. On day 7, her hemoglobin was 11.8 g/dL with a total bilirubin of 1.4 mg/dL. She developed mild pancreatitis and received FFP with vitamin K for supratherapeutic coumadin levels. Her hemoglobin progressively decreased and by day 10 had fallen to 4.6 g/dL with an increase of bilirubin to 9.0 mg/dL. She was transported to an urban shock trauma hospital (facility #3) where her admission hemoglobin was 3.6 g/dL. Emergency transfusion was ordered. The compatibility testing sample, which was grossly hemolyzed with a positive direct antiglobulin test, was forwarded to the blood collection center's reference laboratory. Facility #1 was contacted for the patient's transfusion and antibody identification history. RBCs negative for E, K, and S were distributed and transfused while the immunohematologic evaluation was in progress, but the patient experienced a fatal cardiac arrest. A new anti-Fy b and a warm autoantibody, in addition to the previously known antibodies, were subsequently detected. Anti-Fy b was detected in the eluate after adsorption to remove the warm autoantibody. One of the two units of RBCs transfused was Fy(b+). Antibodies to cephalosporins were not present and the patient's admission blood cultures were negative. Conclusions: This is believed to be the first case of a fatal delayed hemolytic transfusion reaction due to an anti-Fy b alloantibody.

A P Cozac Jr., A Zanelli, T Rodrigo, D Covas, Hemocento de Ribeirão Preto, Ribeirão Preto, Brazil; L Castilho, Hemocento da UNICAMP, Campinas, Brazil Background: The alloantibody development after RBC transfusion is a common event in politransfused patients. The autoantibody formation after alloimmunization is also described. We reviewed the group of bThalassemic politransfused patients searching for the development of autoantibody after alloimmunization. Study design: 30 bThalassemic patients at polytransfusion therapy were analyzed. Each antibody identification was reviewed. Results: 17 (56%) out of 30 patients were alloimmunizated. Six (35%) out of 17 also developed warm autoantibody and 2 out of 6 developed WAIHA. Patients who developed only one or two alloantibodies didn't present autoantibodies. On the other hand, all patients with 3 or more alloantibodies developed autoantibodies with different specificities. Some of the autoantibodies developed showed changes in its specificity during the time. They sometimes make the identification of a new alloantibody difficult. After the anti-D development, three RhD+ patients were confirmed as D partial by genotyping. The number 6 patient developed a severe and acute AIHA (IgG) firstly by anti-e presenting hemoglobin at level of 3 g/dl. Corticosteroid was introduced with success and the RBC phenotype transfused was R 2 R 2 . After its suspension, the AIHA relapsed and at that time the autoantibody specificity was anti-E. Corticosteroid was reintroduced and the RBC phenotype transfused was rr. At the third relapse the anti-e reappeared and the introduction of IVIG therapy was effective. This therapy was introduced to avoid the corticosteroid side effects after long term use in children. This patient is waiting for a not-related bone marrow donor. The bone marrow transplant was indicated due to his hiperesponsive characteristic after blood transfusion. The number one patient also developed a severe and acute AIHA (IgG) by a "nonspecific" autoantibody. The corticosteroid therapy was conducted with success. enzyme or ZZAP treated cells without the addition of enhancement media. Study design and methods: A validation study was performed prior to implementation of this testing protocol. For this study, we reviewed laboratory records from a five year period, 1998-2003. Patient samples on which allo or autoadsorptions were performed were analyzed. Results: During this 5 year period, 751 samples were submitted for autoantibody investigation, representing 412 patients. In 609 (81%) of the cases, alloadsorptions were performed. In the remaining 143 (19%) cases, autoadsorptions were undertaken. A total of 2085 adsorption events were performed, a mean of 3 adsorptions per sample, with a range of 1-8 adsorptions performed. By reducing the sequential adsorbing time from 30 minutes to 10 minutes, time savings of 41700 minutes (695 hours) was achieved over 30 minute incubations (and proportionately greater savings over 60 minute incubations). Another advantage is the adsorbed plasma could be tested by a variety of IAT techniques (saline, LISS, PEG, gel, solid phase). Our alloantibody detection rate was consistent with previously reported data. We detected 149 new alloantibodies and reconfirmed 110 previously identified alloantibodies. In only 31 cases, we were unable to completely remove the autoantibody with this procedure, resulting in additional testing time. Conclusion: Reducing the incubation period during allo and autoadsorptions from 30 to 10 minutes is just as efficient at removing autoantibody for detection of underlying alloantibody and reduces workload in most cases. The resulting improvement in turn around time benefits the patient with critical transfusion needs.

Background: Gel and column agglutination techniques are well established diagnostics in blood group serology. Ease of handling, stable endpoint, and no need for washing steps in the indirect Antiglobulin Test (IAT) are key strengths of these systems. However, weak positive results are sometimes difficult to discern from negatives. The purpose of this study was to develop an agglutination format with distinct areas for positive and negative reactions, which allows for an IAT without washing steps. Methods: A plastic chip, similar in size with an ID-Microtyping Card (DiaMed), containing an intrinsic microcapillary system, was constructed with the following top to bottom design: 1) Reaction chamber; 2) reagent channel with prefilled reagents; 3) capillary system; 4) flash-sized chamber. The capillary zone and the flashsized chamber are the areas of positive and negative reactions, respectively. Forward Typing: 10 ml of diluted whole blood are pipetted into the reaction chamber of a chip carrying reagent channels filled with anti-A or anti-B. The chip is centrifuged in an ID-Centrifuge. Antibody Screening with IAT: 25 ml of 0.8 % Screening Cells and 10 ml of patient plasma are pipetted into the reaction chamber of a chip containing Coombs reagent. The chip is incubated for 15 minutes at 37°C and then centrifuged as above. Positive and negative results are recognized as haemagglutinates that are retained within the capillary system or as a button of red cells in the negative chamber. Results: 20 blood samples (including Ax and A3) have been tested in a prototype chip containing anti-A or anti-B reagent. 10 patient plasma containing irregular antibodies against blood group antigens and 10 patient plasma of blood donors without detectable irregular antibodies have been tested in a prototype Coombs chip. All results were in agreement with the results received with similar ID-Cards. Further experiments have demonstrated the usefullness of the new device also for other applications, such as reverse typing. Conclusions: The results of this feasibility study indicate that this technique may overcome disadvantages of gel and column agglutination techniques, i.e. ease of interpretation of very weak positive reactions and lot-to-lot variances.

G Garratty, R M Leger, P Hunt, A Co, American Red Cross Southern California Region, Los Angeles, CA Background: Hematologists often need help to confirm an immune etiology, ie, autoimmune hemolytic anemia (AIHA) in patients with hemolytic anemia but a negative (neg) DAT. The most common causes for AIHA associated with a neg DAT are RBC-bound IgG below the threshold of the antiglobulin test, RBC-bound IgM and IgA that are not detectable by routine reagents, and low affinity IgG that is washed off the RBCs during the washing phase for the DAT. The direct Polybrene test has previously been shown to be sensitive for detecting low levels of RBC-bound IgG; washing RBCs with cold Conclusions: The non-hemolytic autoantibody in alloimmunized patients was more common than it has been described before. AIHA should be ruled out when there is the suspicion of diagnosis of hemolytic transfusion reaction in patients with multiple alloantibodies. Background: Current rapid tests for blood typing do not have stable endpoints. Sophisticated techniques, which provide more objective and stable results, are rather slow and need a centrifugation step. All currently existing blood typing methods have in common that only one parameter at once can be determined. The purpose of this study was to develop a blood grouping format with multi-parameter testing in a single assay, providing a stable endpoint, but without centrifugation, suited both for a laboratory environment and for point-of-care situations at the patient bedside. Methods: A lateral flow device was designed with a separation membrane equipped in a cassette housing having a central application zone and 2 equidistant detection areas printed with parallel lines of antibody reagents directed against blood groups A, B, AB, D and C, Cw, c, E, e, and K, respectively. Further, both detection areas contain a flow control spot and an auto control spot. 50 ml of EDTAanticoagulated blood is diluted with 200 ml of a buffer solution. 100 ml of the resulting suspension are pipetted into the application zone, followed by 300 ml of a washing buffer. Results can be read after 5 minutes. Positive results are recognized as distinct red bands, negative results are recognized by the absence of the respective band. In an alternate test format, positive results are shown by a plus (+) symbol and negative results by a minus (-) symbol. Results: Bloods of 1415 donors, previously typed for the respective blood groups with the Olympus PK-80, have been tested with the new lateral flow test. The results for all antigens were in agreement with those received from PK-80. Conclusions: A simple, rapid and flexible method for blood typing with stable end-point is presented, allowing for: 1) Miniaturisation. 2) Parallel testing of multiple blood groups in the same assay. 3) Application in nonlaboratory environments without the need of electricity, such as military context, at collection sites for blood donation, and point-of-care testing at the patient bedside.

B K Graves, S R Pierce, Community Blood Center of Greater Kansas City, Kansas City, MO Background: Detecting alloantibodies by adsorption studies in patients with autoantibodies represents a significant percentage of a reference laboratory's workload. Many patients with autoantibodies are severely symptomatic and require transfusion in an urgent manner. The classic sequential adsorption procedures using enzyme or ZZAP treated cells are time consuming, with incubation periods of 30-60 minutes per adsorption. Methods that employ LISS or PEG decrease the adsorption time, but test methods with the adsorbed serum are limited due to the concomitant presence of the enhancement medium. For 5 years, our laboratory has performed sequential adsorptions using reduced incubation periods (10 minutes) against saline or LISS has been shown to enhance detection of low affinity antibodies. Study Design: Specimens referred for a DAT-neg AIHA workup were tested: with a panel of antiglobulin sera (AGS) consisting of anti-IgG, anti-C3, anti-IgM and anti-IgA, and a 6% albumin control (tube method). The anti-IgG is a commercial blood bank reagent; the anti-C3 is a standardized homemade reagent containing stronger C3d activity than commercially available anti-C3. The anti-IgM and anti-IgA were standardized for RBC agglutination. Additional tests performed were: anti-IgG and a control after RBCs were washed with 4C LISS; direct Polybrene test; and IgG-gel test. Results: Since January 1997, 191 (48%) of 398 specimens tested gave a positive (pos) result with at least one DAT method. DAT using our panel of AGS was pos for 175 (44%) cases [IgG: 55 (14%) total, 15 (4%) IgG alone. C3: 148 (37%) total, 115 (29%) C3 alone. IgA: 9 (2%) total, 5 (1%) IgA alone. None were pos for IgM]. The number pos/number tested for the other DAT methods were: 4C LISS washed 64/388 (16%), direct Polybrene test 28/330 (8%), and IgG-gel 49/384 (13%). In 5 (1%) cases, the 4C LISS wash was the only pos test (IgG-gel was neg). No cases were pos by IgG-gel alone. Direct Polybrene was the only pos test in 8 (2%) cases. Two of the IgA alone cases were also pos by direct Polybrene test. Conclusions: Our panel of AGS provide a pos result in 44% of cases [14% IgG total/4% alone, 2% IgA total/1% alone, 37% C3 total/29% alone]. IgG was detected by another technique in an additional 3% of cases. The pos anti-IgG results by tube reflect better technique in the reference lab vs the referring lab; the pos anti-C3 results reflect a stronger AGS used by the reference lab; anti-IgA was not used by referring labs. These results show there is value in using a lab where special reagents are available and the 4C LISS wash and direct Polybrene tests to detect low affinity and small amounts of RBC-bound IgG can be performed. Although the IgG-gel card would be expected to detect low affinity antibodies since the RBCs are not washed prior to testing, not all of the low affinity IgG detected by the 4C LISS wash were detected by gel tests.

T Rainer, B Israel, S Caglioti, D Figueroa, Blood Systems Laboratories, Tempe, AZ Background The 14 th edition of the AABB Technical Manual states that the Vel antigen is a high-incidence antigen that is unaffected by protease and sulfhydryl treatment. We have reason to believe that sulfhydryl treatment has an effect on the Vel antigen as evidenced by a case referred to our laboratory. Case Report A 67 year-old Caucasian female patient presented at a hospital with symptoms of anemia and renal vascular hypotension. Her transfusion history indicated that she had received multiple units in 1959. Several years ago, according to the patient, an attempt was made to find compatible units for the patient with no success. The case was referred to our laboratory and the patient was found to be O, D+ with a negative direct antiglobulin test. The serological picture revealed a strong antibody reacting 2+ s to 3+ s at 37C and at the antiglobulin phase. Low-ionic strength solution (LISS) was used as an enhancement media. The reactivity was present with all cells tested and the autocontrol was negative. Further characterization of the antibody showed similar reactivity using enzyme-treated cells (0.1% ficin) and no reactivity using 0.2M dithiothreitol-(DTT) treated red blood cells. Many rare frozen droplet red blood cells were tested that had antigens that were destroyed with DTT-treatment prior to testing Vel negative cells. Reactivity was negative with four Vel negative cells tested using both LISS and polyethylene glycol (PEG) methods. The patient was antigen typed as Vel negative. All other clinically significant antibodies were ruled out using Vel negative cells or the DTT-treated cells. Based on the unusual reactivity demonstrated by the anti-Vel, we tested twelve different examples of anti-Vel against two known Vel positive cells and one Vel negative cell. Using LISS enhancement on both sets, one set of these cells was DTTtreated, the other set tested neat. One of the twelve antiseras failed to react with the positive controls, one reacted with the negative control. Both were disqualified from the study. Four of the remaining ten antiseras showed a decrease in reactivity between the neat cells and the DTT-treated cells, while the remaining six antiseras showed no change in reactivity. Conclusion Contrary to the statement in the AABB Technical Manual, we found that sulfhydryl treatment (O.2M DTT-treatment) can have an effect on the Vel antigen. Our experience suggests that DTT-treatment may reduce the Vel antigen reactivity enough to use DTT-treated Vel positive cells to rule-out other clinically significant antibodies in samples with anti-Vel.

R C Jewett-Keefe, J Block, Carolinas Medical Center, Charlotte, NC Background: Reagent red cell antigenic stability may vary between manufacturers due to differences in manufacturing processes, preservative solutions or diluents. Our study evaluated antigenic stability of reagent red cells from three manufacturers over time. Study Design: Samples with known antibody specificities (N = 34) were tested in gel technology (Ortho Clinical Diagnostics, OCD) with reagent red cells from the following manufacturers: OCD 0.8% Resolve Panel B, ImmucorGamma Panocell 16, and Medion Diagnostics Data-Cyte Plus. Immucor Panocell 16 and Medion Data-Cyte Plus cells were diluted to 0.8% with MTS Diluent 2 the day of testing. Cells used for testing were homozygous for the corresponding antigen to the antibody. Three phases of testing were performed: Phase I; the week the cells were received, Phase II; the week of expiration, Phase III; two weeks post expiration. Reaction strengths were graded and then scored using the Race and Sanger Scoring System. Antibodies that did not exhibit expected results were classified as IgG or IgM with 0.01 M DTT (Sigma Chemicals) using the PeG (ImmucorGamma) method. Results: See table below. Anti-E (N = 1) not detected by Immucor or Medion was unable to be classified as IgG or IgM. Anti-E (N = 2) not detected by OCD were IgG. Anti-D (N = 1) not detected by Medion was unable to be classified. Anti-Jk a (N = 1) not detected by Medion or OCD was unable to be classified. Anti-Jk a (N = 1) not detected by OCD was classified as IgG. The decrease in overall antibody score between Phase I and Phase III testing was 50 points for OCD, 37 points for Immucor, and 19 points for Medion.

* #ID = Number of antibodies identified with each phase and manufacturer.

Conclusions: Variations in reagent red cell manufacturing processes, preservative solutions or diluents may have an effect on antigen expression over time. As seen by the decrease in score of the OCD cells, long-term storage in enhancing diluents may have a negative effect on the stability of some antigens. (1)] with negative and positive antibody screening tests (A/S) were tested by 2 independent transfusion services (TS). One mL of plasma was removed from each sample and replaced with one mL of HBOC to approximate a sample drawn from a trauma patient transfused with HBOC. ABO (forward & reverse group) and Rh type were performed by standard test tube methods. A/S were performed using a microcolumn gel system at one TS or by polyethylene glycol (PEG) tube method at the other TS. Results were compared to the pretransfusion test results on unaltered samples and/or samples where one mL of plasma was replaced with one mL of normal group AB plasma in lieu of HBOC. Results: ABO/Rh test results on HBOC test samples were consistent with results on unaltered and/or plasma diluted control samples. There were no ABO/Rh typing discrepancies. A/S results on HBOC samples versus plasma diluted control samples yielded comparable results (See table) . One anti-D* sample with < 1+ reactivity was unable to withstand dilution so that both the test and control A/S were negative. Background: Reagent red blood cells (rRBC) of the required phenotypic profile for complex antibody identification studies are not always available on in-date 0.8% commercially-prepared panels. Use of expired rRBC is often necessary to identify or exclude antibody specificities. Accurate results with these cells depend on the integrity of the antigens. There are no large scale published studies documenting the stability of various red cell antigens on the expired rRBC when the same commercially-prepared 0.8% panel is studied over time in gel tests. Study Design: Twenty-three donor-derived red cell antibodies (Ab) of common specificities in the RH, KEL, FY, JK, MNS, and LE systems with initial reactivity £2+ in tube tests were tested against a commercially prepared 0.8% rRBC panel. Gel tests were performed according to the manufacturer's instructions on in-date rRBC and at 1 month intervals to 3 months past expiration. Reactivity with expired antigen positive and antigen negative rRBC was compared to in-date cells. Acceptable results were defined as reactivity with expired rRBC that was consistent with the reactivity of the same in-date rRBC. Results: Twenty-one of 23 antibodies gave acceptable results with all cells tested. (D, C, c, E, e, K(2), k(2), Fy a (2), Fy b (2), Jk a (2), Jk b , S(2), M(2), N). For all 21 antibodies, the reaction strength with antigen positive cells at 3 month post expiration was within one reaction grade of the same in-date cells. One example of anti-N reacted only with M-N+ cells during in-date and all post-expiration testing. An anti-Le a was non-reactive with one of three Le(a+) at two months post-expiration and two of three Le(a+) cells at three months post-expiration. Testing of the same plasma with in-date Le(a+) cells showed that the decreased reactivity was due to antibody deterioration rather than antigen deterioration. Conclusion: Except for the Le a and N antigen reactivity described above, the antigens studied were well detected at in all tests with expired rRBC, regardless of dosage. The unexpected non-reactivity of anti-Le a and one example of anti-N is a function of the antibody used in testing rather than deterioration of the Le a or N antigen. Tests with 0.8% cells requiring the presence or absence of the antigens studied were valid in gel tests for at least 3 months after the manufacturer's assigned expiration date.

S Linauts, C Lammers, Puget Sound Blood Center, Seattle, WA Background: Donor antibody screening is generally conducted using manual or semi-automated methodologies. In this study, we evaluated the use of the fully automated Olympus TANGO to determine if comparable or improved sensitivity and specificity could be obtained with the TANGO as compared to our current test of record, the Immucor Capture-R® Ready Screen (Pooled Cell) solid phase system (SP). Method: Two populations of Conclusions: Although the presence of HBOC may cause plasma in blood samples to be markedly red in color, this study shows that serologic pretransfusion testing by standard methods on these samples yields accurate results consistent with the results from control samples. The results of this in vitro testing correlate well with our findings in clinical trials where samples were collected and tested after HBOC transfusion. If HBOC become licensed, they may be widely used in emergency settings. Transfusion services would need to be prepared to process samples from patients who have received blood substitutes.

J Santi, J T Fueger, J L Gottschall, S T Johnson, Blood Center of Southeastern Wisconsin, Milwaukee, WI Background: Prenatal antibody titers are reported to be about 60% accurate in predicting when additional measures are recommended to monitor the fetus. Poor correlation exists between the level of titer and the effects on the fetus. Because more invasive procedures carry a risk to the fetus the accuracy of prediction is important. Study Design: Surveys were sent to 126 hospital and reference laboratories throughout our state and 1 additional reference laboratory. Survey questions were aimed to obtain information regarding result reporting and test methods used for titration. There were 72 responses with 27 agreeing to perform antibody titrations. To assess variation in actual performance and results 2 different anti-c plasma samples were sent for testing in each lab. Samples tested by 2 technologists in our immunohematology reference laboratory were used as the benchmark. Titrations were done by preparing serial dilutions in 6% albumin and tested against R 1 r (DCe/dce) red cells using a 30 minute, saline IAT with anti-IgG. The titer of sample A was1 and sample B was 8. Simulating a mother with a significant change in titer, samples A and B of known anti-c specificity were sent to each laboratory with a request to perform titration studies using their standard protocol. Results: Twenty two of 27 laboratories reported results. Phenotype of test cells and number of labs using them included rr (7), R 2 R 2 (9), R 1 r (1), r'r (3), and unknown (2). Conclusions: Eight different methods with 4 different cell types were used in 22 laboratories. Among laboratories using the same method, wide variation in results were reported. Titer results also varied, with some labs reporting a score instead of a titer. Only 9 of the 22 labs reported a significant change in titer. These results demonstrate the need for improved training and standardization. samples were tested. The first included 1611 random samples that were tested by SP on day 0 or day 1. SP was performed using a Hamilton AT+ to pipette samples into plates. Plates were washed using a CSW100 washer, indicator cells were manually added, plates were centrifuged in a Sorval T6000 D, and reactions were read with an IBG Multireader Plus. The samples were subsequently tested on day 1 using the Olympus TANGO and a pooled screening cell. Positive results from either system were verified in tube, using a two-cell screen with Ortho Antibody Enhancement Solution (AES). Results from the tube testing were considered to be final. The second set of samples included 59 frozen samples with previously identified weak antibodies. These were first retested with the two-cell AES tube test to ascertain the strength of the reaction after storage. The samples were then tested on the TANGO using the pooled cell reagent. Antibodies tested included Rh, Fy a , Jk a , K, as well as several M, Le a , and I. Results:

Discrepant results in the tube testing included two specimens with increased MCHC due to cold agglutinins, two control specimens and one sample with a MCV > 105 fL. One specimen (*) with an increased MCHC due to cold agglutinins demonstrated both unexpected forward and reverse grouping by tube testing but had an interpretable automated blood group. In the automated method, three specimens with MCVs > 105 fL, one with a MCV < 70 fL, one giant platelet sample, one specimen with an increased wbc > 50 and a control specimen were implicated in discrepant results. CONCLUSION: There is not a significant increase in discrepant results in specimens with abnormal hematologic parameters comparing the automated Ortho ProVue TM method with conventional manual techniques.

G Robertson, J Bodiya, L Vaughn, S Caglioti, Blood Systems Laboratories, Tempe, AZ Background: Solid phase technology is used for unexpected red cell antibody screening at a large regional donor testing facility. Two formulations of indicator cells used with the solid phase technology were evaluated to determine whether the new indicator cells formulation, introduced by the manufacturer in March 2003, was more effective for high-volume donor screening than the original formulation. A 3-cell antibody screen is performed by gel or tube on samples with positive or equivocal results obtained with solid phase to reduce the number of samples submitted for antibody identification. Reducing the number of samples requiring 3-cell screening improves turnaround-time for initial testing results and reduces the compliance risk inherent in additional sample routing and testing. Method: Results obtained through routine donor testing over four months (345,569 samples) using three DIAS Plus systems with original formulation indicator cells were compared with four months (507,532 samples) of routine donor testing results using the same three DIAS Plus systems with the new formulation indicator cells. Performance was evaluated based upon rates of equivocal and positive reactions, both of which require additional testing prior to result release for product labeling. Additionally, the two indicator cell formulations were evaluated for frequency of reported undetected clinically significant antibodies. Results: The average equivocal rate across all three instruments decreased from 0.82% (2817/345,569) with the original indicator cells to 0.17% (843/507,532) using the new formulation indicator cells, a statistically significant improvement at p < 0.0065. The average reactive rate was 0.70% with original indicator cells to 0.72% with the new formulation indicator cells, which was not significantly different. The frequency of reported undetected antibodies (1 with each indicator formulation) did not change in the two time periods evaluated. Conclusion: The use of the new formulation indicator cells improved the specificity of the solid phase antibody screen assay, which reduced the amount of unnecessary additional testing performed. The end result was an absolute 0.64%, and a 79% relative reduction in the number of samples requiring additional testing, which, annualized, is a reduction of approximately 7680 samples that would no longer require 3-cell testing. The new formulation of indicator cells improved specificity with no loss of sensitivity for unexpected antibody screening on the DIAS Plus system.

M J Drew, L L Cardine, E C Paluch, C O Tindall, G J Pawlak, Henry Ford Hospital Blood Bank, Detroit, MI Background: Our transfusion service serves a 903-bed tertiary care hospital, supporting trauma, cardiovascular surgery, solid organ & stem cell transplant & neonatal ICU. Manual tube methodology is used for ABO/Rh testing, antibody screening, & antibody identification (ID). Our facility participated in a pre-market clinical trial comparing the Olympus TANGO TM Automated Blood Bank Analyzer with tube methodology. Methods: The Olympus TANGO TM is designed to perform blood group determinations, antibody (Ab) screening, antibody ID, antibody titration & compatibility testing using Random samples: 98.7% correlation was obtained; SP identified 13 (0.81%) positive results that were not confirmed by tube methodology and therefore considered false positives. TANGO identified 7 (0.43%) false positive results. Each system failed to detect one antibody demonstrable as a 1+ in tube. Known Antibody Positive: Although the reactivity of 13 samples had degraded below AES tube detection levels, TANGO identified 9 of these as positive. TANGO detected 25 of 30 samples testing weak to 1+ with the AES tube screen. TANGO detected all samples testing 2+ or 3+ in tube. Conclusion: TANGO sensitivity with a pooled screening cell was comparable to the SP pooled cell antibody screen assay as well as to AES tube methodology using a two-cell screen. Better specificity was seen with the TANGO than with SP. These results, together with the labor advantages of a fully automated testing system, suggest TANGO to be a viable alternative to performing antibody screening in a blood center setting.

S N Nahirniak, H McConnell, K Hamacher, T Richardson, Capital Health, Edmonton, AB, Canada; G Clarke, Dynacare Kasper Medical Laboratories, Edmonton, AB, Canada BACKGROUND: Although it is known that hemolysis, icteric and lipemic samples may lead to discrepant or "No Results Determined" results when analyzed on the Ortho ProVue TM automated blood bank system, the effect of other common hematologic problems on the ProVue's operation and generation of results has not been studied. Thus this study was designed to determine the effect of selected hematologic sample variables on the ability of the Ortho ProVue TM to produce valid and accurate results in Type and Screen testing. METHOD: A total of 317 samples, including 76 age matched normal controls, underwent parallel type and screen testing. Our reference method was a tube ABO and Rh with a manual gel antibody screen using the Anti-IgG Rabbit Anti-IgG MTS TM card. For the automated analysis, the same gel card was utilized in for the antibody screen but A/B/D Monoclonal and Reverse grouping cards were employed for the ABO and Rh typing. The selected sample variables included: a) white cell count > 50 ¥ 10 9 /L (n = 6); b) Platelet clumping (n = 36); c) giant platelets(n = 9); d) elevated platelet count > 700 ¥ 10 9 /L (n = 9); e) MCV > 105 fL (n = 85); f) MCV < 70 fL (n = 87) and g) increased MCHC due to cold agglutinins (n = 8). All flag variables were confirmed by morphologic assessment before inclusion into the study. A reaction grading of > 1+ or < 1+ from the manual method were not considered significant discrepancies. RESULTS: No discrepancies between the automated and manual gel screens were identified. A total of 12 out of the 317 specimens, including three normal controls, demonstrated discrepant results between the manual and automated methods (Table 1) .

hemagglutination & solid phase technology. For the trial, ABO/Rh typing & Ab screening on patient samples & donor ABO/Rh confirmations were performed on the TANGO TM . Tube testing results served as the reference method. Results: ABO/Rh typing was performed on 875 patient samples, 794 of which (91%) were concordant with the reference method. On 66 samples (7.5%), the ABO type was not determined, due to weak agglutination of reagent cells in reverse typing, or to mechanical issues. On 11 samples (1.4%), Rh type was not determined, due mostly to mechanical issues. Four (4) samples testing D negative on the TANGO TM were weak D positive with the reference method. Of 1039 donor samples tested, ABO/Rh types were successfully determined on 1019 (98%). Antibody screening was performed on 874 patient samples, 838 of which (96%) were concordant with the reference method. On 47 of 874 samples (5%), Ab screening results were indeterminate on the TANGO TM , due mostly to pipetting errors and other mechanical issues. Optimal throughput time was obtained with batch sizes of 16-32 specimens. Conclusions: The TANGO TM would fit well into a busy hospital transfusion service. Hemagglutination provides a familiar, inexpensive, and reliable method for determination of ABO type. Improvement in detection of weak D would be optimal for screening donor samples. Solid phase technology provides good sensitivity, particularly with weakly reactive antibodies. Software improvements in the areas of result reporting and review, tracking progress of runs on screen & readability of on screen display would enhance usability of the instrument. Background. We evaluated the ability of a new Windows compatible software developed to perform RBC antibody (Ab) identification by a walkaway instrument that uses cassette technology and image processing. Methods. The software for Ab identification (Resolvigen 3) imports the data from the walkaway instrument (AutoVue Innova, Ortho), which was designed to automate immunohematology tests utilizing cassettes (BioVue, Ortho) and to perform result interpretation. The software reads images of wells that are stored locally and are available for browsing. The operator selects commercial panels (Ortho) for Ab screening or identification (3% and 0.8% suspensions), type of RBC treatment (untreated or ficin-treated RBCs) and method (indirect antiglobulin test and saline at 37°C). The reaction strengths of the wells are measured (0, ±, 1+, 2+, 3+, 4+). The software checks all non reacting and reacting RBCs, compares the scores and the results of different test conditions and screening and identification panels. For each procedure the software suggests Ab specificity and characteristics, Abs that cannot be excluded and additional tests that the operator could perform to identify the specificity, including Abs to high and low incidence antigens. In this study we performed a preliminary evaluation of Abs that were identified by a skilled operator and by the software using untreated and ficin-treated 3% and 0.8% panels, the manual cassette method (skilled operator) and the automated system and software. We tested 20 samples from not immunized patients, 24 samples from 24 patients with 28 Abs previously identified by manual methods; of these, 14 were tested using both 3% and 0.8% panels; 10 using 3% or 0.8% panels. Results. We observed negative results with all negative samples and complete agreement in 22 immunized patients. The software identified correctly 26 Abs. In 2 patients the software was unable to identify 2 anti-Jk a that reacted only using untreated RBC panels. Conclusions. The preliminary evaluation of Resolvigen 3 used in combination with the walkaway instrument shows potential value to RBC antibody identification. Addional investigations are necessary to complete the validation of the system.

Background Depending on a patient's diagnosis, he or she might be transfused repeatedly. In the laboratory, this means that some patients tend to be evaluated repeatedly for identification of the same antibody. This study was done to get a better idea of the incidence of new antibodies rather than the prevalence of known antibodies. Methods Laboratory records of antibodies identified between May 1, 2002 and May 1, 2004 were reviewed retrospectively. Review included notation of the ordering service. Antibodies that had been previously identified at this institution were not included in the review. Anti-D antibodies due to RhIg therapy, inconclusive antibody evaluations, and autoantibodies with no specificity were also excluded. Results A total of 228 new antibodies were identified in 189 patients (129 female, 60 male). Six women and two men were found to have new antibodies on more than one occasion. In women, 22 patients were found to have antibodies during prenatal evaluations ordered by the obstetric service. The other most common ordering services were surgery (69 patients), internal medicine (other than oncology, 60 patients), and oncology (33 patients). The youngest male patient was a newborn with an unexpected anti-D antibody (not due to RhIg). The youngest female was a 2 year old with an anti-K antibody. A 9 year old girl was found to have an anti-Jk a antibody. The oldest patient was a 96 year old woman with an anti-E antibody. Seventeen patients in their eighties were identified with a variety of antibodies. The most common antibodies are listed in the table.

Conclusions The distribution of the types of new antibodies in this patient population is similar to the prevalence of antibodies previously reported in American populations. Clinically significant new antibodies were found in both very young and very old patients.

M Heintz, M Bahl, J Mann, M Yohannes, J Levitt, University of Iowa, Iowa City, IA Background: The TANGO (Olympus America, Inc, Irving, TX) is a fully automated benchtop blood bank analyzer designed to perform ABO, Rh typing and antibody screen (AbScn) using hemagglutination and solid phase technology. We compared results of patient blood samples tested in parallel with TANGO and our tests of record. Methods: 1108 samples from patients with type & screen orders were tested for ABO/Rh on the ROSYS Plato TM (Immucor, Inc., Norcross, GA) using hemagglutination dilution (HD) strips, or by tube test using Immucor blood grouping reagents. 934 samples were tested for AbScn on the ROSYS Plato, or manually using Immucor Capture R Ready Screen. Antibody identification (AbID) was performed by polyethylene glycol (PEG) enhance tube technique. Results:

All blood grouping discrepancies (63/1108 = 5.7%)were due to no type determined (NTD) by one or more method as explained by serological, sample or mechanical problems. An additional 22 samples had NTDs by both methods. For ABO typing, 85 of 1108 (7.7%) samples tested on TANGO had NTD results, 24 of 746 (3.2%) on ROSYS Plato and 0 of 362 by tube test. For Rh typing, 7/1108 (0.6%) samples tested on TANGO had NTD results, 1/746 (0.1%) on ROSYS Plato and 0/362 by tube test. 18 samples with clinically significant antibodies identified had positive (pos) or indeterminate AbScn on TANGO and Capture R. 4 samples tested weakly pos on TANGO, but negative (neg) on Capture R with 1 room temperature antibody identified and 3 with unidentified reactivity. 1 sample had a passively acquired auto-antibody that was neg on TANGO and pos on Capture R. The TANGO system had 11 false pos ABScn results compared to 16 false pos with Capture R with a specificity of 98.8% and 98.3% respectively. Conclusions: Blood grouping and AbScn testing results are comparable to the tests of record. TANGO did not yield any incorrect ABO/Rh results or fail to detect any clinically significant antibodies detected by the test of record. TANGO results may have been adversely affected due to the delay in testing of 1 to 3 days.

S Smith, Medic Regional Blood Center, Knoxville, TN Background: Since conventional reagents have not been manufactured or validated for use on the Olympus PK7200 instrument, each donor center must determine the optimum antisera dilution for each lot of ABO and Rh reagents. This involves time-consuming dilution studies with each new antisera or bromelin lot. The Olympus ® PK ® 7200 Blood Grouping Reagents (PK BGR) are manufactured ready-to-use ABO, Rh antisera and A1, B reagent red cells. These reagents eliminate the need for dilution studies and any reagent preparation. We evaluated the benefits of PK BGR to replace diluted conventional reagents (DCR) at our facility. Study Design: EDTA samples (N = 1,406) were tested in parallel with the PK BGR and DCR. Testing was performed over 5 consecutive days with samples tested within 24-72 hours of collection. All results were compared and a no type determined rate (NTD) of less than 5% was expected. Results: The Olympus PK BGR had a total of 19/1406 NTDs due to ABO forward or reverse grouping discrepancies, while the DCR had 21/1406. Both the Olympus PK BGR and the DCR had 2/1406 discrepancies between the ABO forward and reverse types on the same samples, one of which was an A3. Each method also had 3/1406 system control errors where the control reagent for the test was indeterminate or positive. and allow for an increase in the serum/cell ratio as well as increased room temperature incubation time. Other factors that may be a consideration are final red cell concentration in the Gel Card System. Given the 42% NTD rate obtained with the B Cells, effect of increased diluent concentration on the B antigen over storage time may also be a consideration. Overall, the gel card method is provides an objective determination for NTD samples. Background: Blood banks require a sensitive, specific, and efficient method to detect clinically significant RBC antibodies. Solid phase antibody screening methods (e.g., red cell adherence assay, Capture-R Ready ID, Immucor, Norcross, GA) are popular due to high sensitivity and automation. However, high sensitivity detects "false positive" antibodies of questionable clinical significance and leads to additional testing. We use Capture-R for RBC antibody screening. Positive samples then are tested for RBC antibody identification (ID) using polyethylene glycol (PEG) for enhancement, as PEG has high sensitivity (IDs nearly all clinically important true positives) and high specificity (IDs few clinically unimportant false positives). Methods: We studied positive rates of Capture-R vs. PEG methods and categorized RBC antibodies identified by first testing 33,564 consecutive blood samples from 3/1/03 to 4/30/04 by Capture-R Ready ID method. Negative samples were not tested further. Positive samples were tested by a tube method (Panocell-10® 11-cell antibody panel, Immucor Inc., Norcross, GA) with PEG enhancement. Antibodies reacting in both assays were true positives; those reacting in Capture-R only were considered false positives. Results: Capture-R was positive in 1084/33,564 (3.2%) of samples. As the "gold standard", PEG confirmed true positivity (i.e., ≥1 cell reacting) in 710 Capture-R positive samples (65.5%); 374 Capture-R positive samples (34.5%) did not react in PEG (i.e., false positives). The 710 true positive RBC alloantibodies were: True Positive Alloantibodies Identified (% of total)* Conclusions: The Olympus PK BGR performed at a NTD rate of 1.6%. The diluted conventional reagents performed at a NTD rate of 1.9%. The ability to eliminate dilution studies, reagent preparation and decrease the NTD rate justified the switch to the Olympus PK BGR.

J Lowther, J Swisher, J Addington, R Plumlee, C Cousins, Gulf Coast Regional Blood Center, Houston, TX Background The current protocol at this institution requires that samples resulting in No Type Determined for ABO blood type (NTD) by the Olympus PK 7200 ® method be resolved by Gel Card testing. An NTD determination includes, per institutional SOP, weak or suspect reactions of £ 1+ reaction strength. Study Design/Methods Samples demonstrating NTD by the automated PK 7200 ® method were subsequently tested by the Ortho Microtyping System TM . Results This study included a total of 205,591 blood type determinations. Of this total number, 3123 required Gel Card resolution. There were 413 NTD determinations by the Gel Card method, which were submitted to the Reference/Consultation laboratory for final resolution by tube method. These NTD samples were further divided into 7 categories. Of these categories, weak reverse typings account for 56% of the NTD, with the weak B cell results comprising 42% of that total. The Consultation/Reference final determination correlates 100% with the forward typing in these instances. January 2003-October 2003 Gel Card NTD by Category.

Conclusion The above data suggests that the Gel Card procedure may not allow for detection of weak reverse groups. Consultation and Reference techniques (tube tests) use the standard 3%-5% final red cell concentration * Table: Antibodies exclude passive antibodies (e.g., Rh immunoglobulin or transplacental) -n = 70; autoantibodies (panreactivity on the 11-cell panel) -n = 95; and inconclusive antibodies (e.g., reactivity ≥ one cell, £11 cells and no defined antigenic pattern) -n = 60. Conclusions: With PEG as the gold standard, only 2/3 of reactions by Capture-R were true positives. Annually, this increased reactivity of Capture-R led to 9,925 additional PEG panels with negative results at an estimated cost of $49,625 (supplies and labor). Because of ease and automation, Capture-R is popular as a screening test, but a more specific method is recommended to ID truly significant antibodies.

Galileo is an automated machine for Blood Grouping and red cell antibody detection. The onboard method for antibody detection is Capture R -Ready Screen and Ready-ID. The user is constrained by the manufacturer's antigen matrix. 1 To widen the available pool of antigens, provide choice and assist in the identification of difficult antibodies or mixtures, the Galileo Antiglobulin Cross-Match facility was utilised as an antibody screening and identification tool. This uses the Capture-Select plate, the wells of which are coated with a binding agent to promote the adhesion of a red cell monolayer; produced by Galileo as part of the cross-match process without user intervention. Antibody Screen: A commercially available 3 cell antibody screening set 2 presented as a 3% suspension in modified alsever's solution was identified by bar code labelling and presented as "donor samples". Plasma samples to be tested were barcode labelled as "recipients" and processed blind against the screening cells using the antiglobulin crossmatch script. Antibody Identification: A panel of red cells prepared as above was offered as donor samples and a selection of antibody bearing plasma samples tested against this panel and using Capture ready ID. 500 plasma samples shown to be free of unexpected antibody activity by routine methods 3 were tested using the modified cross-match. Approximately 100 samples known to contain antibodies were tested blind of the specificity by the same method. These samples were tested in parallel using Capture-R ready screen on Galileo. Plasma samples were tested using the Identification panel and Capture-R ready ID. No antibody bearing samples gave negative results. 3 samples gave false positive results. 4 samples gave? results, 3 reactions of which should have been negative. All antibody free samples gave negative results. The cross-match facility on Galileo is a valuable resource widening the choice of red cells available to the operator investigating the presence of red cell antibodies. References: 1 Immucor Gamma Norcross Georgia USA 2 NBS laboratories Cambridge Engaland 3 Capture-R ready screen / ready ID Immucor Gamma Norcross Georgia USA and Diamed ID system, Diamed GB Dalkeith Scotland.

S Hawk, R Tyler, R Russell, K Pearson Smith, E Jones, K Thompson, St. Luke's Hospital, Kansas City, MO Background: We evaluated the feasibility of replacing conventional tube procedures in our hospital transfusion service with ID-Micro Typing TM manual gel technology. Methods: Five technologists participated in a comparative study of gel and tube technology. Tube procedures consisted of ABO/Rh typing with monoclonal antisera, antibody screening and identification using a standard PEG-IAT procedure. After training by the manufacturer, gel testing was performed according to manufacturer's instructions. Routine patient samples were compared by both methods. These included 76 samples submitted for ABO/Rh typing and antibody screening, 47 for antibody screening only, 15 antibody identification panels, and 11 cord blood samples for ABO/Rh typing and direct antiglobulin testing (DAT). Results: ABO/Rh typing by both methods demonstrated 100% concordance. Nine of 11 cord blood DATs (82%) were concordant. Two cord bloods demonstrated weakly positive DATs by the tube method but were negative by the gel method. These weak positive DATs were due to maternal/fetal ABO incompatibility. Of the 123 samples submitted for antibody screening/identification, results of 92 samples (75%) were concordant by both methods. Seventy-six samples were negative and 16 samples were positive with both methods. Antibodies detected in these samples included those in the Rh, Kell, Duffy, Kidd, Lewis, P1 and MNSs systems. Results of the remaining 31 samples were discordant. All demonstrated reactivity by the tube methods but were negative by the gel method. Twenty-five of the 31 were identified as Sd a or nonspecific. The 6 remaining samples included two examples of C, two examples of K, one example of K + S, and one example of Jkb + Fyb + S. No samples were positive by the gel method and negative by the tube method. Conclusions: The five technologists summarized pros and cons of the gel system as follows: Pros: 1. Reactions were easy to read and could be held for future review. 2. Gel eliminated many nonspecific reactions seen with the PEG method. 3. Gel required small volumes of reagents and samples. 4. Gel provided more free time to perform other duties. Cons: 1. Turn around times with gel were too long for stats. 2. Gel required excessive use of disposables. 3. Specialized equipment required additional space. 4. Red cell dilutions were cumbersome, time consuming and confusing. 5. Gel would be more difficult for cross-trained personnel. 6. Gel technology is expensive. 7. Gel technology failed to identify clinically significant antibodies. The results of this study, especially the undetected antibodies, produced a consensus decision to retain the tube technology. Background: Js b is a high frequency antigen in the KEL6 blood group system. Anti-Js b is relatively rare, but has been associated with delayed hemolytic transfusion reactions and hemolytic disease of the newborn. We report a man with multiple RBC alloantibodies including anti-Js b , who received repeated, apparently successful, transfusions of Js b positive (+) RBCs. Case Report: The patient was a 50-year-old blood group A, Rh negative African-American man with sickle cell disease who presented with anemia (Hgb = 5.7 g/dL) and gastroenteritis and had been transfused 20 years earlier. His Ab screen was negative and he was given 2 units of RBCs, raising his Hgb to 6.9. After 2 additional units, his Hgb = 9.1 and he was released to home. A year later he returned with an Hgb of 4.5. Again, his Ab screen was negative and transfusion of 4 units of RBCs increased his Hgb to 9.6. However, 2 weeks later he presented with chest pain and Hgb = 4.3. Request for immediate transfusion was delayed when the Ab screen and DAT were positive, consistent with a delayed hemolytic transfusion reaction. Further evaluation revealed anti-Js b , Jk a , S, and suspected anti-E and Fy a in his serum. Anti-Jk a and Js b were eluted from his RBCs. Only 1 phenotype-matched donor was identified through the AABB Rare Donor File. The patient's Hgb continued to drop and ischemic EKG changes developed. The patient and his physician consented for transfusion of incompatible RBCs. He was started on high-dose dexamethasone, subsequently Epogen, and after receiving 2 units of RBCs negative for D, C, E, K, Fy a , Jk a , S and Hgb S, symptoms improved although Hgb remained <5. After another 2 similar units and the unit from the phenotype-matched donor, Hgb = 7.9. He was discharged on Epogen, folic acid and vitamins. A month later he returned with worsening shortness of breath and Hgb = 5.6 despite treatment with Epogen and hydroxyurea. Two units of phenotype-matched Js b +, incompatible RBCs raised his Hgb from 5.1 to 7.9, and another 2 to 10.0 g/dL. No further transfusions have been required. Conclusion: This patient with multiple RBC alloantibodies was repeatedly transfused with Js b + incompatible RBCs and tolerated all transfusions well without evidence of acute hemolysis. Background: Antibodies that react with the common products of RHCE, such as anti-Rh18 (Hr/Hr S ), can cause hemolytic disease of the newborn (HDN) as well as severe hemolytic transfusion reactions. While variant RHCE phenotypes are common in some ethnic groups, serologic identification is difficult. Recent studies indicate that individuals with anti-Rh18 antibodies can have varying RHCE genotypes, and that the sera and cells from patients with different genotypes may cross-react. Case Report: A 26-year-old pregnant Puerto Rican female (G6P2, 2 therapeutic and 1 spontaneous abortion), with no history of transfusions or children with HDN, presented at 31 weeks gestation with anti-E and anti-Rh18. Her red cells were phenotyped as: O, weak D+, C-, E-, c+, e+, hr S -. A search of the United States National Rare Donor Registry did not yield any compatible donors; therefore, a unit of the patient's blood was collected one week before her planned C-section. The baby was E-, but had a positive DAT. The eluate reacted with all cells tested, but there was no clinical evidence of hemolysis, and the baby did not need to be transfused. Molecular testing revealed that the mother was homozygous for an Rh haplotype in which the variant/weak RHD allele DAR is linked to the RHCE variant ceAR. The ceAR allele is one of at least three RHCE variants that can lead to formation of anti-Rh18 (Noizat-Pirenne et al. Blood (2002) : 100, 4223). Conclusions: This report describes a Hispanic woman with anti-Rh18 who is homozygous for DAR-ceAR. This genotype is prevalent in the South African Black population (Hemker et al. Blood (1999) : 94, 4337), but its presence in this individual indicates that it may be found in other ethnic groups. This case illustrates the clinical difficulties in the management of pregnancy and transfusion in these patients. In addition, as the multiple molecular backgrounds underlying these phenotypes are not all serologically compatible, molecular testing is critical for the identification of compatible donors.

C Lomas-Francis, M E Reid, R Øyen, New York Blood Center, New York, NY; R Yomtovian, University Hospitals, Cleveland, OH; C McGrath, American Red Cross Blood Services, Cleveland, OH; P S Walker, Blood Centers of the Pacific, San Francisco, CA Background: Patients of African-American descent with variant RHCE genes may make alloanti-e-like antibodies including anti-hr B . Once anti-hr B is identified, transfusion of hr B -blood is often achieved by using R 2 R 2 (e-hr B -) blood. Ultimately, hr B -patients may make anti-E and/or a "broad-spectrum" anti-Hr B (-Rh34). If such patients require frequent transfusion, additional alloantibodies are likely to be present making antibody identification and provision of suitable blood complicated. We describe here 2 multiply transfused D+ patients with Sickle Cell Disease with a history of anti-hr B , (plus other alloantibodies), who appeared to have made anti-Hr B , but where the additional reactions were finally shown to be due to anti-D. Case studies: Patient #1 had made anti-hr B , -C, -E, -S, -K, and -Fy a ; patient #2 had made anti-hr B , -C, -E, and -K. After multiple transfusions of crossmatch compatible appropriate antigen-negative blood, the sera of both patients became strongly incompatible by the IAT: only Rh null RBCs and some hr B -RBCs lacking other relevant antigens were compatible, whereas D-RBCs were incompatible. The cause of the incompatibility was determined to be anti-D and not, as had been supposed, anti-Hr B (anti-Rh34). The compatible hr B -samples were either D-or were identified as having the partial D phenotype DIIIa. RBCs from both patients were determined to be DIIIa by DAK typing and PCR-RFLP analyses. Anti-D made by a DIIIa person does not react with other DIIIa RBCs and explains the compatibility of some D+ hr B -RBCs. Conclusion: Given the reputed immunogenicity of D antigen, it surprised us that anti-D was made after anti-hr B and other alloantibodies. Many transfusions of hr B -D+ RBCs (including R 2 R 2 ) were tolerated before anti-D was produced. The partial D phenotype, DIIIa, is not readily identified because DIIIa RBCs are strongly agglutinated by reagent and monoclonal anti-D. Because DIIIa RBCs are usually VS+ and hr B -, patients who make anti-hr B are likely to have the DIIIa phenotype and be at risk of making anti-D. The presence of anti-D may not be readily apparent when a panel of D+ hr B -RBCs are tested because some, but not all, may be DIIIa. DNA analysis is showing that altered RHCE genes are frequently paired with genes encoding partial D phenotypes. Thus, it is essential to recognize that providing compatible blood products for patients with anti-e-like antibodies directed at high incidence Rh antigens can be confounded by anti-D.

E Isaak, B Lechien, American Red Cross, Saint Louis, MO Background: Warm autoantibodies are associated with a number of medical conditions. In most cases these antibodies are directed against epitope determinants in the Rh system, although it is possible for them to show some degree of specificity. Autoantibodies directed against specific antigens in the Kell group have been reported, but they are exceedingly rare. Previous reports of individuals with autoanti-Kp b have indicated that these antibodies are associated with depression of Kell antigen expression. We present the first published example of autoanti-Kp b without Kell antigen depression. Case Report: The patient is a 65 year old female with poorly controlled diabetes. She had previously undergone one below the knee amputation due to diabetic complications, and she was admitted for a similar amputation of the other lower extremity. There was no history of transfusion for the past three months. Initial typing showed that she was A positive with a positive screen and a weakly positive DAT with polyspecific reagent and with IgG. Evaluation of the serum and eluate led to the identification of an autoantibody with anti-Kp b specificity. A phenotype on the patient was performed with the following results: C+, c-, E-, e+, M+, N-, S+, s+, K1-, Fy(a+b+), Jk(a+b-), Le(a-b+), P 1 +, k(+), Kp(a-),Kp(b+), Js(a-), Js(b+). Both autologous and allogeneic autoadsorptions were performed, and both techniques successfully showed the absence of underlying alloantibodies. Evaluation of Kell antigen system expression was assessed from the reaction strength of the patient's cells with antisera against k, Kp(b) and Js(b); in all three cases the reaction strength was greater than 2+ on a scale of 0 to 4+. Conclusion: Previous reports of autoantibodies with anti-Kp b specificity have been reported over the past three decades. These autoantibodies with anti-Kp b specificity occur in patients who are not only phenotypically Kp(b) positive, but also in some cases phenotypically Kp(b) negative. In the published literature all these autoantibodies with anti-Kp b specificity are transient, and they are associated with concomitant depression of Kell system antigen expression. In the case presented the patient had strong Kell antigen expression despite the presence of the autoanti-Kp b . The underlying mechanism behind the formation of this autoantibody and its relation to Kell antigen expression remain unknown. Background: MNS9(Vw) is a low frequency antigen of the MNS system (former Miltenberger subsystem). The incidence of the MNS9 antigen is around 0.057% in Caucasian populations. In contrast to the rarity of the antigen, anti-MNS9 is a common antibody in deliberate searches (around 1% of the sera). It is frequently associated with other antibodies directed against low frequency antigens. Anti-MNS9 has caused sometimes severe hemolytic disease of the newborn (HDN) but the reports are rare. The responsibility of the antibody in hemolytic transfusion reactions is controversial. We report a case of significant HDN associated with anti-MNS9. Case report: the mother was a 41 years old, gravida 4, para 3, caucasian woman. She had 2 children from a first husband. 2 pregnancies occured from a second husband: the first one ended in an amniocentesis followed by therapeutic abortion (23 weeks); the second one is described in this report. Her blood type was A, D+C+E-c+e+, K-. The antibody screening was negative during the pregnancy that was devoided of significant complications (6 normal ultrasonographies and 1 amniocentesis at 17 weeks). At birth the hypoxemic newborn (36 weeks gestation) showed a bicytopenia (Hb: 4.8 g/dl; Ht: 13.1%; Platelets: 26.10 9 /l) associated with a major hepatosplenomegaly (splenic size 7 cm). The total bilirubin level was 201 micromoles/l (indirect, 149 micromoles/l). The blood type was O, D+C+E-c+e+, K-; the direct antiglobulin test (DAT) on fetal cord blood was positive (IgG: 3+). Blood samples of the mother, father and newborn were referred to our laboratory. The antibody screening of the mother's serum showed only a weak anti-E. The compatibility test of the serum with the father RBC's (type O, E-) was strongly positive. The DAT on the newborn RBC's was positive (IgG: 2+). The eluate gave a strong reaction with the father's RBCs. The identification of the antibody in the mother's serum and the newborn's eluate showed the presence of anti-MNS9 (3 different MNS: 9 RBC's tested). The father was typed MNS: 9. The anti-MNS9 in the mother's serum was IgG (titer 128). The newborn was treated with phototherapy and two transfusions were required: the first (50 ml) on the day of birth, the second (40 ml) 3 weeks later. Two weeks after birth the splenic volume and the platelet count were normal. The Hb level was 7 g/dl, the reticulocyte count 171801/mm 3 . Conclusion: this is a new documented case of a significant severe HDN associated with anti-MNS9. This is a rare event although the antibody is common but not detected on routine antibody screening tests. This case also indicates that the potential responsibility of this antibody should be investigated in hemolytic transfusion reactions when the antibody screening is negative. Background: There are very few case reports of hemolytic disease of the newborn secondary to anti-Rh17, an antibody to a high incidence red cell antigen, found in the serum of people with the rare D-phenotype. Case Report: A gravida 5, para 2 Ecuadorian female, born in November 1974, was seen in the obstetrical clinic in late August 2003. She was O, Rh positive with a positive antibody screen and a negative direct antiglobulin test. Antibody identification showed the presence of an anti-Rh17 by multiple techniques. Her erythrocytes typed D+C-c-E-e-Rh17-, and her most probable phenotype was D-/D-. Her antibody titer by IgG antiglobulin tube technique was 1 : 128 (with the critical value for obstetrical notification for irregular antibody titers being at least 1 : 16). High risk follow-up was recommended, but amniocentesis and periumbilical blood sampling were declined by the patient. The parents were not consanguineous. One sibling was available, but not compatible. Therapy with iron and erythropoietin was discussed as was autogeneic donation. Compatible rare units were searched for as well, but to no avail. One infectious diseases marker negative autogeneic unit was donated in March 2004 (with Blood Bank physician approval) and stored as frozen red blood cells. By this time, the patient's anti-Rh17 titer had risen to 1 : 256, where it remained until delivery. A 3,065 g female child of 37 weeks gestation was born via normal spontaneous vaginal delivery in early April 2004: the cord blood showed icterus and moderate hemolysis; the direct antiglobulin test was 3+ positive with IgG specific antiglobulin reagent; the eluate was reactive against all erythrocytes of standard Rh phenotypes. The neonate was O, Rh positive, C+c-E-e+. At birth, the hemoglobin/hematocrit were respectively 8.6 g/dL and 27.0%; the indirect bilirubin was elevated at 9.5 mg/dL. Three double volume exchanges were required for rising indirect hyperbilirubinemia, with the respective pretransfusion values being 21.7 mg/dL, 17.7 mg/dL, and 8.2 mg/dL. The first double volume exchange used the mother's reconstituted compatible, irradiated, deglycerolized autogeneic unit (after checking to make certain that the mother did not need it); the second and third double volume exchanges used "incompatible" (by full IgG antiglobulin crossmatch with polyethylene glycol against the mother's plasma) arbitrarily selected rr O reconstituted red blood cells. Within a week, the neonate was discharged to be followed in clinic. Conclusions: This is an exceptional case involving hemolytic disease of the newborn secondary to a very rare alloantibody, in which incompatible blood was used successfully for two of three double volume exchanges. Background: The controversy of providing ULR continues. It is, however, agreed that prestorage leukoreduced products are preferable to poststorage (bedside leukofiltration). We provide leukoreduced products only for selected patients. We reviewed the effectiveness of this process by two questions, 1) How frequently are standard RBCs issued with a leukoreduction filter? and 2) How frequently do patients receive standard RBCs emergently or before Blood Bank is notified that leukoreduced products are needed? Methods: Our computer records from January and February 2004 were reviewed.

Objective: To define the profile of red cell units (RBCs) utilization and to evaluate the transfusion appropriateness in a 622-bed university hospital Blood Transfusion Service. Methods: Over a 3-month period the list of discharged patients was daily checked to identify patients who had been RBCs transfused during hospital admission. Patients under 16 were excluded. Charts for patient's age, gender, diagnosis, comorbidities that affected oxygen delivery, number of units received and haemoglobin (Hb) level at discharge were reviewed. The discharge Hb level, coupled with age and comorbidities, was used to determine the appropriateness of RBCs transfusion. Three different levels were evaluated: Hb12 g/dL, for strict, intermediate and excessive transfusion criteria. Results: Among 4333 (2256 male, 2077 female) discharge summaries reviewed, 385 (232 male, 153 female) transfused patients were identified. The median patient age was 68 (range 16-96) with 61 older 80 years. A total of 1377 RBCs units were administered. Almost half (44%) of the transfused patients had comorbidities that affected oxygen delivery. The surgical patients represented 55% of all patients. The discharge Hb level was determined in all but 14 patients and mean Hb level was 10.1+/-1.3 (10.0+/-1.4 male, 10.2+/-1.3 female p = 0.00).

Sixty units of RBCs were either issued emergently to patients designated to receive leukoreduced products or to patients who were subsequently so designated. No adverse affect, immediate or long term, was known to be associated with transfusion of these units. Since leukoreduced RBCs cost approximately $20/unit more, providing all patients leukoreduced rather than standard units would have increased the cost approximately $30,780 although there may be no associated increased benefit. Results: During these two months 2850 units of RBCs were transfused, (Table) 46% were leukoreduced RBCs. The vast majority, >95%, were prestorage leukoreduced. Conclusion: A review of the RBCs transfused during Jan-Feb, 2004, revealed 46% were leukoreduced (>95% prestorage leukoreduced). There were 60 units of standard RBCs transfused to patients who may have benefited from leukoreduced products, but 44 units were transfused to a single patient. Universal rather than selective leukoreduction would have had increased costs to $30,780. Further studies are needed to determine if ULR is cost effective.

M Pujol, J R Grifols, Centre de Transfusió i Banc de Teixits, Badalona, Spain; L Perez, Facultat de Ciències, Bellaterra, Spain Background: Since no single transfusion threshold can be applied in all settings, transfusion should be given on a case-by-case and unit-by-unit basis. Table: Age, comorbidities, surgical indications and number of units transfused for each Hb level group. (*p = 0.002, **p = 0.00). Conclusions: Most RBC transfusions were concentrated in male patients over 65 years of age. However transfusion trigger was not different for women and men. Ninetytwo patients (25%) were identified exceeding the strict criteria level. Among them, 29 received RBC transfusion clearly in excess of their needs. An intermediate Hb level was associated with a subset of patients who were older and presented more comorbidities than the patients whose discharge Hb level was below 10 g/dL. A transfusion of a single unit could have been appropriate in most of these patients. There was greater prevalence of inappropriate transfusions among surgical indications. Discharge Hb level coupled with age and medical comorbidities can be used to determine the appropriateness of transfusion.

Perioperative Blood Salvage-The More Efficacious Autologous Deposit?

G Singbartl, C Buetner, AIT-ENDO-Klinik Hamburg GmbH, Hamburg, Germany Background: Peri-operative blood salvage (PBS) and preoperative autologous deposit (PAD) are established blood conservation measures. PAD can only be applied in elective surgery and carries a high discarding rate. PBS, however, can be applied both in elective and emergency cases, but is considered an expensive measure. So far, no data have been published analyzing/comparing efficacy of either measure. Methods: Prospective analysis of data from 693 orthopedic pat. with PAD concerning increase in total RBCmass (+RBC). After calculating each pat.'s +RBC due to PAD, blood-loss (BL) that could be covered for by re-transfusion of +RBC was calculated, using following re-transfusion concept: Re-transfusion of +RBC to maintain hct min at the desired level (either 24, 21, or 18%) during ongoing BL, and securing normovolemia by additional colloid infusion. These data were compared to a PBS-model with a 30% recovery of RBC lost during surgery (PBS30), until reaching hct min. The same mathematical re-transfusion concept was applied to PAD. Statistically significant differences (*) between PAD and PBS (by t-/U-test according to Gaussian distribution) were adopted when p < 0.05. Results: Tab. 1 summarizes the relevant data (693 pats., 282 males; 62.7 ± 10.8 yrs.; pat.'s estimated blood vol. 4.6 ± 0.8 l; pat.'s RBC mass 1.8 ± 0.4 l). Conclusions: Comparison of PAD and PBS demonstrates PBS to be superior to PAD, regardind +RBC and substitution of blood-loss at different hct min. The lower hct min, the more efficacious is PBS, and the lower the percentage of patients taking advantage from PAD. Background: As 60-70% of transfusions occur in the operating room (OR), decreasing blood donations in the United States may result in delay or cancellation of surgery. Thus efforts to improve blood utilization should focus on surgical blood usage. The Maximum Surgical Blood Ordering Schedule (MSBOS) system has been in use since the late 1970's as a guide for the preparation of crossmatched packed red cells (PRBC) for various surgical procedures. Subsequent technical advances in both OR and blood bank have decreased the apparent need for presurgical crossmatching, but no detailed update to the MSBOS has been undertaken. Thus we have studied surgical blood use at our university medical center to determine where a type and screen can safely replace a preoperative crossmatch. Materials and Methods: We retrospectively reviewed all blood bank and medical records for surgical procedures performed > 20 times per year at our 400 bed medical center between 10/02 and 9/03 for which the MSBOS recommends that 2 Units PRBCs be crossmatched preoperatively, to determine how many units were crossmatched pre-op, transfused in the OR or transfused within 72 hrs post-op. We calculated the average number of units crossmatched prior to each type of procedure (a type & screen counted as 0 units) as well as the average used in the OR, within 24 hours of surgery, 24-48 hrs post-op and 48-72 hours post-op. Results was performed on cardiothoracic (CT) surgery. Three hundred CT surgical cases performed during the three-month period were reviewed. CT cases represented conventional surgical and minimally invasive approaches; adult and pediatric; cardiac, aortic aneurysms, and thoracic procedures. In this audit, the number of RBC units crossmatched was defined as the preoperative order or the preoperative plus added-on orders originated perioperatively from the operating room. Hb levels available postoperatively within 12 hours were reviewed as an indirect measure of assessing the transfusion trigger. Results: The overall crossmatch to transfusion (C/T) ratio was 2 : 1 (1,660 units crossmatched/830 units transfused). The C/T ratios of individual surgeons ranged from 1.5 to 5.6: averages of 1.9 (1.5-2.4) for cardiac and 3.8 (2.2-5.6) for thoracic surgeons. In 130 cases (43%) the numbers of crossmatched units (XM) were same as the maximum surgical blood order schedule (MSBOS) and in 156 cases (52%) XM were higher than the MSBOS. In the former (XM = MSBOS) group, 43% of patients received transfusions (15% were transfused same as the MSBOS and 28% fewer than the MSBOS) and 57% did not receive transfusions. In the XM > MSBOS group, transfusions were given in 84% of the patients (TX > MSBOS, 45%; TX = MSBOS, 15%; TX < MSBOS, 24%) and no transfusions in 16%. Overall, the current MSBOS met or exceeded the perioperative transfusion needs in 76% (no transfusions required in 34%), but the remaining 24% cases required more RBC units than the MSBOS. The transfused patient group (N = 197, 66%) had a mean postoperative Hb level of 11.6 g/dL, while the non-transfused group (N = 103, 34%) had 11.2 g/dL. Conclusions: The institution's current MSBOS was adequate to meet or exceed the perioperative transfusion needs in 76%, while 24% of cases required more RBC units than the MSBOS. The postoperative Hb levels were not significantly different between the transfused and the non-transfused groups: 11.6 vs. 11.2 g/dL, respectively. The finding that 62% of cases required fewer transfusions than the MSBOS (28%) or none at all (34%) implies that implementation of a uniform procedure for assessment of a patient's bleeding risk would improve blood ordering practices.

A M Sakashita, A T Vacarini, M A Quilici, A Teruya, S Mies, J M Kutner, Hospital israelita Albert Einstein, São Paulo, Brazil Background: A liver transplant program presents one of the greatest challenges to a blood bank, demanding maximal transfusional support. Patients undergoing orthotopic liver transplantation (OLT) are susceptible to massive blood loss due to preexisting liver disease, and could require large amount of blood transfusion. This study reviews blood components usage during our first 2 years performing OLT. Methods: Data of transfusion therapy in 272 patients undergoing OLT between January 2002 and March 2004 were reviewed. Results: There were a total of 177 male, and 95 female patients, with median age 49 yrs (range 8-78 yrs) who were submitted to 281 OLT. Intraoperative autologous blood recovery device was utilized in 61 transplants, with a median reinfused volume of 283 mL (range 216-8013 mL). Blood components usage is described in table below: Discussion: Colectomies, total hip replacements and compression screw insertions can all be managed efficiently with a preoperative type and screen: less than 0.1 PRBC was used in the OR (on average) for these procedures. Cardiovascular procedures had a higher rate of intraoperative transfusions, but 72% (174/241) of the coronary artery bypass grafts did not require any intraoperative transfusion. A type and screen or crossmatching of only one PRBC would suffice for these procedures. An efficient blood bank can make PRBC available in the OR in <15 minutes with a pre-op type & screen on record. Studies such as ours may help in modernizing the MSBOS in medical centers where this is the case. Surgeons may thus be convinced of the safety of operating without routinely having two units of PRBCs immediately available in the OR.

Y Choo, L Rudon, C F Whitsett, Department of Pathology, Mount Sinai School of Medicine, New York, NY Background: Our hospital transfusion committee regularly conducts transfusion audits to monitor the blood ordering and transfusion practices of different clinical services. The aim of this quality program is to improve the utilization of blood products and blood bank resources. Methods: An audit There were 14 patients who required a surgical review, and blood components utilized are described below:

Conclusion: Our data indicate that transfusion therapy in a liver transplant program is still a challenge to the blood bank, in order to provide best support to the patient.

T Tasaki, Iwate Medical University, Morioka, Japan; H Ohto, T Igari, T Ogata, Fukushima Medical University, Fukushima, Japan; A Suwabe, T Higuchi, Iwate Medical University, Morioka, Japan; A Ishida, M Handa, Keio University, Tokyo, Japan Background: Leukoreduced blood is useful for the prevention of nonhemolytic transfusion reaction, especially for patients in need of frequent transfusion; however, the significance of such blood used for those undergoing surgery remains obscure. To clarify this issue, the correlation between the number of leukoreduction filters used in our hospital and the frequency of transfusion-related adverse events was retrospectively investigated. The influence of blood transfusion on postoperative infection was also assessed for patients operated upon for colorectal cancer. Method: The number of leukoreduction filters used in our hospital between January 1998 and December 2003 was obtained from records of the administrative department. Adverse events due to blood transfusions were investigated using the records of the blood transfusion service, and details of postoperative infection were obtained from the records of the clinical department. Results: In 1998, almost 600 leukoreduction filters for red cell components were used, but the number gradually decreased to about 200 in 2003. However, the number of reports of adverse events remained essentially unchanged throughout the period, averaging around 30 per year. Of the 152 patients who underwent surgery for colorectal cancer from January 2001 to December 2003, 15 received homologous blood. Inflammatory parameters such as CRP and WBC levels were higher for those transfused patients than for untransfused patients until around postoperative day 14. However, no significant correlation between blood transfusion and postoperative infection could be demonstrated; 3 of 15 transfused patients had postoperative infection and the rate for untransfused patients was 18/137 (p = 0.46). Conclusion: It seems that leukoreduction of red cell components for operations is ot required for the prevention of nonhemolytic transfusion reaction, although problems such as the transmission of infectious disease through leukocytes remain to be resolved.

R G Cable, W B Ells, R L Edwards, American Red Cross Blood Services, Farmington, CT Background: Development of HLA-or platelet-specific antibodies in transfused patients may cause refractoriness to platelet transfusions. The use of leukoreduced blood products has been shown to reduce this outcome. Study Design: Since 1994, our blood center, which provides all blood components for regional hospitals, has screened referred patient samples for the presence of these antibodies using a solid phase red cell adherence assay (Immucor, Norcross, GA) which detects IgG antibodies against HLA antigens on platelets or platelet-specific antigens. Screens are performed by incubating patient sera with 8 randomly selected apheresis platelets. One or more positive reactions are defined as a positive screening result. If the screen is positive, screen negative "crossmatched" apheresis platelets are selected for transfusion. During this period, the routine use of prestorage leukoreduced (PLR) blood components was introduced in our blood service region. From 1994-98, PLR products were not available, but there was ~25% use of bedside leukoreduction filters in hospitals. From 1999-01, PLR products were gradually introduced, from an initial distribution of 15% LR RBCs in early 1999 to 85% LR RBCs and LR platelets by late 2001. Universal prestorage leukoreduction (ULR) began in 2002. We performed a retrospective analysis of initial antibody screens, grouped into these three time periods corresponding to the three phases of PLR. Statistical analysis (chi square) was performed. Results: The proportion of positive screening results was significantly reduced following the introduction of ULR (Table) , but was not affected during the phase-in period. The annual number of negative screening results remained relatively constant within the three time periods (69, 53, 62) while the annual number of positive screening results decreased (55, 44, 26) . Conclusions: We conclude that ULR may have substantially reduced the frequency of HLA and platelet antigen alloimmunization in platelet recipients in our blood services region. However, the incidence of platelet refractoriness for other than immunologic reasons appears to have remained constant. # p-value n.s. compared to 1999-01, + p-value < .001 compared to 2002-03.

S Zhao, X Chen, J Hu, R-Q Li, Southwest Hospital, Chongqing, China

Background Clinical reactions of blood transfusion generally occur in 2 to 10 percent of transfusion. The febrile nonhemolytic transfusion reactions (FNHTRs) caused by the transfusion of blood leucocytes is the main one of these reactions. In order to reduce the clinical reaction of blood transfusion, the leukocyte filtrated RBC and platelet concentrates are used in our hospital. Methods One thousand and seven hundred forty patients with diseases of blood system, respiratory system, digestive system, circulatory system, trauma, burn and tumor were selected to receive RBC concentrates with leukocyte filtration. Another 100 patients with similar diseases were selected to receive non-filtrated RBC concentrates. Two hundred and fifty patients with acute or chronic leukemia, aplastic anemia, multiple myeloma, thrombocytopenia purpura, diabetes mellitus, cirrhosis of liver, upper gastrointestinal hemorrhage, severe hepatitis, burn and cancer post radioactive or chemical treatment were randomly divided into two groups. The experiment group with 130 patients was selected to receive filtrated platelets concentrate. The control group with 120 patients was selected to receive nonfiltrated platelets. The incidence rates of FNHTRs in all patients were investigated. Results There were only 2 in 1740 patients with FNHTRs occurred with leukocyte filtrated RBC concentrates transfusion. Twelve of 100 patients with non-filtrated RBC concentrates transfusion occurred FNHTRs. The incidence of FNHTRs was separately 0.01% and 12% in filtrated and non-filtrated RBC concentrates transfusion. Twenty-five and 8 patients with FNHTRs occurred separately in non-filtrated or filtrated platelets concentrations. The incidence of FNHTRs was separately 20.83% and 6.15% in non-filtrated and filtrated platelets transfusion. Conclusion Leukocyte filtration of RBC and platelet concentrates can reduce FNHTRs in blood transfusion.

Clinical Outcomes of Reverting from Leukoreduced to Non Leukoreduced Blood in Cardiac Surgery M K Fung, Fletcher Allen Health Care / Univ. of Vermont, Burlington, VT; K F Moore, M Ridenour, W J Mook, UPMC-Shadyside Hospital, Pittsburgh, PA; D J Triulzi, University of Pittsburgh Med Center / Institute for Transfusion Medicine, Pittsburgh, PA BACKGROUND: Proven clinical benefits of leukoreduced (LR) blood components include reduced febrile non-hemolytic transfusion reactions, alloimmunization against HLA-antigens, and cytomegalovirus transmission. Immunomodulatory effects of leukoreduction have also been postulated to play a significant role in the clinical outcome of open heart surgery. We recently demonstrated an improvement in the postoperative length of stay (PLOS) with the use of LR blood compared to a historical cohort of cardiac surgery patients who received non-LR blood. Followup data is now presented to further substantiate the benefits associated with using LR blood in cardiac surgery after a twelve month period in which LR blood again was no longer routinely used. STUDY DESIGN AND METHODS: This study is an extension of a previous prospective case control study in which all patients admitted over a one year period for open heart surgery at a single hospital were given LR cellular blood products (Group 2) and were compared against a historical cohort in the previous year who were given non-LR blood products (Group 1). This current study compares the clinical outcomes of a new cohort of patients admitted for cardiac surgery in the subsequent 12 months where non-LR blood was again routinely used (Group 3) versus the clinical outcomes of patients from Group 1 and Group 2. RESULTS: The mean PLOS increased in patients receiving blood after switching to a policy of using non-LR blood (Group 3, 10.8 days; 95%CI 10.0-11.6; N = 595) versus the group that received LR blood (Group 2, 9.5 days; 95%CI 8.9-10.0; N = 645; P = 0.045). This is in comparison to an earlier cohort of patients who received non-LR blood (N = 501) with a mean PLOS of 10.1 days; 95%CI 9.4-10.7 (P = 0.005, Group 2 vs. Group 1). In comparison, the mean PLOS was unchanged for control patients who received no blood from the three time periods (6.4 vs. 6.2 vs. 6.5 days; P > 0.05; N = 308 for Group 1, N = 296 for Group 2, N = 336 for Group 3). The mean RBC usage was decreased in the group that received non-LR blood (Group 3, 3.7 units; 95%CI 3.4-4.0; N = 595) versus the group that received LR blood (Group 2, 4.0 units; 95%CI 3.7-4.4; N = 645; P = 0.034). Otherwise, in the transfused patients from Groups 2 and 3, there were no significant differences in non-RBC blood usage, estimated blood loss, bypass (pump) time, mediastinitis rates, or operative mortality rate. CONCLU-SIONS: Although the mechanism of the benefits of leukoreduction in open heart surgery remain to be elucidated, we demonstrate an associated improvement in cardiac surgery outcomes with the use of LR blood and an associated worsening in cardiac surgery outcomes when switched back to use of non-LR blood. Background. ABO-incompatible errors in organ transplantation have been reported in 1 in 800 cases (J Heart Transplant 9:376, 1990), 47 times higher than in blood transfusion (Transfusion 40:1207 (Transfusion 40: , 2000 . Effective preventive measures are needed in every transplant program. System design for patient safety must incorporate all relevant professional disciplines, proven error-reduction techniques, and a focus on standardizing the steps in any process. Methods. Our program performs liver, kidney, kidney-pancreas, pancreas, and small intestinal transplants. We developed a checklist* for hospital transplant coordinators which requires verbal readback confirmation* of the deceased donor's blood type when the organ is offered, twoperson ABO confirmation at the procurement site, and two ABO results for both donor and recipient. The coordinator notifies the blood bank of the deceased donor's United Network for Organ Sharing number, or the living kidney donor's medical record number, and the intended recipient. Organs from all deceased donors and living-kidney donors are brought to the blood bank, along with records of donor number and ABO type. The blood bank reviews the donor and recipient information and their ABO records, and enters the donor and recipient ABO types into a computerized crossmatch test. Any unresolved identification discrepancy or ABO-incompatible match is reported immediately to the surgeon. The blood bank issues the organ to the operating room (OR) with a compatibility tag like that used for blood components. When the organ arrives in the OR, a mandatory "time-out"* is conducted to verify donor and recipient identification and blood types aloud. (*Checklists, verbal readback and surgical time-outs are methods recommended in the 2004 National Patient Safety Goals of the JCAHO.) Any intentionally ABO-incompatible organ, such as an A 2 kidney, can be issued after review, but only with express approval of the transplant surgeon and blood bank physician. Results. We have confirmed ABO compatibility for 138 organs to date, including 52 living-donor kidneys. The median stat processing time in the blood bank was 8 min (90% £ 20 min). Six clerical discrepancies in the donor number or the recipient name initially provided to the blood bank were resolved during review. Other benefits include uniformity of practice among coordinators by use of checklists, and documentation of donor-recipient ABO types and compatibility in the recipient's laboratory chart report. Conclusion. We have incorporated recognized successful patient safety principles and blood bank procedures into a multilayered system for confirming organ ABO compatibility in our transplant hospital.

A Odeleye, S F Donnelly, P D Mintz, University of Virginia Health System, Charlottesville, VA BACKGROUND: Patients who require platelet transfusions may manifest refractoriness. This may be caused by nonimmune clinical factors that lead to diminished platelet survival, such as sepsis, fever, disseminated intravascular coagulation, splenomegaly, or administration of certain drugs. Alternatively, refractoriness may be caused by antibody-mediated destruction of transfused platelets, usually as a result of alloimmunization due to HLA antigens found on white blood cells and platelets. The incidence of refractori-ness due to all causes has been estimated to be as high as 15-25%. The CAPTURE-P ® Solid Phase System for the Detection of IgG Antibodies to Platelets (Immucor Inc., Norcross, GA), using fresh patient sera (£48 hours old and stored at 1-10°C, as stipulated by product insert) and donor platelets, is used to crossmatch platelets for transfusion to patients who are refractory to platelet transfusion due to alloimmunization. METHODS: We compared platelet crossmatch results using patient samples <48 hours old with the results of crossmatch with the same sera >72 hours old, stored at 1-6°C, and with the same sera that had been frozen at -30°C and thawed, in a series of 9 samples from 7 patients who had been multiply transfused with platelets and were refractory to platelet transfusion: 5 with hematologic malignancies, one with idiopathic thrombocytopenic purpura, and one with multiple trauma injuries. Each patient's assays were done using the same peripheral blood specimen and the same donor platelets, and the reagents used for each patient were the same lot number each time. The patient plasma was centrifuged to remove fibrin clots and the platelets were washed prior to each set of crossmatches. RESULTS: The results of platelet crossmatching using sera £48 hours old were comparable to the results obtained using sera older than 72 hours, and sera that were frozen and thawed (see Table) . CONCLUSION: These results suggest that sera for platelet crossmatching using the CAPTURE-P ® assay may be stored for at least 72 hours, or may be frozen and thawed, without the danger of platelet units with falsely negative crossmatch results being supplied by the transfusion service. This may be particularly helpful for patients when repeated phlebotomy may be problematic.

T Tasaki, K Gotoh, S Satoh, J Takadate, M Kamitsukue, A Suwabe, O Funato, R Sasaki, K Saitoh, Iwate Medical University, Morioka, Japan Background: Fresh frozen plasma (FFP) is sometimes used in hepatic and/or pancreatic operations. To avoid, or at least reduce the use of homologous plasma during surgery, donation of autologous plasma at 10% of patient's blood volume (BV), in addition to red blood cells, was initiated in December 2003. The safety and the significance of this program were retrospectively assessed between the two periods before and after the introduction of the program. Method: Withdrawal of whole blood (0.1 ¥ BV ¥ Ht/(1-Ht)) from a patient was followed by separation of plasma (0.1 ¥ BV ¥ Ht)* by gentle centrifugation, and residual red cells were reinfused. Whole blood (0.1 ¥ BV) was drawn again, and centrifuged, and plasma (0.1 ¥ BV ¥ (1-Ht))** was obtained. Concentrated red cells (0.1 ¥ BV ¥ Ht) were stored in a refrigerator at 4°C and plasma (* + ** = 0.1 ¥ BV) was frozen at -40°C. Patients were usually administered rEPO (24,000 u/W). Results: Seven patients (4 males, 3 females; mean age 69 years; mean body weight 54 kg) donated on the average 814 ml of autologous red blood cells and 761 ml of plasma over a donation period of 11 days, during which time Hb level fell from 12.3 to 11.8 g/dl, and serum protein concentration from 6.3 to 6.1 g/dl. Vital signs such as blood pressure and heart rate were almost stationary. With two of the seven patients it was possible to avoid the use of homologous blood during surgery. Five were transfused homologous blood, but the units were small. On the other hand, ten patients who had not donated autologous blood (4 males, 6 females; mean age 65 years; mean body weight 52 kg) each received 650 ml of homologous red cell components and 600 ml of FFP. The use of homologous blood could not be avoided in any of these cases. Perioperative liver function, and coagulation parameters (PT, APTT, etc.) changed similarly between the two groups. Conclusion: Autologous plasma at 10% of blood volume and red blood cells could be safely donated at the same time. This program also appears to have contributed to a reduction in the use of homologous blood in hepatic and/or pancreatic operations. However, it was not possible to clarify the clinical benefit in this short-term study.

M I Gyongyossy-Issa, Canadian Blood Services, Vancouver, BC, Canada; C J Carter, A Racaza, University of BC, Vancouver, BC, Canada; D V Devine, Canadian Blood Services, Vancouver, BC, Canada Background: The mechanism of action of IVIG is elusive although a number of hypotheses for its action have been offered: the Fc-mediated splenic blockade effect of IVIG seems eminently plausible, but how that blockade is achieved is still unclear. For immune thrombocytopenia (ITP), the early IVIG-induced improved platelet counts have been ascribed to this Fc blockade, while long-term improvement is thought to be due to immune modulation. Methods: We followed patients with ITP (low platelets) and chronic idiopathic demyelinating polyneuropathy, CIDP (normal platelets), and collected time-course blood samples before treatment, 1 h after each infusion block, then 1, 3 & 7 days after IVIG infusion. The samples were analyzed for plasma viscosity, protein content, fibrinogen levels, circulating immune complexes (CIC), evidence of complement activation by C3a and SC5b-9 levels, as well as conventional blood cell counts. Results: We have new evidence that (CIC) are formed immediately upon IVIG infusion, and that their presence correlates with increased plasma viscosity that is unrelated to fibrinogen levels. These CIC levels take more than a week to return to baseline and plasma viscosity is strongly correlated with CIC levels. This is true for both ITP and CIDP patients. The IVIG-caused immune complexes carry activated C3 (C3b) indicating that complement activation had taken place during IVIG infusion. Parallel increasing and decreasing levels of plasma C3a and SC5b-9 further substantiate that complement activation is a direct consequence of IVIG infusion. An inverse pattern of decreasing red blood cell counts suggests red cell sequestration as a consequence of CIC clearance. Conclusion: Based on these data we conclude that IVIG causes the formation of CIC which, through complement activation, may initiate a transient inflammatory response. We hypothesize that such an inflammatory response may tip the innate immunological balance toward normalcy and thus produce a beneficial effect on the underlying immune dysregulation that initially caused the ITP or the CIDP. As only certain IgG subclasses activate complement, these data do not exclude the involvement of Fc-mediated regulatory events. Background: CD40 ligand (CD40L) is a co-stimulatory molecule needed for cell-mediated & humoral immunity, is biologically active as a heterotrimer & is released from resting platelets (plts) on activation. Soluble (s) CD40L has been associated with plt transfusion (tx) reactions. A 29 y.o. patient (pt) with X-linked hyperIgM syndrome, infections & papilloma virus-associated skin cancers received a trial of plt tx to attempt to replace CD40L, permit immunoglobulin (Ig) switching (IgM to IgG) & generate tumor suppressive immunity. Methods: Plt test dose & apheresis units were prepared (ABO identical, leucodepleted & irradiated, but not HLA-matched). sCD40L in plt rich plasma (prp) & supernatant (sn) was measured by ELISA: monomers & trimers (R&D), trimers only (Biogen), pt mutant vs wild-type (in-house assay) after collection, activation & storage. Assessment of serum Igs & in vitro APC, B & T-cell activity at baseline & post-therapy. Results: sCD40L was stable (prp) or increased (sn) in plts stored up to 6d; with levels (mean, days 0-6, in pg/mL) higher in prp (mono + trimers 21901; trimers only 16192) than sn (mono + trimers 13443; trimers only 4414). Plt CD40L release was not increased 1-4 h after stimulation with epinephrine (performed on d2, mean pre-/post-stimulation, in pg/mL): in prp: mono + trimers 26628 vs 22734; trimers only 16985 vs 17391; in sn: mono+trimers 12589 vs 12415; trimers only 4560 vs 4715). Pt baseline IgG was low but not absent (reflecting concurrent IVIG therapy) & IgM was high; pt normal CD40L was v. low & mutant CD40L v. elevated vs controls. Test dose & plt tx (6 apheresis plts over 8 wks) were all well tolerated. Pt CD40L increased rapidly after infu-sion but declined to baseline by 3-4 h. Impaired T-cell (proliferation to tetanus immunisation; IFN g & IL2 production) & dendritic cell (IL-12) activity was not corrected by sCD40L in vitro or after tx. No clinical improvement seen following therapy. Conclusions: CD40L is expressed on plts & accumulates in supernatant during storage. Increased expression reported following activation by thrombin, ADP & collagen was not seen after epinephrine stimulation in our study. Plt tx in this patient was safe but had no measurable in vitro or clinical effect on hyperIgM-related complications. The patient continues on IVIG & other supportive therapy. If provision of functional CD40L via plt tx can benefit X-linked hyperIgM syndrome, greater amount, frequency and/or duration of therapy may be required.

The Effect of Ethnicity on the Ability to Provide High Quality HLA Matched Apheresis Platelet Products J L Jacobson, R Musoko, New York Blood Center, New York, NY Background HLA matched apheresis platelets are frequently ordered for platelet refractory patients. Our blood donor base does not ethnically reflect many of our hospital patients. This discrepency led us to question whether the lack of diversity in our donor base was impacting the quality of HLA matched platelets we were able to provide. Methods The ethnicity of the apheresis platelet donor base was analyzed using the ethnicity data provided by the donors on their health history registration form. The hospitals requesting HLA matched platelets during 2003 were retrospectively asked to provide the ethnicity of the recipients. The ethnicity of the donor base was compared to the ethnicity of the recipients. The quality of the HLA matchgrade of the platelets provided was analyzed. Results During calender year 2003, the apheresis platelet donor base was composed of roughly 12,000 donors who had donated in the past year. 87.7% of the donors identified themselves as White. 4.8% of the donors identified themselves as Hispanic. 4.3% of the donors identified themselves as Black. 3.1% of the donors identified themselves as Other. Of the 12,000 donors, 2,970 had HLA typings. During 2003, 85 patients received a total of 838 HLA matched platelet products. 52.9% of the patients were identified as White. 8.2% patients were identifed as Hispanic. 9.4% of the patients were identified as Black. 10.6% patients were identified as Other/Unknown. Of the 838 HLA matched platelet products, 2.3% were match-grade A, 78.2% were match-grade B,19.6% were match-grade C. The results of the HLA match-grade by ethnicity are shown in the table below. Conclusions Grade A HLA matches are infrequently found within our existing HLA typed donor pool and when provided were only available to White recipients. Finding a quality HLA matched product for a non-White recipients is challenging. 28.9% of White recipients, 42.9% of Hispanic recipients, 37.5% of Black recipients, and 33.3% of Other/Unknown recipients recieved grade C matched HLA platelet products. The current HLA typed apheresis donor base is inadequate to provide high qualtiy HLA matched platelets and is currently being grown with a target of 15,000 HLA typed donors in order to make better matches available to the community.

D Benomar, D Lemau, EFS Ile de France, Paris, France; C Boulat, Hopital Necker, Paris, France At our center over three years 232 granulocyte concentrates (CGs) prepared by apheresis have been transfused in eight patients (aged 2 to 19 years) presenting septic granulomatosis. Seven of these patients had visceral aspergillosis resistant to anti-infectious treatment. The two-year-old child received 31 CGs at a provincial site. Granulocyte transfusion was pursued until there was a clear improvement in the infectious state. In the 19-yearold with MacLeod syndrome and antipublic antibodies, it was apparent straightaway that transfusion was not possible.

134A SCIENTIFIC SECTION TRANSFUSION 2004-Vol. 44, Supplement At the Necker Hospital, patients received on average 2.48 ¥ 10 10 leukocytes (or white blood cells) at each transfusion. The minimum dose was 0.5 ¥ 10 10 per 10 kg bodyweight. The leukocytes were collected using a COBE Spectra apheresis system. The mean differential blood count of donors after stimulation (7.5 mg of oral dexamethasone the day before) was 10.35 (±2.9) ¥ 10 10 leukocytes. In 173 (±13) minutes, 9 842 (±974) ml of blood were treated. Erythrocytes were removed from the CGs and the final hematocrit was approximately 1%. Despite this essential precaution, and in spite of every good clinical safety of the transfusions, four patients developed antierythrocyte antibodies and five others anti-HLA antibodies. Doctor D. Background: Anti-D development after Rh(D)-mismatched allogeneic hematopoietic stem cell transplantation (AHSCT) is a risk to be considered when defining the transfusional support in this group of patients; thus, transfusion of Rh(D)-negative components is recommended. In the present age of profound immunosuppressive conditioning regimens, the risk of developing Rh(D) alloimmunization is poorly defined. We analyzed anti-D formation after Rh(D)-mismatched AHSCT in a group of patients transplanted at our institution. Study design: We retrospectively reviewed 333 AHSCT performed between January 2000 and May 2003. The stem cell source of 126 (38%) was allogeneic and 25 (20%) patients underwent a Rh(D)mismatched AHSCT. Supportive therapy included Rh(D)-negative red blood cells and platelet concentrates (PC) prepared from whole blood donations using the buffy coat method (approx. 80% of total), and PC collected by apheresis (over 20%). PC from Rh(D)-positive donors were transfused without anti-D immunoglobulin administration. Unexpected erythrocyte antibodies screening was performed by LISS-IAT using the tube test. Results Background: Thrombotic microangiopathy (TMA) is a recognized complication of malignant hypertension (HTN). Such patients have blood pressures ≥200/140 mm Hg but the condition is defined by the presence of papilledema. We report two patients with severe HTN (systolic ≥180 or diastolic ≥120), TMA, thrombocytopenia, renal failure and in one case neurological changes (4 of 5 manifestations of the TTP pentad). Case Reports: A 50-year-old male with HTN presented with blurred vision, dizziness, headache, confusion, renal failure and a TMA (with PLT = 39 ¥ 10 9 /L and LD = 2781 normal < 600 U/L). On presentation, BP was 214/133 mm Hg and an ophthalmic exam demonstrated no papilledema. With HTN control over 7 days, his platelet count rebounded (220 ¥ 10 9 /L), LD declined (1730 U/L), and mental status improved. A 60-year-old female with diabetes, HTN, Lupus, mild chronic anemia and thrombocytopenia presented with abdominal pain, SOB, renal failure and a TMA (with PLT = 83 ¥ 10 9 /L and LD = 2929 U/L). Blood pressures were 180-210/89-111 mm Hg and ophthalmic exam demonstrated no papilledema. With HTN control over 8 days, her platelet count rebounded (147 ¥ 10 9 /L), and LD declined (1624 U/L). Discussion: Although in both cases a diagnosis of TTP was considered because of overlap with the classic diagnostic pentad, neither received plasmapheresis. TTP is a diagnosis of exclusion, where there is no other likely diagnosis to explain the TMA. In cases of severe HTN (with or without papilledema), the diagnosis of TTP should be held in abeyance until the effect of HTN control can be assessed.

S Kankirawatana, S T Huang, D Arndt, M B Marques, University of Alabama at Birmingham, Birmingham, AL BACKGROUND: Thrombocytopenia is a common finding in patients admitted to intensive care units, and is often multifactorial. Frequently encountered reasons for thrombocytopenia in this population include medications, massive transfusion, infection and consumption. Heparin-induced thrombocytopenia (HIT) is a life-threatening condition that requires unique management, not dependent on platelet transfusions. We report the case of a platelet refractory patient who was unlikely to be alloimmunized despite the finding of multiple crossmatch incompatible platelets. CASE REPORT: A 33year old male with no significant past medical history was hospitalized for third-degree burns covering 46% of his body surface. On admission, he was in critical condition and required resuscitation. A complete blood cell count revealed normal hemoglobin and white blood cell and platelet counts. His hospital stay was remarkable for several surgical procedures for burn management, continuous need for intensive care and massive fluid support. By hospital-day 6, he had required eight units of fresh frozen plasma but no PRBCs or platelets. However, his hemoglobin and platelet count had trended down to 8.2 g/dL and 81 ¥ 10 9 /L, respectively. After receiving two units of PRBCs and one unit of apheresis platelets, his platelet count rose to 116 ¥ 10 9 /L. On the next day his platelets had fallen to 80 ¥ 10 9 /L and continued to rapidly decrease to 30 ¥ 10 9 /L, when he was given two more units of apheresis platelets. In the following 24 hours, despite transfusion of another four units, his count dropped to 12 ¥ 10 9 /L. Based on the clinical diagnosis of platelet refractoriness, platelet crossmatching was ordered. However, the test revealed that all crossmatched units (13/13) were incompatible with the patient's serum. Concurrently, an ELISA test for antibodies to platelet factor 4 antibodies was reported positive (hospital-day 10). With the diagnosis of heparin-induced thrombocytopenia, no more platelet transfusions were given. The patient soon developed fever and hemodynamic compromise. Despite antibiotics and aggressive ventilatory, circulatory and fluid support, his condition progressively deteriorated and he expired of presumed sepsis on hospital-day 12. CONCLUSION: Heparin-induced thrombocytopenia must be included in the differential diagnosis of decreasing platelet count in severely ill patients. We speculate that the presence of platelet factor 4heparin-IgG complexes in the patient's circulation led to platelet refractoriness and incompatible platelet crossmatchings by inducing platelet aggregation via Fc receptors.

A E Cocco, University Hospitals of Cleveland, Cleveland, OH; R A Yomtovian, M R Jacobs, S L Gerson, K A Downes, University Hospitals of Cleveland and Case Western Reserve University, Cleveland, OH Background: Platelet bacterial contamination has traditionally been underrecognized and under-reported. Organisms typically implicated in platelet bacterial contamination constitute normal skin flora. Our institution currently tests for platelet bacterial contamination by screening random donor platelet (RDP) units using pH meter IQ 120 miniLab (IQ Scientific Instruments, San Diego, CA) followed by microbiologic plate culture (MPC) of the pooled units (RDP-p). We report a case of platelet bacterial contamination with an unusual organism initially described as a FNHTR to red blood cells (RBC). Case Report: The patient (pt) is a 53-year-old male with metastatic melanoma who presented on 03/02/04 for outpatient transfusion for anemia (hematocrit: 24.9%) and thrombocytopenia (platelet count: 8 ¥ 10 9 /L). The pt received the RDP-p followed by a unit of packed RBC. The transfusion data is listed in the table below: hospitals in North Carolina. Our study suggests that actual practices may not reflect current transfusion recommendations. Further research is needed to determine if our results are indicative of national problem and to elucidate possible reasons for this discrepancy. 1995 (15-517, median 99) . Two of these patients developed CSAs (-Fy a , -Jk b , -S). Conclusion: There was a significant rate of alloimmunization pre introduction of the matching protocol. Patients who entered the program with a previous transfusion history had a low incidence of new CSA following implementation of the matching protocol. Patients receiving only phenotypically matched blood had a markedly low incidence of new CSA. This matching protocol for Rh (C, c, E, e) and K is highly effective in reducing the incidence of alloimmunization, even in chronically transfused and exchange patients.

E C Wong, D Stockwell, M J Bell, C S Smith, Children's National Medical Center, Washington, DC; R S Shirey, L Shaak, Johns Hopkins Hospital, Baltimore, MD; N L C Luban, Children's National Medical Center, Washington, DC Background: Ceftriaxone-mediated hemolysis is a rare occurrence with significant mortality risk when encountered. There have been only 10 reported cases of ceftriaxone-mediated hemolysis in pediatric pts, but none reported in pts with hemoglobin SC disease. Case Report: Pt is a 17 yo BF who presented to an outside ER with acute onset of shortness of breath and chest pain. Past medical history was significant for hemoglobin SC disease with multiple hospitalizations for vasoocclusive crises/acute chest syndrome. There was no previous history of txn. Significant physical findings included bilateral coarse breath sounds without rales or wheezes and tenderness over the sternum. Admission labs revealed a WBC of 23,000/mL, Hb 8.7 g/dL and plt 686,000/mL. Ceftriaxone was initiated because of chest x-ray findings suggesting bilateral pneumonia/acute chest syndrome. During the 1st hospital day, pt required NS fluid boluses to maintain adequate urine output and face mask to maintain O 2 sats >93%. On the 2nd hospital day, pt became confused and was noted to have dark red urine with no RBCs noted on urinalysis. Over the ensuing day, 7 packed RBC txns were required to maintain Hb >7 g/dL. Upon arrival to our facility, she had a temp of 37 C, pulse 128/min, RR 30 breaths/min, O 2 sat of 99% on 40% FiO 2 (face mask) and a BP of 125/80. Continued hemoglobinuria was noted and peripheral blood smear revealed 3+ spherocytosis. Direct antiglobulin test was positive for C3 only and antibody screen was negative. Ceftriaxone was continued 1 gram q 12 for the next three days with vancomycin added to the antibiotic regimen. Hemoglobinuria was treated with mannitol and fluids with vigorous Key: txn = transfusion.

During the RDP-p transfusion no adverse events occurred. However, 17 minutes into a subsequent RBC transfusion, the pt developed chills, rigors, and an increase in temperature (>1°C) and blood pressure (100/60 to 136/76); the transfusion was terminated after infusion of about 50cc. The RBC unit was returned to the Blood Bank. The transfusion reaction work-up was negative, including gram stain and MPC. Later that evening the pt had a temperature of 40°C and intermittent chills and contacted his physician. On 3/3/04, the culture from the RDP-p obtained at issuance was positive with the organism subsequently identified as Streptococcus bovis (5 ¥ 10 6 CFU/ml). The clinician was notified, blood cultures were obtained, IV ceftriaxone was started and the pt was admitted to the hospital. Blood cultures drawn on 3/3/04 were negative. The pH of each of the RDP in the pool was ≥ than 6.6 (our cutoff value for platelet screening.) Conclusions: If not for a culture to validate our pH testing, this case of bacterial contamination would never have been detected and the reaction classified as FNHTR. Bacterial contamination of platelets continues as an important transfusion risk.

A N Afenyi-Annan, University of North Carolina at Chapel Hill, Chapel Hill, NC; R Lottenberg, University of Florida, Gainesville, FL; T R Konrad, University of North Carolina, Chapel Hill, NC Background: Transfusion therapy is the mainstay of treatment for patients with sickle cell disease (SCD). The blood bank plays a key role in the delivery of this care. Transfusion recommendations for SCD patients from the National Institutes of Health include the use of simple, exchange, and chronic transfusion, patient red cell phenotyping, and leukoctye reduced, sickle cell trait negative, and antigen matched red blood cells (RBCs). Actual blood bank practices for SCD patients however, are largely unknown. In this study, we describe the current practices of hospital-based blood banks in the state of North Carolina for SCD patients. Methods: We conducted a cross-sectional survey of hospitals in North Carolina that provided blood bank and/or apheresis services on-site (n = 106). Surveys were mailed to the medical director and laboratory supervisor at each hospital and included items on hospital and respondent demographics, volume of SCD patients, offered services, testing, and blood products. Results: Three mailings yielded responses from 85/106 hospitals (80.4%). Surveys from 76/106 hospitals were complete and available for analysis (70%). The most frequent respondent was the laboratory supervisor (71%). The sample consisted of predominantly community hospitals (90%) whose blood banks saw a wide range of SCD patients per month (0 to 30; median = 1). All blood banks offered simple transfusion services; fewer offered exchange transfusion (16%) or chronic transfusion (7%). Few had specific policies/procedures for SCD patients (29%). Patient RBC phenotyping testing was commonly performed (76%). However, specialized blood product use was variable. The majority always provided leukocyte reduced RBCs (86%) but less than half always provided sickle cell trait negative RBCs (42%). Only 38% provided antigen-matched RBCs; of these, about half provided matched products prophylactically (17%). Those providing antigen-matched RBCs were more likely (p =< 0.001) to be in a hospital with a trauma center (84% vs. 25%), have policies/procedures for SCD patient work-up (77% vs. 19%) and RBC delivery (72% vs. 16%), perform patient phenotyping (50% vs. 4%), and provide sickle cell trait negative RBCs (65% vs. 16%). Conclusions: Considerable variation exists in blood bank practices for SCD patients among maintenance of urine output. However, by hospital day 4, the hemolysis continued necessitating another transfusion of 3 units of packed RBCs. Respiratory failure, renal failure, and liver failure developed despite aggressive treatment and support. Expiration occurred on the 19 th day of hospitalization. Retrospective analysis of blood bank specimens from outside hospital and at our institution are seen below. Additional serum samples from 4/19, 4/20 and 4/23 demonstrated 3+ to 4+ agglutination at 37 degree C with ceftriaxone and 1+ to 2+ at the antiglobulin test phase. Tests without the drug were negative. Conclusions: Clinicians need to be aware of this rare and often fatal complication of ceftriaxone in children.

this clinical situation, we analysed the transfusion results in people with AIHA transfused in a single institution. METHODS: Before tranfusion, the autoantibody was eluted and identified, and the patient was phenotyped for the main Rh, Kell, Duffy, Kidd, Lewis and MNSs antigens. An autologous and/or differential absorption was also performed in order to exclude allloantibodies. The cross-match was done with the absorbed serum. All these techniques were performed by the gel-test method. The laboratory evaluation of patients were performed up to 72 hours after transfusion. RESULTS: From October 2000 through April 2003, 155 new cases of AIHA were enrolled. From these, 27 patients-6 men and 21 women, aged 3-83 (median: 52) were transfused. In total, 143 red cell transfusions were administrated, with an average of 5.3 transfusions per patient (1-40). Nine patients (33.3%) presented an alloantibody. From the 27 patients, 21 (77.8%) presented a favorable transfusion outcome, with an average post-transfusion rise of 1.5 g/dL on hemoglobin concentration. In five patients (18.5%) there was a decrease on post-transfusion hemoglobin level in at least one transfusion. The mean decrease was 0.5 g/dL. In one patient (3.7%) there was no change in hemoglobin level after transfusion. Seven patients (25.9%) presented 12 episodes of transfusion reaction. There were 5 febrile nonhemolytic transfusion reactions, 4 hemolytic transfusion reactions, and 3 cases of circulatory overload. There was no transfsuion-related death. CONCLUSIONS: Our data suggest that the transfusion is safe and well tolerated in AIHA patients, provide that red blood cells transfused are compatible with alloantibodies eventually detected.

A M Svensson, S Bushor, M K Fung, Department of Pathology and Laboratory Medicine, Fletcher Allen Health Care and University of Vermont, Burlington, VT Background: RBC transfusions in a patient with a history of autoimmune hemolytic anemia can represent both a laboratory and clinical challenge. The development of high titer low avidity antibodies, and antibodies to high frequency antigens may further impair the ability to identify compatible donor RBCs. Not infrequently, incompatible RBCs must be used and the need to increase oxygen carrying capacity conflicts with the desire to avoid exacerbating the autoimmune hemolytic process with RBC transfusions. Case Report: A 66-year old Caucasian female with coronary artery disease and a history of refractory autoimmune hemolytic anemia recently developed anemia requiring multiple RBC transfusions. The patient had maintained adequate RBC counts with erythropoietin and prednisone therapy for the past 16 months, which made it possible to perform phenotyping at the time of presentation. With worsening anemia, she developed angina that was treated with RBC transfusions in an outpatient setting. However, her angina increased as her RBC counts decreased, leading to hospital admission for further management of her hemolytic anemia and angina. She subsequently required multiple incompatible RBC units despite increased prednisone therapy. RBC units could be tailored to the patient's phenotype (negative for K, C w , Fy a , S and P1). However, increasing reactivity, up to 4+, was noted in cross-matching and her anemia remained refractory. It was determined that a second antibody, presumably directed against a high frequency antigen was present, however this antibody could not be identified at the level of a national reference laboratory. Following coronary artery stent placement the angina symptoms abated. Shortly thereafter, hemolysis decreased promptly following an attempt with high dose intravenous immunoglobulin (IVIG) therapy, in spite of previous nonresponsiveness to IVIG. Variation between IVIG preparations may explain this inconsistency. Conclusions: This case demonstrates the usefulness of early patient phenotyping in accelerating hemolytic anemia to aide in donor RBC selection and the importance of optimizing the patient's clinical condition to avoid ischemia. In refractory hemolysis, repeated attempts with IVIG treatment may be of value despite previous refractoriness to this treatment. L Prihoda, M Plett, Gulf Coast Regional Blood Center, Houston, TX Background: Our Reference Laboratory performs antibody identification and provides antigen-negative RBCs for 89 local hospitals, ranging from large teaching facilities to small rural hospitals. We surveyed our clients' policies for transfusion of patients with known alloantibodies (Ab), sickle cell disease (SCD) or warm autoantibodies (WA). Methods: Eighty-nine local hospitals were surveyed on current transfusion policies for patients with selected problems. Special techniques used for problem resolution in WA cases were also evaluated. Results: Responses from 26 facilities, ranging from <100 beds to >800 beds, showed the expected variation in special techniques used. Chloroquine diphsophate, EDTA/Glycine and hypotonic wash are rarely used in phenotyping for patients with WA or SCD. Warm or cold autoadsorptions are performed by five facilities, while allogeneic adsorptions are performed by only one facility. As expected, all provide antigen-negative, crossmatch-compatible (XM-comp) RBCs for patients with clnically significant Ab. Seven facilities routinely or occasionally provide XM-comp RBCs for patients with known room temperature reactive Ab only. Eleven facilities indicate they would request RBCs negative for all identified Ab, regardless of clinical significance. When no underlying alloantibodies are detected, transfusion policies for WA patients varied, with 18 of 26 (69%) giving least incompatible RBCs. Others provide RBCs compatible with adsorbed serum (1) or full phenotype-matched (1). The remainder relies on the pathologist's or treating physician's decision for what to transfuse. Thirteen of 26 indicated that a written policy exists. For SCD patients with no clinically significant Ab, 46% give XM-comp, 15% match for Rh and Kell antigens, and two routinely match for all clinically significant antigens, while the remaining 8 have no specific policy for transfusion to sickle cell patients. Conclusions: Policies for transfusing patients with clinically significant Ab are well established. If a clinically insignificant antibody is present (anti-P1, anti-M, anti-Le a , etc.), it is not necessary to provide antigen-negative RBCs for transfusion. Elimination of this requirement would reduce costs for the patient. While policies for transfusing SCD and WAIHA patients vary, most facilities give XM-comp or least incompatible, respectively. Only half of the facilities surveyed indicated that they had written policies for one or the other of these special transfusion situations. In order to reduce confusion and ensure that all staff follow the approved protocols, written policies should be established.

L Amorim, T A Miranda, J Godoy, F Azevedo, S L Castilho, HEMORIO, Rio de Janeiro, Brazil BACKGROUND: The finding of compatible red blood cells for transfusion is usually very difficult in Autoimmune Hemolytic Anemia (AIHA) patients, since most of these patients present an autoantibody with a public specificity. Therefore, transfusion should be avoided in this situation, unless it is absolutely necessary. There is a theoretical risk that the transfusion of incompatible red blood cells result in massive hemolysis and enhancing of autologous red cell destruction. To evaluate the outcome of transfusion in cal manifestations. Current serologic diagnosis of autoimmune hemolytic anemia (AIHA) is based on the finding of a positive direct antiglobulin test (DAT). However, since DAT using tube technique may fail to detect red blood cell (RBC)-bound globulin when less than 300 molecules immunoglobulin (Ig) G per cell or less than 60 molecules complement 3 (C3) per cell were present. Some AIHA patients had all the usual hallmarks of AIHA but a negative DAT. We utilized a direct flow cytometric erythrocyte immunofluorescence assay for diagnosing suspected AIHA, because it is a more sensitive method of quantifying RBC-bound IgG or complement. Our transfusion medicine physician is responsible for consultation and interpreting clinical and laboratory data, and he must receive training in clinical hematology and laboratory. Methods: Transfusion medicine physician was consulted for the diagnosis of AIHA. We studies 26 AIHA patients. We measured RBC-bound IgG and C3d by flow cytometry. The increased percentage (>3% over normal control) of fluorescence-activated RBCs was considered positive. D+ RBC sensitized with anti-D was used as positive control. We also performed DAT by tube technique with a polyspecific (anti-IgG and anti-C3d) antihuman globulin (AHG) reagent. If positive, DATs with specific anti-IgG or C3d were performed. Results: By flow cytometric DAT in 26 AIHA patients, 14 patients with AIHA exhibit both immunoglobulin and the C3d on the RBC, and 6 patients exhibit only IgG, 6 exhibit only C3d. Eight out of 26 AIHA patients had negative DAT with tube technique, but flow cytometry demonstrated their RBCs were coated with IgG alone (2 patients), complement alone (4 patients), or both (2 patients). Conclusion: Transfusion medicine physician with clinical evaluation and the utilization of direct flow cytometric erythrocyte immunofluorescence assay plays an important role in the diagnosis of AIHA.

bags" provided there are such bags in storage. Patients requiring more than 490 ml probably will be benefited as well, because they could receive 2 "jumbo bags" (instead of 3 standards) or 2 "jumbo bags" and 1 standard (instead of 4 standard). Conclusion: The Trima® system will help TM patients to reduce the number of different donor exposures and the consequential decrease of disease transmission risks. A new approach to the selection of donors fit for a large-scale production should maximize the benefits in the future.

The Impact of West Nile Virus Single Donor Testing in a Large Metropolitan Area a Retrospective Study J Swisher, J Lowther, R Plumlee, Gulf Coast Regional Blood Center, Houston, TX Background Determining the criteria for the management of West Nile Virus (WNV) testing in 2004 required analysis of many operational parameters. Based on an assessment of equipment, staff, and reagents and consumable supplies a plan was devised to perform Individual Donor NAT Testing (ID-NAT) with the next day's unit collection following a positive result per the entire service region. No retrospective testing would be performed and ID-NAT would continue until testing demonstrates a lack of positive results for a 7-day consecutive period. A retrospective study was performed using the 2003 WNV season data to determine the impact ID-NAT under this plan would have on testing resources and release of blood components to inventory. Study Design Daily testing from June-September 2003 for both local and out of region client data were reviewed and evaluated against positive WNV results. Using data from 2003 and the above-mentioned trigger it was determined how many samples would have been tested by mini-pool (MP) and by ID-NAT. Results During the 4-month WNV season, a total of 107,527 samples were tested. The following table lists the MP and ID-NAT testing that would have resulted if the proposed trigger had been applied. Background: Thalassemia Major (TM) patients require regular transfusion of red blood cells (RBC) in order to keep an adequate hemoglobin level. Great efforts have been made in order to improve the quality of transfusions in TM patients. Furthermore, blood banks often struggle with the shortage of blood donors. We have begun to use the Trima® system in order to get large volumes of packed RBC (conveniently called "jumbo bag") aiming to reduce the number of exposures to different blood donors. This study has tried to analyze the potential benefits of that new way of blood collection. Study design: Thirty-seven collections were carried out using a disposable kit designed to collect platelet and RBC. The maximum volume of RBC that can be harvested by using this kit is 380 ml (mean hematocrit 70%). In our service, the mean volume of the "jumbo bags" was 341 ml, standard deviation 34, and range from 241 to 378 ml. The mean hemoglobin level was 70 g per bag, standard deviation 8, and range from 54 to 81. Those bags were transfused in 16 patients in accordance with availability and compatibility tests. Some patients received a combination of 1 "jumbo bag" with 1 or 2 standard bags, others received 2 "jumbo bags" and for one single patient 1 "jumbo bag" was enough. We compared how many standard packed RBC bags usually the patient have been transfused with how many bags have the patient received when using "jumbo bags". Results: We realized that in 7 patients the availability of "jumbo bags" did not reduce the number of exposures. But in 9 patients there was a decrease in the number of exposure to different donors ranging from 33% to 50% compared to the usual number of bags transfused. This preliminary data showed us that the reduction in the number of bags from different donors depends on the volume every patient was transfused with, on the volume of the "jumbo bags" and on the availability of large volume packed RBC. Recipients requiring volumes between 390 ml and 490 ml (packed RBC with hematocrit 70%) need 2 bags regardless of whether "jumbo" or standard. Recipients, who require transfusions up to 380 ml of RBC, will take advantage of the "jumbo ID-NAT would have accounted for between 68%-98% of the samples tested on 16 separate days during the peak season in July-August, a statistic not reflected in the overall month by month averages. Conclusions The 2003 data and analysis have given us a better decision-making quality with regard to our current plan proposal. While this plan is conservative and exceeds current thinking with regard to start triggers and extent of individual donor testing, it provides the highest level of control and management of testing. It also avoids the need to establish complex formulas for identifying "hot spots", identifying donors that must be ID-NAT tested, and allows for the rapid switch from pooled to ID-NAT testing. It may not be possible to sustain this plan during high prevalence periods, based upon the estimated 320% increase over current resources in reagents and consumables and 150% increase in current staffing resources these data suggests. A secondary approach taking into consideration the number of positives, prevalence and regional divisions will be available as a fallback. Background: The FDA approved use of investigational nucleic acid amplification tests (NAT) for West Nile Virus (WNV) for blood donations in June 2003. We began using the Procleix® WNV assay (Gen-Probe Incorporated) on individual donations on July 1, 2003, with IRB approval. Methodology: All donor samples were tested individually using the FDA approved algorithm. Samples testing repeat reactive were sent to three other laboratories for additional testing. Donors with repeat reactive results were contacted and, following informed consent, enrolled in further research studies. Verification of infection was determined by reactivity in an alternative NAT test, followed SP371 Sensitivity and Specificity of the Procleix ® WNV Assay on the Automated TIGRIS TM System, and Compatibility with the Procleix ® Ultrio Assay M L Deras, J Macioszek, L Stringfellow, C Giachetti, J Linnen, Gen-Probe Incorporated, San Diego, CA Background: The risk of West Nile virus (WNV) transfusion transmission has been reduced by nucleic acid testing (NAT) of donated blood. Individual donor testing (IDT) could improve NAT efficacy and safety compared to testing minipools but implementation of IDT using the current semi-automated system challenges the capacity of most testing facilities. The TIGRIS System, under clinical validation at Gen-Probe, is ideal for IDT because it fully automates NAT. A WNV TIGRIS Assay Definition Module (ADM) has been developed and we present here the sensitivity and specificity of the WNV Assay on TIGRIS and compatibility of the Procleix WNV and Ultrio Assays when tested on a single TIGRIS instrument. Methods: Sensitivity was determined using a WNV Lineage 1 viral standard prepared by Health Canada, and a clinical sensitivity panel from Blood Systems Research Institute. Specificity was tested using fresh plasma samples from normal blood donors. Inter-assay compatibility was validated by sequentially testing the Procleix Ultrio Assay with negative and positive HIV-1, HCV, and HBV samples, followed by testing of the Procleix WNV assay with negative and positive WNV samples. Results: Probit analysis of the results with a viral panel at 100, 30, 10, 3, 1, and 0 copies/mL determined that the 95% detection level was 8.3 copies/mL (95% confidence interval 6.6-11.9 copies/mL) for the WNV Assay on TIGRIS. Testing of clinical sensitivity panels yielded results comparable to those obtained using the semi-automated WNV Assay. The WNV Assay on TIGRIS was used to test 6,407 negative normal donor samples in which 3 false positive results were observed for a specificity rate of 99.95%. When WNV and Ultrio Assays were tested in consecutive runs, each met specifications for sensitivity and specificity. The system correctly verified reagents, expiration dates, inventory consumption, and processed the results of each assay independently. Conclusions: The WNV Assay on the TIGRIS System is sensitive and highly specific. There are no practically significant differences in analytical performance of the assay on TIGRIS when compared to the semi-automated assay. The WNV Assay was shown to be compatible with the Ultrio Assay on TIGRIS. These characteristics suggest that the WNV Assay on TIGRIS will be an effective and flexible solution for fully automated NAT.

Single Donor Testing Using the Roche Taqscreen TM West Nile Virus Assay J Swisher, J Lowther, R Plumlee, C Cousins, Gulf Coast Regional Blood Center, Houston, TX Background The current method of testing for West Nile Virus (WNV) at this institution is performed utilizing mini-pools of 6 samples. Due to the desire to detect lower than expected levels of virus in donor samples, a workflow study was performed to determine the feasibility of performing individual sample nucleic acid testing (ID NAT). Study Design Over a period of 7 days, approximately 300 samples/day that had been previously tested in pools of 6 were relabeled and tested again as individual samples. The amount of time required to process was monitored and results for ID NAT were compared to the results of the original pool. Results Over the 7-day period, 2,399 samples were tested. It required 11 hours a day to process the 300 samples as compared with the 5 hours it would have required to test the mini-pools. All 2,399 samples were non-reactive in both the pool and the individual runs. Conclusions At this institution, we average 1,050 samples/day for WNV testing. Using the current mini-pool system, it takes 9 hours to complete the testing using 3 TaqScreen TM systems. If these samples were tested as individuals, it would take almost 28 hours to complete testing on the same three systems. Accomplishing ID NAT on 50% of our projected year-2004 collections will require 2 additional TaqScreen TM test systems and the associated increase in staffing, reagents an d consumables to support their operation. It should be noted, however, that this solution is not without it's problems and may also cause considerable strain in the areas of manufacture, laboratory staffing and cost per unit-which would eventually be passed along to the patient.

S Pichuantes, S Nguyen, H Huang, D Chien, B Phelps, Chiron Corporation, Emeryville, CA Background: West Nile virus (WNV) is an enveloped, positive singlestranded RNA virus of approximately 11 kb that is naturally transmitted between birds and other vertebrates through mosquitoes feeding on infected animals. Although humans are incidental hosts, the fast spread of human WNV infections in the US, along with the confirmation that the virus can be transmitted through blood transfusion, led to the implementation of nucleic acid amplification tests. Blood centers started routine screening in July 2003 of all donations, so as to quarantine and retrieve potentially infectious blood components. Objective: Clone WNV fragments into vectors suitable for in vitro transcription of viral RNA to allow preparation of transcripts from different viral genomic regions. The resulting transcripts could be further characterized and used as RNA standards to compare the assay performance between different WNV Nucleic Acid Tests. Methods: RNA extracted from cultures of Vero cells infected with WNV strain NY385-99 was used to synthesize, by RT-PCR, a number of DNA fragments spanning the full-length WNV genome. The fragments were cloned into a shuttle vector and their nucleotide sequences verified by DNA sequencing. WNV nucleotide fragments purified from the different shuttle plasmids were cloned into pGEM-4z for further in vitro transcription of different segments of the WNV genome. Construction of a recombinant plasmid encoding for a full-length non-infectious WNV (nucleotide deletion that causes frame-shift and creates a stop codon at residue 576) was also pursued. Results: A total of thirty-one RNA transcripts spanning the entire WNV genome were prepared in vitro using the WNV recombinant plasmids. Conclusion: The availability of a set of WNV transcripts synthesized from specific genomic regions will facilitate the preparation of well-characterized RNA standards to assist in the performance evaluation of different WNV nucleic acid tests. Background: The Procleix WNV Assay is an investigational nucleic acid test (NAT) that has been used in the US to screen donated blood for West Nile virus (WNV) since June 2003. Prior to the 2004 WNV season, one assay kit component (Probe Reagent) was modified with the aim of improving clinical specificity. This study compared assay performance using the modified Probe Reagent formulation versus performance with the original formulation. Methods: The modified formulation was used at four blood centers: American Red Cross (ARC), Charlotte, NC; Oklahoma Blood Institute (OBI), Oklahoma City, OK; Robertson Blood Center (Hood), Ft. Hood, TX; and Indiana Blood Center (IBC), Indianapolis, IN. At each site, the specificity of the Procleix WNV Assay using the modified formulation was compared to results generated from previous testing using the original formulation at the same center. To assess analytical sensitivity, serial dilutions of WNV were tested followed by probit analysis. Results: Over 42,000 test results were collected for the evaluation, including about 12,000 results from pooled testing (16 donations per pool) and over 30,000 individual donation (ID) sample results. Combined pooled and ID results represented testing of over 230,000 donations. In pooled testing using the modified formulation at ARC and IBC, 2 false positive results (0.017%) occurred in about 12,000 tests. In previous pooled testing using the original formulation, 50 false positive results (0.42%) occurred in 12,000 tests. In ID testing with the modified formulation at OBI and Hood, 12 false positive results (0.04%) occurred in 30,000 tests compared to the original formulation testing in which about 46,000 ID samples were tested with 183 false positive results (0.40%). Analytical sensitivity results with the modified formulation showed sensitivity that was comparable to the original formulation, with 95% detection at about 11 copies/mL. Conclusion: Analysis of data from the field evaluation of the modified formulation demonstrated substantially improved specificity compared to results using the original formulation. Analytical sensitivity results indicate that the increase in specificity was achieved without a loss in sensitivity.

C L Cameron, J Reeves, Canadian Blood Services, Ottawa, ON, Canada; N Antonishyn, Provincial Laboratory Services, Regina, SK, Canada; P Tilley, Provincial Laboratory for Public Health, Calgary, AB, Canada; T Alport, B Eurich, Canadian Blood Services, Regina, SK, Canada; D Towns, Canadian Blood Services, Calgary, AB, Canada; D Lane, Canadian Blood Services, Winiipeg, MB, Canada; J Saldanha, Canadian Blood Services, Ottawa, ON, Canada

Background WNV was first detected in North America in 1999, with the first cases of transfusion associated transmission (TAT) of WNV occurring in 2002. In response to this threat to the safety of the blood supply, Canadian Blood Services began testing all donations using a commercial assay in July 2003. Prior to the availability of the commercial assay, an in-house NAT assay was developed and used under an Investigational Testing Application (ITA) to test donations from Ontario for one week. The in-house assay was used subsequently to confirm WNV infection in all donations found to be positive using the commercial assay. Study The in-house assay consisted of a manual viral RNA extraction (DNA mini kit, Qiagen) from 200 ml of donor plasma, to which an internal control (Armoured WNV RNA, Ambion) had been added (400 copies per extraction). A reference reagent (Health Canada) was extracted in each assay for quantitation. Ten ml of viral RNA was subsequently reverse transcribed and amplified (Quantitect Probe RT-PCR kit, Qiagen) by quantitative RT-PCR (MX4000, Stratagene) using primers (positions 10668 and 10770) and a fluorescently tagged probe(position 10692) (Integrated DNA Technologies) (Lanciotti et al.) from the 3¢ noncoding region of the WNV genome for the target RNA, and primers and a probe located at 9971, 10069, and 9998 respectively in the NS5-2 region for the internal control. Data were analysed using the MX4000 software. Available post donation plasma samples (£4 days, £14 days, and £28 days) were tested using the in-house assay and a serological assay for IgG and IgM antibodies to determine seroconversion. The WNV RNA titres of follow up samples were also determined. The RNA titres ranged from 4.8 ¥ 10 4 copies/ml to below the level of quantitation (1000 copies/ml) and the highest titre for any one donor was not necessarily seen on the index sample. Conclusion Twelve of the 14 donations that tested positive using the commercial assay also tested positive using the in-house assay. All available follow up samples tested positive for IgG and WNV IgM antibodies. Two of the 14 samples that were positive for WNV RNA in only one of 38, and one of 39 repeat tests were positive for IgG and WNV IgM antibodies at the time of donation indicating a viremia at the limit of detection for the in-house NAT assay. These two samples were identified by single donor testing in a high prevalence area in Saskatchewan. and 52,043 were tested by WNV (49%). Samples were tested across 2,059 assay runs: 1,075 (52%) for Mx and 984 (48%) for WNV. Assay performance criteria included: valid, invalid, suspect assay runs (control failures or exceeding the 10% rule) and MPT true positive (TP), false positive (FP), true negative (TN) and invalid results. Results: There were 94% (1,015/1,075) Mx valid runs. The 6% (60/1,075) failed Mx runs were due to: 18% (11/60) invalid controls, 50% (30/60) operator errors and 32% (19/60) suspect runs. There were 95% (935/984) WNV valid runs. The 5% (49/984) failed WNV failures were due to: 37% (18/49) invalid controls, 53% (26/49) operator errors and 10% (5/49) suspect runs. For the overall Procleix run perfor-mance evaluation, there is a slight statistically significant difference (p < 0.02) in the Procleix Mx and WNV performance with WNV having fewer preventable and suspect runs, but a higher number of invalid calibrators. For the Mx MPT analyte results there were 1.9% (990/53,464) TP; 0.3% (171/53,464) FP; 97% (51,855/53,464) TN and 0.8% (448/53,464) internal control (IC) failures. For the WNV MPT analyte results there were 0.4% (220/52,043) TP; 0.2% (101/52,043) FP; 99% (51,537/52,043) TN and 0.4% (185/52,043) IC failures. The positive predictive value and percent accuracy was 98.1% for Mx MPT and 99.58% for WNV MPT. Within the valid Procleix assay runs, the WNV analyte performance is significantly better than Mx (p < 0.0001) with WNV having fewer FP, invalids and TN results. Conclusions: The data demonstrate that the Procleix assay systems using semi-automated instrumentation can be used effectively in the performance of Mx and WNV RNA in a high volume laboratory. The Procleix Mx and WNV assays have an overall low incidence of run failures, with 94% and 95% successful runs, respectively. The variation in the assay and analyte performance demonstrates that Procleix WNV has a higher level of assay run and analyte result reliability when compared to the Mx assay. There is a critical need for nucleic acid detection tests as an indicator for WNV due to low viral titers. Here we describe the results from a validated assay, and an improved assay based on Target-Capture PCR technology for qualitative and quantitative detection of WNV RNA. Methods: A magnetic bead-based protocol was used to isolate WNV RNA and a related internal control RNA in a single tube using a semi-automated method from 0.5 mL of plasma. The capsid region of isolated targets on the beads was amplified by real-time PCR technology. A blinded panel of 25 WNV samples distributed by Blood Centers of the Pacific (BCP) was tested by both the validated and improved assays. The validated assay has been used at Bayer Reference and Testing Laboratories (BRTL) to confirm WNV positives. In collaborations with BCP and Bonfils Blood Center, the quantitative Target Capture-PCR assays have been used to determine viral load in WNV positive samples to track the 2003 epidemic of infected blood donations. Results: The results with the BCP WNV panel for analytical sensitivity by the validated assay indicate 100% detection of 30 copies/mL, with the improved assay detecting 10 copies/mL. The assay has high specificity and tolerates a variety of anti-coagulants, interfering substances, and plasma from pathological conditions. The validated assay implemented at BRTL confirmed 682 of the 3811 possible WNV positives (17.9%) between July 2003-January 2004. The profile of positives mirrors the profile of the reported cases. Eighty samples from January 2004 tested at BRTL were also tested by the improved assay. Despite the increase in sensitivity between the validated and the improved assay, no gain in detection of WNV positive donors was observed. The quantitative assay estimates WNV in the range of 10 2 -10 10 copies/mL. The results from quantitative assay performed on blinded samples from Bonfils closely tracks the results from BRTL. Conclusion: The Target Capture-PCR WNV assay is a rapid, sensitive, easy to use, accurate assay for WNV RNA detection and quantitation. (6) and in individual donor specimens at multiple facilities. Study Design: The clinical study was conducted at 11 blood centers in the US. Additional testing was required for specimens that were WNV positive on individual testing. Testing included IgM, IgG, alternate NAT, and WNV quantitation. In addition, these donors were eligible for enrollment into a follow-up study. Results: The first site initiated testing on June 17, 2003. Through November over 234,000 pools of six have been tested representing over 1.4 million specimens. Eight-nine individual speci-141A 2004-Vol. 44, Supplement mens have been identified as WNV RNA positive. Of these, 74 (83%) have been enrolled into the follow-up study. Based on follow-up testing, 73 were confirmed as WNV positive and 1 was confirmed negative. Fifteen were not enrolled and of these, 14 were confirmed to be WNV positive based on additional testing of the index donation; 1 was WNV negative. Seventy-two samples have been quantitated and the median viral titer is 19,500 copies/mL and a mean of 51,348 copies/mL (range 200-760,000). Conclusion: The COBAS TaqScreen WNV test has successfully interdicted WNV RNA positive specimens thereby increasing the safety of the blood supply.

W W Andrews, P Yan, W S Lee, C Harrington, B Phelps, Chiron Corporation, Emeryville, CA Introduction-We have developed a RIBA ® SIA assay to measure the relative amounts of serum IgA, IgM and IgG to the WNV PreME envelope protein. These Ig specific bands are coated with Rabbit-a-IgA, a-IgM and a-IgG respectively. The RIBA assay also measures total a-WNV Ig with a band coated with the recombinant PreME protein. After exposure to the serum or plasma all bands are probed with a recombinant PreME-HRP conjugate to detect bound anti-WNV specific immunoglobin. An a-WNV monoclonal is coated on the strip to give a high control and a cutoff control band, as in the well known HCV RIBA SIA. This allows the assay to give semiquantitative result when read visually. Study-Evaluation of the RIBA assay with 100 WNV IgM positive samples (confirmed on PANBIO WNV IgM assay) showed 100% correlation with the ELISA in terms of IgM reactivity. The RIBA assay also shows good correlation with an in-house WNV ELISA on WNV seroconversion panels. This ELISA assay was constructed as an Ag sandwich format assay using the unlabelled PreME Ag on the plate and the same Ag as an HRP conjugate. In general, the same bleed came up positive on the RIBA as on the ELISA. The ELISA is more sensitive on some panels, however. We used these two immunoassays to test 111 WNV NAT yield samples from a US blood center in a high prevalence area for WNV in 2003. 26 of the 111 samples (23.4%) of the samples were ELISA positive. 14 of these samples had virus detectable on an in-house quantitative WNV NAT assay, the other 12 were negative. The 14 NAT detectable samples had a viral load range of 34 to 714 copies/ml (average = 177 copies/ml). In contrast, the 85 immunonegative samples had viral loads of 0 to 85,049 copies/ml (average = 7,326 copies/ml). Conclusions-We have tested both blood bank WNV NAT yield samples and WNV seroconversion samples with an in-house WNV RIBA and ELISA assay. The results are consistent with data that show most WNV immunoreactive samples from blood centers to be either nonviremic or of low viral load.

Allele Produces Variable Phenotypes Depending on the Co-Inherited ABO Allele

Supplement TRANSFUSION

Excluding the two true positives, the range of S/CO in the second tes run for the other 20 initially reactive samples was 0.01-6.23, with a mean of 1.99. Discussion: This assay showed a sensitivity of 100% and a specificity of 99.3%. However, while the negative predicative value was 100%, the positive predictive value was <1%. Changes in specimen preparation and technical experience affected the number of RR (but not confirmed) results from a high of 59 in July to 28 in August

Gen-Probe Alternative WNV Assay: A TMA-Based Confirmatory Assay for

Background: Worldwide group B subtypes are uncommon although they are relatively common among Koreans. We sequenced exons 6 and 7 of 12 donors with the B 3 phenotype and discovered a novel B var allele characterized by a 547 G>A polymorphism. To elucidate the inheritance of this new allele, family studies on two donors were performed and revealed the interesting serological finding that the B var allele varies in phenotypic expression depending on the co-inherited ABO allele. Materials and Methods: From amongst 169,605 donors collected at the Gwangju-Chonnam Red Cross Blood Center, South Korea, between July 2002 and February 2003, 6 A 1 B 3 and 6 B 3 donors were selected for analysis. Serological evaluation, exons 6 and 7 sequence analyses and cloning were performed using standard protocols. To elucidate the prevalence of the B var allele, allele-specific PCR for the 547 G>A mutation was performed on 10 donors with full B antigen expression and also on 14 additional donors with weak or unusual expression of A or B antigens. For the family studies, the parents of 2 unrelated A 1 B 3 donors were also evaluated as described above. No siblings were available for analysis. Results: From the initial 12 donors, all B 3 (6/6) and 2/6 A 1 B 3 donors demonstrated consensus B303 exons 6 and 7, while 4/6 A 1 B 3 donors demonstrated A101/B var or A102/B var genotypes. The B var allele was not detected in any (0/10) of the donors with full B antigen expression. Analysis of the 14 additional donors revealed that B var was present in 4/6 A 1 B 3 donors and was not detected in B 3 (4), A x B (2), A 1 B x (1) or A x (1) donors. Family study: The A 1 B 3 propositi had slightly different genotypes (A101/B var and A102/B var ). Both fathers demonstrated uncomplicated B phenotypes with B var /O01 genotypes, while both mothers demonstrated full A antigen expression with slightly different genotypes (A101/O02, A102/O01). Cloning studies revealed that the B var allele is inherited in a straightforward Mendelian manner. Conclusion: The finding of full B antigen expression in the setting of the B var /O01 genotype was unexpected considering that the A 1 B 3 phenotype was produced when the B var allele was co-inherited with an A101 or A102 allele. Thus the natural phenotype of the B var allele appears to be uncomplicated B. The 547 G>A mutation is predicted to cause an amino acid substitution Asp183Asn which is part of the disordered loop of amino acids. Perhaps the competition for H antigen between the enzymes encoded by the A101/2 alleles and the B var allele results in the B 3 phenotype when co-inherited. Kinetic studies of the enzymes involved are pending and may elucidate the basis for the variable antigen expression.

M Tanaka, N Yamashita, J Takahashi, F Hirayama, Y Tani, H Shibata, Japanese Red Cross Society Osaka Blood Center, Osaka, Japan BACKGROUND: The blood group P system has three rare phenotypes (p, P1k, and P2k) in addition to two common phenotypes, P1 and P2. The p individuals are very rare. Individuals of the p phenotype lack globotriaosylceramide (Gb3), lactosylceramide (CDH) and P1 antigen, presumably as a result of deficiency in the enzyme a 1,4-galactosyltransferase (Gal-T) responsible for the synthesis of Gb3 from CDH. Recently, cDNAs for Gal-T has been isolated. To explore the molecular basis of the p phenotype in Japanese donors, therefore, we analyzed of Gal-T gene. METHODS: Amplifications of the cording region in Gal-T gene for DNA sequencing were performed to use the original oligonucleotide primers. Constructs for the gene expression were prepared by expression kits and transfected into Namalwa cells by electroporation. The expression of Gb3 on the cells were flowcytometrically analyzed using anti CD77. Allele Specific PCR (ASP) was newly designed to detect Gal-T gene and performed on 50 normal Japanese donors and 4 Japanese donors with the p phenotype determined on the basis of serological methods. RESULTS: Four different alleles of the p phenotype in Japanese donors were homozygous for single base insertion (Gal-T/ins) which caused a reading frame shift and a loss of the stop codon, and resulted in a gene product with an additional 92 amino acids. The results of ASP designed to detect Gal-T/ins, all of 50 normal samples revealed the 328 bp band for the normal Gal-T gene (Gal-T/wild) and four novel mutants had no band. In the flowcytometrically analysis of Namalwa cell transfectants, Gal-T/wild gene caused definite Gb3 expression. In contrast, transfectant cells with Gal-T/ins did not express Gb3 at all. CONCLUSION: In this study, we found 4 novel Gal-T/ins mutants with the p phenotype in Japanese donors and designed an ASP to detect this type of mutant. If fetal blood group p determination is required and serologic tests fail or cannot be performed, this ASP is useful as one method for the blood group p typing.

Background: A large scale study of the serology and genetics of Korean blood donors with A subtypes was undertaken. The A var allele characterized by 784 G>A polymorphism and was identified only in A weak B donors (Seo DH et al. Kor J Transfus Med 2003; 14; 212) . Several donors whose phenotype did not match their genotype were also discovered. Methods: From the 169,605 donors collected at the Gwangju-Chonnam Red Cross Blood Center between July 2002 and February 2003, 150 donors with group A subtypes were identified (Table 1 ). In two donors it was not possible to differentiate the A m from the A el subgroups; they were classified as A m or el . Cis-AB donors were excluded by performing allele-specific PCR (AS-PCR) for the cis-AB01 allele. Serological evaluation including use of anti-A 1 reagent, exons 6 and 7 sequence analyses of selected donors and AS-PCR were performed using standard protocols. AS-PCR for the A var allele was also performed on 10 A 1 donors. Results: Serologic and genotypic results of the donors with weak A antigen expression are presented in the Table. Microscopic mixed field hemagglutination with murine monoclonal anti-A was observed on forward typing of the A weak B cells associated with the A var allele. Weak anti-A was present in these donor's sera. The A var allele was not detected in A 1 or A subtype donors other than A weak B. The frequency of the A 2 phenotype amongst A donors [A 2 /(A 1 + A 2 + all A subtypes)] was 0.02%. The frequency of A 2 B amongst AB donors [A 2 B/(A 1 B + A 2 B + all AB subtypes)] was 0.60%. Interestingly, 10 donors with the A 2 B phenotype demonstrated consensus A101 or A102 alleles. Conclusion: Our work extends the initial findings on the A var allele by demonstrating its presence in A weak B donors exclusively. This allele has not been reported outside of the Korean population and AS-PCR studies on additional donors with weak A antigen expression will elucidate its frequency. Unlike Caucasians the A 2 phenotype is about 28 times more frequent amongst Korean AB donors than A donors. In Japan the A201 allele is more frequent amongst A 2 B donors than A 2 donors accounting for the A 2 phenotype imbalance. This allelic imbalance was not seen amongst Koreans. Complete ABO allele sequence analysis will be performed on the 10 A 2 B donors whose phenotypes did not correlate to their genotypes; perhaps other mutations within the allele will explain these discrepancies. This protocol determines the presence of exon 7 and intron 4 of RHD, the 37 bp insert in RHDC exon 4, the exon 2 of RHCE (specific for allele RHc), and intron 2 of RHCE (specific for allele RHC). Study Design: DNA of one hundred forty-three (143) blood donor samples of the following phenotypes were tested using the protocol described above 1 : R 1 R 1 (14), R 1 r(16), R 1 R 2 (16), R 1 R z (3), R 2 R z (2), R z R z (2), R 2 R 2 (11), R 2w R 2 (1), R 2 r(10), R o r(17), rr(5), r¢r(16), r≤r(3), r¢r≤(3), r≤r≤ (3), rr(C D)(2), D IVa (18), D III (1). Further, all these DNA were also tested by PCR for RHD-CE hybrid exon 3 as described elsewhere 2 . Results: All RHD and RHc phenotypes/genotypes were concordant but some discrepancies were found between phenotype and genotype of RHC allele. Eight (8) samples typed as r¢r were genotyped as rr. According with Tax et al 2 this could be due to r¢ s allele present in some African population. So, we tested all the 143 blood donors samples for the hybrid exon 3 as found in an r¢ s (Cde s ) allele 2 . As suspected, the hybrid exon 3 was found in these 8 discordant samples for RHC phenotyping/genotyping and all 18 D IVa samples. Additionally, from these 8 discordant samples, we could test 3 of them by hemagglutination with anti-VS sera, and they were all positive. Conclusion: the 8 discrepant results for RHC were due to the presence of the r¢ s allele 2 . So, we must include in our routine PCR for RHC and RHc genotyping the hybrid exon 3 for interpretation of results. The racial background from these 8 individuals were reviewed and 50% of them declared not be aware of any African ancestry, whereas the 50% remaining had some African background. In multiracial society where a high degree of miscegenation exists, care shall be paid for application of methods usually vali- Background: The Procleix® WNV Assay is an investigational amplified nucleic acid test (NAT) for screening blood donations for West Nile virus (WNV). For confirmation of initially reactive results obtained with the Procleix WNV Assay, we developed the Gen-Probe Alternative (Alt) WNV Assay, an investigational assay that uses primers and probes that target a different region of the WNV genome than the Procleix WNV Assay. Both WNV assays are based on Transcription-Mediated Amplification (TMA) and use the same semi-automated instrumentation and assay procedures as the FDA-licensed Procleix HIV-1/HCV Assay. Methods: To assess the analytical sensitivity of the Alt WNV Assay, we tested serial dilutions of WNV Lineage 1 and Lineage 2 virus from Boston Biomedica Inc. and an RNA transcript encoding a portion of the WNV genome. To assess specificity, we tested more than 1550 normal negative donors, problematic samples, and samples containing other blood borne viruses. We also determined the reproducibility of the Alt WNV Assay in panels tested by three operators, with three instrument sets, and with two kit lots. Lastly, a preliminary assessment of clinical sensitivity was conducted by testing clinical samples in the Alt WNV Assay. Results: For determination of analytical sensitivity, we found that the Alt WNV Assay is capable of 100% detection of WNV RNA down to 30 copies/mL for both WNV Lineage 1 and 2 viruses; results were comparable to results obtained with the Procleix WNV Assay. Analytical specificity in negative donors was 99.8%. The reproducibility study showed limited variability between operators, instruments, or kit lots. Of 208 donations that were confirmed WNV positive either by a kinetic PCR assay carried out at the Bayer Reference Testing Lab (Berkeley, CA) or by IgM at Focus Labs (Cypress, CA), the Alt WNV Assay detected 200 (96%) of the samples. Conclusions: These results demonstrate the feasibility of using the Gen-Probe Alt WNV Assay to confirm results obtained with the Procleix WNV Assay and other WNV NAT methods.

R R Gammon, Association of Independent Blood Centers, Lake Park, FL; P Simmonds-Brown, C K Williams, M A Soto, BloodnetUSA, Lakeland, FL Background: Screening for West Nile Virus (WNV) by nucleic acid testing (NAT) may improve blood safety by decreasing the incidence of transfusion transmission. We attempted to determine how our donor population compared to the total US (incidence 0.014%, 6/03-12/03). 1 Methods: Beginning 7/1/03 plasma pools of 16 donations were tested by our NAT laboratory using a Transcription Mediated Amplification (TMA) assay. Reactive results were confirmed by testing an alternate sample by TMA, by testing using a validated alternate NAT methodology-polymerase chain reaction (PCR) and by a licensed serological method that detected IgM antibodies to WNV. TMA reactive donors were entered into a substudy and, after IRB approved informed consent was obtained, three Plasma Preparation Tubes were collected and tested for WNV by TMA, PCR and for anti-IgM production. Study participants were not compensated. All testing was performed under IND with strict adherence to the manufacturer's insert. Results: As of 4/15/04, 438,020 specimens from 19 collection sites in the Eastern US were tested. At index donation, forty-three donors were WNV NAT reactive by TMA. 3/43 (6.98%) were reactive by TMA on the alternate sample, reactive by PCR and were classified as confirmed positive. 0/43 were IgM + at index donation. 40/438,020 (0.009%) initially reactive TMA results were not reproducible on alternate specimens or by PCR and were considered to be false positives. 25/43 (58.1%) returned for follow-up bleeds. 2/25 (8.0%) seroconverted (1 IgM and IgG+, 1 IgM+). 1/2 seroconverters (IgM+) remained viremic by TMA and PCR, but did not return for additional follow-up bleeds. 6/43 (14.0%) TMA reactive donors had a signal-to-cutoff ratio (S/CO) of >20, mean 25.6 (range 20.85-32.70) at index. 3/6 with a high S/CO were reactive by TMA (initial and alternate sample) and by PCR, these donors presented in September and October 2003. Specificity of pooled testing was 99.7% (23,784/23,855) . Sample volume was adequate for testing; but, a shipment delay resulted in the three thawed specimens that were unsuitable for PCR testing. Conclusions: In the first 9 1/2 months of WNV NAT testing we report high specificity of pooled testing. Three confirmed cases of WNV viremia (TMA+ on initial and alternate specimen, PCR+) were detected in 1 per 146,007 donations, for an incidence 0.0007% which was well below that of the total US donor population. Our confirmed cases occurred in the fall months only. We saw a benefit in adding WNV NAT to the menu of blood testing since it successfully interdicted viremic blood donations from a volunteer blood donor population. 1. MMWR 2004; 53:281-284.

J Swisher, J Lowther, R Plumlee, C Cousins, Gulf Coast Regional Blood Center, Houston, TX BACKGROUND Currently, the method for testing for West Nile Virus employs the use of donor pools in the primary screening process. Due to the desire to detect lower than expected levels of virus in donor samples, a proposal to test donors singly has been issued. Single donor testing for this assay will increase testing time and may potentially delay test result posting and release of units to inventory. To minimize this risk, a trigger must be identified to determine when this testing is required and if it can be deployed by any demographic factors. Major metropolitan areas provide additional challenges to this objective due to increased donation area and mobility of donors. STUDY DESIGN West Nile test records for units collected in a major metropolitan area in 2003 were reviewed retrospectively. Demographic data from these records were evaluated for commonalties in the positive donors. The compiled data were then compared to the current cut-off of 1 : 1000 cases as the trigger to begin performing single donor testing. RESULTS Over the 2003 review period, 47,138 donors from a major metropolitan area were tested for the West Nile Virus. Of this total, 7 cases or 1.49 per 10,000 donors were identified. In a metropolitan area of 3,737 sq. miles, all cases were located within 11.25-mile radius of each other. Conclusions First review of the data seems to indicate a reasonable trigger is facility location according to the positive cases identified. Further evaluation must take into account that many people donate during work hours at facility locations near their jobs and that in large metropolitan areas it is not unusual for people to commute long distances to work. As it is more likely to be bitten by mosquitoes at dawn or dusk, people are likely infected near their residence rather than employment, making this a questionable trigger. This leads to indicating residence zip code as a possible trigger. While this data is available, it is routinely not available within the system for up to 12 hours, long after testing has begun. Delaying testing to determine those donors requiring single donor testing based on zip code again causes result posting delays and delay of released units to inventory. If single donor testing is truly necessary to provide the level of sensitivity necessary for this assay, then both of these potential triggers may have potentially undesirable results. This may imply that single donor testing on all units is the best alternative to avoid missing donors from a potentially "hot" area and delaying test results or unit release.