key: cord-0006334-3fx79jdg
authors: nan
title: Neurocritical Care Society 8(th) Annual Meeting
date: 2010-08-24
journal: Neurocrit Care
DOI: 10.1007/s12028-010-9426-2
sha: 9fb824a6dcf5348448fc90691a9d0fd2bab05bc4
doc_id: 6334
cord_uid: 3fx79jdg

nan

The introduction of therapeutic hypothermia in addition to improved cardiopulmonary resuscitation practices may increase the probability of surviving a cardiac arrest. Current predictors of outcome were established prior to the use of therapeutic hypothermia. The accuracy of prognostic indicators in this setting is largely unknown. The aim of our study was to evaluate the accuracy of predictors in patients treated with hypothermia after cardiac arrest.

The optimal diagnostic evaluation for spontaneous intracerebral or intraventricular hemorrhage (ICH or IVH) remains controversial. We compared the diagnostic utility of early MRI in addition to non-contrast head CT (CT).

Therapeutic hypothermia after cardiac arrest (HACA) is the most powerful therapy to improve survival and neurologic outcome and is emphasized in guidelines since 2003 1 ; yet the majority of physicians in the US report that they have not used HACA. 2 We investigated the utilization of HACA in United States (US) hospitals in 2007.

We used a national sample of US hospitals (2007 Nationwide Inpatient Sample, NIS, representative of 20% of inpatient admissions in the US) 3 to identify adult ( investigated if hypothermia was administered. We assessed individual patient and hospital characteristics in univariable analyses, and significant candidate variables were then analyzed in a multivariable stepwise logistical regression model to determine factors associated with hypothermia use. P-value < 0.05 was considered significant.

26,915 adult patients were coded as cardiac arrest (ICD-9 diagnosis code = 427.5) in the NIS for 2007. The mean age was 66.6 years, with 54.7% men, and 63.4% mortality. Therapeutic hypothermia (ICD-9 procedure code = 9981) was administered to 92 (0.35%) patients. In multivariable analyses, the following were independently associated with hypothermia use: age (adj. OR 0.98, 95% CI 0.96-0.99, p<0.001), western states region (adj. OR 2.45, 95% CI 1.57-3.83, p<0.001), teaching hospital (adj. OR 2.39, 95% CI 1.50-3.80, p<0.001), emergency department admission source (adj. OR 2.43, 95% CI 1.34-4.41, p<0.001), and large hospital size (adj. OR 1.72, 95% CI 1.03-2.89, p=0.040).

Fewer than 1% of patients with cardiac arrest received hypothermia in this large sample. Factors independently and significantly associated with a higher use of therapeutic hypothermia include patient factors (younger age and admission through the emergency department) and important hospital characteristics (western states, teaching hospital status, and large hospital size).

Cardiac arrest often results in devastating brain damage due to a transient global hypoxic-ischemic injury. It is unknown whether the human brain remains vulnerable to secondary injury in a delayed fashion after initial resuscitation.

We sought to determine whether intracranial multimodality monitoring (MMM) can guide postresuscitation care by detecting secondary injury due to hypotension, hypoxia, seizures, hypoglycaemia, or excessive hyperventilation.

We placed two multimodality intracranial monitoring probes in the right frontal lobe of 8 cardiac arrest patients admitted to the neuro-ICU between October 2008 and April 2010. All patients were comatose and treated with therapeutic hypothermia (Alsius or Arctic Sun system) for 24 hours and were rewarmed at a rate of 0.2 °C thereafter. ICP, CPP, brain tissue oxygen pressure (PbtO2, Licox/Integra) and surface EEG were monitored in all patients; micro dialysis (CMA Inc.), cerebral blood flow (CBF, Hemedex), and depth intracortical EEG (ICE) were monitored when possible.

Mean age was 53 (range 27-85); 5 of 8 were men. Initial rhythm was VF in 1, PEA in 1, and asystole in 6. Mean time to ROSC was 23 minutes. Patients underwent an average of 66 (range, 20-90) hours of monitoring. At the start of monitoring mean PbtO2, CBF, and LPR were near or beyond critical thresholds despite normal CPP, ICP, and brain glucose levels. Electrographic seizures occurred in 4 of 8 patients. Six patients (75%) experienced multiple discrete episodes of brain tissue hypoxia below <20 mm Hg during the course of monitoring, and 3 had reductions below 10 mm Hg. Factors contributing to brain tissue hypoxic events included systemic PaO2 desaturation (N=3), hypotension (N=3), seizures (N=1), and paroxysmal hyperventilation (N=1). PbtO2 levels remained stable, CBF trended upward, and LPR trended downward over time in all patients. Two of 4 monitored patients experienced late LPR elevations in response to hypoxia or hypotension. Impaired auto regulation (a positive correlation between CPP and PBtO2) at normal MAP levels occurred in 3 of 8 patients. One patient experienced critical brain hypoglycaemia due to systemic hypoglycaemia. One patient survived to discharge following commands, 3 were discharge alive in coma, and 4 patients died.

Traumatic brain injury is a common neurointensive care problem. Increased glucose metabolism, termed hyperglycolysis, is associated with worsened outcome and cell death. Glycaemic control after TBI may be important, but the optimal range of glycaemic control is presently unknown. We hypothesized that intensive glycaemic control; glucose 80-110 mg/dl would be associated with a reduction in glucose metabolism as compared to the hyperglycaemic state.

Five patients with severe TBI underwent, in a randomized fashion, a pair of quantitative fluoro-deoxy-glucose F18 brain positron emission scans, first under baseline conditions of intensive glycemic control (80-110 mg/dl) followed by a second scan under hyperglycemia (120-150 mg/dl). Systemic arterial glucose was measured hourly, and kept steadily within range, for 12 hours preceding each PET scan. Cerebral microdialysis in normal appearing brain was also done at 0.3 uL/min, and hourly values during PET scanning were analyzed using a within-subjects design. Quantitative PET scans were co-registered, and regional values as well as subtraction values were obtained.

PET scans were done within 5 days of injury. Whole brain PET glucose metabolism was 27 ± 30% higher under conditions of intensive glycaemic control as compared with the hyperglycaemic condition in the same patient (p = 0.06). The extent of change in blood glucose did not correlate well with the increase in brain metabolism. In some cases, elevation of blood glucose did not result in increases in micro dialysis glucose concentrations. In contrast, the extent of change in brain micro dialysis glucose did correlate with the increase in whole brain metabolism (r = 0.78). Pericontusional regions exhibited the least, and normal appearing brain regions exhibited the greatest difference in glucose metabolism between the intensive glycaemic and hyperglycaemic control states. No net reduction in lactate/pyruvate ratio occurs during intensive glycaemic control.

Tight glycaemic control is associated with increase in glucose metabolism in normal regions of the brain. Pericontusional tissue exhibited little change in metabolism despite varying systemic glycaemic control. These data suggest that when intensive glycaemic control reduces brain glucose levels, the brain up regulates glucose uptake in a compensatory fashion. This up regulation may not be beneficial.

Head injury is associated with capillary leak in the brain. Pressure in the interstitial tissue is increased by this leakage causing compression of capillaries and further dysfunction of the microcirculation. Edema is an obvious after effect of this damaged microcirculation. The smaller molecules found in traditional colloid and crystalloid fluids exacerbate the problem by rapidly leaking into the interstitial space. BPZ-302 is a combination therapy consisting of a narrow-range medium weight hydroxyethyl starch and high dose ascorbic acid delivered in hypertonic solution. In this study we will test the hypothesis that BPZ-302 will limit the extent of neuronal loss post TBI.

Adult male Sprague-Dawley rats were anesthetized. The rats was than intubated and ventilated. After achieving a proper anaesthesia a 10-mm diameter stainless steel disc with a thickness of 3-mm was affixed to the skull. A sectioned brass weight of 450g was dropped from a height of either 1 (mild TBI) or 2 m (severe TBI) onto the center of the metal disk to induce varying TBI. Animals are allowed to survive for 72-h. After perfusion and cryoprotection, the brain was extracted and embedded in paraffin for tissue sectioning. Total count of alive and dead neurons was obtained and percent of dead neurons was calculated. ANOVA test with Bonferroni correction were utilized for the analysis.

In the severe TBI group, treated rats with BPZ-302 had significantly less neuronal death than control Hextend rats (20.76% vs. 44.06%, p<0.0001). In the mild TBI group, treated rats with BPZ-302 had significantly less neuronal death than control Hextend rats (16.00% vs. 33.77%, p<0.0001).

A therapeutic approach tailored to prevent capillary leak in TBI may have some neuroprotective effect.

Neurocrit Care (2010) 13:S1-S231 Our objective was to compare protocols throughout the US to assess the degree of variability, with attention to the definition of cardiac death and the means by which it is pronounced.

We contacted organ procurement agencies throughout the US and requested copies of current DCD protocols. We collected details on trigger points for DCD, exclusion criteria, determination of death, methods for determination, mandatory observation time after death until procurement, and time allotted in the operating room before returning to the unit or floor.

Nineteen of 36 (53%) agencies provided examples of protocols. We obtained 64 DCD protocols from 16 states in the US. The methods by which a determination of death should be made were not specified in 28 (44%) protocols Fortyfour (69%) specified electrocardiographic rhythms as part of the cardiac death determination. Among these, 6 different combinations of cardiac rhythms were used. The mandatory observation time from death pronouncement to organ procurement was 5 minutes in 18 (28%), 2 minutes in 2 (3%), 3 minutes in 1 (2%), none in 33 (52%), and unclear in 10 (16%). Most of those with no time mandated had a built-in waiting time in the definition of death.

There is considerable variability among DCD protocols in the US. Major differences exist in the definitions of death, methods recommended for determination of death, and the waiting time required from death pronouncement to organ procurement. In order to improve the DCD process and to provide consistent care, standardization of DCD protocols may be needed.

Neurocrit Care (2010) 13:S1-S231 

The objective of this study is to identify a reliable method of performing apnoea testing, as part of the brain death examination, in patients without brainstem reflexes receiving ECMO. ECMO provides extra-circulatory support to patients in cardio respiratory failure who would otherwise be expected to die. Many studies have reported brain death as a potential complication of ECMO, but none have cited how apnea testing was performed in these patients.

This retrospective review identified adults 18 years or older treated with ECMO at our institution (2001) (2002) (2003) (2004) (2005) (2006) (2007) (2008) (2009) (2010) .

Loss of all brainstem reflexes was identified in three cases. Apnoea test was not performed as it was deemed "difficult," leading to withdrawal of care. Ancillary tests such as cerebral flow studies may not document absent flow due to the ischemic nature of the injury. We suggest the following to perform an oxygenated apnoea test on ECMO. Prerequisites include eliminating the influence of neuromuscular blocking agents and sedation with a body temperature at least 37 degrees centigrade. Cardiac output with a pretest systolic blood pressure of > 90 mm Hg should be maintained through the ECMO circuit. The patient should be placed on adequate continuous positive airway pressure (CPAP) through the ventilator or anaesthesia bag, with an inline manometer and an end tidal CO 2 device to aid in the detection of respiration. Adequate oxygen concentration should be provided through the blender circuit at the lowest oxygenator sweep flow possible. A baseline arterial blood gas should be obtained before addition of CO 2 to the oxygenator from a baseline PaC0 2 of 40 mm Hg to a target of 60 mm Hg.

Apnoea testing is essential in determining brain death but is not employed consistently in ECMO-treated patients. Apnea testing using the above protocol may assist in better decision making for patients at risk of brain death.

In a previous study we compared dual brain death (DBD) to single brain death (SBD) examination, initiated in our hospital 4 years ago, regarding organ donation consent. The SBD exam includes an apnoea test and a cerebral blood flow confirmatory test only in patients with catastrophic neurological injuries. In this study we examined if SBD exam leads to more physiologically normal patients, which may expedite the transplantation process or harvest more transplantable organs

The hospital and Gift of Life of Michigan databases were screened for BD patients over 42 months. SBD or DBD exams have been optional in the new hospital policy and performed based on physicians' preference. We compared the two groups with chi-square and Wilcoxon tests (alpha 0.05).

Ninety-three out of 103 patients had complete records. Forty patients had a SBD and 53 DBD exams. No difference was found in age, sex, race, serum potassium, bicarbonate, BUN, creatinine, blood gases, use of hypertonic solutions or pressors/inotropes, heart rate, time from admission to 1 st exam, apnea test completion, decoupling, consent rate, time from consent to operating room and number of organs recovered or transplanted. A SBD exam was performed more often by the neuro-service and in the Neuro-ICU. At the time of the 2 nd BD exam, 35% of patients were on increased pressors/inotropes. Patients with SBD exam had higher incidence of diabetes insipidus, higher serum sodium and chloride at the time of death, higher systolic blood pressure and higher utilisation of confirmatory tests

Although the consent rate and the organs transplanted after a SBD was comparable to a DBD exam, we did not confirm that SBD-examined patients were more hemodynamically stable or physiologically normal nor that SBD exam expedited the potential transplantation process.

The role of continuous EEG monitoring in patients undergoing therapeutic hypothermia after cardiac arrest has yet to be defined. According to recent literature, up to one third of post-arrest patients treated with hypothermia may have seizures or status epilepticus. Questions remain about whether continuous EEG will provide information to guide treatment, change clinical outcomes, and possibly assist in prognosis.

We conducted a retrospective review of the 22 patients who underwent therapeutic hypothermia at our institution over a 10-month period. Continuous EEG monitoring was performed on all but two patients. The EEG patterns, clinical seizure activity, treatments administered, and clinical outcomes were assessed. For patterns that prompted antiepileptic therapy, we reviewed whether those patterns changed with treatment and the clinical outcome that followed.

Results 20/22 patients undergoing hypothermia were monitored with continuous EEG and their predominant pattern during hypothermia and re-warming was considered. In 2 patients, the decision was made to pursue comfort care before an EEG was performed. 7/10 patients with encephalopathic EEG patterns had good clinical outcomes. Discharge directly to home or to acute rehabilitation was considered a good clinical outcome. 0/10 patients with clinical seizure activity, epileptiform discharges, or burst suppression survived to discharge. 4/7 patients who were given antiepileptic therapy showed changes in their EEG patterns; however, all 7 patients died.

Our data suggests that continuous EEG monitoring of patients undergoing therapeutic hypothermia assists in prognosis, with encephalopathic patterns predicting better outcomes. While monitoring does lead to antiepileptic therapies in some cases, this may not change universally poor clinical outcomes that are likely associated with malignant EEG patterns.

Therapeutic hypothermia (TH) is recommended by the American Heart Association for post cardiac arrest to improve neurological outcomes. During hypothermia serum potassium levels are observed to decrease although it is unknown to what degree or how much potassium to replace. The purpose of this study was to compare serum potassium levels and potassium replacement requirements in post cardiac arrest patients undergoing TH with patients who did not receive hypothermia.

Serum potassium levels and potassium replacement amounts were analyzed every 6 hours for the first 24 hours after cardiac arrest in patients without TH (n=33) and with TH (n=48) via retrospective chart review. Potassium was replaced using a standard potassium replacement scale.

Although more potassium was replaced in the cooled group (mean = 92.5 ± 81.8 mEq) compared with in the noncooled group (mean = 71.7 ± 63.2 mEq), statistical significance was not reached (p= 0.22). Most potassium replacement occurred in 12-18 hrs and 18-24 hrs into cooling. Serum potassium levels were not significant different between the cooled and non-cooled groups at Baseline ( 

We found a trend towards more potassium requirement in cardiac arrest patients who undergo TH than in non cooled patients. Using standard potassium replacement scales is sufficient in achieving adequate potassium levels throughout the cooling phase.

Neuron specific enolase (NSE) is a serum marker of brain injury. Elevated NSE > 33 μL/ml drawn 24-72 hours after cardiac arrest (CA) predicts poor outcome with a reported 0% false positive rate. 1, 2 The prognostic value of elevated NSE in CA patients treated with hypothermia has not been validated. 1 

We retrospectively reviewed the medical records of all patients with CA who received therapeutic hypothermia at our institution from November 2006 to April 2010. NSE values were obtained after CA per American Academy of Neurology guidelines. We measured outcome by Pittsburgh Cerebral Performance Category (CPC) score at time of hospital discharge or death, as well as Glasgow Coma Scale motor subscore (GCS M ) 72 hours after CA. Good outcome was defined as CPC 1-2, or GCS M

Forty-seven patients received therapeutic hypothermia after CA. Six patients failed to reach target temperature and 6 patients did not have recorded NSE values, thus were excluded from analysis. Of the remaining 35 patients, 22 had NSE < 33 μL/ml, 13 of whom had good outcome (13/22, 59%). Thirteen patients had NSE values >33, of whom 11 (11/13, 85%) had a poor outcome as measured both by GCS motor subscore and CPC at discharge. However, 2 patients achieved initial good neurologic outcome as measured by GCS M , following commands 3 days after CA (false positive rate = 13%). One patient expired due to respiratory failure despite having achieved good neurologic outcome, thus measured by CPC score at discharge, 1/13 (7%) had a good outcome despite elevated NSE.

NSE > 33 μL/ml is not an absolute predictor of poor neurologic outcome in patients treated with hypothermia after cardiac arrest. The previously reported 0% false positive rate for predicting poor outcome after cardiac arrest rate was not validated in our population treated with hypothermia.

Magnetic resonance imaging (MRI) may prove useful in assessing the comatose patient's prognosis. After injury, MRI findings evolve temporally and spatially, but little is known about these changing patterns and their relevance to a patient's outcome. We report the MRI data from 9 patients who suffered a global hypoxic-ischemic insult, and subsequently had serial MRI studies at different time points. Their imaging findings are described in the context of their clinical outcomes. This is the largest serial imaging study of this patient population to date.

Diffusion weighted imaging (DWI) and fluid attenuated inversion recovery (FLAIR) MRIs were conducted at two time points for 8 patients and three time points for one patient, all of whom were comatose after cardiac arrest. Each of the four cortices, the basal ganglia, thalamus, brainstem, cerebellar hemispheres and vermis of each study were read by 2 stroke neurologists as either "normal" or as demonstrating hyper intensities related to the recent insult. Neurological exams were recorded on days 0, 1, 3 and 7. Outcomes were determined by the modified Rankin Scale (mRS), with a poor outcome defined as a mRS score

In the acute (<24h) period, both imaged patients exhibited changes on DWI and FLAIR in the cerebellum and basal ganglia. In the early subacute period (days 1-5), cortical abnormalities predominated, with a shift to more white matter changes in the late subacute period (days 6-12). We observed widespread imaging abnormalities in the 7 patients with poor outcomes, and partial or full resolution of DWI abnormalities in the two patients with good outcomes.

Our findings support previous reports of temporally dependent MRI patterns after global hypoxic-ischemic injury, suggesting widespread, persistent changes may associate with a poor prognosis. Further research is necessary to clarify these patterns and their significance.

Shivering is common during targeted temperature management (TTM) after cardiac arrest, but few reports describe shivering intensity or pattern. We present a graph of shivering based on EMG power recording, categorize the patterns, and compare demographics and outcomes between those with robust vs. minimal shivering.

We prospectively recorded electromyographic (EMG) power and bispectral index (BIS) from the forehead using the Aspect A2000 monitor during TTM in 112 adults. The average EMG each minute was plotted consecutively from start of TTM until rewarming was started, and the BIS1 was recorded after the first dose of neuromuscular blockade (NMB). Clinically-recognized shivering was treated independent of the EMG by bedside nurses using intermittent NMB, and outcome at hospital discharge was assessed using the Cerebral Performance Category (CPC), with CPC 1-2 classified as a good outcome. Retrospectively, EMG patterns were reviewed independently by 2 investigators blind to all patient data and scored as little-or-no shivering (NO), or robust shivering (SH). Demographic and clinical data were compared between groups.

21 EMG patterns were classified as NO, 81 as SH; 10 could not be classified. Patients in the NO group received fewer NMB doses (2 vs. 6). 28% (6/21) of NO patients attained a good outcome, compared to 52% (42/81) SH patients (p=0.047 Yates Chi-Square). NO patients had a longer time to ROSC (28 vs. 18 mins) and a lower BIS1 at start of TTM (10 vs. 26) suggesting more severe brain injury.

During TTM after cardiac arrest, continuous EMG recording provides a reliable marker of shivering intensity and NMB use. Patients with minimal shivering are less likely to make a good recovery, perhaps reflecting more severe brain injury and impaired thermoregulation capabilities.

We tested the hypothesis that patients with more severe brain injury after cardiac arrest experience less shivering during therapeutic hypothermia (TH). We used four novel methods to quantify shivering, since optimal quantitative evaluation of shivering has not been established.

Methods 39 patients were treated with TH followed by controlled warming and normothermia. EMG was measured each minute for the first 24 hours using a single frontotemporal surface electrode and the Aspect A2000 (BIS) monitor. We evaluated shivering intensity using 4 different methods: the Bedside Shivering Assessment Scale (BSAS), number of doses of neuromuscular blockade (NMB) administered by nurses in response to shivering, "shiver minutes" when EMG>34dB (shivering) or <27dB (received NMB), and number of "shivering episodes" -2 hour blocks in which EMG>34 or NMB administered or BSAS>0. Each of these variables was compared to one of 2 markers of severity of brain injury: the discharge Cerebral Performance Category (CPC) and the "BIS1," described as the BIS score immediately following the first dose of NMB.

Each increase of one shivering episode was associated with a 35% increase in the odds of a good neurological outcome (CPC 1-2, OR 1.35, p= 0.05). There was no significant association between other measurements of shivering and the markers of severity of brain injury. 

"Shivering episodes," a combined index of EMG and the clinical identification of shivering by nurses, correlated with severity of brain injury as measured by the Cerebral Performance Category at hospital discharge, In seven other models, the lack of association between brain injury and shivering may be real, or may be due to errors in measurement or methodology.

Mild therapeutic hypothermia for survivors of cardiac arrest worldwide has been underused, at least partially because of the perception that the implementation of this therapy is technically difficult at the community level. Our objective was to demonstrate that the application of therapeutic hypothermia is technically feasible in a community-based setting in a developing country.

Retrospective cohort study of patients treated with therapeutic hypothermia after cardiac arrest in a community hospital in São Paulo, Brazil. After return of spontaneous circulation (ROSC), survivors of cardiac arrest were treated with mild therapeutic hypothermia using ice and cold saline infusions in order to cool patients to 32 to-34 degrees Celsius within 6 hours of cardiac arrest to achieve goal temperature within 4 hours of hypothermia initiation and to maintain goal temperature for 24 hours. Esophageal temperature was monitored for all patients during hypothermia and rewarming.

From January 2009 to April 2010, 19 survivors of cardiac arrest (14 males, mean age 33 ± 25 years) were managed with therapeutic hypothermia. Initial rhythm was most commonly ventricular fibrillation (13 patients), followed by asystole (4 patients), ventricular tachycardia (1 patient) and pulse less electrical activity (1 patient). The mean time from ROSC to initiation of therapeutic hypothermia was 89 ± 50 minutes and the mean time from ROSC to goal temperature was 51 ± 32 minutes. Complications during treatment with hypothermia included: arrhythmias (36%), pneumonia (16%). and gastrointestinal bleeding (11%). Overall survival at hospital discharge with good neurologic outcome (Glasgow outcome scale 4-5) was seen in 42% of the patients.

Mild therapeutic hypothermia can be implemented safely using a simple protocol and locally available resources in a community-based setting in a developing country.

Accuracy of blood glucose measurements from various sites during therapeutic hypothermia (TH) is poorly characterized. We hypothesized that body temperature might affect the agreement of capillary (CG) and arterial glucose (AG) levels in cardiac arrest (CA) survivors, and evaluated this question prospectively.

IRB approval and informed consent were obtained, and TH performed after CA by protocol. Concurrent AG and CG specimens were analyzed using point of care (POC) testing (Precision Exceed Pro, Abbot Diabetes Care Inc.), and additional laboratory testing (Cobas 501, Roche Inc.) was performed at intervals. Agreement between samples was defined as difference < 15mg/dl when AG <75mg/dl, and <20mg/dl when AG > 75mg/dl. Agreement was determined using the method of Bland and Altman, and strength of association by Pearson's correlation coefficient.

Overall correlation between POC and laboratory AG was 0. 

Hyperglycaemia is more prevalent at T<34ºC than T>36ºC. Capillary glucose values were frequently discordant from arterial glucose values at T<34ºC, and when discrepant, capillary values were higher. Clinicians should use caution titrating insulin therapy in cardiac arrest survivors during TH, as capillary glucose values may overestimate the serum glucose level by a significant degree.

Approximately 350,000 cardiac arrests occur annually in the United States. Only a subset of patients survives with good outcome. Hypoxic-ischemic brain injury is the main cause of cognitive disability. Data on long-term cognitive outcome is limited.

Consecutive comatose post-cardiac arrest patients were enrolled prospectively. Survivors underwent neuropsychological assessment at 6 months using the Expanded Halstead-Reitan Battery

Medical care is critically needed in underdeveloped countries throughout the world, and medical mission trips are a valuable means of providing such care. In November 2009, a team of expert neuroscience clinicians and support staff embarked on a life-altering experience by participating in the first West African neurosurgical mission to Nigeria and Ghana. This poster describes their experience from the perspective of the neurocritical care nurse practitioner on the team. Details of cases performed, education provided, and lessons learned from this intensive, rewarding experience and recommendations for future neurocritical care medical mission trips are discussed.

Fifteen medical volunteers from neurosurgery, neuroscience nursing, anaesthesia, and neuroradiology were assembled. Preparations included obtaining licensure/visas, immunizations, and medical supplies and arranging for travel and lodging. An OR list of potential cases was distributed in advance to all. A comprehensive educational activity was organized to offer formal instruction to the West African neurosciences medical community.

Seventeen surgeries were performed in Nigeria over 5 days, and 10 were performed in Ghana over 4 days. The team worked well under suboptimal conditions considering that the GOR and ICU availability and ICU medical coverage were not as originally promised. Surgeries were performed on pediatric and adult patients from 8 months to 65 years old and included craniotomy for tumor, vertebral spinal procedures, endoscopic third ventriculostomy, and myelomeningoceal repair.

Although the impact of this trip seems minuscule when compared to the total number of patients who require neurosurgical intervention in Nigeria alone, the lives of 27 people were drastically changed. The education provided via hands-on instruction and the didactic activities have already created changes in practice that will ultimately enhance the quality of care provided. However, a formal evaluation with recommendations for future endeavours and ongoing communication are vital for a lasting result.

To assess the value of the practice of obtaining frequent electrolyte measurements in patients with extended stay in a neuroscience intensive care unit (NICU).

We identified consecutive patients 18 years or older, admitted to the NICU between January 1 and July 31, 2009, with length of stay orus laboratory measurements and haemoglobin levels, and recorded electrolyte replacement orders and red blood cell transfusions.

Ninety-three patients were included in the study (54 men, mean age 54 years, range 18-85 years). Mean length of stay was 10.4 days (range 5-36 days). Sodium and potassium were the electrolytes most frequently measured (averages of 14.1 and 13.1 per patient, respectively). More than 75% of the results were within normal range for all electrolytes measured and critical values were extremely uncommon. The number of phlebotomies for electrolyte measurements was strongly associated with the degree of haemoglobin drop (p<0.0001). Electrolyte panels were ordered much more often than individual electrolytes.

Electrolytes measurements are very frequent in the NICU, but results are most often normal and only exceptionally critical. The phlebotomies required for these tests significantly worsen haemoglobin levels. A more conservative use of electrolyte measurements can result in reduction of blood loss.

The primary aim of this study is to assess the value in the practice of obtaining frequent electrolyte laboratory data in patients with an extended stay in a neuroscience intensive care unit (NICU).

Ninety-three patients, 18 years of age or older, admitted to a NICU for a minimal of 5 consecutive days between January 1 and July 31, 2009, were included in this retrospective review. Laboratory data gathered included potassium, sodium, magnesium, ionized calcium and phosphorus measures, noting the type of test ordered, whether individual or a part of an electrolyte panel. Electrolyte results were recorded to assess number of abnormal values, and haemoglobin values were collected to determine if a decline occurred during the ICU stay. Average laboratory costs were estimated.

Electrolyte panels were drawn substantially more often than individual electrolytes which added to patient costs. The majority of the laboratory results (75%) were within normal limits for all of the electrolytes values. Only 7 (0.6%) of the potassium and 1 (0.2%) of the phosphorus results were deemed critical. Eighty-nine percent of the potassium results were drawn as part of the electrolyte panel, and 84% of sodium results were drawn as an electrolyte panel. A statistically significant relationship was found between the change in haemoglobin and the number of laboratory draws obtained (p<0.0001). The estimated average patient costs for electrolyte panels during this study period exceeded $2200 per patient. Replacing half of these electrolyte panels with single sodium or potassium orders would have resulted in savings greater than $100,000 in our population.

A more conservative use of electrolyte measurements can result in substantial cost savings in this patient population. Obtaining only single electrolyte measurements instead of electrolyte panels is one way to achieve cost savings in any patient population. The effect on safety with a reduced frequency of laboratory monitoring in this patient population needs further research.

Intracranial pressure (ICP) data is required for clinical research in acute brain injury (ABI). In ICU's without continuous data acquisition (CDA), ICP must be obtained from the EMR. EMR data is typically recorded every 15 minutes or hourly by nursing. CDA appears to be the "gold standard" for experimental physiologic monitoring. However, EMR data is cleansed of erroneous values by nursing, while these are included in CDA. Prior studies have shown that mean ICP EMR values are valid, but they miss SBI events (ICP>20mHg) 1 . EMR data may also be extrapolated to approximate CDA 2 . The goal of our study is to examine the validity of EMR and CDA data as a research tool in ABI.

Five patients with ABI requiring ICP monitoring were randomly selected for CDA using ICU Pilot ® software sampling at 1 minute intervals for 24 hours. EMR ICP data, recorded by nursing who generally charted values every 15 minutes, was retrospectively collected for the same period after discharge. Mean ICP, ICP distribution, and SBI events (ICP > 20mmHg) were calculated and compared between CDA and EMR.

CDA and EMR data yielded a total of 7407 and 408 observations, respectively. The average number of observations per patient was 1481.4 and 81.6, respectively. The mean sampling rate of EMR was 1 observation per 20 minutes. The mean ICP values were 13.40mmHg and 13.38mmHg (p < 0.993). Analysis of ICP distribution was virtually identical. Importantly, the total number of SBI events recorded by CDA was 119 whereas EMR documented only 33.

The results of this pilot study indicate EMR data is sufficient to examine mean ICP and trends. However, EMR data significantly underreports the frequency of SBI's. Therefore, EMR data is insufficient for research data collection even at a sampling rate of every 20 minutes. 

Health care providers' perception of prognosis may influence decisions regarding aggressiveness of care in the intensive care unit. We aimed to evaluate whether differences exist in health care providers' perception of prognosis in neurologically critically ill patients.

Patients who were intubated with a neurological illness for (NI), general intensivists (MICU), nurses (RN) and residents (RES) who were caring for the patient were asked to predict 6-month: 1) functional outcome rated by the modified Rankin Scale (mRS 0-6) and 2) quality of life (QOL) rated by a simple scale: poor, fair, good, or excellent. Differences between groups were analyzed using the Friedman, Wilcoxon signed rank and McNemar tests.

For the 189 prospectively enrolled patients, 153 (81%) and 149 (79%) mRS and QOL predictions were completed by all four health provider groups, respectively. The majority of patients had a diagnosis of hemorrhagic or ischemic stroke and status epilepticus. Comparing all groups, there were significant differences in mRS and QOL prediction scores (p<0.001). After adjustment for multiple comparisons, the only significant differences were between the NI and other health care providers (p<0.001). The NI were more optimistic in predicting functional outcomes, particularly with a higher frequency of mRS scores optimistic in predicting QOL, particularly with a higher frequency of fair to excellent responses compared to MICU, RN and RES (p<0.001).

At our institution neurointensivists are more optimistic in their predictions of long-term outcome in neurologically critically ill patients than general intensivists, nurses and residents caring for the same patient. 

Two anonymous, one-time surveys: The first was mailed to survivors of patients who died in 2008 in two adult, rural, tertiary care ICUs. The companion survey was emailed to intensivists in those ICUs.

109/374 surveys from survivors were returned (28% response rate). 22% received bereavement support. 91% found it helpful or very helpful. 70% felt that bereavement support "showed that our loved one mattered." Interestingly, of those who did not receive bereavement support, 51% would have liked support and 49% were OK with not receiving support.

18/40 adult intensivists responded (45% response rate) 33% of the intensivists provided bereavement support for families which they universally experienced as comforting or very comforting for families. Of those who did not provide bereavement support the most commonly cited reasons were the discontinuity of care and that bereavement support was not necessary.

This is the first study to describe adult intensivist bereavement care. Most intensivists do not provide bereavement support The experience of providing bereavement support was well received by the families and a good experience for the intensivists providing the support.

This study is a description of bereavement support in two tertiary care ICUs. Many more questions are raised about what constitutes quality bereavement support, who needs it, how should it be delivered and how we should train future intensivists.

Future research may help elucidate the role of bereavement support as part of family-centered care in the ICU. There are no studies to date describing whether neurointensivists provide bereavement support. Our research goal was to describe bereavement care in the Neuro ICU.

An anonymous, one-time electronic survey was sent to members (physicians and other healthcare providers) of the Neurocritical Care Society.

151/1028 members responded (19% response rate). 50% of the responders provided bereavement support. Of those who did, 57% used the phone. 17% attended a wake or a funeral. 79% found the bereavement support to be comforting or very comforting for the family. 75% provided bereavement support to comfort the family and 50% to show that the family/patient mattered.

The 50% of providers who did not render bereavement support did not for many reasons: not their role (38%), lack of knowledge/skill set (25%), discontinuity/fragmentation of care (38%).

52% of all respondents wanted more training in bereavement care.

This is the first study to describe adult neurointensivist bereavement care. The experience of providing bereavement support was perceived to be well received by the families and a good experience for the intensivists providing the support. Many more questions are raised about what constitutes quality bereavement support, who needs it, how should it be delivered, and how we should train future neurointensivists. In addition, there may be institutional practices which provide barriers to bereavement care.

Future research may help elucidate the role of bereavement support as part of family-centered care in the Neuro ICU.

Recent studies have shown worsened mortality and increased length of stay in teaching hospitals in many diseases. The hospital course of subarachnoid hemorrhage (SAH) is complex and involves dedicated neuro -intensive care unit admission and care. We hypothesized that the outcomes will be better in a teaching hospitals due to better specialized care and adherence to evidence based protocols.

A retrospective study utilizing the Nationwide Inpatient Sample 2003 to 2007 was performed. Using the ICD-9-CM code 430, all adult records (age considered to be teaching if they were either approved by Accreditation Council for Graduate Medical Education or were in Membership in the Council of Teaching Hospitals or had a ratio of full-time equivalent interns and residents to beds of .25 or higher. Multivariate logistic regression was performed which adjusted for demographic characteristics and Elixhauser co-morbidity index. The primary outcomes studied were mortality, LOS and total charges. The procedures studied were aneurysm repair, intracranial pressure monitoring and ventriculostomy.

A total of 163,160 estimated discharges with SAH from 2003 to 2007 were evaluated. 71% were admitted to a teaching hospital and 29% were admitted to a non teaching hospital. The mortality of patients was lower in teaching hospitals when compared to non teaching hospitals (Odds Ratio (OR) 0.72; 95% Confidence Interval (95%CI) 0.66-0.78). There were significantly more procedures performed in the teaching hospitals than non teaching hospitals (OR 2.94; 95%CI 2.44-3.54). The length of stay in SAH was 3.7 days longer in teaching hospitals (95% CI 3.00-4.36). They also incurred $34,167 higher charges when compared to nonteaching hospitals (95%CI $23,495 -$44,839).

Our results show that SAH associated mortality was significantly lower in teaching hospitals although the LOS and associated total charges are higher.

Anemia is common in the NeuroICU. Etiologies include post-operative bleeding, phlebotomy, and decreased production. Blood transfusion risks include infection, TRALI, and transfusion-related reactions/complications. To limit blood-product exposure, restrictive blood administration strategies have been recommended. Since September 2008, patients admitted to our NeuroICU have been co-managed by neurointensivists with adoption of a simple blood conservation program.

A strategic program to improve blood utilization was implemented. This included: 1) lectures and didactic sessions given to the neurosurgical house staff on the risks/benefits of blood transfusions; 2) decrease volume of routine phlebotomy; 3) transfusing based on physiologic parameters, not hemoglobin levels; 4) transfusing single unit of PRBCs with re-assessment.

We retrospectively compared PRBCs transfusions the year prior to the addition of neurointensivists to the first year after the neurointensivists program with the blood conservation program.

The pre-program transfusion rate was 0.45 units/admission. The transfusion rate after establishment of the program was 0.30 units/admission; a 14% decrease in PRBC transfusions. This reduction was in light of a 22 % increase in admission rates. During that time there was also a reduced LOS, ventilator days and mortality (4.2% to 2.7%).

Based on the admission rate and PRBC administration rate, this blood conservation strategy resulted in a $25,000 saving in just PRBC cost alone. Not included is the cost savings of blood typing and subsequent lab, administration cost and cost of blood-product related complications. How the program affected these metrics is unknown.

Critical neurological illnesses remain a leading cause of death in the US yet, little is known about the use of palliative care consultation for patients in neuro-critical care settings. We sought to understand how our NICU utilized the Palliative Care Service.

We retrospectively analyzed patient characteristics from five years of palliative care consults at Duke University Hospital. Data collected included demographics, discharge disposition, time between admission and consultation, time between consultation and discharge, length of stay, diagnosis, and types of recommendations made by the palliative care consultant. Individual characteristics were grouped to identify trends among neuro-critical care patients.

Our use of formal Palliative Care consultations has doubled each year since 2005. The average age of referred patients was 62. A typical patient was hospitalized for 11 days prior to a palliative care consult being initiated. Patients' had a disposition, on average, within five days following the consultation. The most common reasons for consultation in order of frequency were to help facilitate communication with the family, facilitate end of life discussions, and to assist with symptom management and disposition issues. Of patients receiving a palliative care consultation, 73% of patients died in the hospital.

This study demonstrates the perceived value of a Palliative Care Consultation in the NICU as shown by the steadily increasing utilization of the service. Palliative care consultations are known to result in improved communication, often reducing the length of stay and should be considered earlier in the course of illness. This study generates a number of questions for future research including identifying which patients would benefit from a palliative care referral and when, as well as the impact of palliative care on symptom burden, satisfaction, and quality of life.

Management of the critically-ill neurological patient is complex and dynamic. Goals of care are aimed at optimizing neurological function by reducing and preventing secondary injury. Advances in technology and understanding of brain physiology through multimodality neuromonitoring have provided specific parameters that can be used as general guidelines in the prevention of secondary brain injury which may potentially improve outcomes.

A survey was performed to determine the ongoing educational needs of nurses working in a neurological intensive care unit.

Identified educational needs include essential therapies for acutely brain injured patients as well as key indicators of cerebal resuscitation. These include: therapies to manage ICP; therapeutic modulation techniques; the Bedside Shivering Assessment Scale (BSAS); multimodality neuromonitoring parameters as well as "normal" levels and critical elements influencing oxygen utilization and delivery. Subsequently, a simple concise BRAIN (Brain Resuscitation: Advanced Indicators in Neuromonitoring) card was developed which integrates the aforementioned treatments as well as monitoring parameters while allowing for flexibility in individual patient responses and needs.

The BRAIN card will be incorporated into yearly nursing neurologic competencies and integrated into the orientation of new nurses. It will be updated annually by consensus of NICU staff (Nurses, Fellows and Attendings) to account for the rapidly changing literature. It can act as a reference for bedside nurses to quickly access information which may help novice nurses become familiar with the therapies and parameters. Experienced nurses may use it as a teaching tool. Furthermore, it may serve as a checklist to determine patient needs which aims to enhance continuity of care and increase collaboration with the critical care team.

On January 12 Haiti was struck with a 7.0 Richter scale earthquake resulting in death, massive injuries and thousands homeless. A disaster in the setting of existing poor socio-economic and health conditions coupled with inadequate health care delivery systems adds to the challenges of acute care delivery and management of ongoing health issues. Within 8 days of the disaster the USNS Comfort was docked two miles off of the coast of Port-Au-Prince and served as a tertiary health care facility. Seven days later 30 volunteers from several medical disciplines and specialties flew onto the deck of the Comfort to serve on a 21-day medical mission. During those three weeks and four more to follow the Comfort treated close to 900 patients who presented with a host of injuries; the majority were neurotrauma, fractures, paresis, and traumatic brain injury, surgical and non-surgical alike. There were many peripheral injuries as well as exacerbation of chronic issues such as CHF, stroke and kidney disease. An inordinate amount of patients became "do not resuscitate" due to the resources and nature of the situation. Whereas, the Comfort had been in port reorganizing the ship just prior to the earthquake, several infrastructures and supply chains were not fully up to par.

By way of personal experience, interviews and data provided by the US Navy, USAID and Project Hope, an effort is made to provide a snapshot of the scope of the problem, severity of injuries, logistics of acute and chronic care, as well as the dedication of Navy and civilian health care workers. Neurological injuries incorporated a large extent of the injuries.

850 plus patients were treated during the 7-week mission that included complex spinal and craniotomy surgical procedures.

The acute care, consultation and subsequent after care is vital in for those so much less fortunate then ourselves. There is hope for a bright future and ongoing vitality. The brilliant effort of the Navy to prepare the ship "on the fly" was nothing less than genius. The ingenuity of Navy medicine and the undying effort of military and civilian members were awe-inspiring Financial Support: None

Sodium management is a mainstay of therapy for many patients in neurocritical care units. However, not all neuroICU practitioners believe that aggressive maintenance of sodium goals is indicated, and if so, how sodium goals are best achieved. Strategies to protocolize target sodium goals may be useful, if neurointensivists consistently agreed that pre-defined sodium goals were appropriate, achievable, and beneficial.

Data was collected for 100 consecutive days for every patient in a 27 bed academic neuroICU. The patients were assessed for whether or not they needed any sodium goal, and if so, if that goal was to be normal (Na of 135-145 meq/L), hypernatremia (145-155 meq/L) or "other." Secondary endpoints included consistency in category assignments amongst the four attending neurointensivists and response rate in order to assess feasibility with moving forward with the next phases of data collection.

1995 data points were collected, out of a possible 2700. 1840 (92%) were felt to need a sodium goal. Of those, 74% were assigned a "normal" goal and 18% were assigned a "hypernatremia" goal. Only 7 patients were assigned to the category "other," so the pre-defined categories fit almost all patients. Opinions differed amongst the four physicians (no goal: 2-37%; normal: 50-88%; hypernatremia: 7-25%). Data points were missing in 24% (n=644; additional missing data accounted for by 61 empty beds), with a disproportionate number on weekends/holidays (n=324).

Despite four different physicians providing care for them, a vast majority of patients in one neuroscience ICU were thought to need specific management of sodium (either eunatremia or hypernatremia). Future questions to be addressed by this study include whether patients with set goals were meeting the goals, if protocolization of sodium management is feasible and would improve maintenance of sodium goals, and ultimately if setting and maintaining sodium goals would improve outcomes in neuroICU patients.

Poster 29

Traditionally, M&M conference was an academic exercise using a punitive approach to case review. Audience participation was limited to physicians in a single department, with exclusion of other health-care providers. Recent evidence suggests that a multidisciplinary approach may be more effective for identifying and correcting system-based errors.

The neurocritical care team redesigned M&M conference to focus on multidisciplinary and system-based issues, rather than using the traditional punitive approach. First, a literature search using Medline and Google Scholar was conducted to evaluate reported M&M structures. Next, a tool was developed to identify important M&M cases for discussion. Voluntary reporting of potential cases was encouraged. After case selection, all involved health-care providers were invited to either present or attend the conference. An experienced physician uninvolved in the case moderated the M&M using a specific process based on the literature. After each case, an action plan was created, and instituted by the appropriate services. The results of the action plan were discussed at the subsequent meeting, and further modifications were made as indicated.

The new Neurocritical Care M&M process has been utilized for one year, on a monthly basis. The meeting was wellattended by neurologists, neurosurgeons, neurointensivists, neurointerventionalists, nurses, pharmacists, and hospital administrators. Case presentations were multidisciplinary and interactive. To date, 35 patients, with 15 system-based errors were presented. 13 of those issues have been remediated to date. Issues successfully resolved relate to interdisciplinary communication, consultation of appropriate services, and errors in drug administration. Solutions included refinement of ICU protocols, highlighting deficiencies in communication, and educating health-care personnel on patient-care issues.

Transitioning from a traditional punitive-based M&M to a structured multidisciplinary systems-based approach has resulted in a more open and productive environment for enhancing patient care.

Telestroke consultation extends acute stroke care to hospitals and regions that lack stroke expertise, but costs of technology may be prohibitive. We sought to develop a Telestroke model that focused on excellent system-based processes rather than expensive technology.

We developed a protocol to facilitate development of the hub (telestroke team) and spoke hospital relationship. Hospital systems were analyzed with regards to stroke volumes, referral patterns, local EMS systems, and existing technologies. Educational support was provided by the Neurocritical Care telestroke team to hospital physicians, nurses, and local EMS providers. A shared outcomes database was developed to monitor program quality, and reviewed quarterly.

Connectivity between the first hub and spoke hospital began December 2009. During a six-month period, tPA was given to 11 patients, compared to no patients during the prior six months. Thrombolysis rates for all ischemic stroke patients arriving to the spoke hospital increased from <1% to 20% during the same period. Average time from telestroke consultation to rtPA administration and triage time to rtPA administration was 31 min (12-75) and 76 min (37-110). All patients receiving rtPA at the spoke facility were transferred to the hub hospital; no transfer-related complications occurred. In cases where endovascular treatment was deemed necessary, time from triage at spoke hospital to endovascular intervention at Hub hospital averaged 206 min (183-237). The cost of technology for the spoke hospital was $1500, for the telestroke computer and web camera. Hiring of additional IT personnel was not necessary due to the simplicity of the system.

The Neurocritical Care TeleStroke program resulted in a substantial increase in the number and rate of patients receiving thrombolysis with rtPA at the spoke hospital, with limited technology costs.

The use of IV rt-PA remains low (1-3%) 1 , despite improvement in clinical outcome up to 4.5h after symptom onset 2 .

Visualizing ischemic as opposed to infracted tissue may aide in expanding rt-PA eligibility. Some MRI based studies have suggested a role of perfusion imaging in identifying candidates for thrombolysis 3, 4 . However, CTP is far more accessible, and 64 slice CT permits near whole brain coverage, which could significantly improve CTP utility. We sought to determine the feasibility and utility of 64 slice CTP in IV rt-PA treated stroke patients.

Retrospective study of all IV rt-PA treated patients between 7/2006-6/2010 who received pre-treatment CTP. Preand post-rt-PA CTP, CT, and MR images were reviewed. Outcome was determined by initial and discharge NIHSS. Safety was evaluated by symptomatic ICH rate.

117 patients had a pre-treatment CTP. 9 also received IA rt-PA. 50% demonstrated perfusion mismatch (pre-and post-average NIHSS 10.3, 6.9); 42% no perfusion defect (NIHSS 7.5, 3.1); 8% matched defect (NIHSS 8.8, 9.9). 8 patients received rt-PA at >4.5h (range 275-390 min); none had a matched perfusion defect. Follow-up MRI demonstrated reduced final infarct volume in 4/8 patients, 7 showed improved/stable NIHSS, and 1 had sICH in the contra lateral hemisphere. 2/117 patients had sICH, but their perfusion defects were contra lateral to the ICH.

CTP is feasible and useful in determining eligibility for IV rt-PA. Perfusion mismatch and lack of perfusion defect may be predictive of improved clinical outcome post-IV rt-PA. CTP may also be useful in expanding the time window for IV rt-PA administration, and in identifying patients with low risk of bleeding complications post-thrombolysis.

Tracheostomy is a common procedure in critical care patients. Advantages of a short tracheal tube compared to a long orotracheal tube are the avoidance of laryngeal lesions and sinusitis, facilitation of nursing care and physiotherapy and the reduction of analgosedatives. The optimal point in time for tracheostomy is still unknown, but it is commonly done not later than 2-3 weeks and after one or several failed extubation trials. Studies in different sets of critical care patients have suggested advantages of early tracheostomy: less pneumonias and other complications, better patient comfort, less analgosedation, shorter duration of ventilation and of ICU stay. These questions have not been looked at in non-traumatic neurocritical care patients, although these patients might have a special weaning benefit by early tracheostomy, being mainly compromised in securing their airway, but not necessarily in breathing.

Intubated non-traumatic neurocritical care patients with ischemic stroke and intracerebral or subarachnoid hemorrhage were screened for the trial. If an in-house score based on lesion volume, lesion extension, dysphagia and aspiration on admission, certain concomitant disorders and additional clinical estimation suggested ventilation demand for about 2 weeks, patients were eligible. After obtaining informed consent from the legal representative, patients were randomized into group 1 -tracheostomy within the first 3 days after intubation -or group 2 -continuation of orotracheal intubation with consecutive weaning and extubation or tracheostomy between day 7 and 14 after intubation. Tracheostomy was done as percutaneous dilatation by neurologists. Details of the study design, endpoints and inclusion/exclusion criteria will be presented.

This pilot trial aims at inclusion of 60 patients. So far, we have included 35 patients and will be able to present an interim analysis at the time of the meeting, when 2/3 of the study population should be included.

We hypothesize that early tracheostomy is of benefit for cerebrovascular neurocritical patients through reduction of analgosedation allowing earlier neurological assessment and weaning, avoidance of infections and other complications, shorter ventilation time, shorter ICU stay, earlier mobilization, physiotherapy, rehabilitation and maybe even outcome. We will present the study design, the literature context and first results.

Neurocritical care patients with severe space-occupying -so-called malignant -middle cerebral artery (MCA) infarcts have a poor prognosis even with aggressive medical treatment. Randomized trials showed that early hemicraniectomy reduces mortality from about 70% to 20% and improves outcome in survivors without increasing the risk of being very severely disabled. However, only patients up to 60 years have been included in these trials. It is presently unclear whether patients older than 60 years -who represent about 50% of all patients with malignant MCA infarcts-would also benefit from this therapy. Data from observational studies suggesting that older patients may not profit from hemicraniectomy are inconclusive, because these patients have generally been treated later and less aggressively.

DESTINY II is a national multicenter randomized controlled trial including patients 61 years and older with malignant MCA infarcts. Patients are randomized to either maximum conservative (i.e. intensive care) treatment alone or in addition to early hemicraniectomy within 48 hours after symptom onset. The trial uses a sequential design with a maximum number of 160 patients to be enrolled (ISRCTN 21702227). The primary aim is to assess whether decompressive surgery reduces mortality without increasing the number of very severely disabled -i.e. completely dependent and bedridden -survivors. Outcome is assessed by clinical scores and questionnaires directed at quality of life. DESTINY -R is a national registry of patients with malignant MCA infarcts; DESTINY -S is a national survey on the current treatment and views on treatment of patients with malignant MCA infarcts, both irrespective of the DESTINY II trial.

Currently, over 10 sites in Germany are enrolling patients into DESTINY II, with the number of initiated sites still growing. DESTINY -S has yielded over 800 responses so far in Germany and might be extended internationally.

We postulate clarification of the effect of hemicraniectomy in malignant MCA infarct patients older than 60 years by DESTINY II. We anticipate elucidation of the current practice of treating malignant MCA infarct by DESTINY -R and -S. Trial designs, literature context and perspectives will be presented.

In aging, astrogliosis and the development of the glial scar is accelerated. This change has been hypothesized to be responsible for impaired functional recovery in older individuals. TGF is implicated in glial scarring, we have therefore been investigating whether TGF signalling after stroke is increased in aged animals, and whether this difference is responsible for accelerated glial scar formation and impaired functional recovery.

To look at the response to TGF in real time we used reporter mice that give a real-time readout of TGF responses after stroke via the expression of luciferase (SBE-lucRT mice). We performed distal middle cerebral artery occlusion (dMCAO) strokes on 5 and 18 month old mice. To determine which cells are responding to TGF after stroke, brain sections were double-labelled with anti-pSMAD2, a marker of TGF signalling, and markers of different cell types.

We found that TGF signalling after stroke in the old mice was substantially higher than in the younger animals. In line with previous reports, astrogliosis was accelerated in the older mice. Reactive astrocytes showed increased pSMAD2 immunoreactivity after stroke demonstrating that they respond to TGF after stroke. To test if TGF signaling in astrocytes is a mediator of astrogliosis we constructed transgenic mice that are designed to decrease TGF signalling in reactive astrocytes via a dominant negative TGF receptor mutation (Astrocyte DN-TGF RII mice). We performed dMCAO strokes on 3-4 month old transgenic mice and their wild type littermates. In our DN mice at 14 days after stroke we observed significantly decreased staining of GFAP. Interestingly, this alteration in astrocytic activation also correlated with a significantly larger immune response 7 and 14 days after stroke, characterized by increased numbers of classically activated macrophages.

Together these findings show that after ischemia, astrocytic TGF signalling modulates astrogliosis, limits proinflammatory immune responses and increases with age.

Combined full dose IV and IA thrombolytic therapy theoretically takes advantage of both modalities, but little is known about the efficacy and risks of this approach. We retrospectively evaluated the efficacy and safety of this method at our center.

A retrospective chart review was conducted of all acute ischemic stroke patients receiving full dose (0.9 mg/kg) IV r-tPA and IA thrombolysis. Treatment characteristics and clinical outcomes were assessed.

From July 2007 to June 2010, 21 patients received combined IV and IA r-tPA therapy; fourteen were men and seven were women. Median age was 63 (range 43-94). One patient on a continuous therapeutic heparin infusion received 0.45 mg/kg of IV r-tPA, while all others received 0.9 mg/kg. The mean dose of IA r-tPA was 20 mg (range 6-52 mg). Seventeen (81%) had angioplasty, mechanical clot disruption or retrieval performed in addition to IA r-tPA. Revascularization was achieved in 95%. Reocclusion occurred in one patient. Mean time to IV r-tPA and IA r-tPA administration was 161 minutes (range 55-865) and 512 minutes (range 244-2182), respectively. Twelve (57%) patients received IA r-tPA > 6 hours from symptom onset. Symptomatic intracerebral hemorrhage (ICH) occurred in 1 (5%) patient. Two patients died, one from ICH occurring 22 days after thrombolytic therapy, and one was made comfort measures after having progressive ischemic strokes from a cardioembolic source. The median NIHSS prior to endovascular therapy was 15 (range 3-27); median NIHSS on discharge was 7 (range 0-24). The median modified Rankin score (mRS) on discharge was 3 (range 0-6). Six (29%) patients had a discharge mRS of < 2. Three (14%) patients were discharged home, and twelve (57%) were discharged to acute rehabilitation.

Sequential IV and IA r-tPA administration for acute ischemic stroke apppears to be safe, associated with a high recanalization rate, and good clinical outcomes.

Financial Support: Dr. Jack Rose is on the Speakers Bureau for Boehringer Ingelheim and EKR Therapeutics.

Malignant hemispheric stroke is associated with a high mortality rate in untreated patients. Pooled results from three randomized trials comprising 93 patients aged less than 60 years showed a survival benefit with decompressive craniectomy. The benefit of craniectomy in patients older than 60 remains unknown though previous nonrandomized studies have suggested a worse outcome in older patients

We conducted a retrospective chart review looking at in-hospital survival, and mortality at the time of follow-up following decompressive hemicraniectomy during the study period of January 2003 to December 2009. Specifically we compared outcomes of patients with dominant (left) middle cerebral artery territory infarcts compared with nondominant (right) middle cerebral artery territory infarcts. We also compared outcomes in younger patients (aged 18-60 years) with those in older patients (aged > 60 years).

A total of 30 decompressed patients were included. The average age was 57.6 ± 12.6 years. There were 24 patients <60 years old (average age 47.7 ± 9.9) and 6 >60 years old (average age 70.0 ± 4.3). There were no significant differences in the presence of co-morbid conditions (hypertension, diabetes, atrial fibrillation, previous stroke), baseline NIHSS scores, or timing of surgery. The in-hospital mortality rate in patients <60 years old was 16.6% (4/24) compared with 0% (0/6) in older patients (p=0.028). At follow-up, the mortality rate was 20.8% (5/24) in younger patients and 50% (3/6) in older patients (p=0.18). The incidence of hemorrhagic transformation was higher in younger patients who were decompressed (7/24 versus 0/6, p=0.002).

While previous studies have suggested a worse outcome for older patients undergoing decompressive craniectomy, we observed that the in-hospital mortality rate of patients > 60 years was comparable with younger patients. The mortality rate at follow-up was 20.8% (5/24) in younger patients vs. 50% (3/6) in older patients (p=0.18). Patient selection may be a more important factor than chronological age in determining outcome after hemispheric stroke. More studies are needed to determine the role of decompressive craniectomy in elderly patients.

In 1900 Albert Terson described subhyloid hemorrhage in the setting of subarachnoid hemorrhage (SAH). [1] . The syndrome has come to include vitreous and retinal hemorrhages, with intraocular hemorrhages occurring in 10% or more of patients with SAH. While intraocular hemorrhages are more common in high grade SAH patients (Hunt and Hess IV and V) and are associated with poorer outcome [2] , the mechanism of the hemorrhage has remained unclear. A case is presented that supports one of the potential mechanisms for Terson hemorrhage.

Clinical evaluation and imaging studies of a patient with SAH were reviewed.

A 55 year old woman was unresponsive on admission (Hunt and Hess V). CT imaging demonstrated extensive SAH throughout the basal cisterns with thick clot consistent with criteria for Fisher Group 3 ( Fig. 1 ). Increased attenuation was seen extending out the optic nerve-sheath complex and into the posterior globes bilaterally (Fig. 2 ).

The most commonly proposed mechanism for subhyloid hemorrhage suggests that increased pressure in the subarachnoid space is transmitted down the optic sheath leading to retinal vein compression, venous hypertension and hemorrhage [3] . An alternate proposed mechanism suggests that intracranial pressure (ICP) is transmitted directly through venous channels rather than along the optic sheath [4] . Our case provides evidence of direct extension of subarachnoid blood into the globe as the mechanism in some cases of Terson Syndrome. The blood is clearly seen within the sheath and is continuous with the intraocular collection. The high ICP associated with aneurismal rupture likely forces subarachnoid blood along the nerve-sheath complexes into the globes. An extrusion mechanism is also supported by a recent case report detailing a patient with SAH studied with magnetic resonance imaging [5] . Other mechanisms for hemorrhage may also exist, as Terson Syndrome can be seen in cases of intracranial hypertension without SAH. 

The use of rtPA in acute ischemis stroke is associated with a 6-9% risk of intracranial hemorrhage (ICH). An rtPA induced coagulopathy has been proposed as a mechanism for post-thrombolytic ICH after acute ischemic stroke. We investigated an association between a rise in INR post-rtPA and hemorrhagic conversion of acute ischemic stroke.

We identified in our stroke database patients who received IV rtPA for ischemic stroke during the period November 2003 to November 2007 with a post rtPA INR performed within15hrs. Demographics, history of hypertension, DM, preadmission aspirin and statin use, NIHSS and pre rtPA INR were analyzed using t-test and fisher exact score were applicable. All subjects also had a repeat CT to rule out ICH. The clinical utility of INR in predicting ICH was tested using the receiver operator characteristic curve

Of our 24 patients there were 7 ICH and 17 non-ICH. There was no significant difference in demographics and baseline clinical characteristics. The mean post-rtPA INR in the ICH group (1.46 vs 1.18, p=0.003) and the change in INR for those with ICH was significantly greater than those without ICH (0.34 vs 0.08, p=0.003). Using the ROC curve, a post thrombolytic INR of 1.28 within 15hrs of rtPA predicted post thrombolytic ICH (AUC 0.83, CI 0.62-0.95, p=0.0014).

This study indicates that INR may remain elevated for up to 15 hours post IV rtPA. Furthermore, our data suggests that a persistently increased INR significantly increases the risk of post-thrombolytic ICH

While the problem of stroke (ischemic and hemorrhagic) in the patients with sickle cell disease (SCD) has been known for more than 75 years, preventive and treatment strategies are just now being tested. Furthermore, silent cerebral infarcts in children and young adults with SCD remain under recognized by physicians and families though they are the most common form of neurologic injury in children.

We describe a 21-year-old male with SCD, admitted to medical ICU with difficulties weaning from ventilator support secondary to poor mental status. A review of the literature was conducted to identify level of knowledge of stroke risk among patients, their caregivers and healthcare providers.

Brain MRI imaging revealed extensive acute ischemic strokes, severe leucoaraiosis and numerous small hemorrhagic infarcts. His medical history was pertinent for frequent episodes of acute chest syndrome, severe anemia and single exchange transfusion. Neither his caregiver nor the patients were aware that stroke is potential complication of the SCD. He never received TCD screening for first-time stroke. There is evidence in the literature, that there is a lack of knowledge in patients, their caregivers and physicians on the need for screening and treatment of stroke complications in SCD.

This illustrates that physicians taking care of patients with SCD need a preemptive investigative approach which can be important for providing preventive treatment protocol. The advent of TCD and brain MRI has greatly increased the ability to detect early CNS disease in patients with SCD. Stroke, which if diagnosed early, can be effectively managed with aggressive transfusion therapy. No data is available to show the cost-effectiveness of the neuroimaging in the evaluation of patients with SCD and suspected stroke, however, our case provides example of lack of awareness that SCD is associated with increased risk of stroke.

Use of to tissue plasminogen activator (rt-PA) in acute lacunar strokes is controversial. NINDS trial did not show any difference in outcome of lacunar strokes as compare to other stroke types. Some experts believe that rt-PA is not the right treatment choice for lacunar strokes as the pathogenic mechanism of stroke may not lead to reperfusion. We share our center's experience in the treatment of lacunar stroke patients.

We identified cases of ischemic stroke eligible for intravenous (IV) rt-PA by time criteria with NIHSS score <8 admitted to Saint Luke's Stroke Center from March, 2006 through April, 2010. Outcome was defined based on discharge disposition. Discharges to home with or without home health or inpatient rehabilitation were defined as good outcomes. Discharge to nursing homes or hospice homes or death were defined as poor outcomes. Univariate analysis including t-test, Chi-square, and Fisher Exact test was used when appropriate.

Twenty mild ischemic strokes (female to male: 3:1 and a mean age of 59±15 yrs ranging from 34 to 91 yrs) were included in this study. Forty percent patients were treated with IV rt-PA. There were no deaths in both treated and untreated group. Good outcome was seen in 95% patients who either went home or to an in-patient rehab facility. Only 1 patient went to nursing home that was not treated with rt-PA.

Small vessel lacunar strokes result in good outcome. Randomized controlled trials are needed to test the safety and efficacy of IV rt-PA in this patient population.

Financial Support: Dr. Rymer is on the speaker's bureau for Genentech, Inc.

Cerebral hyper perfusion syndrome (CHS) after carotid stenting is defined as nonthrombotic neurological deficit attributed to reperfusion of brain previously exposed to low cerebral blood flow. Incidence varies from 1-6.9% but associated mortality/morbidity is significant. Control of blood pressure during and after procedure is believed to be mainstay of pharmacological management but short and long term outcomes of lowering BP and degree to which BP can be lowered safely is not known. (1.29%)

We reviewed literature to assess if aggressive BP control correlates with incidence of CHS. Pubmed search for MeSH term "cerebral hyper perfusion and stent" revealed 63 articles about CHS in carotid stenting. Manual review excluded case reports and revealed seven publications addressing BP control.

Four studies with 3616 patients in total had 47 patients with CHS (1.29%) where no relationship between CHS and BP was noted. Three studies with total 1106 patients had 10 patients (0.9%) with CHS all of whom were noted to hypertensive compared to others. Only one study reported intra-procedural BP. None of these were randomized or controlled. Previous history of hypertension and impaired auto regulation was noted to be a risk factor in few studies. Relative risk for a patient having CHS in the BP controlled studies was 0.69.

Though reported studies were limited in design, the results elucidate the fact that patients who undergo carotid stenting do need closer monitoring to prevent and treat CHS but post procedural aggressive BP control may not be sufficient by itself. Some degree of high blood pressure over a primed brain may predispose to risk of hyper perfusion. Intra-procedural control of blood pressure, absence of auto regulation pre-procedurally and optimum values of blood pressure control in patients after carotid stenting need further investigation.

Poster 42

Patients with large hemispheric infarction (LHI) and cerebral edema are often treated with osmotic agents. While these agents reduce intracranial pressure (ICP), their ability to improve cerebral blood flow (CBF) and restore oxidative metabolism (CMRO 2 ) remains unclear. To more fully understand the cerebrovascular impact of these therapies, we compared the effects of mannitol and hypertonic saline (HS) by positron emission tomography (PET).

Nine patients with edema and midline shift after LHI were randomly assigned to receive an equi-osmolar bolus of 20% mannitol or 23.4% saline. 15 O-PET imaging was performed before and one hour after treatment to measure CBF, cerebral blood volume (CBV), oxygen extraction fraction (OEF) and CMRO 2 . Regions representing infarct, peri-infarct, and remaining ipsilateral and contralateral hemispheres were drawn on head CT scans and co-registered with PET images. Two-tailed paired t-tests were used for continuous variables and Pearson's coefficient to compare change in CBF and baseline mean arterial pressure (MAP).

Baseline CBF (ml/100g/min) in the infarct core, peri-infarct region, remaining ipsilateral hemisphere and contralateral hemisphere in the mannitol group was 5.0±3.9, 25.6±4.4, 35.6±8.6 and 45.5±2.2, and in the HS group was 8.3±9.8, 35.3±10.9, 38.2±15.1 and 35.2±12.4. There was a trend for CBF to rise in the peri-infarct region and contralateral hemisphere after mannitol to 30.2±5.6 (p=0.049) and 57.6±21.7 (p=0.098), that was not seen after HS. The change in CBF following osmotic agent in non-ischemic regions was strongly correlated with MAP (r=0.89, p=0.001). CBV, OEF and CMRO 2 did not change after either agent.

While neither osmotic agent improved cerebral metabolism, mannitol tended to improve CBF in non-ischemic tissue. This response was highly correlated to blood pressure, raising the possibility that at higher perfusion pressures osmotic agents may effectively raise CBF in patients with cerebral edema.

Cerebral hyper perfusion injury (CHI) is often seen with revascularization of large extra cranial vessels. With the advent of revascularization strategies of intracranial vessel stenosis, reperfusion injury has also been noted in this group. It is important to outline angiographic predictors and clinical parameters of CHI in patients that undergo intracranial revascularization. We describe four patients with chronic severe intracranial arterial stenosis due to atherosclerotic disease (3 patients) and vessel dissection (1 patient) and subsequent hyper perfusion injury after revascularization.

A retrospective review of the patients with CHI that underwent revascularization for intracranial stenosis was performed.

May 2010. Four patients (4/196; 2.06%) with subsequent CHI were identified. Three were men; mean age was 60.7 years (range 41-77). Arterial stenosis was greater than 90% for every lesion. Residual stenosis was between 20% and 26% for all lesions. All patients underwent stent placement with balloon angioplasty. No intra-operative complications occurred. Symptoms of CHI were noted within 24 hours of the procedure in all four patients. Three patients had intracranial hemorrhage as a complication. One patient had cerebral edema and seizure. Three of the four patients had systolic blood pressure exceeding 140 mmHg in the postoperative period of 24 hours. Permanent neurologic deficit occurred in 2 patients. There were no mortalities.

Cerebral hyper perfusion injury is an uncommon complication of stent placement and angioplasty for intracranial vessel stenosis. Patient with severe stenosis(>90%) are at increased risk. Patients should be monitored in the intensive care unit for 24 hours. In our experience symptoms occurred within first 24 hours.

Strict blood pressure management in the post-operative period may reduce the risk of CHI.

Poster 44

The activation of inflammatory cascades triggered by ischemic stroke may play a key role in the development of infections. We sought to define markers of immunodysfunction and identify early predictors of infection during the hospital course of stroke patients.

Patients admitted with an ischemic stroke within 24 hours of onset were prospectively enrolled. Temperature, white blood cell count (WBC), C-reactive protein (CRP), procalcitonin (PCT), interleukin-6 (IL-6), lipopolysaccharide-binding protein (LBP), and HLA-DR-expression on monocytes (HLADREM) were measured on day 1,3, and 5. The patients were continuously monitored for predefined infections. Outcome was assessed by modified Rankin Scale (mRS) on day 90.

Ninety four patients were included in the study. Infection was encountered in 39 patients (42%). Patients with infection were older (75.5±8.9 vs 69.2±1.6 years, P=0.006), had a higher National Institute of Health Stroke Scale Score (NIHSS, 13.3±0.9 vs 8.3±0.6, P<0.0001), more infarctions in the insula (19 vs 12, P=0.006), and higher stroke volumes on diffusion-weighted imaging on day 1 (45.6±11.9 vs 11.3±3,5 cm 3 , P=0.002). The maximum temperature on day 1 and 3, WBC, IL-6, LBP, and HLADREM on day 1, 3, and 5, CRP on day 3 and 5, and PCT on day 5 were significantly higher in patients with infection. Age (OR 1.1, 95% CI 1.01-1.1), NIHSS (OR 1.2, 95% CI 1.03-1.3), and maximum temperature on day 1 (OR 3.4, 95% CI 1.4-8.3) predicted the development of infections in the multivariate analysis. Infection was an independent predictor of poor outcome (mRS 3-6) at 90 days.

Higher temperature, WBC, IL-6, LBP, and HLADREM may be early markers of immunodysfunction in patients with ischemic stroke. Severe stroke, higher admission temperature, and advanced age were shown to be early predictors for the development of infections after stroke. 

Current guidelines for extubation do not adequately account for neurologic dysfunction in acute ischemic stroke (AIS). Glasgow Coma Scale (GCS) 8, presence of a cough and less frequent suctioning correlate with extubation success in brain-injured patients. We hypothesized that patients with middle cerebral artery (MCA) AIS may have more specific predictors of successful extubation.

Retrospective cohort study of patients discharged between January 2004 and December 2008 with MCA AIS requiring intubation. Patients were classified as having a successful extubation or failing extubation (defined as requiring reintubation within 48 hours or tracheostomy without extubation attempt). We excluded patients who self-extubated, were extubated terminally or died prior to an extubation attempt. Baseline clinical characteristics and outcomes were collected.

Of 71 patients with MCA AIS requiring intubation, 35 patients were successfully extubated and 10 failed. Patients were well-matched with respect to admission GCS and National Institutes of Health Stroke Scale. Those patients who were successfully extubated had a higher extubation GCS (median 10 [IQR 9-11] vs. 8.5 [7-10], p=0.01) and a higher score for best eye response of the GCS (median 4 [3] [4] vs. 2.5 [1] [2] [3] , p=0.004). All patients had a cough at time of attempted extubation. There was no difference between groups in the prevalence of less frequent suctioning (78% vs. 50%, 0.53). Those patients who successfully extubated had fewer total ventilator days (median 3 [2] [3] [4] [5] vs. 5 [3] [4] [5] [6] [7] [8] [9] , p=0.03). No differences were seen in median ventilator days until first extubation attempt (2 [1] [2] [3] [4] 

The total GCS and particularly the best eye item of the GCS may be important predictors for successful extubation in patients with MCA AIS.

Poster 46 

While the influence of hemispheric laterality on functional outcome after acute ischemic stroke (AIS) has been demonstrated, outcomes regarding respiratory support and liberation from the ventilator have not been adequately investigated. We hypothesized that right middle cerebral artery (MCA) AIS patients fail extubation more often than left MCA AIS patients because of underestimation of the National Institutes of Health Stroke Scale (NIHSS) and the additional symptom of neglect.

Retrospective cohort study of patients discharged between January 2004 and December 2008 with MCA AIS requiring intubation. Baseline clinical characteristics and outcomes were collected. Failed extubation was defined as the need for reintubation within 48 hours of an attempt at extubation or tracheostomy without extubation attempt.

Of 71 patients with MCA AIS requiring intubation, 37 had stroke on the left and 34 on the right. Patients were wellmatched for age, gender and medical history. Left MCA AIS patients had a higher median NIHSS (22 [IQR 20-26] vs. 16.5 [13-19], p<0.001) and a higher incidence of aphasia (97% vs. 12%, p<0.001). Right MCA AIS patients had a higher median admission Glasgow Coma Scale (GCS) score (14 [9-15] vs. 10 [9-11], p=0.004) and a higher incidence of neglect (71% vs. 14%, p<0.001) and eyelid opening apraxia (21% vs. 0%, p=0.004). Forty-four patients had an attempt at extubation, and 2 patients had a tracheostomy without an extubation attempt. Right MCA AIS patients failed extubation more often (30% vs. 16%) although this was not significant (p=0.30). When adjusted for age and admission, right MCA AIS patients were at a nonsignificant increased risk for failed extubation (OR 4.71 [CI 0.7-33], p=0.12).

Right MCA AIS patients have clinical differences referable to lateralization of stroke symptoms and, although not significant, may fail extubation more often.

The use of IV rt-PA is recommended up to 4.5 hours from stroke symptom onset. Brain diffusion-weighted imaging (DWI) is important for the evaluation of ischemic changes in patients with acute ischemic stroke but it is not available before thrombolysis for most patients to avoid treatment delay. This study aimed to evaluate the safety of IV rt-PA in patients with negative DWI obtained after thrombolysis.

We studied 48 patients treated with IV-tPA within 4.5 hours after stroke symptom onset. The patients were identified in our stroke center registry from January 2009 to February 2010. Information on patients' demographics, clinical characteristics, neuroimaging, treatment complications, and outcomes was obtained and analyzed. The study was approved by the IRB.

We found 19 patients (40%) who had negative DWI on follow-up imaging. In retrospect, 11 patients had TIA and 8 patients had stroke mimics, including Todd's paralysis, hemiplegic migraine, and somatoform disorder. We found significant differences between patients with positive and negative DWI in median age (61.5 years versus 51 years; P<0.05) and median admission NIHSS score (12.5 points versus 6 points; P<0.002). Among patients with positive DWI, two patients had symptomatic intracerebral hemorrhage. The only complication among patients with negative DWI was gingival bleeding in one patient.

Our results suggest that the administration of IV rt-PA to patients with negative DWI is safe. Further study is warranted.

Acute ischemic stroke (IS) due to basilar artery thrombosis (BAT) causes high mortality and severe disability. Early neurological assessment and timely initiation of thrombolysis might improve outcome. BAT is often difficult to diagnose due to wide spectrum of presentation, including jerky convulsion-like limb movements and decreased level of consciousness. Not infrequently, emergency physicians intubate some patients due to airway compromise, even before arrival of stroke neurologist. We aimed at evaluating the role of CT angiography (CTA) of brain and cervical arteries in early diagnosis of acute BAT and facilitating decision-making in initiating thrombolysis in intubated patients.

Consecutive acute IS patients presenting from 2007 to 2009 at our tertiary-care facility, with history of sudden deterioration in level of consciousness within 6hours of symptom-onset and intubated before assessment by neurologist were included. All patients underwent limited fast-track clinical evaluation, brain CT and CTA.

Thrombolytic therapy, mainly intravenous tissue plasminogen activator (IV-TPA), was administered to 161 (8.4%) of 1917 acute IS patients admitted during the study period. Acute BAT contributed 10.9% (208 cases) to our patient cohort. Five cases (3males, mean age 72yrs) of acute BAT with severe symptoms and airway compromise were intubated early, sedated and paralysed before any assessment by neurologist. CTA showed BAT in all. IV-TPA (0.9mg/Kg body-weight) was initiated at 236±40minutes in 4 patients while 1 received intra-arterial thrombolysis initiated at 13hours. There was no intracranial hemorrhage and mean length of hospital stay was 11.8days. Despite severe strokes at presentation, good functional recovery at 3months (modified Rankin scale-mRS 1) occurred in 2 patients; mRS 4 in one, and two died (including the patient treated intra-arterially).

In acute ischaemic stroke due to BAT, intubated before proper neurological assessment, CT angiography is helpful in confirming the diagnosis and facilitating therapeutic decision-making for initiating thrombolysis.

In patients with impaired conscious levels, the significance of periodic patterns on electroencephalogram (EEG) is controversial. We aimed to determine if attempting to terminate such patterns influences outcome.

In this retrospective study from 2007-2009, consecutive patients with impaired consciousness and periodic EEG patterns comprising either periodic lateralized epileptiform discharges (PLEDs), bilateral PLEDs (BiPLEDs), generalized periodic epileptiform discharges (GPEDs), triphasic waves or repetitive discharges were included. Those with clinical seizures within the preceding 24 hours of EEG were excluded. Age, underlying etiology, past history of structural central nervous system (CNS) disease such as stroke, tumor, cranial surgery, intracerebral hemorrhage, and GCS on admission, were recorded. Death upon discharge was used as a measure of poor outcome.

4246 patients underwent EEG during the study period. 64 met the inclusion criteria; median age was 67years (range 24-94) and 26 were males. 37 (57.8%) had triphasic waves, 17 (26.6%) PLEDs, 7 BiPLEDs, 2 GPEDs, 1 repetitive discharges.

18 patients received abortive therapy and of these, 35.3% died on discharge, compared to 27.3% of those who were not treated (p=0.538). Neither etiology nor type of periodic pattern affected outcome. Those without a history of CNS abnormalities were at higher risk of death on discharge compared to those who did (40.6% vs 17.2%, p = 0.046). Low GCS at admission was an independent risk factor for death on discharge (p=0.03, OR= 0.80, 95% CI 0.66 to 0.98).

Treatment of periodic EEG patterns does not independently improve the clinical outcome of patients with impaired conscious levels. Their occurrence in patients without prior CNS disease is associated with a poorer outcome and may be a marker of severe acute neuronal damage. Lower GCS scores at admission are independent predictors of poor prognosis.

Spontaneous intracranial hemorrhage (S-ICH) often leads to significant morbidity and mortality. Clinical scoring systems can predict the mortality and functional outcome after ICH (FUNC score). FOUR score is a recently developed coma scale that incorporates brainstem reflexes. Our hypothesis is that the combination of FUNC and FOUR scoring system will be a better prognostication indicator for S-ICH patients.

A retrospective chart review of all S-ICH patients (greater than 18 years of age) from January 2008 to May 2010 was conducted. ICH volume was calculated using the ABC/2 formula. FUNC (<5 or were calculated and dichotomized from admission data. Statistical analyses included chi square test and relative risk assessment.

Total number of patients was 92. Mean age of presentation was 64.6±15.6 yrs (Range: 27-95 yrs). Male to female ratio was 1.6:1. Most common risk factors were hypertension 81/92 (88%) and coagulopathy 26/92 (28%). FOUR score (<13), FUNC score (<5), ICH volume (> 60ml) and coagulopathy were independent predictors of mortality (P <0.05) ( Table) .

Combination of the FUNC and the FOUR scoring system is a better prognostication indicator for S-ICH patients.

Poster 52

Didem Aksoy 1 , Chitra Venkatasubramanian 1 , Roland Bammer

Blood-brain barrier (BBB) injury after spontaneous intracerebral hemorrhage (ICH) may contribute to perihematomal edema (PHE) formation. We determined the severity of BBB injury (i.e. permeability) and relationship with PHE severity (i.e. diffusivity) using dynamic contrast-enhanced (DCE) MRI and apparent diffusion co-efficient (ADC) maps.

Eight prospectively enrolled patients from the Diagnostic Accuracy of MRI in Spontaneous intra-cerebral Hemorrhage (DASH) study had DCE MRI using 0.1 mM/kg Gd-DPTA on a 1.5T GE scanner one week after ICH onset. Regions of interests (ROI) covering the perihematomal rim were placed on the permeability map ( Figure 1 ), and control ROIs were placed on homologous locations in the contralateral hemisphere. CineTool software (GE Healthcare, Waukesha, WI) was used to estimate BBB permeability measured by the forward leakage rate (K trans ). Wilcoxon signed rank test was used to analyze K trans from the lesion and control ROIs. ADC maps were co-registered with FLAIR images using MIPAV software. The perihematomal rim was outlined on FLAIR and mean ADC values for these ROIs were calculated. Correlation between ADC values and BBB permeability values were estimated using Spearman Rank test.

At one week, BBB permeability in the perihematomal rim was significantly increased compared to control ROI (0.044 min -1 Vs 0.004 min -1 ). ADC values in the perihematomal rim were increased indicating vasogenic edema (1001 (877-1143) x 10 -6 mm 2 /sec) and were strongly correlated with the corresponding BBB leakage rate (R=0.58, p<0.0001) ( Figure 2 ). Also, for each patient, mean ADC values in the PHE correlated with mean BBB leakage (R=0.74, p=0.041). 

Perihematomal BBB permeability is substantially increased at one week after ICH onset and strongly correlates with the severity of vasogenic perihematomal edema. Therapies aimed at limiting BBB injury might attenuate the formation and evolution of perihematomal vasogenic edema.

Financial Support: Sandeep Gupta is an employee of applied science laboratory of General Electric Systems.

Neurocrit Care (2010) 13:S1-S231 S87

Poster 53

Anne-Sophie BERAUD 2 , Ryan SNIDER

Unlike in subarachnoid hemorrhage, changes in cardiac function in the acute phase of spontaneous intracerebral hemorrhage (ICH) are ill defined. We sought to characterize left ventricular systolic and diastolic function in the acute and chronic phase of ICH.

Consecutive patients admitted with non-traumatic ICH were prospectively recruited. A complete bedside transthoracic echocardiography (TTE) was obtained within four days of symptom onset. A repeat echocardiogram was pursued in patients with acute left ventricular systolic or diastolic dysfunction. Left ventricular systolic and diastolic functions were defined according to the recommendations of the American Society of Echocardiography.

One hundred and forty seven (147) 

Diastolic dysfunction is common in patients with acute ICH and most likely a result of long-standing hypertension potentially worsened by acutely elevated blood pressures. In contrast, left ventricular systolic dysfunction is rare. Almost half of patients with acute ICH and left ventricular systolic and/or diastolic dysfunction experience improvement over time. These findings may have therapeutic implications.

Warfarin-associated intracerebral hemorrhage (wICH) has over 50% mortality and is associated with secondary thromboembolic complications (TE). Prothrombin complex concentrate (PCC) reverses warfarin-induced coagulopathy but may be associated increased TE.

All wICH patients with INR>1.4 are treated with intravenous vitamin K and fresh frozen plasma (FFP). Starting 2008, all wICH with INR>1.5 received adjunct PCC. We identified consecutive PCC-treated wICH patients and examined their incidence of EKG changes, symptomatic myocardial infarction (MI), and venous thrombosis following PCC use. We compared TE rates in PCC-treated subjects to control wICH subjects treated with FFP and vitamin K alone.

We analyzed 27 PCC-treated and 34 control wICH subjects. 

PCC use in wICH is not associated with additional TE risk compare with vitamin K and FFP alone. PCC is associated with more rapid INR correction but not wICH mortality in this study population. A prospective randomized study is necessary to determine whether PCC use reduces morbidity and improves outcome in wICH patients.

Raised intracranial pressure due to cerebral edema leading to herniation is the main cause of death in patients with intracranial hemorrhage (ICH). Few case reports have suggested higher mortality a subset of patients who are on hemodialysis and are admitted with ICH . We hypothesized that the in hospital mortality of patients with intracranial hemorrhage (ICH) was higher in patients undergoing dialysis.

A retrospective analysis utilizing the Nationwide Inpatient Sample was performed. Using the ICD-9-CM codes all adult records (age 18 or more) with a discharge diagnosis of ICH (431) were identified from 2000 to 2007. Dialysis (39.95, V56.x, 585.6) and acute renal failure (584.1-9) were similarly identified. New dialysis was defined as patients with ICD code of ARF with a dialysis or a procedure code of venous catheterization for renal dialysis (38.95). Multivariate logistic regression was performed using survey commands in STATA. The regression was adjusted for age, sex, race and other co-morbidities. The primary outcome studied was all cause in-hospital mortality.

A total of 668,868 estimated discharges with ICH were studied during years 2000 to 2007. Out of these, 16,146 (2.5%) were on dialysis.

In ICH patients not on dialysis, the all cause in-hospital mortality was 29.5% while it was 37.5% in those on dialysis.

After adjusting for age, sex, race and other co-morbidities, the in-hospital mortality was found to be significantly higher in patients with dialysis (OR 1.24; 95%CI 1.13-1.36).

Using nationally representative data, this observational study reveals increased in hospital mortality of patients with ICH who are on hemodialysis.

Little data exists on influence of gender on outcome after ICH. The objective of this study was to study whether a relationship exists between gender and functional outcome/mortality after Intracerebral Hemorrhage (ICH).

This was a prospective consecutive cohort study of 245 adult patients presenting to the Emergency Department with spontaneous intracerebral hemorrhage (ICH), between January 2006 -December 2008. Patients with an extradural hemorrhage, subarachnoid hemorrhage, recurrence of hemorrhage were excluded. Statistical analysis performed using JMP 8.0.

Our cohort was 51.1% female and as expected significantly older than the males There were no differences in severity of stroke on presentation(NIHSS), volume of ICH or difference in the proportion with intra-ventricular extension ( 

Female gender appears to be associated with increased likelihood of poor functional outcome, with a higher likelihood of not being independent in activities of daily living after ICH.. This may be useful 1) as a hypothesis generating study looking at underlying pathophysiologic mechanisms in women and 2) in guiding organization of support systems and discharge planning.

Outcomes in spontaneous intracerebral hemorrhage (ICH) are dramatically worse in the setting of anticoagulation (AC). We characterized the contribution of coagulopathy to ICH mortality at our center, and evaluated practice patterns and success at correction of coagulopathy.

An IRB approved, retrospective review of all ICH admission from 2007-2009. We compared demographics, medical history, clinical course, and radiographic and laboratory data of noncoagulopathic patients to those receiving warfarin, aspirin, other antiplatelet therapy, or with thrombocytopenia or creatinine > 1.8. Differences in continuous or categorical variables or proportions were estimated by t-test or Mann-Whitney U test as appropriate.

Of 288 

More than half of patients admitted with ICH had preexisting coagulopathy. Coagulopathic patients presented with greater hemorrhage severity, were more likely to suffer hematoma expansion, and were more likely to die. Rapid reversal of coagulopathy was rarely achieved within the time windows most frequently associated with hematoma expansion, emphasizing the need to explore protocol-driven therapy and the use of novel agents.

AHA guidelines for management of ICH recommend monitoring ICP in patients with clinical deterioration or suspected elevated ICP while reducing systemic hypertension to maintain CPP >60mmHg. ACCELERATE evaluated aggressive blood pressure (BP) management in patients with ICH using clevidipine, an IV antihypertensive with a t1/2 of about 1min that directly reduces systemic vascular resistance and is titratable to effect. CPP and ICP were assessed in a subset of ACCELERATE patients.

Patients with acute ICH presenting within 12 hours of symptoms and with SBP>160mmHg were prospectively enrolled, including a subset of patients requiring ICP monitoring (n=7). Patients were treated with IV clevidipine, started at 2.0mg/h and titrated every 90 seconds until SBP between 140 to 160mmHg. Initial CPP and ICP results are presented (n=6; one patient was excluded from this analysis due to prior treatment with IV mannitol).

Six evaluable ICP-monitored patients (mean age 60 years; 5 men) received clevidipine for at least 3 hours. At baseline, median GCS was 8.0, median NIHSS 20, mean hematoma volume 36.3mL and mean SBP 170mmHg.

Mean SBP at 30 min was 145mmHg. Median time to target SBP was 3.0 min. Mean time-to-infusion from symptom onset was 9.1 hours.

Pooled 24-hour monitoring showed mean CPP ranged from 82.9mmHg to 101.0mmHg and mean ICP ranged from 7.3 to 18.0mmHg. CPP values were consistently >60mmHg with no apparent trends. No meaningful increases or other changes in ICP were observed during rapid BP reduction, 24 hours of clevidipine administration or through 3 hours post-study drug.

During aggressive BP management with clevidipine in ACCELERATE patients requiring ICP monitoring, CPP was maintained in an optimal range. In spite of clevidipine being a potent vasodilator, we did not observe any meaningful increases or other changes in ICP. 

Early and aggressive lowering of systolic blood pressure (SBP) in intracerebral hemorrhage (ICH) is a current research focus; recent studies support its safety [1] and ability to attenuate hematoma expansion. [2] Clevidipine is an IV antihypertensive with rapid onset and rapid offset that reduces systemic vascular resistance without affecting venous dilation, and is titratable to precise BP effect. Hematoma volume was evaluated in ACCELERATE patients.

Patients with ICH presenting within 12 hours of symptoms and with SBP>160mmHg were prospectively enrolled and treated with open-label IV clevidipine, started at 2.0mg/h and titrated every 90sec until SBP then titrated to maintain SBP between 140 to160mmHg. Initial study results include intracerebral hematoma volume assessed by blinded readings from baseline and 24-hour CT scans.

Thirty-five patients (mean age 63.5 years; 27 males) with baseline median Glasgow Coma Scale (GCS) score 14, median NIH Stroke Scale (NIHSS) score 12, mean SBP 186mmHg, and mean DBP 85mmHg received clevidipine. Mean time-to-infusion from symptom onset was 5.5hours. Mean average clevidipine infusion rate for the first 30min of treatment was 7.9mg/h. Mean on-drug clevidipine infusion duration was 28.1 hours.

In 32 patients meeting all study eligibility criteria, median time-to-target-SBP ( patients achieved target SBP within 30min, and 31/32 (96.9%) achieved target SBP without receiving additional or alternative IV antihypertensives. Thirty patients had evaluable serial CT scans, overall showing minimal hematoma volume change (-0.2mL; Table) . Pyrexia (n=7) and headache (n=5) were the most commonly reported adverse events (AEs). Hypotension was reported as a mild or moderate AE in 3 patients and resolved with dose reduction or drug discontinuation. 

Patients with acute ICH overall showed minimal intracerebral hematoma volume change after SBP reduction with clevidipine. Clevidipine was effective and safe for rapid BP reduction in this cohort of critically ill patients.

Poster 61

Introduction Intraventricular hemorrhage (IVH) is a neurological emergency with mortality reported as high as 80-100%. The primary goals of treatment are controlling intracranial pressure (ICP), cerebral spinal fluid (CSF) diversion, and removing blood from the ventricles commonly via an external ventricular drain (EVD). A new intervention, EVD plus thrombolytic, has shown benefit in some patients by expediting the removal of IVH from the ventricles. The purpose of the study is to evaluate the safety and efficacy of an EVD plus recombinant tissue plasminogen activator (rt-PA) for the treatment of IVH.

A retrospective review of the electronic medical record was conducted for patients admitted to Mayo Clinic and St. Luke's hospitals in Jacksonville, FL who received intraventricular rt-PA for IVH from January 1, 2004 to September 30, 2009. Outcome was assessed by Glasgow Outcome Scale (GOS) and safety by the presence of intracranial bleeding and infection. IVH volumetrics was assessed by the Le Roux score (0-16).

Twenty-seven patients received rt-PA for IVH. The median dose was 2 mg (range 0.3-8) and a median of 2 doses (range 1-17) were given for treatment. Two-thirds of patients had a favorable outcome (GOS 3-5). IVH volumetrics of hematoma burden decreased by a median Le Roux score of 10 (range 3-16) prior to rt-PA to 4 (range 0-16) 24 hours after rt-PA. There were no central nervous system bacterial infections out of 101 intraventricular injections. Incidence of EVD catheter tract hemorrhage after rt-PA was 46.7%, but half of these EVDs appeared to have some fenestrations in the parenchyma.

An EVD plus intraventricular rt-PA appears relatively safe when correctly placed within the ventricle and expedites IVH removal. The most common complication was EVD tract hemorrhage. Larger prospective studies should address optimal rt-PA dose and frequency, and have larger enrolment for safety analysis. 

Mean age was 56.5 ± 16; 8 of 17 were men. MMM was initiated within 24 hours from symptom onset. Etiologies of ICH were hypertensive hemorrhage in 11 (65%), vascular malformation 2 (12%), and Moyamoya disease 1 (6%). Median GSC on admission was 4 (range 3 -12), Location of hematoma was deep in 14 patients,and lobar in 3 patients. Eleven patients survived to discharge. Daily averaged PRx plots peaked on ICH days 3 and 4 and progressively normalized thereafter. Daily mean PRx values were significantly higher among those who died than in survivors throughout the entire 10 days monitoring period (P< 0.05). Mean CPP for non-survivors were consistently higher than their CPPopt values while mean CPP for survivors were close to or lower than their CPPopt.

Abnormal pressure autoregulation, as reflected by higher PRx values, is associated with mortality after ICH. Daily PRx monitoring may be useful determining prognosis in patients with ICH. 

A composite validated scale called the ICH score by Hemphill et al incorporates multiple variables to estimate 30-day mortality. A similar clinical grading scale in warfarin-related intracerebral hemorrhage (WICH) patients in predicting 30-day mortality would be of value for physicians caring for these patients. We hypothesized that application of the ICH score would show similar increasing mortality; with increasing score and possibly in comparison to patients not on warfarin.

We retrospectively reviewed from December 2002 to May 2010 all cases of WICH at the Mayo Clinics in Rochester, MN, Phoenix, AZ, and Jacksonville, FL. The ICH score was abstracted from the patient's medical record. ICH volume was obtained on CT scans at presentation. One-month mortality was obtained from the medical record from the date of initial presentation.

Over an 8-year period, 146 patients with WICH were identified. A trend in increasing 30-day mortality was observed with increasing WICH score: [WICH score=0 (15.6% mortality), WICH=1 (28.6% mortality), WICH=2 (43.4% mortality), WICH=3 (52.9% mortality), WICH=4 (100% mortality), WICH=5 (100% mortality)]. Also, in WICH patients, the 'WICH Score' 30-day mortality was higher at lower relative ICH scores, suggesting that warfarin anticoagulation increases mortality independent of other clinical or radiographic variables. INR values at cutoffs (>3 or <3) were not statistically significant at a high p value (p>.30).

WICH patients have increasing mortality with increasing ICH score, but not associated with initial INR. Whether warfarin leads to larger ICH or indicates more comorbid conditions that increase the risk of death during hospitalization requires further study.

Platelet transfusion has been proposed and as a potential therapy for acute intracerebral hemorrhage (ICH), but there are few data to verify a biological effect. We tested the hypothesis that platelet transfusion improves measured platelet activity.

Forty-five patients with ICH who received one of 48 platelet transfusions. We prospectively measured platelet activity (VerifyNow-ASA, Accumetrics, CA and PFA-100, Siemens AG, Germany) before and after platelets were administered and recorded acute adverse events (hypotension, new fever, rash, radiographic pulmonary edema or respiratory distress).

Platelet transfusion increased platelet activity from a baseline of 473 ± 49 ARU by 89 ± 105 ARU (P<0.001) on the VerifyNow-ASA assay. Platelet transfusion reduced the PFA-100 (epinephrine agonist) closure time from 224 ± 70 sec. by 68 ± 79 sec. (P=0.001). Among patients who were known to take clopidogrel, P2Y12 inhibition decreased from 31 ± 21% inhibition to 9 ± 9% inhibition (P=0.01). Eight (17%) had any adverse event, hypotension in five and new fever in three. Two patients had paradoxical worsening of platelet activity with platelet transfusion.

While an imperfect agent, platelet transfusion usually increases platelet activity. Other platelet activating agents (desmopressin, Factor VII, etc.), and platelet activating therapy vs. placebo, should be tested.

The size of perihematomal edema (PHE) surrounding intracerebral hemorrhage (ICH) is important as it may lead to mass effect and increased intracranial pressure. Hematoma location may affect PHE volume because of structural differences in the brain tissue surrounding the hematoma. We aimed to determine the effect of hematoma location on PHE volume.

Consecutive patients presenting with a spontaneous supratentorial intraparenchymal hemorrhage were prospectively enrolled. Patients had an initial MRI and up to 3 additional weekly MRIs when feasible. ICH volume (HV) and PHE volume were determined using the FLAIR sequence of the first MRI while assuming a constant HV on subsequent MRIs. MRIs were grouped into four time-intervals: (0-48)±8 hours, and (3-7) and 6, large lobar ICH had 15, 10, 6 and 6 and large deep ICH had 8, 6, 9 and 5 MRIs measured in the four timeintervals, respectively. Lobar hematomas were a mean of 10.4cc larger than deep hematomas (p=0.048). ANOVA showed that PHE volume was associated with hematoma volume (p<0.001) and time from symptom onset (p=0.001) as well as their interaction (p=0.01), but not with location (p=0.767).

PHE volume correlates with hematoma volume and time from symptom onset, but does not appear to be affected by hematoma location in the supratentorial compartment.

Studies suggest neurocritical care (NCC) improves intracerebral hemorrhage (ICH) outcomes [1, 2] , but it is unclear what accounts for this effect. We hypothesized that hiring a neurointensivist (NI) would improve mortality and increase appropriate use of early CMO orders [3] .

We retrospectively reviewed 289 ICH admissions during three years to a 604 bed tertiary referral center with 53 critical care beds, but no dedicated neurological ICU (NICU). A NI was hired at the midpoint. Severity of illness in the two epochs was well matched according to ICH score and most demographics. Other clinical data included coagulopathy, Charlson Comorbidity Index, transfer status, and time of hospital admission. These factors were evaluated in a stepwise multiple logistic regression model. We used ROC curves to compare appropriate use of early comfort measures -often in patients with very high and rarely in those with low ICH scores.

There was no statistically significant difference in in-hospital mortality between the two epochs [29% vs 29%, p=0.97]. In a categorical data analysis, factors associated with mortality were ICH score (OR 4.74, 95%CI 3.08-7.29), Charlson Comorbidity Index (OR 1.34, 95%CI 1.04-1.73), and transfer status (OR 0.33, 95%CI 0.14-.0.74), but not coagulopathy (OR 1.76, 95%CI 0.77-4.51), admission period, or time of hospital admission. Appropriate matching of early CMO status to ICH score did not change between admission periods.

These findings are at odds with prior investigations suggesting that NCC improves ICH outcomes. This may have been due to lack of a NICU or specialized nursing, the expected start-up time required to establish a NCC program, to lack of NI access to most ICH patients, or other factors. Nonetheless, our findings indicate that hiring a NI without development of essential NCC infrastructure may not improve outcomes.

More than one third of patients with stroke admitted to a rehabilitation unit have a PEG. Prolonged dysphagia is an independent predictor of mortality in the acute stroke setting. There is limited data on the predictors of prolonged dysphagia, and therefore the need for placing a gastrostomy for enteral feeding in patients with ICH. The aim of our study is to define such predictors

We reviewed medical records of patients seen in our institution between 2005 and 2009, using ICD-9 codes for ICH, for admission clinical and radiological parameters and the placement of PEG tube. Statistical analysis used T-test for continuous, and fisher's exact test for categorical variables

Of 346 patients identified, 124 were excluded for incorrect coding, 55 for incomplete admission data, 51 for comfort care measures status, and 29 for early death. Twenty-five of the remaining 87 patients had prolonged dysphagia and required a PEG. As compared to those without prolonged dysphagia, patients with prolonged dysphagia had a larger mean hematoma volume in ml (15.5 ± 14.6 vs. 41.4 ± 25.0, p < 0.001), and were more likely to have capsuloganglionic hematoma (27.4% vs. 56%, p = 0.010), intraventricular hemorrhage (22.5% vs. 56%, p = 0.005), hydrocephalus (3.2% vs. 52%, p<0.001), and midline shift (6.5% vs. 48%, p < 0.001). Patients with an admission GCS < 8 were significantly more likely to need PEG (1.6% vs. 44%, p < 0.001). Ninety four percent of intubated patients required PEG

We could potentially predict need for PEG early in the course of ICH based on admission CT findings and GCS, both of which are prognostic factors in ICH. Early PEG provide a more definitive mean for feeding, and would decrease length of hospital stay by allowing timely discharge to rehabilitation facilities or nursing homes 

Transfer of critically-ill patients from external Emergency Department (OSH-ED) has the potential of delaying the admission to the Intensive Care Unit (ICU). The effect of OSH-ED transfer on hospital outcomes of ICH patients has not been studied.

We designed a retrospective cohort study using a prospectively compiled and maintained registry (Cerner Project IMPACT). ICH patients admitted to our ICU from our ED and OSH-ED within 24 hrs of stroke between 2003-2008 were selected for the analysis. Data collected included demographics, admission physiologic variables, Glasgow Coma Scale (GCS), APACHE-II, scores; and total ICU and hospital length of stay (LOS). Primary outcome was functional status at hospital discharge and secondary outcomes were ICU and hospital LOS. Poor outcome was defined as death or severe disability at hospital discharge. To assess for the impact of OSH-ED transfer on primary and secondary outcomes, demographic and admission clinical variables were used to construct logistic regression models using the outcome measure as a dependent variable.

A total of 296 patients were selected. The mean age was 65±14 years, of which 47% were male, 63% were white, and 66% were transferred from OSH-ED. The median hospital LOS was 6 days (Interquartile range [IQR]=4-11) and median ICU-LOS was 2 days (IQR=1-4). Overall hospital mortality was 37%. Transfer from OSH-ED was associated with a 75% probability of death or poor outcome at hospital discharge. Multivariate regression analysis showed that APACHE-II (OR, 1.2; 95% CI; 1.1-1.3), GCS <12 (OR, 2.8; 95% CI; 1.8-4.1), and OSH-ED transfer (OR, 1.7; 95% CI; 1.1-2.5) were independently associated with poor outcome. OSH-ED was not significantly associated with secondary outcome measures.

This data suggests that in ICH patients, OSH-ED transfer is independently associated with poor outcome at hospital discharge. Further research is needed as to identify the potential causes for this effect. 

In phase 1, the agreement of emergency POC and CL INR measurements was determined. In phase 2, stepwise OAC reversal was performed with PCC using a predetermined dosing schedule. Concordance of POC and CL INR measurements during reversal and time gain due to POC were determined.

In phase 1 (n=165), Bland-Altman analysis showed close agreement between POC and CL (mean INR deviation 0.04). In phase 2 (n=26), POC caused a median initial net time gain of 24min for the start of treatment with PCC. Median time for POC-documented complete OAC reversal was 28min compared to 120min for CL. Bland-Altman analysis between POC and CL revealed a mean INR deviation of 0.13 during stepwise PCC administration. POC tended to slightly overestimate the INR especially with higher INR levels. Remarkably, POC-guided reversal led to a median reduction of 30.5% of PCC dose compared to the a priori dose calculation. Hematomas enlarged in 20% of patients.

POC INR monitoring is a fast, effective, and economic means of PCC dose-titration in OAC-ICH. Larger studies examining the clinical efficacy of this procedure are warranted.

Hypoalbuminemia is associated with worse nutritional status and poorer outcomes in medical and surgical patients as well as those with ischemic stroke and severe head injury (1) (2) (3) (4) . In contrast, others have found that hypoalbuminemia is a relatively poor predictor of hospital mortality in medical and surgical patients (5) . We tested the hypothesis that albumin is associated with morbidity and mortality in patients with ICH, a previously unstudied population.

We prospectively identified patients with ICH admitted from 2006 to 2010 and recorded each patient's ICH Score (6), if the patient developed pneumonia (by CDC criteria), and modified Rankin scores (mRS) at 14, 28 and 90 days with a validated questionnaire. Albumin levels were electronically retrieved. Multivariate logistic regression models were used to analyze associations between albumin levels and outcomes.

We identified 117 patients (50.4% male, average age 64.9±14.4 years) whose initial and mean albumin levels were 3.04±0.82 and 3.05±0.67 g/dL, respectively (mean±SD). Twenty-one patients developed pneumonia, the risk for which was inversely associated with mean albumin levels (OR 0.319 per g/dL, 95% CI 0.144-0.707, P=0.005) but not associated with ICH scores (P=0.348). The risk of death or dependency (mRS 4-6) at 3 months was inversely associated with mean albumin levels (OR 0.167 per g/dL, 95% CI 0.060-0.462, P=0.001) and directly associated with ICH score (OR 2.413 per point, 95% CI 1.475-3.947, P < 0.001).

In patients with ICH, lower mean albumin levels are associated with an increased risk of developing pneumonia and 90-day death or dependency, independently of the ICH score. Poor nutrition, malnutrition during ICU care, or albumin loss are potential mechanisms for increased infection and worse outcomes. Future studies will need to investigate if nutritional intervention or albumin supplementation can ameliorate this effect.

Perihematomal edema (PHE) after intracerebral hemorrhage (ICH) may influence outcome. We compared the relationship between hematoma volume (Hv) and PHE volume (Ev) in large and small hematomas over time.

Methods 119 consecutive patients with spontaneous ICH were prospectively enrolled. Patients had a baseline MRI and two additional weekly MRI's when feasible. Hv and Ev were determined using the FLAIR sequence. Patients were grouped by timing of the MRI (0-2, 3-7 and 8-14 days). Timing of maximum Ev and the effect of time on the relationship between Ev and Hv was compared between patients with small ( 30) and large (>30cc) hematomas. Data was analyzed using Mann-Whitney U test, t test and linear regression.

119 patients with 202 MRI's were included. At baseline, median Hv was 10.3cc (3.8-29.7) and Ev 18.9cc (9.9-45.7). Ev correlated with Hv at all time intervals (R 2 =0.75, 0.73 and 0.73). Ev in patients with small hematomas (N=89) at 3-7 days was significantly higher than at 0-2 days (p<0.001) but was not significantly higher compared to 8-14 days (p=0.71). In contrast, for large hematomas, Ev at 8-14 days was larger than at 0-2 days (p= 0.01) and had a strong trend towards being larger compared to 3-7 days (p=0.06). Furthermore, for small hematomas the correlation between Hv and Ev decreased over time (R 2 =0.67, 0.58 and 0.46) whereas for large hematomas it increased (R 2 =0.21, 0.28 and 0.37).

Edema volume in patients with intracerebral hemorrhage reaches its maximum towards the end of the first week if the hematoma is small and in the second week if the hematoma is large. Furthermore, the correlation between hematoma volume and edema volume decreases over time for small hematomas and increases for large hematomas. Introduction VTE remains a significant cause of morbidity and mortality in critically ill patients; however clinicians are often reluctant to initiate early pharmacologic prophylaxis in patients with ICH due to the fear of BE. We sought to describe the BE and associated risk factors related to the early use of pharmacologic prophylaxis in ICH patients.

ICH Patients admitted to the NSICU over a 4-year period prescribed pharmacologic prophylaxis were included. A major BE was defined as clinically overt blood with a decrease in Hgb by >2gm/dL and/or a transfusion of 2U PRBC or increase in ICH size; minor BE were all other BE. Since the introduction of new agents such as levetiracetam, little data is available to guide clinicians in the treatment of SE & marked variation in clinical approach is likely. This study aims to describe the common characteristics of patients who present with SE and their outcomes.

Retrospective, multicenter, observational study at 20 American academic medical centers. Sites reviewed the last 10 to 20 cases of SE at their institution. Data was submitted via secured on-line platform and analyzed for trends in initial management of SE. Demographic data, drug selection and timing, and other pertinent data were recorded.

Data collection and analysis is on-going. Fifteen of seventeen patients (88%) had a history of seizures prior to presenting with SE. Other attributed causes include traumatic brain injury, ischemic stroke, drug withdrawal, and infection (1 each). Patients were primarily male (65%), with a mean age of 41 years (SD 15) and a mean BMI of 29kg/m 2 (SD 6). Twelve patients presented with generalized convulsive SE (71%); 3 with complex partial SE, 1 with simple partial SE, and 1 with non-convulsive SE. The median time to SE resolution was 1 day (ranging as high as 22 days). Two patients died (without seizure cessation), one was ventilator dependent, while the rest were discharged from the hospital (3 to a long-term care facility, 1 to an acute rehabilitation facility, and 10 to home).

Generalized convulsive SE in male patients with a seizure disorder appears to be the most common patient-type presenting with SE. SE cessation may occur quickly or may be delayed for a prolonged period of time.

Introduction Phenobarbital (Pb) is a potent anticonvulsant drug (ACD) for Status Epilepticus (SE). Common side effects include sedation and respiratory depression; tolerance may develop despite high serum levels. High dose Pb (HDPb) has been used to control refractory SE (RSE), since tolerance occurs and there is less hemodynamic instability than with other ACDs. We report the use of HDPb following control of RSE with high dose suppressive therapy (HDST).

Twelve children received HDPb, defined as a Pb serum level >60ug/ml. RSE was diagnosed when SE continued despite treatment with two ACDs. We typically use HDPb while weaning HDST (other barbiturates) to prevent seizure recurrence and give additional Pb doses when seizures occur.

All children (M/F, 6/6 mean age of 8.5±5.7 years) underwent continuous EEG monitoring for RSE. Eleven children had acute symptomatic (AS) RSE from presumed encephalitis; one had remote symptomatic (RS) RSE from cerebral dysgenesis. All received HDST by continuous infusion, using midazolam first, followed by a barbiturate (pentobarbital or thiopental), and EEG demonstrated a burst-suppression pattern during HDST (thiopental in 2 or pentobarbital in 10).

The mean serum Pb level was 161±37.6ug/ml (range: 80 to 195). Hypotension occurred during HDST, but not specifically with HDPb. Seizures recurred during HDST weaning process in 8 and repeat Pb boluses were used. Except for one, barbiturates were tapered within 11-94 days after Pb initiation and extubation was possible at levels ranging from 60-150 ug/ml. One child died of sepsis while on only HDPb and the child with RS RSE had care withdrawn. Ten went to rehabilitation with various degrees of disability.

HDPb therapy is useful to maintain seizure control in RSE when tapering HDST. Tolerance develops to HDPb, allowing the tapering of HDST agents that require intubation and the need for continuing ICU care.

In the PICU, children with acute neurological disorders are commonly placed on an antiepileptic medication (AED) for seizure prophylaxis. (Fos)phenytoin has traditionally been used for this purpose but levetiracetam is increasingly being used as an alternative.

Data were prospectively collected from October 2008 to May 2010 on 53 children who received seizure prophylaxis while in the PICU. Diagnoses for which seizure prophylaxis was indicated included severe traumatic brain injury, intracranial hemorrhage and supratentorial neurosurgery. Patients with a prior history of seizure(s) were excluded as were patients with seizure on presentation. All AEDs were chosen according to physician discretion. Levetiracetam was administered as the sole agent in 12 children. Twenty eight children received (fos)phenytoin only. Thirteen patients received levetiracetam and (fos)phenytoin over their hospital course. Twelve patients were initially given (fos)phenytoin and switched to levetiracetam at the discretion of the treating physician and in 1 patient the reverse was true. Twenty had continuous EEG monitoring during their PICU course of which 11 received (fos)phenytoin only, 1 levetiracetam only and 8 both.

Of the 53 children who received seizure prophylaxis, 4 had a subsequent seizure during their hospital course (1 levetiracetam only, 1 (fos)phenytoin and 2 both). Adverse drug reaction prompting a change in therapy did not occur in any patient. Two patients died, one received levetiracetam prophylaxis and the other received fosphenytoin. Of the 44 patients followed for at least 6 months, 2 developed epilepsy (1 (fos)phenytoin and 1 levetiracetam).

Children treated with levetiracetam and (fos)phenytoin for seizure prophylaxis had an equivalent number of breakthrough seizures. Both medications were sufficiently tolerated allowing for a complete the treatment course. A similar number of patients in each group went on to develop epilepsy within 6 months.

Numerous anticonvulsant agents are now available for treating status epilepticus (SE). Since the introduction of new anticonvulsants which might be helpful in this setting, little data is available to guide clinicians in the initial treatment of seizures or SE. This study aims to describe the current practice trends in initial treatment of SE in the USA.

Retrospective, multicenter, observational study at 20 American academic medical centers. Sites reviewed the last 10 to 20 cases of SE at their institution. Data was submitted via secured on-line platform and analyzed for trends in initial management of SE. Demographic data, drug selection and timing, and other pertinent data were recorded.

Data collection and analysis is on-going. Preliminary data from 15 patients with available data suggests the most commonly used agent used in the initial management of SE was intravenous lorazepam (11/15, 73%). Two other patients received intravenous diazepam as initial therapy. Thirteen (87%) patients required more than one drug for treatment. The most common second agent was phenytoin (6/13, 46%). Levetiracetam (23%) and valproate (15%) were also used relatively frequently as second anticonvulsant for SE. The majority of initial doses were given in the emergency department (ED) within a median of 12 minutes (2-60) from arrival. Nine patients received benzodiazepines prior to admission by emergency personnel or an outlying facility. Seizures in eight patients (53%) abated after the second agent.

Benzodiazepines (lorazepam) and phenytoin continue to be cornerstones of initial SE management. Nearly all patients required more than one drug for treatment. The most common second line agents were phenytoin and valproate, along with newer agents such as levetiracetam. Approximately half of the patients required further treatment for SE after the initial two anticonvulsants.

Numerous anticonvulsant agents are now available for treating refractory status epilepticus (RSE). Little data is available to guide clinicians in the treatment of RSE, particularly with regard to newer agents such as lacosamide and levetiracetam. Agents such as ketamine and lidocaine have been used in case reports to treat RSE, but their role remains unclear. This study aims to describe the current practice trends in the treatment of RSE in the USA.

Retrospective, multicenter, observational study at 20 American academic medical centers. Sites reviewed the last 10 to 20 cases of SE at their institution. Data was submitted via secured on-line platform and analyzed for trends in initial management of SE. Demographic data, drug selection and timing, and other pertinent data were recorded. RSE was defined as SE occurring for > 30 minutes despite therapy with at least 2 first-line anticonvulsants.

Data collection and analysis is on-going. Preliminary data available from 15 patients suggests that seven (47%) of patients received 3 anticonvulsants for refractory SE. Four patients received phenytoin (after failing benzodiazepine therapy in addition to another first-line agent), while others received a variety of anticonvulsants including valproate, phenobarbital, topiramate, levetiracetam, and/or lorazepam in various combinations. Overall 4 patients (27%) received pentobarbital or propofol infusions for burst suppression.

Management of RSE is highly variable among each institution and often within each institution. Pharmacologic induction of coma via propofol or long-acting barbiturate continues to be a treatment option, though numerous other anticonvulsants are typically given prior to initiation. 

Patients were grouped based upon whether or not they were taking -channel blocking AED at baseline. Efficacy measures included change in seizure frequency and >50% responder rates. Treatment emergent adverse events (TEAEs) and discontinuations were also evaluated.

Of 1,308 patients in the overall safety population, 82% were using -channel blocking AED. In this subgroup, adjunctive lacosamide significantly reduced seizure frequency at all doses (P<.01, median percent reduction per 28 days for 200 mg/day, 400 mg/day, 600 mg/day: 33.3%, 39.0%, 42.7%) compared with placebo (18.9%), similar to results in the overall pooled population. In the subgroup not taking traditional sodium-channel blockers (18%), efficacy appeared more pronounced (38.0%, 62.5%, 79.0% reduction for 200 mg/day, 400 mg/day, 600 mg/day vs 28.0% for placebo), with significance for 400 mg/day and 600 mg/day. Similar results were observed for discontinuation rates were lower in the non-sodium channel subgroup.

In this post hoc exploratory analysis, adjunctive lacosamide demonstrated significant seizure reduction over placebo regardless of concomitant AED primary mechanism of action. Further study is warranted to confirm whether certain combinations may produce additional benefit or improved tolerability.

Financial Support: All authors are employees of UCB, Inc., UCB Pharma S.A., or Schwarz Biosciences (a division of UCB, Inc.)

Onyx (ev3, Irvine, California) embolization of cerebral arterio-venous malformations (AVM) has become increasingly common. We aimed to explore the risk of seizures after onyx use.

A retrospective review was conducted of patients with AVM who received onyx embolization between 2006-2009 at our facility. Baseline demographics, clinical history, seizure history, use of anti-epileptic drugs (AED), and AVM characteristics were recorded. Details of onyx treatment and were assessed.

Of 20 patients who underwent onyx embolization, the median age was 35 (range 11-67). The initial AVM presentation was SAH in 10% (N=2), and intracerebral hemorrhage in 40% (N=8). The median Spetzler Martin grade was 3 (range 1-5) and the median size was 35 mm (range 10-60). A history of seizure was present in 50% (N=10) of patients preembolization and 12 (60%) patients were on long term AEDs prior to treatment. The median number of embolizations was 2 (range 1-4). The median obliteration rate after the first and final embolizations were 60% (range 35-100%) and 90% (range 50-100%), respectively. Seizure post-onyx embolization occurred in 45% (N=9). The median time to seizure post-onyx was 23.5 days (range 0.3-180 days). Four patients with seizure had no prior seizure history (20%). Two patients (10%) had no other identifiable cause for seizure other than recent onyx embolization. Of these 2, one patient did not receive AEDs prior to embolization and one had a subtherapeutic AED level at the time of embolization.

New onset seizures post-onyx embolization are not uncommon. AED prophylaxis may be warranted.

Neonatal seizures typically signal underlying brain injury. Mortality is high, and survivors have high rates of neurological disability. There is accumulating evidence that seizures may disrupt brain development, and lead to longterm deficits. Routine EEG monitoring lasting 30 minutes is considered standard of care for evaluating neonatal seizures in many centers.

From July 2008 to June 2009, the Neonatal Neurocritical Care Service at UCSF evaluated 155 newborns. Indications for monitoring using long-term conventional video-EEG (i.e. >24 hours) included suspected clinical seizures and/or neonatal encephalopathy. Seizure etiology was determined based on review of clinical charts, laboratory evaluation, and neuroimaging.

Ninety newborns were monitored >24 hours. Demographics were as follows: male sex 52%, preterm <37 weeks gestation at birth 22%, transferred from a referral center 89%. Of the monitored newborns, 32 (36%) had electrographic seizures, 19 (21%) had clinical seizures but no EEG seizures during monitoring, 9 (10%) had spells that were not thought to be seizure, and 30 (33%) with encephalopathy did not have seizures. Of the 32 children with electrographic seizures, 24 (75%) had subclinical seizures (including 11 who never had a clinical seizure). In addition, 21 (66%) did not have seizures within the first 30 minutes of recording, and 27 (84%) had single or recurrent seizures lasting longer than 30 minutes. Cause of seizures included: perinatal hypoxic-ischemic encephalopathy (53%), stroke (14%), intracranial hemorrhage (8%), and infection (8%).

Newborns with seizures and encephalopathy frequently have subclinical events, as well as seizures with onset after 30 minutes of recording and/or that recur over longer than 30 minutes. Clinical observation and routine 30 minute EEG are insufficient to capture ongoing seizures. With accumulating evidence that seizures may harm the developing brain, adequate monitoring for seizures is imperative for appropriate treatment. 

Lacosamide is a new antiepileptic drug (AED) with a novel mechanism of action. Several case reports suggest lacosamide may have a role in status epilepticus (SE). The purpose of this case series is to describe the use of lacosamide in RSE at our institution.

Prospective observational study of all patients admitted to the neurosciences intensive care unit with RSE who received at least one dose of lacosamide from October 2009 to May 2010.

Eight patients received lacosamide after failure of at least two other agents. On average, lacosamide was started 5 days (range 0 -14 days) after the onset of SE. The most frequently used dosing regimen was an initial intravenous dose of 200 mg followed by 200 mg every 12 hours. Most patients had received 3 (range 2 -4) AEDs prior to lacosamide. Levetiracetam was used prior to lacosamide in all cases. None of the 7 patients evaluated responded to lacosamide according to our predefined criteria (resolution of electrographic seizure within 4 hours of administration for patients not in burst suppression; or, no seizures within 24 hours of emergence from burst suppression). One patient was unable to be evaluated because she was in burst suppression during the time she was on lacosamide because it was discontinued when the patient developed angioedema after receiving two doses. Care was withdrawn in two of the eight patients for reasons unrelated to lacosamide. Lacosamide was continued at discharge on all surviving patients except in the case of angioedema.

This is the largest case series to date describing the use of lacosamide in patients with RSE. Despite the novel mechanism of action, we observed no evidence that lacosamide is effective in RSE; however, our sample size was small. Further study is needed to determine the effect of lacosamide in SE.

Status epilepticus (SE) is a life-threatening neurological emergency. Conventional therapy involves benzodiazepine (first-line), phenytoin/fosphenytoin or valproate (second-line) and anti-epileptic drugs (AEDs) with general anesthetic effects (third-line) if refractory. The non-sedating (second-line) AED levetiracetam has recently gained popularity. We examined SE therapeutic practice at a large academic medical center.

Adults presenting with SE (continuous seizures > 5 minutes) at two university emergency departments (ED) from August 1999 to October 2007 were retrospectively studied.

Of 128 patients with 145 separate SE admissions, (mean age 54.5 ± 15.8 years) most presented with convulsive (91%), and generalized tonic clonic (86%) seizures. In the field and ED, 83% of patients required first-line AEDS, 70% second, 57% first and second, 18% third and 9% all lines. During all treatment, patients received a mean 3.4 ± 1.6 AEDs; 85% received first-and second-line AEDs, 35% third and 28% all three. Most common first-line AEDs were lorazepam (61%) and diazepam (27%); second-line were phenytoin (50%), valproate (14%) and levetiracetam (19%). Frequency of levetiracetam use increased from 8% (2000) to 71% of patients in 2007, but was used in the field/ED in only 2%. Those receiving levetiracetam more likely used it before (p<0.001); had continuous EEG (p=0.020); were intubated longer (p<0.001) and trended toward more frequent third-line treatment (46% v. 30%, p=0.059). Of patients receiving third-line AEDs, 41% had an EEG on day of admission and 43% had continuous EEG monitoring. Intubated (p<0.001), hypoxic (p=0.008) and patients with seizure recurrence (p<0.001) more likely received third-line AEDs.

Using a current definition of SE, treatment is effective with first-and second-line AEDs in a majority of cases. Use of third-line AEDs occurs frequently without immediate or continuous EEG monitoring. Use of levetiracetam is increasing as a SE therapy, but is largely reserved for prior users, post-ED management and for more resistant SE.

Mortality and morbidity remain frustratingly high in status epilepticus (SE). While seizure cessation is the paramount goal driving escalation of anti-epileptic therapy, use of anti-epileptic agents with general anesthetic effects (third-line therapy) requiring intubation and hemodynamic management may adversely impact outcome. A systematic evaluation of patient-and treatment-related outcome determinants was undertaken.

Adult patients admitted for SE (continuous seizures > 5 minutes), to two emergency departments of an academic medical center between August 1999 and October 2007 were retrospectively analyzed for factors affecting morbidity, mortality and Glasgow Outcome Score (GOS) at discharge.

Of 128 patients with 145 separate SE admissions 58 (45%) were female, with mean age 54.4 ± 15.8 years. SE was convulsive in 129 cases (89%) at presentation. SE etiologies were sub-therapeutic AED (43%), remote or progressive CNS disease (17%), alcohol or other drug (13%) and acute CNS disease (12%). Ninety-seven patients (67%) had previous seizures, and 32 (23%) prior SE. Seventy-nine patients (54%) had good outcomes (GOS<3). On univariate analysis, good outcome was associated with previous seizures (77% vs. 54%, p=0.004) and previous SE (35% vs. 11%, p=0.001.) Poor outcome (GOS>2) was associated with hypoxia (100% vs. 0%, p<0.001), acute CNS disease (21% vs. 5%, p=0.003), intubation (70% vs. 52%, p=0.006), and third-line AED (51% vs. 21%, p<0.001). Independent predictors of poor outcome were age (p<0.001), hypoxia (p<0.001), and third-line AED use (p=0.008). A history of SE independently predicted a good outcome (p=0.007).

A history of status epilepticus is associated with better outcome, while increased age and third-line drug therapy are independently associated with severe disability and mortality in SE. Further investigation should focus on whether third-line anesthetic AEDs are markers of worse SE or induce additional morbidity.

Lacosamide (LCM) is a novel antiepileptic drug with favorable properties including IV formulation, minimal hepatic metabolism, minimal protein binding and no adverse respiratory or hemodynamic effects. Little is known of LCM safety in critically ill patients with regard to hemodynamic profile and adverse events.

We retrospectively reviewed electronic medical records for patients treated with LCM at a single tertiary-care academic medical center (July 2009 -January 2010). Data collection included age, sex, dosing and documented seizure activity (clinical diagnosis/EEG). Safety was assessed by pre-/post-drug laboratory and hemodynamic parameters. Adverse drug reactions and drug interactions were reported by clinical documentation. 

LCM demonstrated a favorable hemodynamic safety profile for critically ill patients. Twelve percent reported adverse events, resolved in all cases with LCM discontinuation. Future prospective studies with hematological monitoring are needed to support the safety of LCM in critically ill patients.

Introduction cEEG is linked to cerebral metabolism and sensitive to cerebral ischemia and hypoxia. The severity of cerebral ischemia can be seen on cEEG as changes in morphology, amplitude, or frequency and may detect neuronal dysfunction at a reversible stage.

We present a case of focal cerebral edema noticed on cEEG twenty-four hours prior to clinical signs of increased intracranial pressure.

Sixty-six year-old male admitted to the medical ICU for respiratory distress and a NSTEMI requiring anticoagulation. He developed confusion and right-sided weakness. CT head showed a left parietal intracerebral hematoma (50 x 37 mm) with 9-mm midline shift. Workup did not reveal underlying tumor or vascular malformation. He developed left eye deviation with right arm twitching. cEEG showed focal status epilepticus from the left parietal-occipital region without clinical signs. He was loaded with lorazepam, phenytoin, levetiracetam and phenobarbital with cessation of his seizures. Due to significant medical co-morbities, he was not a candidate for aggressive intervention. One week later, cEEG showed a developing asymmetry with decreased power recorded from the left hemisphere. This decreased power progressed to burst suppression in the left hemisphere, followed by suppression of the right hemisphere. Complete suppression of the EEG in both hemispheres was noted six hours prior to his development of clinical signs of herniation, and the medical ICU team was notified. He was later found to have a left dilated, non-reactive pupil. Repeat CT head confirmed 16-mm shift with diffuse cerebral edema. Medical interventions for decreasing intracranial pressure were unsuccessful and his family withdrew care.

This case illustrates the utility of cEEG in monitoring events of cerebral function in addition to detecting seizure activity. It serves a vital role in early recognition of neurological complications from brain injuries that may not be noticed clinically, which is paramount to early intervention.

Introduction Standard brain tissue oxygen monitors are threshold-based, triggering actions based on presumption of imminent ischemia [brain oxygen or PbtO2 < 20 mmHg]. Some published practice guidelines suggest that seizure is a potential culprit when PbtO2 crosses this threshold. The evidence for this assumption is slim.

Data were collected as part of a prospective observational database. Brain oxygen monitors and continuous scalp EEG (if resources available) were placed by clinical protocol in patients with aneurysmal subarachnoid hemorrhage (aSAH) or traumatic brain injury (TBI) and a Glasgow Coma Scale of < 8. Eight of these patients had one or more discrete EEG-detected seizures during an overlapping monitored period (both brain oxygen and EEG) and were selected for review. Probability of seizure given PbtO2 value < 20 mmHg (and the inverse) were calculated by linear regression, clustered by patient.

There were 274 distinct seizure episodes and 1797 PbtO2 measurements in 8 patients (5 aSAH, 3 TBI). 2 of 8 patients had oxygen monitors contralateral to seizure focus (both TBI). 9.4% of seizures (17/180) were followed by declining PbtO2 values to <20 within subsequent 30 minutes (95% CI=0-26%); these occurred in 2 patients. Conversely, 2.5% of the PbtO2<20 measurements (8/315) had a seizure in the preceding 30 minutes.

We found that PbtO2 below the ischemic threshold in brain-injured patients did not correlate with seizure. This calls into question the assumption within some published practice guidelines. Focal brain oxygen response to seizure would rather be expected to parallel the complex spatiotemporal hemodynamic changes described by fMRI and intrinsic optical spectroscopy. The failure of brain oxygen monitors to detect seizure may be in their threshold-based interpretation. We propose prospective research for more effective strategies of interpreting patient data. In tro d u c tio n Previous smaller studies have evaluated the indications and usefulness of sEEG [1] [2] , with some arguing that neurological consultation is needed prior to obtaining a sEEG [3] . The primary objective of our study was to evaluate the utility of sEEG in detecting discrete seizures or status epilepticus as a function of ordering physician and clinical indication.

We retrospectively reviewed all consecutive sEEG reports done over 1 year at a tertiary care institution. All reports were reviewed for the parameters listed in 

In our tertiary care institution sample, the rate of finding status epilepticus or seizures among sEEG is 5.1%. In our sample, the best clinical predictors of finding NCSE or discrete seizures on a sEEG are overt, continuous seizures or witnessed seizure without return to baseline. Additional monitoring methods, such as continuous EEG, should be considered in critically ill patients for improved yield. Furthermore, our data suggests that perhaps neurological consultation should NOT be mandatory prior to obtaining a sEEG.

Various anticonvulsant agents are now available for treating status epilepticus (SE). Clinical pharmacists (CP) have been shown to optimize the care of patients receiving phenytoin therapy, demonstrating reductions in seizure activity, duration of therapy, and cost. Since the introduction of new anticonvulsants which might be helpful in this setting, limited data is available regarding the impact that CPs may have in managing this complex clinical condition. This study aims to describe the impact that a CP can have in this setting.

Retrospective, multicenter, observational study at 20 American academic medical centers. Sites reviewed the last 10 to 20 cases of SE at their institution. Data was submitted via secured on-line platform & analyzed for trends in initial management of SE. Demographic data, drug selection and timing, & other pertinent data were recorded.

Data collection and analysis is on-going. Preliminary data available from 15 patients suggests the median time to the patient receiving the first dose an initial medication is lower in the presence of a CP (6 vs 37.5 minutes). This reduction in time delay is consistent in those receiving a second medication with presence of a CP (1.5 vs 22.5 hrs). With regards to specific therapy, those patients receiving phenytoin therapy were less likely to need re-dosing when the initial dose was determined with the assistance of a CP (33% vs 50%).

Initial data suggests that having a CP involved in the acute management of a patient with SE decreases the time to medication administration and optimizes initial dosing. Optimizing the speed and treatment of SE has been shown previously to decrease neurologic damage and improve overall outcomes.

Poster 90 

The electroencephalogram (EEG) is underused or not available in the Emergency Department (ED). Criteria for EEG in the ED have not been established. This study sought to determine whether EEG performed within 30 minutes of referral by an ED physician helps establish diagnosis and/or changes management; and which clinical features/diagnoses are useful in selecting patients for ED EEG.

Single-center prospective cohort intervention study 1 day/week, of sequentially referred adult patients presenting with clinical seizure activity or AMS. Standard EEGs (20 minute 16-channel) were performed by an EEG technician using a commercially available cap. EEGs were immediately reported to the ED physician and a utility survey completed.

Over 1 

Rapid availability of standard full-montage EEG in the ED is feasible and helps establish a diagnosis in about half of patients with AMS, but rarely changes management. While witnessed seizure activity is most likely to prompt a request for ED EEG, non-neurologic causes of AMS are more likely to be associated with positive utility.

Financial Support: IBT performs R&D on an EEG monitoring system similar to that used in this study.

Therapeutic hypothermia (TH) is widely accepted as the first successful intervention that can improve neurologic recovery in patients who are hemodynamically stable after resuscitation from a V-fib arrest. In this case report, we document full neurological recovery after its application for an unwitnessed asystolic arrest of at least 43 minutes.

A 16-year-old male with polysubstance overdose was last seen awake at 03:00. Eleven hours later he was found on his sofa unresponsive and apnic. Bystander CPR was not initiated, and EMS documented asystole upon arrival. ROSC occurred 43 minutes later in the ED with an initial bladder temperature of 88.9 F. In spite of contraindications and the lack of supportive data, we elected to continue with TH. At initiation he had a GCS of 3, myoclonic jerks, and profound hemodynamic instability.

The patient's clinical course was complicated by refractory septic shock, pneumonia, ARDS requiring HFO, AKI requiring CRRT, shock liver, coagulopathy, and pancreatits. He was extubated on day 11, with a GCS of 4.

Remarkably, by day 13, his GCS was 12. MRI at that time was within normal limits except for a small lacunar infarction in the right basal ganglia. MR Spectroscopy was only minimally abnormal with mild elevation of the choline peak and mild depression of the NAA peak. Evidence of the massive neuronal loss and metabolic disarray that is common in cases of severe anoxia was not seen. By day 21, the patient was transferred to rehabilitation. He was discharged home on day 35 without deficits.

Although TH was applied under circumstances which were excluded in the original studies, the imaging findings and clinical outcome for this patient suggest that TH may be neuroprotective in a wider patient population.

To determine the degree to which microgliosis and astrocytosis affect recovery after murine models of acute CNS injury.

Utilizing traumatic brain injury (TBI) and intracerebral hemorrhage (ICH) models, microgliosis and astrocytosis was examined in the presence of a selective microglial inhibitor, TT-301 (1mg/kg, i.p.) administered 30 min and then daily after injury, with their effect on neuronal survival assessed at 10 and 28 days after injury. To determine microglial modulation's effect on functional improvement, rotorod (RR) through 7 days and Morris water maze (MWM) at 28 to 32 days after injury was performed.

Reduction in microgliosis was seen at 10 but not 28 days in mice treated with TT-301 after injury with TBI; there was no reduction in astrocytosis during this period. Improved neuronal survival was confirmed at 10 days but was not seen at 28 days in mice treated with TT-301 after injury with TBI. These findings correlated with improved neurological rate of recovery via increased RR latencies through 7 days after injury and long-term recovery via shortened MWM latencies at 28 through 32 days after injury with TBI in mice treated with TT-301. Finally, neurological improvement was extended in ICH-injured mice treated with TT-301 demonstrated by increased RR latencies through 5 days after injury with reduced cerebral edema and no change in hematoma volume at 24 hr after injury.

Modulation of microgliosis, but not astrocytosis, by daily administration of TT-301 improves the rate and ultimate neurological recovery after murine models of acute CNS injury.

Hypothermia and hyperthermia are common occurrences in intensive care units. As benefits of therapeutic hypothermia unfold, we investigate the effect of temperature manipulation under the influence of isoflurane on the physiological response of the non-injured brain using EEG and somatosensory evoked potentials (SSEPs).

Fourteen rats were divided into two groups, based on electrode placement, at either frontal-occipital (FO) or primary somatosensory (SS) cortical locations. Neural signals were recorded during normothermia (T = 36.5-37.5 o C), moderate hypothermia (T = 32-34 o C) and mild hyperthermia (T = 38.5-39.5 o C), with temperature maintained using surface cooling and re-warming. Burst-suppression ratio (BSR) was used to evaluate EEGs and amplitude-latency analyses were used to assess SSEPs. Repeated measures ANOVA with Bonferroni correction was to study differences between different temperature phases.

The electroencephalic Burst-suppression ratio (BSR) in both groups (n=7 each) was higher during hypothermia (FO: 0.58, 0.52-0.64 (mean, 95% Confidence Interval); SS: 0.30, 0.18-0.42) than normothermia (FO: 0.16; 0.04-0.28; SS: 0.02, 0.0-0.08) (p<0.001 and p=0.006 respectively). SSEPs showed a marked increase in normalized N10-P15 peakto-peak amplitudes during hypothermia (2.89, 1.75-4.04 times baseline, p=0.02). Peaks were delayed during hypothermia (N10: 10.8, 10.4-11.2msec; P15: 16.2, 15.4-17.0msec) in comparison with normothermia (N10: 9.1, 8.9-9.4msec; P15: 13.7, 13.1-14.3msec) (p<0.001). While no differences in EEG and SSEP amplitudes were seen in hyperthermia, SSEP peak latencies reduced (N10: 8.6, 8.4-8.7msec; P15: 12.6, 12.2-13.0msec) (p<0.001).

Our results show that both spontaneous and evoked neural responses under anesthesia are sensitive to changes in temperature. Hypothermia was characterized by electroencephalic suppression and activation of primary somatosensory response, whereas hyperthermia showed no marked changes except increase in SSEP conduction velocity. Our study highlights that hypothermia acts synergistically with isoflurane to down-regulate basal cortical firing, resulting in hyper-responsive cortical state, leading to amplified SSEP that may be attributed to increased neural recruitment and synchronous cortical firing upon arrival of an afferent stimulus from the thalamus. This work was supported by the American Heart Association (Grant 09SDG1110140).

The incidence of fever and secondary injury in the neuroscience patient population is higher than other populations. The occurrence and duration of fever in TBI patients has been associated with poor outcomes such as length of stay. Under-treatment of fever is common and fever is often mismanaged. The purpose of this study was to identify evidence-based practices (EBP) for normothermic temperature management in NSICU patients and adopt a standardized treatment guideline to optimize nursing practice and outcomes.

A retrospective chart review (12/08-8/09) of fever in TBI patients was conducted to establish baseline data for incidence of fever and identify non-invasive interventions utilized to reduce fever in our 8 bed NSICU. The IOWA model of EBP was used to guide development of a nursing guideline for management of fever in NICU patients.

606 episodes of T>37°C (103 patients) were analyzed. 43% of these episodes were not documented as receiving treatment. When treatment was provided, three interventions or combinations were documented. 73% received pharmacological interventions (acetaminophen only), 19% external cooling measures (fans, ice packs, or cooling blanket) and 10% received a combination of pharmacological plus external cooling. For the 3 interventions, results of ANCOVA showed no significant group differences in time spent >37°C. Suggestive evidence showed that external cooling alone was least effective in mean hours >37°C. An EBP Team formed, developed guidelines and is implementing a practice change to incorporate (1)Normothermia for NeuroProtection and (2) Shivering Management Guidelines in the NSICU.

The concept of normothermia for neuroprotection has not been well adopted into nursing practice. Our future plans include incorporating the 2 EBP protocols to manage temperature elevations in the NICU. Data will then be analyzed to determine if efficacy of treatment options can be established and if nursing interventions related to normothermia have been enculturated into practice.

Fever in neurocritical care patients occurs with an incidence of up to 70% and has a negative impact on neurological outcome. The purpose of this prospective study was 1. to develop a new method for lowering fever resistant to conventional procedures and 2. to be specific and rapid in decreasing elevated temperature of the brain.

Methods 9 patients with 10 episodes of intractable fever and aneurysmal subarachnoid hemorrhage were treated with specifically adapted cooling neck pads. Intractable fever was defined as a temperature > 37.8° Celsius persistent after 2 hours of expanded standard fever management with application of acetaminophen, novaminsulfone, specific washing solutions and ice packs. Temperature values of the brain, blood and urinary bladder were taken close meshed after application of the cooling neck pads up to hour 8. We developed one regular shape of neck pads and one shape for short necks or tracheostoma, each shape covering the carotid triangles for cooling the blood inflow to the brain using the ArcticSun® material.

The brain, blood and urinary bladder temperatures decreased significantly from hour 0 to a minimum in hour 5 ( 

Selective cooling of the neck might be efficient in decreasing intractable fever for several hours, but not for a sustained period.

The benefits of therapeutic hypothermia for brain injuries have already been documented. Systemic procedures have been used, with severe complications. An alternative to avoid such side effects is selective hypothermia for cliquish brain cooling. The aim of this paper was to evaluate the effectiveness of a new method of selective hypothermia using a cooling device with liquid flow through swine nasopharynx.

Ten hybrid swine, weighing 18-25kg, were anesthetized, intubated, and underwent controlled mechanical ventilation. They were monitored with hemodynamic and systemic parameters, along with continuous transcranial Doppler, temperature of right and left brain hemispheres, and PbrO2 monitoring. After proper installation of the nasopharyngeal device, the animals underwent hypothermia for 60minutes, with passive re-heat for 15 minutes.

Brain temperature was reduced in both hemispheres after nasopharyngeal hypothermia. At right, there was a reduction of 38,8 1,1ºC to 37,8 1,1 ºC on the first five minutes, 36,3 0,8 ºC after fifteen minutes, and 35,3 2 ºC after sixty minutes. At left, where the cooling device was located, there was a reduction of 38,8 1,2 ºC to 37,3 1,6 ºC on the first five minutes, 35,6 1,4 ºC after fifteen minutes, and 34,3 1,7 ºC after sixty minutes. Central temperatures (rectal and esophageal) were stable during the whole procedure.

We demonstrated that nasopharyngeal cooling with cold liquid flow is capable of reducing brain temperature Financial Support: None

The effect of hyperglycemia on brain cells of septic shock patients is unknown. The aim of this study was to evaluate the relationship between hyperglycaemia and apoptosis in the brains of septic shock patients.

In a prospective study of 17 patients who died from septic shock, ammon's horn was assessed for neuronal ischemia, neuronal and microglial apoptosis, neuronal Glucose Transporter (GLUT)4, endothelial iNOS, microglial GLUT5 expression, microglial and astrocyte activation. Glycemia was recorded five times a day from ICU admission to death. Blood glucose (BG )> 2 g/l defined hyperglycaemia and area under the BG curve (AUBGC) > 2 g/l was assessed.

Median BG over ICU stay was 2.2 g/l. Neuronal apoptosis correlated with endothelial iNOS expression (rho=0.68, p=0.04) while microglial apoptosis correlated with AUBGC > 2g/l (rho=0.70; p=0.002). Neuronal and microglial apoptosis correlated (rho=0.69, p=0.006), but neither correlated with the duration of septic shock, nor with GLUT4 and 5 expression. Neuronal apoptosis and ischemia tended to correlate with duration of hypotension.

In patients with septic shock, neuronal apoptosis is rather associated with iNOS expression and microglial apoptosis with hyperglycaemia, may be via absence of GLUT5 downregulation. These data provide a mechanistic basis for understanding the neuroprotective effects of glycemic control.

Poster 98

Influenza virus infection of the respiratory tract is associated with a range of neurologic complications. The emergence of 2009 pandemic influenza A (H1N1) virus has been linked to neurological complications, including encephalopathy and encephalitis.

Case report and literature review.

We review case management of a 20 year old Hispanic male who developed febrile upper respiratory tract signs and symptoms followed by a confusional state. He had rapid neurologic decline and his clinical course was complicated by refractory seizures and malignant brain edema. He was managed with oseltamavir and peramavir, corticosteroids, intravenous gamma globulin treatment, anticonvulsants, intracranial pressure management with external ventricular drain placement, hyperosmolar therapy, sedation, and mechanical ventilation. RT-PCR analysis of nasal secretions confirmed 2009 H1N1 virus infection; CSF was negative for 2009 H1N1 viral RNA. Follow up imaging demonstrated improvement in brain edema but restricted diffusion in the basal ganglia. We review the literature and provide a review of the clinical spectrum of neurologic complications of seasonal influenza and 2009 H1N1, and current approaches towards managing these complications.

2009 H1N1-associated acute encephalitis and encephalopathy appear to be variable in severity, including a subset of patients with a malignant clinical course complicated by high morbidity and mortality. Since the 2009 H1N1 influenza virus has not been detected in the CSF or brain tissue in patients with this diagnosis, the emerging view is that the host immune response plays a key role in pathogenesis.

Myasthenic crisis (M.C.) is potentially life threatening complication with 4-5% mortality.It occurs in approxiamately 15-20% of myasthenia graivis(M.G.) patients. As a serious complication its characters needs to be well recognized and investgated as will as it's posible risk factors or diseases namely thymic disorders .Myasthenic crisis is defined as respiratory failure requiring mechanical ventilation. However, this definition would not include patients with severe respiratory or bulbar symptoms not requiring intubation.Those patients were classified in crisis in the new myasthenia gravis faundation of amarica (MGFA) classification in stage (V).

Retrospectively the charts of 63 MG patients followed in the neurology clinics at multiple tertiary level hospitals (all in Riyadh , Saudi Arabia ) were included. Myasthenic crisis were reviewed. Demographic data clinical information, level of ACh-R antibodies and thymus pathology in patients who had thymectomy were recorded.

Thymoma wether benign or malignant was found to be associated with myasthenic crises . This association was statistical significance with P value of = 0.0067 . Gender difference , age or Acetylcholine levels failed to show any significant association with MC.

As noted with other international data, thymoma is strongly associated with more aggressive course of MG, and the development of MC. Patient with MC warrant a full and complete work up to exclude thymomas, which if suspected requires urgent and aggressive attention.

Hyponatremia is a common problem in the brain-injured patient. There are multiple causes of hyponatremia in this population. While determining the cause may be difficult, the effect is thought to be detrimental. Hyponatremia is associated with confusion, seizures and increased cerebral edema. Previously to correct hyponatremia, isotonic and hypertonic saline solutions were used which led to frequent changes in tonicity and infusion volumes and erratic serum sodium. Employing a novel sodium scale using a mixture of 3% NaCl and 0.9% NaCl while maintaining the same infusion volume maintained the serum sodium within a target range of 135-150 mEq/dl.

The sodium scale was started on all brain-injured patients on the Neurocritical Care service. Central access was obtained via a PICC line or central venous catheter. Sodium levels were drawn every 6 hours and nursing staff adjusted the 3% NaCl infusion based on the serum sodium level using this new sodium scale, then adjusted the 0.9% NaCl infusion to maintain the same volume of infusion. This study represents a sample of 10 patients from this group.

Patients in this group included 5 severe traumatic brain injuries, 2 intracranial hemorrhages, 2 subarachnoid hemorrhages and one ischemic stroke. 2 of the patients had increased intracranial pressure requiring frequent hypertonic saline boluses. A total of 482 sodium measurements were obtained. Of these, only 3 (0.6%) of the serum sodium values were less than 135mEq/dl while 37 (7.7%) were greater than 150mEq/dl. Of those values greater than 150mEq/dl, 31 (6.4%) were measured in patients with recalcitrant elevated ICP.

The sodium scale prevented hyponatremia in greater than 99% and hypernatremia in greater than 92% of study patients in this nursing driven protocol.

Constipation is often a side effect in the treatment of severely brain-injured patients. This can be secondary to narcotic-based sedation and/or the critical nature of their illness. Constipation can alter intraabdominal pressure, which can increase intracranial pressure. Neostigmine is an agent that has been shown to induce bowel movements in the critically ill and is often used in the surgical intensive care unit. There have been no published studies looking at the safety of neostigmine in the brain-injured patient.

A bowel regimen was started on brain-injured neurocritical patients at our center. The last step in the regime was neostigmine 5mg IV over 12 hours or until a bowel movement was achieved. This was only started after standard interventions had failed. In this study all patients had intracranial monitors.

A total of 8 doses of neostigmine were given to 3 patients. 1 patient with severe TBI, 1 with subarachnoid hemorhage and one with an intracranial hemorrhage. There were no generalized seizures, no bradycardia and no increase in ICP seen. A bowel movement occurred with 5 of the 8 doses. Of the doses that did not produce a bowel movement, these patients were in pentobarbital coma and were therapeutically hypothermic at less than 34 degrees Celsius.

Neostigmine is safe in the brain-injured patient, but may not be as effective in patients in pentobarbital coma and therapeutically cooled.

Cerebral microdialysis is a bedside technology that measures markers of cerebral metabolism and cellular destruction following traumatic brain injury and other primary central nervous system insults, allowing for interventions that are more prompt and the potential for better patient outcomes.

This poster will describe the physiologic rationale, clinical indications, and nursing considerations for this novel neuromonitoring technique. By utilizing a microdialysis catheter and microinjection pumps, samples of interstitial fluid are collected and analyzed for glucose, lactate, pyruvate, glycerol and glutamate. Hourly analysis is performed by the ICU nursing staff using a special microdialysis analyzer (ISCUS Flex) that is housed within the ICU.

Frequent collection of interstitial fluid illustrates how seriously brain cells are affected by the ischemia that often follows traumatic brain injury, subarachnoid hemorrhage, or vasospasm. It also depicts the effectiveness of pharmacological and surgical interventions for patients with increased intracranial pressure and decreased cerebral perfusion. Normal and abnormal values for each of these parameters will be presented in a table format along with recommended interventions. Pictures of the microdialysis equipment used will be illustrated.

Finally, nursing considerations in the bedside management of patients on microdialysis to include patient care considerations, use of equipment, and patient care assignments, will be listed. While this novel technology is used in a mere handful of tertiary brain injury centers in the US, our Level 1 trauma center has used it in over 120 patients. Its promise as a reliable measurement of the status of brain physiology has resulted in its adoption in an increasing number of neurological intensive care units. Any nurse who cares for patients with neurological injuries will benefit learning about the key elements of cerebral microdialysis depicted in this poster.

The carbon monoxide poisoning (CO) is a leading cause of death by poisoning. It may be pure by CO release in incomplete combustion of organic substances, or mixed during fires or inhaling exhaust fumes. Many data exist on poisoned patients with burns but few are available on patients without burns. The specific character of this study is to look at smoke poisoned patients unhurt of any burn. The objective of the study is to compare the patients poisoned to the CO with or without smoke inhalation.

Prospective cohort study realized between 01/01/2006 and 30/12/2009 including all carbon monoxide-poisoned patients having benefited from hyperbaric oxygen therapy except the children. Following parameters were seized: age, sex, date of admission, Sofa, the source of the intoxication, the gravity CO score (0: no symptoms, 1: discomfort, tiredness, headache, 2: nausea, vomiting, asthenia, giddiness, 3: transitory loss of conscience quickly reversible, 4: Organ failure (coma, shock, respiratory distress), 5: death), the initial clinical examination (realized by first aid) neurological, cardiac or respiratory sign, state of consciousness, biology, arterial blood gases, the rate of HbCO, the Murray score and the rate of complication (intubation, pneumonia, shock and death).

405 patients were included in the study. The sex ratio was 61 %, the mean age was 42± 18 years and the global mortality was 5, 9 %. Among the 405 patients 29 % were poisoned by smoke (S group), 60 % by pure CO (C group) and 11 % by exhaust fumes. More than 30% of the exhaust fumes victims were suicide origin. This characteristic is associated with neurological impairment induce by ingested drugs. Then, their neurological status is impossible to link to the CO poisoning. We have therefore decided to exclude this group. The distribution of CO poisoned victims according to the year's months confirmed that the peak of poisoning occurs mainly during the winter. The SOFA score was higher in the S group compared with the C group (0.38 vs 2,17; p<0.0001). A CO score equal to 4 was present in 33% vs 1% respectively in S vs C group (p<0.001). All those patients were ventilated. In the under group of patients having a CO score at 3, 0% (0/101) of CO poisoned patients vs 18.4% (7/38) of smoke poisoned patients were ventilated (p<0.001). These patients were intubated either during transport or in the intensive care and none of them received hydroxycobalamine during the first aid (before intubation). The laboratory data showed in the S group a higher lactates level (2,2 vs 3,9 mmol/l; p=0.005) and lower initial PaO 2 /F I O 2 ratio (365 vs 306; p=0.002). Nine percent of the S group present a Murray score at 4 vs 0 % for C group (p<0.0001). Pneumonia, shock and death were significantly more frequent in the S group (respectively 8,6 % vs 0,8 %, p<0.0005; 11,2 % vs 0;8 %, p<0.0001; and 17,9 % vs 1,2 %, p <0.0001) . As expected the smoke poisoned group has a higher mortality than pure CO group (mortality 17% vs overall mortality 5.9%). At equivalent CO gravity score, mortality and complications are always more frequent in the smoke poisoned group. The smoke poisoned group has a high risk of degradation. Those patients require specific monitoring and support and probably early administration of hydroxycobalamine.

In view of this study, it appears that smoke inhalation with cyanide poisoning is probably under estimate even in patients with moderate intoxication. The initial cyanide blood level could help to discriminate the intoxicate patients A larger utilization of hydroxocobalamine could improve their prognosis.

To evaluate the Transcranial Doppler utility in the following of comatose neurocritical patients.

A retrospective and observational study was designed to investigate the relationship between the Pulsatility Index (PI), flow velocities (FV) and the outcome in neurocritical patients. We reviewed the results of 73 records and evaluations done with the Transcranial Doppler with the Pulsatility Index (PI) and mean flow velocity (mFV) in the middle cerebral artery (MCA) on patients who underwent treatment in our Neurocritical care Unit from January 1st to December 31, 2006.

From the 73 reviewed records, ten fulfilled the proposed inclusion criteria. Good recovery was defined as a positive punctuation change in the Glasgow coma scale while bad outcome as a negative or no different punctuation between two clinical evaluations. We found seven cases with favourable recovery and just three cases with bad outcome. By using the Mann-Whitney statistical test for independent samples with a mean Pulsatility Index (PI) of 3,4 for the bad outcome group and of 1,7 for the good recovery group (Table 1) we did not find significant statistical difference between the two groups. Besides, applying the same test and using the mean flow Velocity (mFV) with a trustworthy interval of 95% we found significant statistical difference between the maximum systolic flow velocity (FV) in middle cerebral artery (MCA) and the neurological outcome, with a mean flow velocity of 113 cm/s for the good recovery group and 33cm/s for the bad outcome group ( Table 2 Conclusions Higher values of the maximum systolic flow velocity in middle cerebral artery were statistically associated to better neurological recovery. Pulsatility Index values PI>1 were associated with no significant statistical difference to bad outcome. 

We describe a Dantrolene responsive extended dysautonomic syndrome (DREDS). This syndrome includes elevated creatinine kinase (CK) levels, persistent fevers unresponsive to antipyretics and not associated with infection, tachycardia, blood pressure (BP) instability, elevated intracranial pressure (ICP), and medically refractory seizures. Dantrolene can be considered as a treatment for dysautonomic syndromes such as malignant hyperthermia and neuroleptic malignant syndrome. However, the symptoms of DREDS are not consistent with previously defined dysautonomic syndromes.

Two cases ("RR" and "TM") that developed similar clinical symptoms of dysautonomia with different presenting diagnoses are described. RR was a 22 year old female who underwent functional endoscopic sinus surgery for septoplasty. TM was a 21 year old male with traumatic brain injury. Both patients developed elevated CK levels, persistent fevers, tachycardia, blood pressure instability, increased ICP, and medically refractory seizures. Their symptoms were persistent, and failed other medical interventions, including beta-blockers, antipyretics, multiple antiepileptic drugs and fluid management.

Upon initiation of Dantrolene therapy, both patients had resolving serum CK, blood pressure stabilization, return of heart rate to baseline, seizure control, fever resolution, and normalizing ICP.

Our two patients demonstrate a Dantrolene responsive extended dysautonomic syndrome (DREDS) that has not been previously described and resolved after receiving Dantrolene. Diencephalic storming is the closest diagnosis for these patients; however, their symptoms were not paroxysmal but persisted with evidence of progressive escalation in symptoms until Dantrolene administration. Dantrolene should be considered in patients who present with persistent dysautonomia that does not respond to medical management for other clinical etiologies or other previously described dysautonomic syndromes.

Catheter-related bloodstream infections occurring in the intensive care unit (ICU) are common, costly, and potentially lethal. Specific prevention strategies and improved guidelines for the use of intravascular devices can decrease the rate of the infection.

A multidisciplinary group of clinicians reviewed C-RBSI rates in the Neuro ICU (NICU) and developed strategies to decrease our C-RBSI rate. Initially, we focused on improving compliance with maximal barrier precautions through the use of a central line (CL) insertion box containing all necessary supplies. Additional interventions were implemented over the next year as part of a larger hospital-wide initiative. These interventions included implementation of a central line bundle, antiseptic-coated catheters, and Biopatch use. Information on C-RBSI prevention strategies were shared with the multidisciplinary team through a variety of educational strategies. The unit's C-RBSI rates were shared frequently with the bedside clinicians, and were compared to the National Healthcare Safety Network (NHSN) mean for similar ICUs.

The NICU's C-RBSI rate in 2006 was 6.24 per 1000 catheter days (NHSN mean for similar ICUs of 3.5). The rate decreased to 3.73 per 1000 catheter days in 2007; this decrease was related to a marked decline in infections following implementation of the CL insertion box. Because the NHSN definition of C-RBSI changed in January 2008, rates cannot be directly compared. However, continued improvement has been noted, with the rate of C-RBSI decreasing from 3.01 in 2008 to 1.85 in 2009 (NHSN mean of 2.5). In April 2010, the NICU marked one year without a C-RBSI.

An evidence-based intervention resulted in a large and sustained reduction in rates of catheter-related bloodstream infection that has been maintained over time.

Posterior reversible encephalopathy syndrome (PRES) is thought to be secondary to a disruption of the blood-brain barrier when a sudden elevation in systemic blood pressure exceeds the autoregulatory capacity of the cerebral vasculature leading to cerebral edema [1] . Cases of recurrent PRES have been described in many conditions including end stage renal disease (ESRD) [2, 3, 4, 5] . To our knowledge, this is the first case in which a patient on chronic hemodialysis (HD) developed PRES that led to recurrent cerebral edema and herniation with each subsequent HD treatment leading to death.

A 20 year old man with ESRD, on HD for three years with the same dialysate, was admitted for PRES. He was made normotensive and became asymptomatic. Two days later, during his scheduled HD he developed a severe headache; HD was aborted, and he returned to baseline. HD was reattempted on day five during which he became unresponsive, developed bilateral extensor posturing and non-reactive pupils. No fluctuations in blood pressure or change in the components of the dialysate were noted. Serial CT scans of the brain showed worsening cerebral edema leading to tonsillar herniation. An EVD was placed and he received 23.4% NaCl with normalization of the ICP, yet he remained comatose. MRI brain noted further worsening cerebral edema while MRV was unremarkable. On day ten he expired.

We believe this to be the first report of recurrent cerebral edema and herniation during HD in a patient presenting with PRES. This may be secondary to failure of blood-brain barrier reconstitution leading to osmotic disequilibrium with each HD treatment and worsening cerebral edema.

In patients with PRES, one should consider delaying HD to allow for return of cerebral autoregulation. Further studies investigating the risks and benefits of delaying HD in such patients are warranted.

Specific neuroimaging findings, such as the "white cerebellum" and "reversal sign" have correlated with poor outcomes in patients suffering global ischemic injury. Here we report an imaging abnormality, bilateral hippocampal hyperintensities observed on fluid attenuated inversion recovery (FLAIR) and diffusion weighted imaging (DWI) magnetic resonance imaging (MRI) that was discovered incidentally in eighteen comatose patients after cardiac arrest, and may portend a poor prognosis in this population.

Methods 80 patients who were comatose subsequent to a cardiac arrest underwent MRI, including DWI. Qualitative regional analyses of these images were performed. Patients were placed into either the "affected" (n=18) or "not affected" (n=62) group based on whether or not they had bilateral hippocampal hyperintensities on DWI or FLAIR. Clinical outcomes were measured by the modified Rankin Scale (mRS) at 6 months, with poor outcome defined as a mRS of

Patients with bilateral hippocampal abnormalities had worse outcomes than those in the not affected group (p=0.014). A greater fraction of patients in the affected group had a poor outcome than in the not affected group (p=0.032). Only one patient in the affected group was alive at 6 months, remaining in a persistent vegetative state, giving this imaging abnormality a specificity of 100% for predicting poor outcome. Imaging in patients in the affected group indicated that these patients suffered more severe cerebral injury, with lower whole-brain apparent diffusion coefficient values (p=0.043), and a greater number of affected regions on both DWI (p=0.001) and FLAIR (p=0.001) than patients in the unaffected group.

Bilateral hippocampal hyperintensities on MRI may be a specific imaging finding indicative of poor prognosis in patients who suffer global hypoxic-ischemic injury.

Cerebral edema and raised intracranial pressure are common problems in neurologic intensive care. Osmotherapy, typically using mannitol or hypertonic sodium chloride (HTS), has become one of the first-line interventions. However, the literature on the use of these agents is heterogeneous and lacking in class I studies. We hypothesized that clinical practice would reflect this heterogeneity with respect to choice of agent, dosing strategy, and methods for monitoring therapy.

An on-line survey was administered by e-mail to members of the Neurocritical Care Society. Multiple-choice questions regarding use of mannitol and HTS were employed to gain insight into clinician practices.

A total of 295 responses were received, 79.7% of which were from physicians. The majority (89.9%) reported using osmotherapy as-needed for intracranial hypertension, though a minority reported initiating treatment prophylactically. Practitioners were fairly evenly split between those who preferred hypertonic saline (54.9%) and those who preferred mannitol (45.1%), with some respondents reserving HTS for patients with refractory intracranial hypertension. Respondents who preferred HTS were more likely to endorse prophylactic administration. Preferred dosing regimens for both agents varied considerably, as did monitoring parameters.

Treatment of cerebral edema using osmotically active substances varies considerably between practitioners. This variation could hamper efforts to design and implement multi-center trials in neurocritical care.

Procalcitonin (PCT) has been shown to have a positive predictive value in identifying significant bacterial infections. Procalcitonin is not increased by viral infections or by mediators in non-infectious inflammatory conditions. PCT in both serum and spinal fluid has shown to be elevated in meningitis. However, there is little data about PCT's predictive value in localized CNS infections. To investigate PCT's value in this cohort, we reviewed all localized CNS infections over a one year period that had PCT performed.

Retrospective review of all localized CNS infections that had PCT's levels obtained over a one year period. Localized CNS infection was defined as cerebral abscess (CA), spinal abscess (SA), cerebral or spinal wound infections (CSW) and ventricultis (V). A positive PCT was defined as > 1 ng/ml using a sensitive PCT assay.

24 patients met criteria: CA-8, SA-1, CSW-10, V-5. All patients with CA had PCT < 1ng/ml. The one patient with SA had a PCT of 3.5 ng/ml and subsequently found to have bacteremia from non-CNS site. All 10 patients with CSW had PCT < 1ng/ml. 4 of 5 with V had PCT < 1ng/ml. The exception had a UTI with bacteremia (PCT 1.6 ng/ml).

1) Excluding menningits, localized CNS infections do not appear to elevate PCT > 1ng/ml based on this limited series. 2) If PCT > 1 ng/ml, either meningitis or a systemic infectious process should be suspected.

Metronidazole, can cause central nervous system (CNS) side effects which are remarkable, but rare and usually reversible. We present a case of metronidazole induced encephalopathy (MIE) along with a systematic review of literature using Pubmed and Google Scholar.

A 53 year old man was admitted for Enterobacter species urinary tract infection. He was also on treatment for Clostridium difficile colitis with metronidazole 1.5 g daily for six weeks and vancomycin for two weeks which were continued. Three days later patient had slurring of speech with rhythmic chewing movements suspicious for seizures. Lorazepam stopped the rhythmic movements however patient was unresponsive and required intubation with intensive care management. Electroencephalogram (EEG) was suggestive of nonconvulsive status epilepticus warranting treatment with phosphenytoin and levetiracetam. Encephalopathy evaluation included cerebrospinal fluid examination and a metabolic profile that were unremarkable except for serum creatinine, 2.0 mg/dl. MRI brain on day 5 showed increased T2 and FLAIR signals within cerebellar dentate nuclei and splenium of corpus callosum suggesting MIE which prompted metronidazole discontinuation. MRI brain after one week showed improvement in previously noted changes. When off sedation, patient was alert and able to follow simple commands. Patient however could not be weaned off the ventilator and later died after family requested extubation and supportive care keeping in view the patient's wishes. MIE literature from Pubmed and Google Scholar was reviewed.

MIE affects men more than women; there is no relation to age or drug dosage. Sensorimotor (weakness, paresthesias) symptoms were reported by patients with onset of encephalopathy. Symptoms usually resolved except for peripheral neuropathy. Typical imaging abnormalities were seen which improved on discontinuing metronidazole.

Based on this review, it appears that the sensorimotor symptoms may indicate impending MIE. These symptoms and signs should prompt immediate withdrawal of metronidazole. MIE can be confirmed with typical imaging abnormalities.

Poster 112

Delirium is a common and serious condition that affects 64-80% of ICU patients. The acute insult and several treatments have been identified as dilirogenics. Delirium is associated with a 3x increase in death and a 10x increase in the rate of cognitive impairment at hospital discharge. Although there are a variety of validated ICU delirium assessment scales, there is scant investigation into the detection and management of delirium in ICU patients with PNI.

We searched Ovid Medline, PsychInfo, EBSCO, and CINAHL (key words: delirium, critical illness, intensive care, assessment, neurological injury). Twenty-seven papers investigated the assessment of delirium in the ICU.

Sixteen of the 27 (59%) papers report the validity and reliability of delirium tools designed for ICU patients. Fourteen of the 16 (89%) excluded patients with PNI and 15/16 (94%) excluded ICU patients with PNI. Of the 8 tools that exist for detecting ICU delirium, CAM-ICU (6/16 (38%) and ICDSC (3/16 (19%)) were the most studied. The DSM IV criteria for detecting delirium was the most widely used comparator (10/16 (63%)). The CAM-ICU (56%-96%) and ICDSC (50%-99%) were found to be most sensitive excluding patients with PNI. Only one study included critically ill patients with PNI. In this study, the prevalence of delirium in patients with PNI was 33% (69% had hypoactive delirium). The CAM-ICU and ICDSC showed sensitivity of 80% and 30% and specificity of 84% and 93% respectively for patients with PNI.

The CAM-ICU and ICDSC are the most widely investigated tools for detecting and monitoring delirium in the ICU and show the greatest accuracy in detecting delirium. However, there is too little evidence to suggest that delirium in patients with PNI can be detected and if a tool exists that can accurately detect delirium in the Neuro ICU.

Hyperglycemia is common in critically ill patients and is associated with poor outcome. Blood glucose (BG) control is essential for these patients to prevent neurological and systemic deterioration. Strict BG control with continuous insulin infusion is warranted but it carries a significant risk of hypoglycemia resulting in poor outcome.

The STG-22 TM (Nikkiso Co. Ltd., Tokyo, Japan) is a closed-loop BG control system that provides continuous BG monitoring and subsequent control of BG levels by the automatic infusion of regular insulin or glucose into the blood. Using this artificial endocrine pancreas system since August 2008, BG level was continuously monitored and controlled in a total of 34 critically ill neurological patients (14 subarachnoid hemorrhage, 6 cerebral infarction, 3 intracerbral hematoma, etc). All of them showed hyperglycemia above 180mg/dL before induction of BG control. Fourteen of them underwent induced hypothermia treatment.

BG level was strictly controlled at the target level previously determined (usually 100 -150 mg/dL). No hypoglycmic events were observed and BG fluctuation was minimized. Difficulties in drawing blood happened in some patients that were resolved by changing the peripheral venous route.

TM is safe and feasible in the setting of neurocritical care.

Edema and increased ICP remain a major cause of death and disability after TBI and stroke. Decompressive craniectomy (DC) is occasionally used as last resort, to control ICP and prevent herniation.

We present two patients who underwent DC and had ICP and TCD monitoring. Patient 1 suffered a severe TBI with a (L) SDH and underwent two (L) sided DCs. Patient 2 presented with a cerebellar ICH and coma. He underwent a suboccipital DC.

Patient 1 showed improved exam, FV and PI only after the second operation despite normalization of ICP/CPP post initial DC. Over the next few days we observed a progressive rise of FV. Patient 2 had a major exam improvement post SOC and demonstrated normalization of FV and PI over both cerebral hemispheres with no hyperemic response.

The intracranial pressure-flow dynamics pre and post DC have only recently been examined. CBF increases significantly post DC (1, 2) with an increase of microvascular blood volume (3) . These findings and the observation of deranged pressure reactivity (4), suggest a state of vasoparesis. FVs in our patient 1 increased significantly and remained elevated for several days. These increases could be attributed to cerebral vasospasm versus a hyperemic state. 1. Based on the normalization of PI, ICP, CPP and clinical improvement, we believe that DC induced a state of vasoparalysis with ensuing hyperemia. 2. Patient 2 had an improved profile over both MCAs after an SOC with normalization of PI and FV, hyperemia was not observed. The effects on blood flow may be different between supratentorial vs. infratentorial DC. 3. Patient 1 only clinically improved post repeat DC despite normal ICP/CPP after the first operation. Clinical improvement was matched by a decrease to normal of the PI. The PI may serve as a useful marker in the assessment of DC effectiveness.

Introduction ADEM is an uncommon and usually self-limited demyelinating encephalomyelitis. In most cases, management is limited to aggressive medical therapy with immunomodulatory treatments such as steroids, IVIG or plasmapheresis. In very rare cases, fulminant ADEM has required neurosurgical intervention including placement of an intracranial pressure monitor. Decompressive hemicraniectomy has rarely been reported as a therapy for cerebral edema in this condition. We now report our experience of a case of fulminant ADEM involving the dominant hemisphere.

A 41 year-old Caucasian man presented with the sudden onset of dysarthria and expressive aphasia one week after an upper respiratory tract infection. MRI and CT data suggested an acute, multifocal demyelinating process involving the left hemisphere, left brainstem, and bilateral middle cerebellar peduncles. The patient was treated with high dose steroids and IVIG with no clinical improvement. Extensive, hemispheric cerebral edema rapidly progressed over 5 days resulting in coma and impending uncal herniation despite sedation, hyperosmotic therapy, and external ventricular drainage of CSF. The patient was taken for emergent hemicraniectomy as a life-saving procedure after a family discussion.

Post-operatively, symptoms of herniation had resolved but the patient remained globally aphasic with moderate right hemiparesis. Biopsy from the left hemisphere showed multiple areas of perivascular demyelination. The patient slowly recovered strength on his right side but remained aphasic at the time of discharge to a rehabilitation hospital. 1 year after injury, the patient is walking independently, using both arms well, and writing to communicate. He recently began driving with supervision. Modified Rankin Scale 2 at one year.

This is the first case report describing the use of hemicraniectomy in dominant hemisphere, fulminant ADEM. Neurointensivists should consider neurosurgical interventions when managing patients with fulminant ADEM even with dominant hemisphere involvement.

To analyze the risk of neurological events, predictive value of potential risk factors, and overall survival of adult patients treated with ECMO.

A retrospective cohort study was performed at the Mayo Clinic of all patients aged 18 years or older treated with ECMO for at least 12 hours (2002) (2003) (2004) (2005) (2006) (2007) (2008) (2009) (2010) . Main outcome measures included death, neurological event (as diagnosed by brain imaging, neurological examination, electroencephalogram, and/or necropsy), and final neurological diagnosis. Potential risk factors included age, sex, reason for ECMO (post-cardiopulmonary resuscitation vs. post-cardiothoracic surgery), pre-surgical/pre-CPR modified Rankin scale (mRS) and cerebral performance category (CPC) scores, duration of ECMO, trough systolic blood pressure, trough arterial oxygen pressure, and peak serum lactate. Statistical analysis was performed using logistic regression and the Cox proportional hazards model with time-varying covariates.

Results 87 adults (35 female (40%); median age 54 years, IQR 31; mean duration of ECMO 91 hours, IQR 100; ECPR, n=14, 16%) were treated. Neurological events occurred in 65 patients (75%, 95% CI 64-83%). Neurological diagnoses included subarachnoid hemorrhage, multiple cerebral emboli, seizures, hypoxic-ischemic encephalopathy, myoclonus, and brain death. Neurological events were significantly associated with serum lactate >3 units/L (HR 1.10, 95% CI 1.01-1.22, p=0.03) and nonlinearly with the duration of ECMO (p=0.04). Overall survival on ECMO was 40% (n=35, 95% CI 30-51%). CPR as a reason for ECMO treatment (HR 22.5, 95% CI 0.94-536), a higher pre-treatment CPC score (HR 2.37, 95% CI 1.23-4.60), and peak serum lactate >3 mmol/L (HR 1.11, 95% CI 1.06-1.16) were predictive of death.

A majority of adult patients treated with ECMO experience neurological sequelae. High serum lactate is significantly associated with neurological events and could be an indicator for major hypoxic ischemic events that involve brain injury. Age and sex should not be used to determine who will benefit from adult ECMO.

The reported incidence of deep venous thrombosis (DVT) in neurosurgical patients has varied from 10-80%. The utility of surveillance lower extremity doppler ultrasounds (LEDs) for asymptomatic DVT has been debated. The purpose of this retrospective chart review is to determine the incidence and risk factors of DVT in the neurosurgical intensive care unit (NSICU) as detected by LEDs.

A retrospective chart review was completed on all adult patients admitted to the NSICU for at least 48 hours who received a surveillance LED during a one-year period. Extracted data included demographics, medical history, presence of chemical prophylaxis, development of DVT, and known DVT risk factors. Data was analyzed to determine the overall incidence and independent risk factors for DVT in the NSICU.

A total of 110 patients met inclusion criteria. The most common admitting diagnoses were aneurysmal subarachnoid hemorrhage (40.9%), and intracerebral hemorrhage (20.9%). DVTs were detected in 11.8% of patients by a median of 6 days post-admission. Independent risk factors significant for DVT included BMI diagnosis (p <0.05). All but one patient were started on chemical DVT prophylaxis with a median time to initiation of 1 day post admission.

This study confirms that despite prompt initiation of chemical prophylaxis, DVT still occur. Our event rate corresponds with previous studies of asymptomatic DVT detection, and was higher than studies assessing symptomatic patients. The necessity or frequency of LEDs remains to be determined; however, it may be reasonable to consider if the length of stay is >5 days in high risk patients (BMI 

Hyponatremia has the potential to worsen cerebral edema in those with acute brain injury. Conivaptan is an intravenously-formulated vasopressin-receptor antagonist that has been shown to raise serum sodium (Na + ) by inducing aquaresis. We have used this novel agent to treat acute hyponatremia in the Neuro-ICU, employing bolus dosing without an accompanying infusion. The natremic and aquaretic response has been variable, and we sought to ascertain whether any baseline or dose-related variables predicted those who would respond best to this novel treatment modality.

We studied all patients given a bolus of conivaptan over a 3-year period in our NICU, excluding those receiving osmotic/diuretic therapy or repeat dose within 12 hours. We retrospectively collected baseline and follow-up Na + values (with response based on peak Na + obtained 4-18 hours post-dose), as well as urine output for 12 hours postdose. We defined responders as those with a rise in Na + by 4mEq/L or more, analyzing the first dose given to each patient separately to exclude any bias from multiple doses.

A total of 214 doses were given to 124 patients. The mean rise in Na + after the first dose of conivaptan was 4.8±4.0mEq/L, with 60% of patients responding (by 4mEq/L or more). Response was strongly associated with urine output (86% responded if 12-hour UOP>3600ml) and degree of baseline hyponatremia (70% if Na + <130mEq/L vs. 19% if Na + >135). There appeared to be a dose-response relationship with all receiving 40-mg responding vs. 20% among 10-mg doses. Few patients exhibited an acute rise of over 12mEq/L.

Conivaptan is effective at acutely raising Na + in the majority of patients with brain injury. As expected, the natremic response parallels the degree of aquaresis, which may be a useful bedside marker of response. Treatment may be less effective at raising sodium in those with milder (or absent) hyponatremia, raising into question its utility to induce hypernatremia (as has been proposed). The observed dose-response effect suggests that higher doses may be required for such a purpose.

Financial Support: Drs. Human and Dhar have received speaking and consulting honoraria from Astellas Pharma.

Patients in general ICU, undergoing mechanical ventilation, are at risk of acquiring VAP leading to possible increased mortality in this population. Usually the neurological patients experience a greater risk for pneumonia but the risk of VAP and its impact on mortality are not well described. Studies defining the specific features related to neurocritical patients should be conducted.

We performed a retrospective analysis over 1 year analyzing data from 62 patients in a neurological ICU. We compared data on the incidence of VAP, length of stay in MV, ICU stay and mortality.Patients with and patients without VAP were compared.

Of the 62 ICU patients who underwent invasive mechanical ventilation, 25.81% (16) of patients developed VAP. No significant differences occurred between the ages of the patients with and without VAP in the same way that mortality was not influenced by the incidence of VAP. We found 25% mortality in patients with VAP and 24% of patients without VAP. The hospital stay was significantly higher in patients with VAP to 29 days on average versus 19 days in the group without VAP.

Patients with VAP have no increased mortality compared to patients without VAP. Despite this, they show an increase in length on mechanical ventilation and ICU stay. These data differ from those found on studies with critically ill patients for non-neurological causes. Studies in the same hospital with patients population from different ICUs (neurological and non neurological) but with standardized protocols to prevent VAP should be encouraged to determine the true impact of VAP in these patient groups.

Introduction 40Hz auditory steady-state evoked potentials (40HzaSSEP) refers to a group of potential of 25ms intervals sine waves by 40 times per second sound stimulus. It is easy to read and its anatomic origins contains cotex and brainstem. 40HzaSSEP has been used in the prognosis of traumatic brain injury and the monitoring of brain death patients, Serafini G et al. found that 40HzaSSEP had high sensibility and specificity for a fetal prognosis. But it has not been studied in neurological critically ill patients.

We retrospected 172 patients admitted to N_ICU of Nanfang hospital from March 2008 to December 2009. 40HzaSSEP, BAEP and lower limb SLSEP was finished within 24 hours. The prognostic value of 40HzaSSEP was analysed for the outcome of patients leaving N_ICU(death or survivor) and after three months. The outcome was evaluated with Glasgow outcome scale---4 to 5 being favourable outcome, whereas 1 to 3 being unfavourable outcome.

The bilateral absence of 40HzaSSEP prognosed strongly death and unfavourable outcome. The prognostic value of 40HzaSSEP showed no significantly difference among diffuse lesions , supratentorial and infratentorial lesions( P, 0.642/0.725). The sensitivity,specificity and accuracy of 40HzaSSEP which prognosed death during staying in N_ICU was 41.4%,97.4% and 78.5% respectively. The sensitivity,specificity and accuracy of 40HzaSSEP which prognosed unfavourable outcome three months after onset was 23.7%,100.0% and 46.0% respectively. The bilateral absence of V wave of BAEP is more sensitive than 40HzaSSEP and lower limb SLSEP to prognose death in N_ICU and three months unfavourable outcome with the sensitivity of 54.5% and 50% for . supratentorial diseases. The lower limb SLSEP is also significantly valuable to prognose outcome, but no better than 40HzaSSEP.

40HzaSSEP is valuable to prognose outcome of neurological critically ill patients in N_ICU with the advantage of easy judgement and waves steady.

Extraventricular drains (EVD) are essential in the treatment of patients with neurological injuries. The use of prolonged antibiotic prophylaxis is not recommended by recent guidelines, however their use remains high. This study determined the impact of prolonged courses of antibiotic prophylaxis on the incidence of ventriculitis and antimicrobial resistance.

Consecutive adult patients that required EVD insertion and received prophylactic antibiotics were identified retrospectively via an institutional database. Patients with a brain abscess or meningitis were excluded. Patients were stratified into two groups: those receiving short course antibiotic prophylaxis ( prolonged prophylaxis (>24 hours). The primary outcome was to compare the incidence of ventriculitis and resistant secondary infections between the short and prolonged prophylaxis groups. A secondary infection included any noncentral nervous system-related infection (e.g., pneumonia). A resistant infection was defined as penicillin-resistant streptococcus, methicillin-resistant staphylococcus, or a gram-negative organism resistant to cefepime, imipenem, levofloxacin or piperacillin-tazobactam. Centers for Disease Control criteria were used to define all infections.

Data from 101 patients were collected (short: n=54; prolonged: n=47). Demographics were similar between groups. The median duration of antibiotic prophylaxis in the short group was 1 (1-1) day and 4 (2-13) days in the prolonged group (p<0.001). The most common prophylactic antibiotic was cefazolin (short: 82%; prolonged: 66%, p=0.075). The incidence of ventriculitis was 13% (7/54) and 12.8% (6/47) in the short and prolonged groups, respectively (p=1.00). There were a total of 46 secondary infections [short: n=22/54 (41%); prolonged: n=24/47 (51%), (p=0.299)]. The incidence of resistant secondary infections was 3.7% (2/54) in the short group and 6.4% (3/47) in the prolonged group (p=0.661).

Patients with EVDs should not receive prolonged antibiotic prophylaxis because there is no benefit in preventing ventriculitis. Although a significant increase in resistant secondary infections was not observed, proper antimicrobial stewardship should be maintained.

The incidence of ventriculitis related to external ventricular drains (EVD) averages 8-10%. Attempts to reduce infection have been peri-operative antibiotics, continuous antibiotics and antibiotics-coated catheters with or without antibiotic prophylaxis. Antibiotic catheters are more expensive and cost/benefit ratios have not been conclusively established.

Our center has traditionally used peri-operative with continuous antibiotics and non-antibiotic coated catheters. To evaluate our practice and to determine if the data justifies continuation of this practice and help determine if a change to antibiotic coated catheters is cost effective.

A retrospective chart review of 166 patients with EVD placement for non-infectious causes. This study was approved by the institutional IRB.

116 patients reviewed. 8.6% did not receive antibiotics. Of the remaining 91.4%, 2/3 received peri-operative and continuous antibiotics; 1/3 received continuous antibiotics only. Based on CDC criteria, the ventriculitis rate in the antibiotic cohort was 17.2%. None of the 8.6% that failed to receive antibiotics developed ventriculitis.

Of the 20 cases of ventricultis, 16 had microbiologic data. All but two were multi-resistant organisms and specifically resistant to the prophylactic antibiotic. Comparing duration of catheters, those that became infected were in 4 days longer than the non-infected group (p > 0.001).

1. Our standard policy of peri-operative and continuous antibiotics is followed only 2/3's of the time. 2. Ventriculitis rates were unaffected by our antibiotic policy.

3. This study reaffirms that catheter duration remains a major risk factor for developing ventriculits. 4. Consistent with other data, antibiotics may have contributed to multi-resistant selection. 5. A change of practice of no antibiotics prophylaxis would result in a reduction of cost and offset the cost of antibiotic coated catheters. 6. Although underpowered, this data suggest, though not statistically significant (p=0.2), that antibiotics may in fact increase risk of developing ventricultis.

Invasive mechanical ventilation is required in one third of patients with Guillain-Barré syndrome (GBS). There are few early indicators of subsequent progression to respiratory failure. We hypothesized that anxiety was associated with respiratory failure.

At admission, anxiety was assessed with State-Trait Anxiety Inventory (STAI) and visual analogical scale (VAS, ranging from 0 to 10). Patients were also asked to declare if breathlessness, pain, weakness or uncertainty was the most stressful. These tests were performed in only GBS patients who were not mechanically ventilated at admission.

Ninety-nine GBS patients were included, among whom 22 (22%) had subsequently required mechanical ventilation. STAI was 47±14 and anxiety VAS 5.3±2.9. Mechanically ventilated patients had at admission a significantly higher disability grade (3.3±0.8 vs 3.8±0.7, p=0.0004) and lower vital capacity (78±23 vs 60±21 % of predicted value, p=0.002). STAI tended to be greater in patients who will require mechanical ventilation (47±14 vs 53±14, p=0.08), but not anxiety VAS (5.2±2.9 vs 5.5±2.9, p=0.77). Patients who will require mechanical ventilation complained more frequently of uncertainty.

Anxiety is frequently intense in GBS patients and needs to be considered by physician in charge.

Ventilator-associated pneumonia (VAP) is associated with a longer duration of mechanical ventilation, intensive care unit length of stay, hospital length of stay, and an increase in mortality. Several recent studies suggest an increased risk for the development of nosocomial pneumonia in patients receiving stress ulcer prophylaxis, more specifically associated with the use of proton pump inhibitors. The data are limited to general medical and critically ill cardiothoracic patient populations. The objective of this study is to compare the incidence of VAP in neuro-critical care patients prescribed a histamine-2 receptor (H 2 RA) antagonist versus a proton pump inhibitor (PPI) for stress ulcer prophylaxis (SUP).

This was a retrospective, cohort analysis of all mechanically ventilated patients admitted to the neuro-critical care unit between July 1 st and December 10 th , 2009 who were prescribed SUP with either an H 2 RA or PPI. A total of 607 patients were identified through a query of the university healthcare database. Of these, 219 were eligible for data collection and analysis. The final data set contained 103 patients in the H 2 RA group and 116 patients in the PPI group.

VAP occurred in 4 (3.9%) patients in the H 2 RA group and 13 (11.2%) patients in the PPI group (OR, 3.1; 90% CI, 1.2 to 8.2; p = 0.04).

The use of PPIs for SUP in the neuro-critical care patient population was associated with an increased rate of VAP development.

Neuroendocrine response to critical illness is divided into acute and chronic phases. Acute phase lasts up to 7 days and is characterized with increased ACTH, cortisol, GH, TTH, T4, LH, prolactin, and decreased T3, testosterone, IGF-I. After one week these levels fall, except for cortisol, which progressively increases. We studied the pattern of neuroendocrine stress response in critically-ill postoperative neurosurgical patients.

Study includes 9 patients: 4 pituitary adenomas, 3 craniopharyngiomas, 2 suprasellar meningiomas. There were 5 females, 4 males. Age ranged from 19 to 61 years old; all were treated with exogenous glucocorticoids. All nine patients had critical illness postoperatively, with varying degrees of delirium, water-electrolytes disturbances, and dynamic ileus; arterial hypotension with vasopressors in 7 patients, and prolonged mechanical ventilation in 5 patients. We investigated GH, TTH, T4, T3, LH, testosterone, prolactin, IGF-I on 8:30 am of 2 nd postoperative day.

On post-operative day 2 levels of GH, TTH, T3, LH, and testosterone were decreased in all patients. Prolactin was elevated in 2 and decreased in the rest. IGF-I was normal in 4 patients and decreased in 5. T4 was normal in 2 patients and decreased in the rest. Mean cortisol was highly elevated at 540 ± 441 nmol/l. Duration ICU stay was 14,1±9,9 days. Outcome at discharge from Institute was assessed using modified Rankin scale: death (1), moderate disability (3), slight disability (4), no significant disability (1).

Postoperative patients with sellar region tumors do not mount the typical neuroendocrine stress response to critical illness. Suppression of hypothalamic and pituitary function may be the result of injury to these structures of exogenous glucocorticoids administration. Careful evaluation of neuroendocrine function and appropriate hormone replacement therapy may improve outcome in this high-risk patient population.

An accurate non-invasive bedside method to detect intracranial hypertension will permit identification of impending neurological deterioration as well as reduce costs and morbidity associated with the use of invasive intracranial pressure (ICP) monitoring by selecting patients most likely to require invasive monitoring and treatment. The optic nerve sheath is a continuation of the dura and an increase in Optic Nerve Sheath Diameter (ONSD) as detected by Optic Nerve Ultrasonography (ONUS) has been shown to correlate with elevated ICP. 1,2,3 Some studies have proposed ONSD>5.2-5.8mm to identify ICP>20mmHg. 1, 2, 3 Methods Patients in the neuroICU with either ventriculostomy catheters or intraparenchymal ICP monitors underwent ONUS using a SonoSite TM portable ultrasound machine and a linear 10-5MHz transducer, performed by the attending neurointensivist. ONSD measurement was performed 3mm behind the retina and was correlated to simultaneous invasive ICP measurement. Comparison was also made between specific radiological signs suggestive of high ICP on CT/MRI scans performed within a 24 hour period and ICP>20mmHg by invasive monitoring.

296 ONSD measurements correlated to simultaneous invasive ICP measurement were performed on 36 patients. Diagnoses included Traumatic Brain Injury, Subarachnoid Hemorrhage, Intracerebral Hemorrhage, Ischemic Stroke and Brain Tumor. ROC curve analysis yielded an optimal ONSD of 4.7mm in identifying ICP>20mmHg-sensitivity 95.18% (95%C.I. 88.1-98.7%); specificity 92.02% (87.5-95.3%); AUC 0.97 (0.94-0.99); p(Area=0.5)<0.0001. Sensitivity of the higher cutoff of 5.2mm proposed by some authors was only 53.01% (41.7-64.1%), with specificity 97.65% (94.6-99.2%). Accuracy of CT/MRI features in identifying ICP>20mmHg was poor-sensitivity 75% (47.62-92.73%) and specificity 28.57% (11.28-52.18%).

ONSD measurement is an accurate, non-invasive, bedside method to identify ICP>20mmHg and can be performed by neurointensivists using commonly available portable ultrasound machines. ONSD>4.7mm has the greatest accuracy, however, institutional validation of ONSD criteria seems to be required, as with other ultrasound techniques.

Approximately 20% of patients with myasthenia gravis (MG) will suffer from a myasthenic crisis during the course of their disease. The prognosis of myasthenic crises has declined in the last 50 years from 40% to 5-10%. We present a 10 year retrospective review of our patients with myasthenic crises.

Cases of myasthenic crisis from 2000 to 2009 were identified from the ICU register and case notes were retrospectively reviewed.

44 patients were admitted in myasthenic crisis during this period. 5 patients had multiple admissions. The mean age at admission was 55 years. 29 were female and 15 male. Mean duration of MG at ICU admission was 2.6 years. 84% of patients were acetylcholine receptor antibody positive and 16% were seronegative (4 patients were anti MuSK positive). 30% of the patients had a previous thymectomy. All cases required ICU admission because of respiratory failure, with 68% requiring mechanical ventilation (mean duration 16 days). Infection was identified as the precipitating cause in 15 cases (35%). 80% of patients were treated with IVIg and 11% received plasma exchange. The mean duration of ICU stay was 18.5 days. There were 3 reported deaths all due to sepsis (7% mortality).

Myasthenic crisis is a serious complication of MG with significant mortality. Sepsis remains a potent precipitant of respiratory failure in MG and contributes to the mortality of such patients. The 7% mortality in this series may reflect the referral patterns to a specialist neuromedical ICU.

ICUAP is a frequent and severe complication of critical illness (ref 1). Its pathophysiology involves unloading that may partly result from sedation. Our objective was to assess whether sedated patients who will develop ICUAP can be early and clinically identified.

Patients requiring midazolam infusion (± sufentanyl) for at least 12 hours were included. Patients with preexisting neurological disorder or treated with neuromuscular blocking agent were excluded. Neurological examination included the Glasgow Coma Scale, the Richmond Agitation Sedation Scale (ref 2), the Assessment to Intensive Care Environment scale (ref 3) and the assessment of biceps and patellar tendon reflexes. Demographic data, diagnosis, SAPS-II, daily SOFA score and sedatives doses were collected. The first neurological examination was performed between the 12 th and 24 th hour of sedation. When patient regained consciousness, limbs muscle strength was assessed with Medical Research Council score (MRC; ref 4). ICUAP was considered when MRC score was < 48/60.

148 patients were included, among whom 81 (54%) patients regained consciousness. At inclusion, tendon reflexes were all preserved or abolished in 54 (68%) and 27 (33%) patients. Thirty (37%) patients developed ICUAP, which was significantly more frequent in patients with abolished tendon reflexes (56% versus 28%, p=0.02). ICUAP correlated with mean SOFA but not with 24 hours-cumulative sedative dose. Sepsis (OR= 3.23, 95%CI [1.03-10.2]; p=0.045) and abolished tendon reflexes (OR (by abolished reflex) =1.81 [1.04-3.13]; p=0.035) were independently associated with ICUAP.

Sedated patients with abolished biceps and patellar reflexes within the 24 hours of sedation are at higher risk to develop ICUAP. Abolition of tendon reflex might reflect muscle tone alteration by midazolam. Titrating sedation to preserve tendon reflex might be a strategy to prevent ICUAP.

We are not aware of any specific MRI findings associated with raised intracranial pressure (RICP). We noted a symmetric pattern of reduced diffusion in the deep gray nuclei (DGN) of two patients with prolonged RICP (Figure) . We aimed to study brain MRI diffusion changes in patients with RICP.

Patients with an ICP monitor and a brain MRI in the Stanford Neurocritical Care database 2003-2009 were reviewed.

In addition to our two index cases we identified eight patients with reduced diffusion affecting the DGN bilaterally. Clinical variables are described in the table. None of the patients had episodes of hypotension or hypoxemia, but all had had prolonged episodes of RICP caused by various pathophysiologic mechanisms. Mean ADC values in the caudate nucleus were lower than in normal controls (p putamen and thalamus. 

The deep gray nuclei, in particular the caudate nucleus, may exhibit reduced diffusion in the context of prolonged RICP. We suspect that these changes develop over days to weeks after onset of RICP. The clinical implications of this newly identified MRI pattern deserve further study. 

Newer monitoring techniques like continuous external cardiac output monitoring are being used to monitor fluid resuscitation. We hypothesized that External Cardiac Output monitoring correlates with hemodynamic parameters and blood markers of peripheral perfusion in neurocritical care patients. We tested its correlation against commonly used measures of volume status in Neurointensive care.

We reviewed charts of 14 patients monitored using the Vigileo system(Edward LifeSciences ).906 hours of monitoring data with the Vigileo Monitor cardiac output (eCO), BUN, creatinine(C), blood pressure, central venous pressure(CVP) values were collected. Cardiac output(Q) was also estimated using pulse pressure as surrogate for stroke volume(Q=2ml Pulse Pressure Heart rate). Diagnoses included subarachnoid hemorrhage, hemispheric stroke, intracranial hemorrhage,meningitis, epidural abscess. 5 patients were consistently on vasopressors, 5 intermittently, 2 on antihypertensives, 2 only on intravenous fluids. All values were evaluated as independent values with Pearson's correlation coefficient.

Average 

Indocyanine green (ICG) dye dilution technique and near infrared extinction measurements (NIRE) offer the feasibility for cerebral blood flow (CBF) monitoring. Otherwise, ICG generates cytotoxic decomposition products, when activated by light. We investigated potential cytotoxic effects of the concomitant use of ICG and NIRE on the involved cell types, in the first instance on endothelial cells.

Toxicity of ICG was determined by exposing human endothelial cells (Ea.hy 926) to different concentrations. Light emitted at 975, 850, 810 and 780nm wavelength was applied for an irradiation period up to 120h. To evaluate a potentiating effect of the combined use of ICG and NIRE, irradiation trials were repeated in the presence of ICG containing growth medium. Cell survival was tested using Crystal Violet test of cell viability.

Neither a 24h exposure to ICG up to c=53.7 mol/l nor continous laser-irradiation for up to 5 days alone affected survival of the Ea.hy culture. Results for combined exposure of the cell culture to ICG (c=53.7 mol/l ) and irradiation can be summarized as follows: Short time exposure to ICG of 2 x 10 min per day with an irradiation period of 5 days yielded a cell survival of 97.4%. If the ICG containing supernatant was removed after 24 hours, followed by 5 days continous irradiation, cell survival raised to 106%. Extending exposure times to ICG with combined irradiation to 64 hours showed no decrease in cell viability. However, optical assessment documented an increasing number of cell death after 70 hours. After 120 hours, cell survival decreased to 31.2%.

With regard to the low dose rate needed for CBF measurement, the high protein binding of ICG and a half-life of only 4 mins, our results document a safe viability profile for ICG and its concomitant use with NIRE for endothelial cells.

Financial Support: None

The "ideal" placement of a ventriculostomy (EVD) is considered the ipsilateral frontal horn. In this study, we analyzed the long-term patency of EVDs placed at the bedside based on the initial placement location and the reason of failure.

We prospectively examined all bedside EVDs placed within 18 months. We categorized those EVDs by location and monitored which catheters became non-functional and needed to be replaced. EVD locations were categorized as: "ideal" (ipsilateral frontal horn), third ventricle, ipsilateral body of the lateral ventricle, contralateral ventricle, partially parenchymal (catheter tip deep in the parenchyma, but draining a ventricle), completely parenchymal.

A total of 142 catheters were placed in 103 patients. Twenty-six patients had more than one catheter inserted: 23 were replaced and 3 were placed at separate times during the same admission. Of the 23 EVDs replaced, 6 were considered a "failed procedure," meaning the catheter was either proven or assumed to be extraventricular and had to be replaced within 24 hours. Seven patients had patency problems related to their catheter being inadvertently removed. Two EVDs were replaced for CSF infection. Only 8/103 (7.7%) catheters truly malfunctioned after an initial patent interval (7 with clotted blood at the tip of the catheter and 1 malfunctioning for an unknown reason). Seven out of 45 catheters (13%) in the "ideal" location had patency problems compared to 4/17 (24%) of those in the 3 rd ventricle, 2/12 (17%) of catheters in the ipsilateral body, 0/8 (0%) of those crossing midline, 2/13 (15.3%) of those parenchymal but draining and 5/5 (100%) of catheters that were completely intraparenchymal.

"Ideal" placement occurs only 44% of the time, yet catheters that are initially successfully placed only fail 7.7% of the time.

Since the CMA 600 cerebral microdialysis (MD) analyzer received FDA approval for clinical use in the United States (2005), cerebral MD has gained increasing acceptance as part of routine multimodal monitoring in patients with TBI, SAH, and stroke. Nonetheless, technical challenges involving placement of cathethers by neurosurgeons and maintenance by the nursing staff slow the widespread use of MD. We report the evolution of our placement techniques for the MD catheters from 2005-2010.

Patients undergoing craniotomy for TBI, ICH, or SAH were preferred candidates for placing MD catheters. CMA 70 catheters were implanted via a 1mm corticectomy 2 cm below the cortex. Preference was for placement of catheters in the frontal and parietal lobes. If a craniotomy was not performed, catheters were inserted via twist drill hole, Camino or ICP express bolt, or via double lumen Licox bolt. Post-operative CT scans documented MD catheter location. Catheters implanted into the brain were perfused with artificial CSF (0.3 ul/minute). Samples were collected hourly, and lactate, pyruvate, glutamate, glycerol, and glucose levels were analyzed with CMA-600 or ISCUS-Flex analyzers.

238 MD catheters were placed in 166 patients. 177 were placed directly into the brain at craniotomy, 32 via burr/twist drill hole, 31 via bolt. Accuracy of catheter placement was one problem encountered. Leakage from the exit site was eliminated by placing sutures around the CMA 70 catheter hub. With the bolt or burr/twist drill methods, pia-arachnoid was opened to prevent migration of catheters into the subdural space. Measurements of the bolt lengths was necessary to ensure proper placement. Peri-catheter bleeds (n=3) were seen with bolt but not open placement techniques.

MD catheters may be safely placed via an open or bolt technique. Careful measurements and securing the catheters will facilitate care by the nursing staff and will increase reliability of the measurements.

The treatment of patients with traumatic brain injury is challenging. In comparison to young patients, elderly patients consistently have poor outcomes after TBI. We placed cerebral microdialysis (MD) catheters in patients with TBI, and performed a retrospective review of 10 young patients (ages 18-30) and 10 elderly patients (ages 70-90) to determine if there were differences in lactate/pyruvate (LPR) and glucose as a function of CPP and ICP.

In the young group, the mechanisms of injury were: motor vehicle accident (n=5), pedestrian vs train (n=1), assault (n=1), fall (n=2), skateboarding fall (n=1). Primary pathologies included: SDH (n=3), contusion (n=2), SAH (n=2), depressed skull fracture (n=1). Surgical treatments included decompressive craniectomies (n= 7) and craniotomies (n=3), both with blood clot evacuation. In the elderly group, the mechanisms of injury were ground level falls (n=8) and falls from a ladder (n=2). The primary pathologies were SDH (n=7) and contusion (n=3). Surgical treatments included decompressive craniectomies (n=3) and craniotomies with blood clot removal (n=7). Microdialysis catheters implanted into the brain at surgery were perfused with artificial CSF (0.3 ul/minute). Samples were collected hourly, and lactate, pyruvate and glucose levels were analyzed with CMA-600. We defined a data set as the hourly microdialysis values with corresponding ICP/CPPs. The LPR is a well-established measure of cerebral stress. We placed an average of 1.89 catheters/patient. Total number of data sets in this analysis was 2999 for an average of 150 sets/patient. Mean ICP for the young and elderly group was <15.0mmHg.

Mortality was significantly higher in the elderly (60%) as compared to the young(0%) group, although LPR and glucose values were similar in both groups. This raises questions about the relationship between cerebral metabolites and outcome.

Oxidative stress after traumatic brain injury (TBI) produces significant pathology. The gold standard measure of oxidative stress measures is the oxidized product of arachadonic acid, 8-Isoprostane (8-IsoP). We sought to determine whether 8-IsoP levels correlated with severity of injury and outcome in patients with TBI.

Patients with severe TBI (GCS<9) who receive ventriculostomies for clinical care were recruited into the study. Samples of CSF, plasma and urine for 8-IsoP analysis were taken at time of ventriculostomy placement (within 6 hours of injury) and 24 hours later. Levels of 8-IsoP were measured by enzyme immunoassay (EIA). Recruitment is ongoing for the study.

Seven TBI patients and seven control ED CSF samples (from patients receiving LP for other reasons) have been recruited. Initial data from 8-IsoP EIA analysis shows ten-fold elevations in 8-IsoP (peak at 200pg/ml) with most patient samples decreasing 50% by 24 hours. The patients with the highest levels had more resultant disability at discharge and the patient with lowest level was discharged with no neurological deficit. Additional patient recruitment, plasma analysis and control patient analysis of 8-IsoP are ongoing.

Oxidative stress causes significant pathology in patients with TBI. Preliminary data suggest that levels are highest in patients with more severe TBI and that patients with lower levels or faster decrease have better outcomes. Research is ongoing, but the basis and timing for successful antioxidant therapies may be found in better understanding the pathophysiology and time course of oxidative stress in patients with TBI.

Financial Support: None

The incidence and significance of thrombocytosis in severe traumatic brain injury has not been well established, although it has been suggested that its presence may be associated with adverse sequelae. The purpose of this study is to investigate the incidence and clinical significance of platelet counts 600,000/uL or greater in severe traumatic brain injury patients.

A retrospective review of a traumatic brain injury database was used to investigate 122 severe TBI patients (GCS motor score 5 or less) between May 2006 and January 2010. Platelet counts during ICU stay, Injury Severity Score, ICU length of stay, mortality, and incidence of thromboembolism (including superficial and deep vein thrombosis, pulmonary emboli) was compared in those with and without thrombocytosis (plt 600k/ul or greater).

122 patients were included. Average age was 32 (range15 to 74). Thrombocytosis developed in 64 patients (52%). There were no incidences of pulmonary emboli. The average ISS was 30.5; avg 32 in those with thrombocytosis vs 29 in those without thrombocytosis. ICU length of stay was longer in those with thrombocytosis; 19.9 days vs 10.9 days (p <.01). Incidence of mortality was less in those with thrombocytosis; 0 vs 7. Incidence of thromboembolism was 16 (13%) overall; 7 (5.7%) in those with thrombocytosis vs 9 (7.4%) in those without thrombocytosis.

Thrombocytosis is a very common finding in patients with severe TBI. The occurrence of thrombocytosis is associated with higher injury severity and longer ICU stay, but lower mortality overall. The presence of thrombocytosis as an independent factor does not increase the risk of developing thromboembolism in patients with severe traumatic brain injury.

Poster 141

The purpose of this study is to explore the various factors affecting brain re-expansion after subdural hematoma evacuation.

A retrospective study of thirty-six patients undergoing subdural hematoma evacuation was performed between 1/2/2009 and 6/1/2010. Evacuation methods included formal craniotomy (28), burr holes (4) or subdural drain (4) . The data was aggregated by type of hematoma; acute (18), subacute (14) or chronic (4) .The records were also sorted by gender, 9 females and 27 males, and by age groups; 0-29 (5), 30-49 (3), 50-69 (7), 70-89 (21). The maximal distance of the displaced brain from the skull was measured in the preoperative and immediate post-operative period. The percent change formula [b-a /ABS (a)] was applied to the measurements obtained from each case. The percent change for each case was compared among groups based on the experimental variable being tested and the average was obtained.

The average percent change in brain re-expansion postoperatively was obtained for the following variables. All subdural types resulted in 18.87% increased expansion for formal craniotomy, 12.11% for burr holes and 16.67% for subdural drain. Chronic and subacute collectively produced 5.88% for formal craniotomy, 9.49% for burr holes and 16.67% for subdural drain. Across all procedures re-expansion was 26.88% for acute subdural hematoma, 0.06% for subacute and 0.20% for chronic. The outcome based on gender for all subdural types was 0.18% for females and 17.87% for males. The re-expansion after acute subdural evacuation alone was 0.19% for females and 28.44% for males. Chronic and subacute subdural evacuations produced 0.17% for females and 0.05% for males. Across all subdural types age-related changes were 0.22% for 0-29 year-olds, 0.46% for 30-49, 0.09% for 50-69 and 0.16% for 70-89, while for acute subdural evacuation 0.37% for 0-29, 0.46% for 30-49, 0.15% for 50-69 and 25% for 70-89. Chronic subdural evacuation, however, resulted in 11.53% change for 0-29, no cases for 30-49 age group, 0.04% for 50-69, and 0.10% for 70-89.

Clinical decision-making based on the postoperative CT scan in patients undergoing subdural hematoma evacuation is complex. Often the evacuated hematoma is replaced by fluid, air, and blood which collect in a space whose volume is a factor of the mechanical elasticity of an individual's brain. Gender, age, and type of procedure have a predictable effect on the immediate re-expansion of the brain. Further studies are needed to ascertain the long-term radiographic changes and ultimate clinical outcome as they relate to the post-operative cranial imaging.

Poster 142

Intracranial hypertension (ICH) is a crucial modifiable risk factor for poor outcome after traumatic brain injury (TBI). Limited evidence suggests that decompressive laparotomy may be an effective treatment for refractory ICH in patients who have elevated intraabdominal pressure.

Case report.

A 17 year old girl sustained severe TBI in a motor vehicle collision. On arrival to the ED, she had a GCS of 3 with 2mm, sluggishly reactive pupils. Initial CT showed diffuse cerebral edema, obliteration of the perimesencephalic cisterns, and right subdural hematoma with 8 mm midline shift. Admission APACHE II Score was 26. Resuscitation requirements in the first 24h included large volumes of blood products and crystalloids.

The opening ICP was 27 mmHg, which was controlled for the next 18h with standard medical therapy, including three doses of mannitol. Thereafter, over the course of 6h, she developed abdominal compartment syndrome (firm abdomen, hypoventilation, declining urine output, rising serum lactate, and a rising bladder pressure) and her ICP became refractory to medical therapy. This culminated in an ICP plateau wave of 35-40 mmHg, severe hypertension, bradycardia, and a bladder pressure of 32 mmHg (upper limit of normal = 12 mmHg). Emergent decompressive laparotomy was performed at the bedside, resulting in an immediate decline of her ICP to 5-10 mmHg.

The ICP began to rise again 36h later, but was responsive to intermittent mannitol boluses. After hospital day 10 she required no active ICP treatment. Two months after her injury she was steadily recovering in an acute rehabilitation facility; her GOS was 3.

Elevated intraabdominal pressure can exacerbate ICH in patients with TBI. Recognition of this condition and treatment with decompressive laparotomy can be useful in patients with ICH refractory to optimal medical therapy.

Frequently, MRI scanning is used to supplement the CT scan of the cervical spine in order to help exclude significant non bony pathology in the blunt trauma patient. This use of MRI has recently been questioned because of the perceived high diagnostic yield of the new helical CT scan. This retrospective study helps clarify the role of MRI of the cervical spine in blunt trauma patients who have a normal helical CT with reconstructions.

A retrospective chart review was done on all Trauma activations between January, 2007 through December, 2008 on patients with a blunt injury mechanism who had both a helical CT scan of the neck on admission and an MRI of the neck within 72 hours to "clear" the cervical spine of injury. Readings of both studies were reviewed to find those who had "normal" CT scan but "abnormal" MRI scan. The charts of those found to have "abnormal" MRI scans were then reviewed to classify the nature of the abnormality and to determine the clinical significance of the abnormality identified on the MRI but not on the CT scan.

A total of 431 patients were identified fulfilling the criteria of having a CT of the neck on admission and an MRI within 72 hours. Forty-two patients had a "normal" read on CT scan but an "abnormal" read on the MRI scan. Of these 42 patients; six had an MRI done to assess clinically evident abnormal neurologic findings, despite a normal CT scan. Twenty-five had the MRI for pain; 9 for mental status changes/unable to clear clinically and 2 for distracting injury, unclear clinical exam. There were 3 deaths: none due to unrecognized/untreated neck injury. None of the patients, except those with neurologic findings on clinical exam, required treatment for their neck except hard or soft collars for comfort. All collars were subsequently removed without further intervention.

Clearing the Cervical spine of significant structural injury can be safely done using helical CT with reconstructions. Adding an MRI scan will add further information regarding ligamentous integrity, but will not identify significant structural problems not already present on the CT scan.

To determine long-term outcomes in elderly patients that underwent evacuation of a traumatic intracranial hemorrhage (ICH).

This is an IRB-approved study using trauma registry data. The following was collected in patients age -2008: age, ISS, admit GCS, chronic vs. acute hemorrhage, ventilator days, mortality (both in-hospital and after discharge), prehospital use of warfarin and/or clopidogrel, and discharge functional independence measurement (d/cFIM). Functional outcome of survivors was determined by a phone interview using an Extended Glasgow Outcome Score (GOSE). Favorable outcome defined as surviving >1 year or a GOSE defined an unfavorable outcome. Analysis was performed using t-test and Chi-square. 

In elderly patients requiring evacuation of traumatic ICH, 2 out of 3 die within 1 year. However, in the remainder, most (22 of 32) have a favorable outcome based on GOSE. This study also demonstrates that in-hospital mortality is not an adequate measure of outcome.

Randomized trials have suggested that hypertonic saline solutions may be superior to mannitol for the treatment of elevated intracranial pressure (ICP), but their impact on clinical practice has been limited, partly by their small size. We therefore combined their findings in a meta-analysis.

We searched for relevant studies in MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), Scopus, and ISI Web of Knowledge. Randomized trials were included if they directly compared equiosmolar doses of hypertonic sodium solutions to mannitol for the treatment of elevated ICP in human subjects undergoing quantitative ICP measurement. Two investigators independently reviewed potentially eligible trials and extracted data using a preformed data collection sheet. Disagreements were resolved by consensus or by a third investigator if needed. We collected data on patient demographics, type of intracranial pathology, baseline ICP, osms per treatment dose, quantitative change in ICP, and prespecified adverse events. Our primary outcome was the proportion of successfully treated episodes of elevated ICP.

Five trials comprising 112 patients with 184 episodes of elevated ICP met our inclusion criteria. In random-effects models, the relative risk of ICP control was 1.16 (95% CI 1.00-1.34), and the difference in mean ICP reduction was 2.0 mm Hg (95% CI -1.6-5.7), both favoring hypertonic saline over mannitol. A mild degree of heterogeneity was present among the included trials. There were no significant adverse events reported.

We found that hypertonic saline is more effective than mannitol for the treatment of elevated ICP. Our meta-analysis is limited by the small number and size of eligible trials, but our findings suggest that hypertonic saline may be superior to the current standard of care, and argue for a large multi-center randomized trial to definitively establish the first-line medical therapy for intracranial hypertension.

Introduction Traumatic brain injury (TBI) results in multiple pathologies, including a significant reduction in cerebral blood flow. It has been demonstrated that endothelin-1 (ET-1) plays a major role in the induction of hypoperfusion following TBI, and that treatment with selective endothelin receptor A (ETrA) antagonists causes a decrease in both hypoperfusion and cellular injury, and improves cognitive outcome.

A rodent model of diffuse traumatic brain injury was used to assess the effects of IV administration of Clazosentan, a novel and clinically relevant selective ETrA antagonist, given at several time points following injury. Cerebral blood flow (CBF) (as measured using arterial spin labeling MRI), cellular injury (as measured using Fluoro-Jade labeling), and behavioral outcome (as measured using a radial arm maze-spatial learning task) were assessed following TBI.

Our results indicate that treatment with Clazosentan is effective in ameliorating hypoperfusion, preserving neuronal integrity and improving cognitive outcome after head trauma. Furthermore, while Clazosentan given 2 hours after head trauma provides a partial improvement of TBI-induced deficits, administration at 2 and 24 hours improves behavior to near sham control levels.

These data suggest that endothelin receptor antagonists with high affinity for ETrA may be effective in improving overall outcome after TBI. Furthermore, these results suggest that selective ETrA antagonists, such as Clazosentan, be incorporated in clinical trials for TBI treatment.

Traditionally outcomes of patients with traumatic brain injury (TBI) have been linked to primary neuronal injury and secondary injury as a consequence of hypoxia and hypotension. Severe sepsis is common in the intensive care units and can exacerbate secondary neuronal injury. However there is paucity of literature on the outcomes of TBI patients developing severe sepsis. We aimed to study the outcomes in this subset of patients using a representative sample of patients in United States.

A retrospective analysis utilizing the 2007 Nationwide Inpatient Sample was performed. Using the ICD-9-CM codes, all adult records (age 18 or more) with a primary discharge diagnosis of TBI (850-854) were identified. Patients with severe sepsis were similarly identified using appropriate ICD-9-CM codes. Multivariate logistic regression was performed using survey commands in STATA. The regression adjusted for age, sex, race and Elixhauser co-morbidity index. The primary outcome studied was all cause in-hospital mortality, length of stay and discharges to long term acute care facilities.

A total of 134,936 estimated primary discharges with TBI were studied. Out of these, 2792 (2.1%) had severe sepsis. The unadjusted mortality in patients with severe sepsis was 28.5% when compared to 7.5% in those who did not. On multivariate logistic regression, severe sepsis was a strong predictor of in-hospital mortality (OR 4.3; 95%CI 3.5-5.

3)

The LOS was 28 days longer in patients developing severe sepsis (95%CI 22.4 -33.9 days). The number of patients discharged to a LTC was significantly higher in patients developing sepsis (28% vs. 50%; p< 0.001).

Using nationally representative data, this observational study shows that severe sepsis is a strong predictor of mortality and worse outcomes in patients with TBI.

Continuous sedation is frequently used in patients with severe traumatic brain injury and cessation of sedation for daily exams is often performed. There is concern that this may cause tissue distress or secondary injury due to increased metabolic demand in patients with severe traumatic brain injury. Cerebral microdialysis can demonstrate tissue injury or distress by low extracellular glucose (<2mmol/l) and elevated Lactate/Pyruvate ratios ( L/P>40), which are biomarkers of cellular brain metabolism.

A retrospective chart review was performed on patients who were monitored with cerebral microdialysis, admitted to the neuroscience intensive care unit at Jackson Memorial Hospital with the diagnosis of traumatic brain injury. Hourly sampling of microdialysis fluid was performed and extracellular lactate, pyruvate (L/P) ratios, and glucose concentrations were recorded. Charts were then analyzed looking for changes of these biomarkers with cessation of continuous sedation and analgesia. Analyte measurements were compared for the hour of sedation/ analgesia change and the two subsequent hours.

Four patient's data were reviewed. Three males (ages 21,57, 60) and one female age 21. One patient had a penetrating injury and three had blunt traumatic injury. 

We did not identify a significant decrease in extracellular glucose or elevation of L/P ratios with daily cessation of sedation in 3 patients. However, our last patient had markedly elevated L/P ratios and decreased glucose with cessation of sedation, indicating cellular distress and tissue vulnerability. This patient's brain neurochemistry appeared to be more volatile as there were also fluctuations in the glucose and L/P ratio with fever. More patients need to be studied to determine if there are specific patient characteristics which distinguish patients with volatile brain chemistry. These patients might have a higher risk for secondary injury and may benefit more from tissue monitoring.

Introduction Traumatic brain injury (TBI) is a leading cause of morbidity and death. The Intrathoracic Pressure Regulator (ITPR) is a device that creates negative intrathoracic pressures of -9 mmHg in between positive pressure breaths. It has previously been shown to increase cerebral perfusion pressure (CPP) and reduce intracranial pressure (ICP) in a porcine model of TBI. The purpose of this study was to validate the long-term effect of the device after a sustained insult.

Eight isofluorane anesthetized pigs were subjected to focal brain injury by insertion of an epidural foley catheter. After target ICP (>22 mmHg) was achieved by inflating the foley balloon, animals remained untreated for 90 minutes. Subsequently, animals were treated with the ITPR (CirQlator™, Advanced Circulatory Systems) for 240 minutes. ICP, CPP and all key hemodynamic parameters were measured. Ventilator settings were adjusted to maintain constant arterial CO 2 tension to control for any chemical effects of CO 2 on ICP. All results are reported as mean ± SEM.

Mean ICP and CPP at baseline were 19 ± 1.4 and 50.3 ± 4.8 mmHg, respectively. A significant reduction in ICP during device use was observed at 60 min (16.5 ± 1.1, p=0.01), 90 min (16.5 ± 1.3, p=0.03), 120 min (15.0 ±1.0, p <0.01) and at 240 min (15.5 ± 1.1, p=0.01). CPP was significantly increased at 30 min (59 ± 6.8, p=0.02), 60 min (57.3± 5.2, p<0.01) and 120 min (59.6 ± 4.8, p <0.01). 48 hour neurological exams in all animals were unremarkable. No pulmonary edema was noted based on chest x-rays and post mortem lung wet to dry ratio.

Treatment with the ITPR demonstrated significant improvement in ICP and CPP without compromising lung function. The ITPR may provide clinical value in optimizing cerebral perfusion in states of elevated ICP in mechanically ventilated patients.

Financial Support: Dr. Metzger and Mr. Matsuura are employed the company that manufactures the device studied, Dr. Keith Lurie is Chief Medical Officer of the company that manufactures the device studied.

Ventriculostomy infections are expensive to treat and can cause serious harm to patients. Appropriate measures can be undertaken to prevent infections using proper technique and management strategies. We describe a protocol that resulted in a 0% infection rate in the Neuroscience Intensive Care Unit (NSICU) over a single academic year.

Methods 215 ventriculostomies were placed using a standard sterile procedure protocol and were maintained with a separate maintenance protocol. 1298 CSF samples were obtained by a neurosurgeon in the NSICU.

No infections were identified during the year that the protocol was enacted. 2 contaminated specimens were identified in 2 separate patients by delayed growth, the identity of the organism, and the absence of positive cultures before or after the contaminated specimens were drawn. One organism was a coagulase negative staphylococcus and the other was a gram positive rod.

Effective prevention of ventriculostomy infections using protocols for placement and maintenance can be obtained in the NSICU. The results are better patient outcomes and markedly reduced costs of medical care.

Introduction Traumatic brain injury (TBI) results in cellular injury ultimately leading to behavioral deficits. Among multiple pathologies, reduced blood flow and microglial activation contribute to the histopathology and behavioral deficits observed following head trauma.

A rodent weight drop-acceleration impact model of diffuse brain injury was used to test the effects of ameliorating hypoperfusion and suppressing microglial activation on the extent of cellular injury and behavior following TBI. Microglial activation was assessed by staining fixed brain sections with HRP-conjugated Isolectin B 4 (ILB 4 ). Blood flow measurements were made using arterial spin labeling (ASL) MRI on a 4.7-T horizontal-bore magnetic resonance spectrometer. Cognitive outcome was assessed using the radial arm maze behavior assay for rodents.

Treatment with BQ123, a selective endothelin receptor A (ETrA) antagonist, reduces the extent of TBI-induced hypoperfusion and cellular injury and improves behavioral outcome. Furthermore, administration of minocycline, a broad spectrum tetracycline antibiotic known to inhibit microglial activation, also improves outcome. Moreover, our results indicate that combined treatment significantly improves outcome over administration of either treatment independently.

The combined treatment with Clazosentan and Minocycline was more effective than single treatment in improving outcome following TBI, suggesting a synergystic effect of this treatment regime. These data suggest that combined treatment approaches targeted at controlling TBI-induced vasoreactivity and activation of microglia may be more effective than single therapies alone.

Packed red blood cell (PRBC) transfusion has been associated with medical complications and increased mortality in the ICU population. We sought to determine the effects of PRBC transfusion on in-hospital mortality in a cohort of traumatic brain injury (TBI) patients admitted to our ICU.

We designed a retrospective cohort study using a prospectively compiled and maintained registry (Cerner Project IMPACT). Patients with TBI who were admitted to our ICU within 24 hrs of injury between 2003-2008 were selected for the analysis. Data collected included demographics, admission physiologic variables, Glasgow Coma Scale (GCS), APACHE-II, scores, and total ICU and hospital length of stay (LOS). Primary outcome was in-hospital mortality and secondary outcomes were ICU and hospital LOS. To assess for the impact of PRBC transfusion on our primary and secondary outcomes, demographic and admission clinical variables were used to construct logistic regression models using the outcome measure as a dependent variable.

A total of 169 patients were selected. Finally, an interaction between PRBC transfusion and hematocrit level was not seen in our models.

This data suggests that in TBI patients, PRBC transfusion is independently associated with prolonged ICU-LOS but not with in-hospital mortality. The effect of the severity of brain injury determined by the admission GCS remains a powerful predictor of in-hospital mortality after TBI.

Recent evidence supports the use of ECMO in adults with severe respiratory failure that failed conventional support. Because of the need for anticoagulation, intracranial hemorrhage has been a contraindication, particularly if it is related to traumatic injury. However, with newer technology that allows less anticoagulation and platelet activation, ECMO use in this population needs re-evaluation. We report a case of VV-ECMO for severe respiratory failure with accompanying traumatic ICH.

Retrospective case review.

A 29 y/o fell from a tree suffering a TBI, numerous rib fractures and pulmonary contusions-GCS 8T. CT revealed an occipital/parietal hemorrhage that was stable over 24-hours based on serial CT scans. On day two, respiratory failure developed with rapid deterioration necessitating VV-ECMO support. After cannulation, the patient received heparin bolus followed by continuous infusion. Approximately 4 hours after initiating ECMO, another head CT was performed demonstrating no interval change. Heparin was maintained on a constant dosing strategy with ACT goals of 150-170sec. Platelet counts were kept >100,000/ L. A polypeptide/heparin coated (BIOLINE , Maquet, USA) circuit was used, engineered to prevent thrombocyte and thrombin adherence to artificial surfaces. Also, sedation was minimal using dexmedatomidine, ziprasidone and propofol. Sedation interruptions were performed every four hours to allow neuro-assessment. The patient was on ECMO 10 days with continued neurologic improvement. At time of decannulation the GCS was 11T. The patient made full neurologic recovery.

1. This case demonstrates that ECMO, with the use of coated circuits along with a minimal anticoagulation regime can be performed in patients with ICH. 2. Traumatic brain injury should no longer be considered an absolute contraindication for ECMO support. 3. Minimal sedation techniques can be utilized to allow ongoing neurologic assessment.

Secondary insults following severe traumatic brain injury (sTBI) are known to be associated with poor outcome. Automated continuously recorded vital signs have been shown to be superior to manual recordings in capturing these events. The purpose of this study was to investigate the use of automated continuously recorded vital signs early after injury to predict outcome following sTBI.

Patients with head AIS prospectively enrolled. Outcome was measured by Extended Glasgow Outcome Scale (GOSE) at 6 months. Continuous, automated, digital data was collected every 6 seconds. Five minute means were calculated and mean values over the duration of monitoring were recorded. Percent time (% time) and pressure time dose (PTD, mmHg*h) above and below thresholds (ICP>20 and >30 and CPP<50 and <60) were calculated and analyzed to compare their ability to predict 30-day mortality and functional outcome by ROC.

60 patients were enrolled. Demographics included: mean age 34 years (range 16-83), mean admission GCS of 6.4 ±3.1, mean head AIS of 4.2 ±0.7 and mean Marshall score of 2.5 ±0.9. The 30-day mortality rate was 13%. 35 patients had good neurological outcomes (GOSE>4). Moderate (CPP<60) and severe (CPP<50) cerebral hypoperfusion and moderate (ICP>20) and severe (ICP>30) intracranial hypertension as measured by %time or PTD were predictive of both mortality (AUC range 0.85-0.75, p<0.05) and poor neurological outcome (0.68-0.78, p<0.05). PTD of cerebral hypoperfusion better predicted poor functional outcome (Figure) . AUC 

Automated continuously recorded vital signs early after injury are useful in predicting outcomes in patients with sTBI. Cerebral hypoperfusion and intracranial hypertension are clearly associated with poor outcome. Automated processing of measures that account for both depth and duration of secondary insults shows promise in the development of predictive models for patients requiring intervention and has potential for the design of real-time bedside early-warning systems.

Management of severe traumatic brain injury (TBI) focuses on minimizing secondary insults, such as edema and ischemia. Both neuron-specific enolase (NSE) and S100-beta (S100b) are well-known markers of brain injury. We sought to determine the association of NSE and S100b in the CSF with traditional clinical markers of severity and poor outcome, namely intracranial hypertension (ICH) and cerebral hypoperfusion (CH).

Patients with head AIS>3, age>14, "isolated" TBI, need for intracranial pressure monitor, and injury deemed survivable were prospectively enrolled. CSF was collected within 24 hours of injury and twice daily for 7 days and analyzed by ELISA. Hourly intracranial pressure (ICP) and cerebral perfusion pressure (CPP) were recorded. Means over 12 hour periods of monitoring were calculated for each patient as were minimum values for CPP and maximum values for ICP. % time and pressure time dose (PTD, mmHg*hour) of ICP>20 and CPP<60 were calculated for each 12 hour period and compared to CSF values of NSE and S100b drawn prior to (PRE), during (MID), and after (POST) the 12-hour period of monitoring.

Of 55 patients enrolled, 23 had CSF collected for analysis. NSE and S100b concentrations in 254 samples were matched with clinical variables recorded for each 12 hour period. Patients had a mean age of 31.9 ±15.1. 87% were male. Mean head AIS was 4.4 ±0.6, Marshall score 2.7 ±0.9, and admission GCS 4.4 ±2.3. Mortality was 17.4%. PRE, MID, and POST S100b levels were associated with mean ICP (r=0.19 to 0.22, p<0.01) and mean CPP (r=-0.21 to -0.23, p<0.001). % time and PTD of ICP>20 were correlated with S100b levels and % time and PTD CPP<60 were associated with elevations of both S100b and NSE in the CSF (Table) .

Both S100b and NSE are associated with clinical measures of TBI severity. These markers of neuronal cell death demonstrate promise as both indicators of impending clinical deterioration and targets of future therapeutic interventions.

Financial Support: None 

We have used mild therapeutic hypothermia in patients with severe traumatic brain injury. In this study we investigated the effects of hypothermia on brain tissue oxygenation.

Brain tissue oxygen tension (PbtO 2 ) in addition to intracranial pressure (ICP), cerebral perfusion pressure (CPP), and jugular venous saturation (SjO 2 ) were monitored in 13 consecutive patients with a Glasgow Coma Scale score of 4 to 6 (mean age 36.9, range 16 to 64 years). Patients were cooled to a target temperature of 35.0 o C. Patients with good recovery and moderate disability on the Glasgow Outcome Scale were regarded as having favorable outcomes.

Seven patients had favorable outcomes and six had unfavorable outcomes. In nine patients, the temperature was 

These results suggest that decreasing the temperature to 35 o C can reduce intracranial hypertension while hypothermia below 35 o C may impair brain tissue oxygenation.

Hypertonic saline is used to reduce intracranial pressure in patients with traumatic brain injury (TBI). Aquaresis instead of the administration of high quantities of Na + might be a safe alternative. We hypothesized that conivaptan would be safe, will maintain stable serum Na + and decrease the overall Na + load.

Open label, randomized, controlled trial enrolling subjects within 24 hours of severe TBI. Patients were assigned to receive a single 20 mg dose of conivaptan (Vaprisol ) (n=5) or usual care (n=5). The primary endpoint was the evaluation of the drug safety profile as indicated by serum Na + increases greater than 1 mEq/hr, serum Na + levels above the target range, and adverse events. Secondary endpoints included: 48-hour Na + load, urine output and fluid balance.

Ten patients were included in the intention-to-treat analysis. The mean age was 50 13 years in the conivaptan group and 48 20 years in the control group (p=0.83). No patients experienced Na + increases >1 mEq/L/h. The average serum Na + on day 2 was 143 6 and 142 7 (p=0.82), with no patients achieving Na+ levels > 155 mEq/L. The amount of Na + administered in the first 48 hours was 818 724 mEq in the conivaptan and 1076 1203 mEq in the control group (p=0.69). 48-hour urine output and fluid balance were not different between the two groups (p=0.20 and 0.06, respectively). There were no serious adverse events related to drug administration.

These data suggest that a single dose conivaptan is safe in non-hyponatremic patients with severe TBI who require increases in serum Na + for the purpose of ICP control. Further studies are needed to establish the effect of higher doses and longer administration on clinically relevant endpoints.

Venous thromboembolism (VTE) is a preventable cause of morbidity and mortality in hospitalized patients, and is estimated to occur in 15 to 40% of neurosurgery patients. We sought to establish the incidence of and risk factors for VTE among patients with non-traumatic subarachnoid hemorrhage (SAH) or who underwent intracranial tumor resection (ITR).

This was a single-center, retrospective study of patients admitted to Saint Louis University Hospital between January 1 st , 2007 to December 31 st , 2009 with non-traumatic SAH or who underwent ITR. We collected baseline demographic data, prophylactic modalities utilized, surgery duration, type and location of VTE, and VTE treatment. Doppler ultrasonography (US) was used as the diagnostic method. Comparisons were made between patients who did and did not develop VTE; logistic regression was used to determine if any variables were significant predictors of VTE development.

347 patients were identified for VTE screening (176 ITR, 171 SAH). Further data were available for analysis for 209 patients (176 ITR, 33 SAH). All patients had sequential compression devices (SCDs) for prophylaxis. Twenty six ITR patients (14.8%) and 4 SAH patients (12.1%) were also on pharmacologic prophylaxis. Doppler US was done in 49.4% and 90.9% of the ITR and SAH groups, respectively. There were 64 deep vein thromboses (DVT); 22 (12.5%) in the ITR group and 42 (24.6%) in the SAH group. One ITR patient developed a pulmonary embolism within 30 days post-discharge. Inferior vena cava filter was the most common VTE treatment used. Age (p=0.022), surgery length (p=0.023), and body mass index (BMI) (p=0.029) were predictors of VTE.

VTE is a common complication among ITR and SAH patients. The incidence of DVT may be higher if diagnosis is systematically sought. Risk stratification is needed to identify patients requiring screening and those who might benefit from combined prophylaxis.

We present a case report in a patient with severe, recurrent, thunderclap with CT evidence of subarachnoid blood and negative work-up for aneurysm. This case is an example of Call-Fleming syndrome with subarachnoid hemorrhage in which trans-cranial dopplers were used for monitoring of cerebral vasoconstriction when angiography did not evidence vasoconstriction. We will review Call-Fleming syndrome and the utility of trans-cranial doppler imaging to assess cerebral vasoconstriction.

A review of the current literature regarding diagnostics, treatment, and morbidity in Call-Fleming (reversible cerebral vasoconstriction syndrome) as well as a review of the data using trans-cranial color-coded sonography and transcranial doppler imaging to assess vasospasm in these cases.

The patient underwent CTA/CTV, catheter angiography, lumbar puncture, and vasculitis work-up which were all negative. His MRI showed T2 and FLAIR hyper-intensities in the posterior frontal lobes as well as subarachnoid blood along bilateral occipital convexities. TCDs were obtained which showed elevated mean velocities.

The use of bedside trans-cranial dopplers are a non-invasive means of assessing vasospasm in Call-Fleming syndrome; even in cases where angiography is negative. Determining the degree of vasospasm based on the data in subarachnoid hemmorhage, we are able to predict a patient's risk of complications related to vasospasm including PRES and ischemic events.

Hemodynamic augmentation is the mainstay of therapy for aneurysmal subarachnoid hemorrhage (SAH)-associated cerebral vasospasm. We describe asymmetric posterior reversible encephalopathy syndrome (PRES), least prominent in the territory of vasospasm, as an unusual complication of this therapy.

Case report.

A 35 year-old woman presented with Hunt & Hess grade II SAH secondary to rupture of a right middle cerebral artery (MCA) bifurcation aneurysm. Blood pressure on admission was 121/78 mmHg. Head CT showed extensive SAH, predominantly perimesencephalic and in right sylvian fissure. The aneurysm was surgically clipped on hospital day (HD) 2, and a lumbar drain was placed. On HD 3, she developed left leg weakness. Conventional angiography revealed moderate narrowing of the right MCA M1 and M2 segments, treated with intra-arterial nicardipine. Phenylephrine was started to maintain a systolic blood pressure (SBP) of 140-180 mmHg. The leg weakness resolved. One day later, she was less responsive with left hemiparesis. Repeat angiography revealed persistent vasospasm in the right MCA, treated with intra-arterial nicardipine and M1 segment angioplasty. The SBP goal was elevated to 160-200 mmHg, with improvement of the drowsiness and hemiparesis. On HD 7 her level of arousal progressively declined, culminating in stupor and two generalized tonic-clonic seizures. MRI showed widespread, yet markedly asymmetric changes consistent with PRES largely sparing the right MCA territory. Phenytoin was initiated, phenylephrine discontinued and SBP goal decreased to 110-130 mmHg. She had no further seizures. Two days later she was fully alert with mild left hemiparesis.

PRES is a rare complication of hyperdynamic therapy that should be considered in the differential diagnosis of delayed neurological decline in patients with aneurysmal SAH-associated cerebral vasospasm. The markedly asymmetric distribution of PRES lesions with sparing of the area affected by vasospasm supports the hypothesis that hyperperfusion underlies the pathophysiology of this disorder.

Clinical trials have demonstrated a reduction in the incidence of vasospasm identified via TCD ultrasound with the use of statins. However few studies have investigated the effect of statins in preventing angiographic vasospasm. In 2006 routine screening angiograms on day 7 of admission became standard of care for SAH patients at our institution, as did routine use of pravastatin 80 mg daily for 14 days. The aim of the present study was to evaluate the efficacy of statin therapy in the prevention of angiographic and clinical vasospasm.

This is a retrospective cohort study of adult patients with high Fisher grade aneurysmal SAH admitted to Neurosurgery/Neurology Intensive Care Unit between January 1, 2006 and December 31, 2008. The primary endpoint was combined clinical or significant angiographic vasospasm. Clinical vasospasm was defined as any new or worsened neurologic deficit which responded to vasopressors or corresponded with the presence of vasospasm on angiogram. Significant angiographic vasospasm was defined as any vasospasm graded moderate or severe or any vasospasm which warranted endovascular intervention. Secondary outcomes included hospital and ICU length of stay and discharge disposition.

Eighty-nine patients were included. Thirty-five (39%) patients received a statin. A numerical decrease in the incidence of the primary endpoint was observed (74% vs 54% p = 0.05). There were no differences in any secondary endpoints. Multivariate analysis which included age, gender, and premorbid statin use revealed only in-hospital statin therapy to be associated with a reduction in clinical or significant angiographic vasospasm (OR 0.41;95% CI 0.16-1.02).

In this retrospective study of patients with high Fisher grade aSAH, a decrease in the incidence of combined clinical or significant angiographic vasospasm was observed. This study was conducted as part of a larger investigation to better elucidate the connection between angiographic vasospasm and cerebral infarction. Future large scale prospective studies are needed to determine the impact of statin therapy on the prevention of vasospasm and cerebral infarction in this population.

Up to a third of aneurysmal subarachnoid hemorrhage patients are active smokers and are at risk for nicotine withdrawal during their admission to the hospital. Nicotine is known to cause vasoconstriction in animal models. The safety of nicotine replacement in subarachnoid hemorrhage in humans is still uncertain and the literature is very limited. Nicotine also has possible anti-inflammatory effects and may have some beneficial effect on outcomes.

Retrospective analysis of a prospectively collected database of aneurysmal subarachnoid hemorrhage patients who were active smokers admitted from 1994 to 2008. Paired controls matched for age, sex, fisher score, treatment epoch, aneurysm size and number, hypertension and medication were analyzed in univariate and multivariate analysis. Primary outcome was clinical and angiographic vasospasm and secondary outcome was Glasgow outcome scores on discharge.

Analysis was performed on 258 patients, well matched for age, sex, gender, fisher scores, hypertension, aneurysm size and number and medications, 87 who were treated with transdermal nicotine replacement. On univariate analysis patients who received nicotine replacement were less likely to have clinical vasospasm (19.5 vs. 32.8%; p = 0.026) and more likely to have better Glasgow outcome scores (81.6 vs. 62.6%; p = 0.005). On multivariate analysis however after correcting for modified fisher scores and clinical severity scores these differences were found to not be significant.

Nicotine replacement therapy is not associated with increased vasospasm. The improved Glasgow outcome scores seen in the patients who received nicotine replacement are likely secondary to less severe clinical disease. Further study with better matched cases and controls for clinical severity and modified fisher scores are needed to investigate whether nicotine replacement has any beneficial effects on vasospasm and clinical outcome. Nicotine replacement is safe in aneurysmal subarachnoid hemorrhage and does not worsen vasospasm.

Cerebral vasospasm after aneurysmal subarachnoid hemorrhage is a potentially severe sequel. The combination of hypertension, hypervolemia and hemodilution is frequently advocated for vasospasm. It is unclear whether hemodilution offers any beneficial effects independent of hypertension and hypervolemia. We developed a finite element model of oxygen transport in the proximal (M1) segment of the middle cerebral artery to evaluate the complex relationship between hematocrit, viscosity, oxygen content, and intra-and post-stenotic blood flow with the goal of evaluating hemodilution for cerebral vasospasm.

A single phase non-Newtonian finite element model based on 3-dimensional incompressible Navier-Stokes equations were constructed of M1 segment of the MCA. Blood density and viscosity were calculated as a function of hematocrit. The model was solved using a time-dependent stabilized finite element method at vessel stenoses ranging from 0% to 90% and hematocrit ranging from 0.2 to 0.6.

A small area of post stenotic recirculation and increased rotational flow is seen with mild (30%) stenosis. Post stenotic eddy formation is seen with more severe (60-90%) stenosis. Volumetric flow was inversely related to hematocrit at mild stenosis (0-30%). With near-complete stenosis (90%), a paradoxical increase in volumetric flow is seen with increasing hematocrit. Oxygen transport across the M1 segment was related to hematocrit at all levels of stenosis with increasing oxygen transport despite the corresponding reduction in blood flow. There was no level of stenosis at which hemodilution appeared to be beneficial.

This simulation suggests that at clinically significant degrees of vasospasm in the middle cerebral artery, hemodilution does not improve oxygen transport across the stenotic segment. To the contrary, our model suggest that ischemia may be worsened by inducing hemodilution.

Delayed cerebral ischemia (DCI) is the most common cause of poor outcome in subarachnoid hemorrhage (SAH) patients in the subacute setting. Reduced vasodilatory response of cerebral arteries contributes. Intravenous Nicardipine reduces the incidence of symptomatic vasospasm, but does not affect outcome. Recently, the combination of hypertensive hypervolemia and vasodilatory (HHVD) therapy via IV Nicardipine demonstrated faster reductions in mean MCA velocities. HHVD could offer an improved means to reduce DCI, improving outcomes.

We retrospectively reviewed our hospital use of HHVD since December 2009. Fisher 3 or 4 SAH patients with acute changes in neurologic examination, and elevation in mean TCD velocities -or-with documented radiographic and clinical vasospasm refractory to intra-arterial angioplasty/IA Nicardipine received HHVD. IV Nicardipine was infused at 1.5-2.5 mg/hr -or-at a dose necessary to maintain mean arterial pressure (MAP) goals for 3-10 days. Data included daily TCD mean velocities of the middle cerebral and basilar arteries, near-infrared spectrophotometry-determined regional cerebral oxygen saturation (rSO2), and Nicardipine dose.

Five patients were reviewed. All 5 were Fisher 3/4, and Hunt-Hess 3 or higher. One patient only had occipital windows for TCD insonation. One patient did not receive rSO2 measurements. Four of 5 patients demonstrated reduced mean TCD velocities correlating to Nicardipine. These achieved significance (p<0.05) in both MCAs and Basilar in two of 4, and in the basilar artery of the remaining 2. However rSO2 values significantly (p<0.05) fell 8-10% in three of 4 patients correlating with Nicardipine administration.

In this small series of HHVD therapy, 4/5 of patients demonstrated some reduction of TCD velocities, and 3/4 demonstrated significant reductions in rSO2 correlating to the use of Nicardipine. Prior to expanded use of HHVD therapy to reduce TCD velocities during vasospasm, the physiologic consequences of this reduction in cerebral oxygenation must be explored.

Global cerebral edema (GCE) is common amongst poor-grade subarachnoid hemorrhage (SAH) patients and associated with poor outcome. Currently no targeted therapy exists largely due to an incomplete understanding of the underlying mechanisms.

This is a prospective observational study including 39 consecutive poor-grade SAH patients with multimodal neuromonitoring. Levels of microdialysate lactate/pyruvate ratio (LPR), episodes of cerebral metabolic crisis (MC; LPR >40 and brain-glucose <0.7 mmol/L), brain tissue oxygen tension (PbtO2), cerebral perfusion pressure (CPP), and transcranial Doppler sonography flow velocities were analyzed.

Median age was 54 years (45-61) and 62% were female. Patients with GCE on admission (n=24, 62%) had a higher incidence of MC in the first 12 hours of monitoring than those without GCE (n=15; 15% v.s. 2%, P<0.05) and during total time of neuromonitoring (20% v.s. 3%, P<0.001). There was no difference in PbtO2 and CPP between the groups, however, in patients with GCE a higher CPP was associated with lower LPR (P<0.05). Episodes of crisis were associated with poor outcome (modified Rankin Score 5 or 6, P<0.05).

In poor-grade SAH patients, GCE is associated with early brain metabolic distress. Optimizing cerebral blood flow and homeostasis early after SAH may prove beneficial for patients with GCE.

Silent infarction is common in poor-grade subarachnoid hemorrhage (SAH) patients and associated with poor outcome. Invasive neuromonitoring devices may detect changes in cerebral metabolism and oxygenation

From a consecutive series of 32 poor-grade SAH patients we identified all CT-scans obtained during multimodal neuromonitoring and analyzed microdialysis parameters and brain tissue oxygen tension (PbtO2) preceding CTscanning.

Eighteen percent of the reviewed head-CTs (12/67) revealed new infarcts. Of the eight infarcts in the vascular territory of the neuromonitoring, seven were clinically silent. Neuromonitoring changes preceding radiologic evidence of infarction included lactate-pyruvate-ratio elevation and brain-glucose decreases when compared to those with distant or no ischemia (P

These data suggest that there may be distinct changes in brain metabolism and oxygenation associated with the development of silent infarction within the monitored vascular territory in poor-grade SAH patients. Larger prospective studies are needed to determine if treatment triggered by neuromonitoring data has an impact on outcome.

Hemorrhagic stroke onset times display circadian fluctuations. 1 This variation is independent of various risk factors for aneurysmal rupture. 2 We investigated the time of hemorrhage and its relationship to traditional risk factors among patients admitted with aneurysmal subarachnoid hemorrhage (aSAH).

Admitted patients with aSAH were prospectively followed through outcomes. Baseline demographics, past medical history, and medical course were abstracted through chart review. aSAH onset time was classified as night (11 p.m. to 5 a.m.), morning (5 a.m. to 11 a.m.), afternoon (11 a.m. to 5 p.m.), and evening (5 p.m. to 11 p.m.). The odds ratio (OR) with 95% CI was calculated for having an aSAH during the morning, afternoon, and evening hours using night as a reference. Multinomial logit models were fitted using aSAH cases across time blocks to determine their associations with different risk factors.

Out of a total of 259 patients with aSAH admitted from 8/06 to 12/09, 251 patients had identifiable aSAH onset times. 38 patients had night, 98 patients had morning, 58 patients had afternoon, and 57 patients had evening onset times. For all aSAH cases, morning onset was significantly more common than night onset (OR=2.58, 95% CI=1.77, 3.75). Non-smokers were more likely to have aSAH in the morning than smokers (p=0.043, RR=3.10, 95% CI=1.33, 7.23). There was a strong trend for men, compared to women, to have a higher likelihood for morning hemorrhage onset (p=0.059, RR=2.62, 95% CI=0.998, 6.87).

We found that there was a significantly higher probability of hemorrhage onset in the morning in our cohort of aSAH patients. Non-smokers were found to have a significantly higher OR of morning onset in comparison to smokers. The association of these risk factors with existing onset patterns should be investigated in future studies. 

Limited data are available regarding the overall impact on clinical outcome after introduction of primary angioplasty and intra-arterial vasodilators for treatment of symptomatic cerebral vasospasm in patients with subarachnoid hemorrhage (SAH).

We performed an analysis of all hospital admissions with aneurysmal SAH over a 6 year period (2002 to 2008), with a total of 146 patients. Primary angioplasty and intra-arterial vasodilators were introduced for treatment of symptomatic cerebral vasospasm in the last two years.

There was no difference between the 89 patients admitted prior to and 57 patients admitted after introduction of endovascular treatment for cerebral vasospasm in regards to age, medical co-morbidities, and admission clinical severity of patients (measured by Hunt and Hess grade and Glasgow Coma scale). A total of 18 (32%) of 57 patients underwent primary angioplasty and/or intra-arterial vasodilator treatment in the last two years. A statistically non significant trend was noted towards a lower proportion of severe disability and death (mRS 5-6) at discharge (45% versus 33%, p=0.09) and 1 year mortality (31% versus 23%, p=0.26). There was no significant difference between the two time periods in regards to length of stay. Results were similar after multivariate analysis.

Primary angioplasty and intra-arterial vasodilators for cerebral vasospasm appear to modestly reduce the rate of disability and death at discharge and 1 year mortality; however it is not statistically significant. The overall effect is partly obscured by low sample size and a multitude of factors that affect outcome in patients with SAH.

Aneurysmal subarachnoid hemorrhage is associated with a 67% mortality and significant disability. Dysfunction of the cerebral vasculature may be a crucial factor that may manifest as increased blood-brain barrier permeability (BBBP), and ultimately vasospasm. The goal of this study was to determine which clinical and radiographic variables are associated with novel in vivo BBBP assessments obtained using Perfusion-CT (PCT) technology.

Patients with confirmed aneurysmal subarachnoid hemorrhage admitted to our institution between January 2007 and July 2008, with a Perfusion-CT scan and subsequent Digital-Subtraction Angiogram (DSA) performed no more than 24 hours apart were included in our cross-sectional analysis. Clinical variables determined from chart review as well as radiographic variables determined by respective neuroradiologist and interventional neuroradiologist image review were analyzed using random-effects generalized linear models to determine the association with increased BBBP in 14 pre-determined vascular territories of these patients.

Eighty-three patients were identified with a mean age of 54, and average time from onset to image of 72 hours. Angiographic vasospasm was present in 22% of PCT-DSA pairs. 1162 vascular territories from these 83 patients were evaluated further for Perfusion-CT parameters and for the presence and severity of vasospasm on angiogram. The mean BBBP increased by severity of vasospasm on DSA, however, in multivariate analysis, only Mean Transit Time (MTT), Cerebral Blood Volume (CBV), and severity of hydrocephalus were significantly associated with BBBP. Increased BBBP was not associated with angiographic vasospasm severity in multivariate analysis.

BBBP assessment by Perfusion-CT may serve as a unique biomarker in addition to angiography in patients with aneurysmal subarachnoid hemorrhage, especially in characterizing microvascular dysfunction, or even in selecting patients for medications with endothelial activity. However, future prospective studies will be required to evaluate its clinical utility.

There is widespread variation in the use of red blood cell (RBC) transfusions in patients with aneurysmal subarachnoid hemorrhage (SAH). An adequately powered randomized controlled trial (RCT) is required to clarify optimal transfusion practices.

We performed a survey involving American and Canadian academic neurointensivists, vascular neurosurgeons and multidisciplinary intensivists who regularly care for SAH patients, in order to determine minimum and maximum hemoglobin (Hb) concentrations which clinicians consider acceptable transfusion thresholds to compare in a RCT. We also assessed how opinions are modified by the presence or absence of delayed cerebral ischemia (DCI) and various clinician characteristics.

The survey was sent to 531 individuals, of whom 282 (53%) responded. In patients with high grade SAH, the majority of respondents expressed willingness to compare a restrictive transfusion threshold < 8 g/dl (92%) and a liberal goal > 10 g/dl (75%) (Figure) . When respondents having a transfusion trigger of 7 g/dl were excluded, 84% expressed willingness to accept an Hb threshold lower than their own. Similarly, 94% were willing to accept an Hb threshold higher than usual. Both lower and upper acceptable transfusion triggers were higher among patients with DCI (7.7 and 10.3 g/dl with DCI vs. 7.4 and 10.1 g/dl without DCI; p<0.0001). In multivariable analysis, clinician specialty (neurosurgeon vs. intensivist) and seniority (years in practice) were associated with reluctance to accept very restrictive transfusion thresholds. Increasing seniority predicted greater willingness to target higher Hb levels; in contrast, use of transfusion protocols was associated with more reluctance.

Clinicians appear willing to compare relatively divergent transfusion thresholds in a RCT. Preferences are influenced by clinician specialty, seniority and use of protocols. A comparison of fixed transfusion thresholds may be less acceptable to clinicians than a dynamic strategy, whereby Hb goals vary depending on the presence or absence of DCI.

Neurogenic myocardial stunning is a frequently seen complication of aneurysmal subarachnoid hemorrhage (aSAH). The resulting left ventricular dysfunction makes treatment of other complications such as cerebral vasospasm more difficult. Stunning is thought to be mediated by a catecholamine surge at the time of aneurysmal rupture. We hypothesized that patients who present while on beta blockers will have a lower incidence of myocardial stunning compared to those not on beta blockers.

A retrospective chart review was performed based on OHSU Neurosurgery admissions between November 2008 and December 2009. 93 patients with aneurysmal subarachnoid hemorrhage were identified based on angiography and/or CT angiogram, of which 53 received transthoracic echocardiograms. Patients were determined to have neurogenic myocardial stunning if their initial echocardiogram showed wall motion abnormalities that had resolved on follow-up echo.

10 of the 53 included patients were known to be on pre-admission beta blockers, of which none developed neurogenic stunning. Of the remaining 43 patients, 11/43 (26%) developed neurogenic stunning (p = 0.07 by Chi square for univariate analysis). Though not statistically significant, the fact that none of the patients on pre-admission beta blockers developed myocardial stunning suggests that beta blockers may prevent stunning.

Pre-admission beta-blockers may reduce the incidence of neurogenic myocardial stunning in patients with subarachnoid hemorrhage. A prospective clinical trial analyzing the potential benefit of starting beta-blockers on patients with aSAH is warranted.

Dual antiplatelets regimen prior to the stent-assisted coiling of the wide-necked aneurysm is necessary to protect stent and to prevent thromboembolic events. However, the safety and feasibility of dual antiplatelets regimen prior to the stent deployment in the acute phase of ruptured wide-necked aneurysm especially in the basilar artery (BA) is not known. Objective of our study is to determine the safely and feasibility of loading doses of aspirin and clopidogrel prior to the stent-assisted repair of the acute ruptured BA wide-necked aneurysms.

Consecutive patients who underwent stent-assisted coiling for ruptured wide-necked BA aneurysm were enrolled from 2007 to 2010. Patient's demographics including the Hunt & Hess grade, Fished scale, use of ventriculostomy catheter, location and size of aneurysm were collected. Complications such as rupture of aneurysm, ventriculostomy associated hemorrhage or systemic bleeding were recorded. Additionally, 30-days outcome measurement was obtained using Glasgow Outcome Scale (GOS).

Eight patients with a mean age of 53 ± 11 years underwent successful stent-assisted repair of the ruptured widenecked BA aneurysm. Hunt and Hess (H&H) I was present in 4, H&H II in 1, H&H III in 1 and H&H IV in II. Two patients received ventriculostomy prior to the procedure. Each patient received loading dose of aspirin (325 mg) and clopidogrel (300 mg) at least 2 hours prior to the stent placement. There was no intraoperative ruptured of aneurysm, hemorrhages or ventriculostomy associated complications. Additionally, there was no intra-operative or post-operative thromboembolic event. Good outcome was observed in 6 (GOS5 in 5, GOS4 in 1) and poor outcome in 2 (GOS 3 in 2).

In our series, loading doses of dual antiplatelets prior to the stent-assisted repair of the BA ruptured wide-necked aneurysm in their acute phase is not associated with rupture of aneurysm, intracranial or systemic hemorrhages. Further study is necessary.

Stent-assisted coiling of intracranial aneurysm is performed either as a single stage or a multi-staged procedure. The objective of our study is to compare the complications of intracranial stent in single stage versus multi-staged treatment of intracranial aneurysm.

From January 2003 to January 2010, consecutive patients treated with Neuroform/Enterprise stent for intracranial aneurysms were prospectively enrolled. Patients' demographics including cerebrovascular risk factors, aneurysms size and locations were collected. Technical and clinical complications as well as clinical outcomes were measured. Data were analyzed retrospectively using SPSS software version 11.5.

87 patients with a mean of 51.2 ± 13.6 years were treated with 87 intracranial stents (Neuroform 72, Enterprise 15), single stage 37 (42.5%) and multi-staged 50 (57.5%). Eight adverse events were observed without any mortality, 6 of which were in the single stage group -rupture of aneurysm in 2, and thromboembolic events in 4. Both rupture occurred in basilar artery bifurcation aneurysms, required ventruculostomy catheter and resuscitations. Aneurysms were secured with complete occlusion. In single stage, asymptomatic intraoperative stent thrombosis developed in one, symptomatic stent thrombosis in one on day 14, transient ischemic attack on day 6 and immediate post operative stroke in one. Only two minor strokes were observed in the multi-staged group, one on post-procedure day 7 with NIHSS 4 and other on day 60 with NIHSS 1. Majority of the patients had good outcomes including those with events.

Our study revealed that single stage stent-assisted coiling is associated with a higher rate of complications than multistaged procedure. However these events had no impact on the outcome of both group. We also observed no interval hemorrhages in multi-staged procedure cases. Therefore, caution should be made in single stage stent-assisted repair of intracranial aneurysm.

Cerebral aneurysms are an important cause of morbidity and mortality which the most severe consequence is their rupture. SAH has the arterial vasospasm and DCI as one of the most worried complications.

After institutional approval and informed consent, a prospective, randomized, single blind study was occurred between February 2008 to November 2009. The study evaluated the Mg use on patients from the 1 to 4 Beds and control group from 5 to 8 beds. The serum measure of Mg was made by colorimetric method in order to reach the measurement between 2,5 to 3,5mg/dl, using a solution of Mg 2%(SG 5%-400ml + MgSO4 10%-100ml/24h), during the first 14 days of event(aneurysm rupture).Admission Criteria: Patients diagnosed with SAH confirmed by CT or cerebral angiography and t < 96h. Exclusion Criteria: Patients with SAH and T > 96h; Patients who presented a vasospasm episode in less than 24 hrs of Mg Solution Infusion.

In a previous study evaluation were analysed a total of 107 patients with n = 55 on Group 1 and n = 52 on Group 2 (table 1 and 2). The main results refer: Group 1 -Vasospasm Frequency 20.0% confidence interval(CI) = 10,4% -33.3%; and Mortality 16.4% in 28 days CI=7.8% -28,8; Group 2-Vasospasm Frequency 51.9% CI=37.6% -66.0%; and Mortality 23.1% in 28 days CI=12.5% -36.8%. The analysis for the vasospasm showed Odds Ratio(OR)= 0.23 CI = 0.098%-0.544% and p-value = 0.0011, and for mortality: OR = 0.65 CI = 0,24% -1.70% and p-value = 0.5284.

According to the outcome, we can conclude that the Group 1 obtained a greater protection on the vasospasm incidence in comparison to Group 2 but showed no difference in mortality. The p-value was significant for vasospasm but still not significant for mortality.

Vasospasm is the main cause of death and cognitive deficits in patients with subarachnoid hemorrhage after rupture of the aneurysm(aSAH). Some trials have shown that statins in the acute phase of aSAH reduces the incidence, morbidity and mortality of cerebral vasospasm.

Was realized a prospective study,randomized,non-blind,with the use of 80mg of SVT(at night) in the first 72h of the beginning of bleeding, and the control group that didn't use SVT,for 21 days,between January to December 2008. Informed consent for all patients.CT scans was performed as control and another CT scan in patients with altered neurological signals. In the presence of changes suggestive of vasospasm or correlation in clinical and CT scans the patients were taken to cerebral arteriography exam followed by angioplasty procedure if necessary.Liver and renal function,LDL cholesterol evaluated weekly, and CK Total evaluated every 3day.Exclusion criteria: liver and renal disease, pregnant elevation of serum transaminases(3 times the value of normality),creatinine 2,5, rabdomyolysis or CK Total 1000U/L.

Were excluded 2 patients with bleeding more than 72hs.There was no significant change in the levels of CK total,renal or liver function.We included 21 patients,11 in the SVT group and 9 in the control group.The mortality was 8 patients(38%),6 patients in the control group and 2 of the SVT group.Vasospasm was confirmed by cerebral arteriography exam in 4 patients in the control group and 1 patient in the SVT group.All the patients who died had scale Fisher IV.

The SVT at a dose of 80mg was effective in reducing the mortality(18,1% against 66%)compared to the group that did not use SVT,and also decrease the incidence of cerebral vasospasm despite the APACHE II higher in the group that used SVT(14,3 vs 10,7).Less morbidity in the SVT group with an average of scale of Glasgow 3,25 vs 2,1

Introduction SIRS is non-specific but frequent finding in patients with aSAH. It may be triggered by infection or aSAH, and may herald inflammatory processes involved with cerebral vasospasm. Vasospasm is a frequent complication of high grade aneurysmal subarachnoid hemorrhage. We hypothesize that SIRS and infection are positively correlated with the incidence of vasospasm.

A retrospective study of all aneurysmal SAH admitted to the intensive care unit at an academic tertiary medical center from 8/2009-5/2010. Data collected include World Federation of Neurological Surgeons grading of aSAH (WFNS), incidence and severity of vasospasm, incidence of SIRS and infection, as well as patient demographics. Vasospasm was identified by clinical notes, transcranial Doppler, CT angiography or cerebral angiogram and classified as mild, moderate, or severe. Infection was defined by clinical, microbiological, and radiographic data.

32 patients were enrolled (23 female). Twenty-four had evidence of SIRS and 21 had vasospasm (4 mild, 3 moderate, and 14 severe). Seventeen out of 21 patients with vasospasm had SIRS, and all 5 patients with infection had vasospasm. Out of 11 patients without vasospasm, 7 had evidence of SIRS. There were no significant associations between the incidence of SIRS and/or infection with vasospasm. There were direct associations between the WFNS and incidence of SIRS (p=0.03) and vasospasm (p=0.07).

The results of our retrospective study fail to demonstrate significant associations between aSAH-related SIRS and infection with vasospasm. However, this is likely due to a relatively small sample size, and a larger study is needed to confirm our hypothesis.

Medical management of cerebral vasospasm following aneurysmal subarachnoid hemorrhage (SAH) includes hypertensive, hypervolemic, and hemodilution ("triple-H") therapy. However, there is little information regarding the indications and guidance used to initiate and adjust triple-H therapy.

A 43-item online questionnaire was emailed to 375 members of the Neurocritical Care Society. Questions were designed to investigate the diagnostic approach to cerebral vasospasm and prophylactic and therapeutic administration of triple-H therapy.

Completed surveys were received from 167 respondents (45% response proportion). Eighty-six percent of respondents worked in hospitals with neurointensive care units (NICUs). SAH patients in hospitals with a NICU had longer ICU stay (P=0.037) and had indwelling central venous catheters for longer (P<0.01). Centers without dedicated NICUs were more likely to induce prophylactic hypervolemia (P<0.01). 27% of respondents (n=45) reported using prophylactic hypervolemia in patients with SAH, while 100% reported inducing hypervolemia for severe symptomatic vasospasm. Twelve percent (n=20) of respondents reported inducing prophylactic hypertension, while all reported inducing hypertension with severe symptomatic vasospasm. Half of respondents relied on the mean arterial pressure and half on systolic blood pressure as the clinical parameter for blood pressure titration. The most widely used agents to induce hypertension were phenylephrine (48%) and norepinephrine (39%). There was little variation in the use of hemodilution therapy comparing patients with or without evidence of vasospasm.

There are substantial differences in the administration of prophylactic triple-H, but there was high agreement on indication for therapeutic use. There was wide variability in the extent of ICU monitoring, diagnostic approach, physiologic parameters and values used as target of therapy. NICU availability was associated with more intensive monitoring. Lack of evidence and guidelines for triple-H therapy might largely explain these findings.

Many neuromonitoring devices provide regional data specific only to the vascular distribution of their placement. Often monitors are placed in the Middle Cerebral Artery (MCA) distribution, irrespective of the brain territory at risk for secondary injury. The risk of DIND is common after aSAH and may occur near or remote from the ruptured aneurysm. The value of regional monitoring for this potentially diffuse process is not known.

All aSAH patients presenting to a single tertiary referral center over a three year period were assessed. Patients with a treated aneurysm who survived > 10 days and suffered DIND were considered. Only those patients receiving Transcranial Doppler (TCD) monitoring and routine neuroimaging were included. Patients were divided into groups of midline and non-midline aneurysms and assessed for likelihood of spasm and stroke within the aneurysmal vessel, ipsilateral hemisphere, and ipsilateral MCA. The threshold for TCD spasm was > 120 cm / second for anterior and > 85 cm / second for posterior circulation vasculature. Comparisons of clinical characteristics were made to determine factors predisposing to remote infarction.

Likelihood Twenty nine patients met study criteria with 15 patients being classified as non-midline aneurysms. The rarity of isolated remote DIND in this study did not allow for adequate assessment of predictive clinical characteristics.

For non-midline aneurysms, DIND occurs ipsilateral to the ruptured aneurysm in 93% and within the same vascular territory in 86% of patients. Midline anterior circulation aneurysms frequently result in Anterior Cerebral Artery infarction. A neuromonitoring device with 100% sensitivity for ischemia placed in the MCA territory ipsilateral to a nonmidline ruptured aneurysm would identify 71% of DIND.

Cerebral vasospasm and stroke are common after aSAH. TCD monitoring is employed to assess the onset of vasospasm and allow intervention to avoid infarction. The duration of TCD monitoring required and the factors impacting ischemic risks are not well established.

A 3 year retrospective analysis of aSAH treated in a tertiary medical center was undertaken. Eligible patients were age 18-85, presenting within 2 days of hemorrhage with documented or suspected aSAH who had undergone 7 days of TCD monitoring. Exclusion criteria included a history of trauma or vascular malformation, absent TCD windows, palliative care implementation, survival < 10 days, or absent routine brain imaging. Patients were assessed to determine if vasospasm onset (MCA/ACA/PCA > 120 cm/sec, Vertebral/Basilar > 85 cm/sec) occurred after 10 days and resulted in stroke.

107 patients met study criteria with 51 (48%) demonstrating vasospasm and 31 (29%) developing stroke. Of those developing infarction, 22 (71%) demonstrated vasospasm while 9 (22%) did not. Two (2%) patients developed vasospasm only after day 10, neither experiencing stroke. Time to vasospasm onset was 5.5 days (SD 2.5) and was not impacted by Modified Fisher Grade (FG) or World Federation of Neurological Surgeons Grading Scale (WFNS). GCS, aneurysm treatment, FG, modified FG, intubation, EVD placement, intraventricular hemorrhage (IVH), sex, Hunt and Hess Score (H&H), and WFNS score were not associated with spasm occurrence. GCS, H&H score, WFNS, EVD placement, intubation, and presence of IVH were associated with likelihood of stroke (p < 0.05).

TCD identification of spasm after day 10 is rare, but present. Stroke is more likely to result from poor detection than from brevity of TCD monitoring. A monitoring protocol, relying exclusively on TCD detection, may result in vasospasm related infarction regardless of monitoring duration. Improved or alternative monitoring is needed to effectively prevent stroke.

Angiographic cerebral vasospasm (CVS) occurs in two-thirds of patients with subarachnoid hemorrhage (SAH) with half becoming symptomatic. Measures that can predict and diagnose relevant CVS after SAH may advance our understanding of its pathophysiology. We studied the time course of dynamic cerebral autoregulation in patients with SAH.

Thirteen consecutive patients with acute non-traumatic SAH and adequate transcranial Doppler windows were prospectively studied within 48 hours of symptom onset. Daily 5-minute measures of continues beat-by-beat bilateral middle cerebral artery (MCA) and mean arterial blood pressure (MAP) recordings were obtained on days 1-10. Transfer function analysis of spontaneous MAP and MCA mean flow velocity (MFV) oscillations were performed in the very low (0.03-0.07 Hz), low (0.07-0.15 Hz), and high (0.15-0.3 Hz) frequency ranges. CVS was diagnosed using conventional cerebral angiography on day 7 (±1day).

Patients with CVS had higher MAP (p=0.001) and MFV (p=0.016) than those without vasospasm. Peak differences occurred on day 10 for the MAP (102±17mmHg vs. 66± 7mmHg, p=0.003), day 4 for the right MCA MFV (112±35cm/s vs. 55±9cm/s; p=0.004) and left MCA MFV (112±37cm/s vs. 65±17cm/s; p=0.017). However, MFV were all <160 cm/s on day 4. CVS was also associated with higher transfer function gains in the very low frequency ranges (p=0.026). Peak differences in gains were observed on day 3 for both the right (1.5±0.8 vs. 0.6±0.4, p=0.04) and left MCA (1.1±0.4vs. 0.4±0.2, p=0.01). Gains in the low and high frequency ranges were also higher in those with CVS.

In this pilot study, dynamic cerebral autoregulation as measured by transfer function gain in the autoregulatory frequency range (i.e. very low frequency) was significantly impaired in the early days after SAH in patients that developed subsequent angiographic vasospasm. This measure may be an additional predictor of CVS.

Reversible Cerebral Vasoconstriction Syndrome (RCVS) mimics aneurysmal subarachnoid hemorrhage (aSAH) with thunderclap headache in >90% and SAH in 42%. RCVS with SAH (RCVS-SAH) is often misinterpreted as aSAH, subjecting patients to unnecessary imaging, invasive procedures and prolonged ICU stay. Our objective was to identify predictors that can reliably differentiate RCVS-SAH from aSAH.

Clinical and imaging features were compared between 515 consecutive aSAH and 35 consecutive RCVS-SAH patients at Massachusetts General Hospital. Data was analyzed using student's T-test, ChiSquare or Fisher's Exact test, as appropriate. A forward stepwise multiple logistic regression model was developed to predict RCVS-SAH.

Univariate analysis: RCVS-SAH patients were younger (43±12 vs. 55±14 years; p<0.001), with a higher proportion of women (87% vs. 71%; p=0.04), prior migraine, depression, alcohol use, COPD and drug abuse (all p < 0.001). RCVS-SAH had lower Hunt & Hess (HH) and Fisher grade (median 2 (IQR 2;2) vs. 3 (1;3); p<0.001 and 2(2;2) vs. 3(2;3); p<0.001). Imaging analysis showed that RCVS-SAH had a higher proportion of CT hypodensities (63% vs. 35%; p<0.001) and angiographic vasospasm (100% vs. 62%; p<0.001); more arteries affected by vasospasm (6 (6;9) vs. 2 (0;4); p<0.001), bilateral involvement (89% vs. 36%, p<0.001), and earlier occurrence of vasospasm (day 1 (1;1) vs. 6 (5;8); p<0.001). Multiple logistic regression showed the following predictors of RCVS 

We identified important clinical and imaging differences between RCVS-SAH and aSAH, which may improve time to correct diagnosis, treatment and resource utilization with a potential to decrease length of stay.

Prompt and aggressive hemodynamic therapy for vasospasm and delayed ischemic neurological deficits (DINDs) after SAH may be associated with a number of complications. Despite raising mean arterial pressure (MAP) far above normal levels, hypertensive encephalopathy has rarely been reported in the context of induced hypertension (IH).

Features of posterior reversible encephalopathy syndrome (PRES) may emerge if MAP exceeds the upper limits of autoregulation, leading to endothelial dysfunction and vasogenic edema.

We report the first case of unilateral PRES as a complication of IH in a patient with focal (contralateral) vasospasm.

A 47-year old woman suffered PCOMM aneurysm rupture with SAH. Six days later she developed mental status changes and left facial droop; angiography confirmed moderate-severe vasospasm affecting the right MCA and bilateral ACA vessels. Symptoms resolved with IH. After vasopressors were weaned, she developed recurrent symptoms including left hemiparesis, prompting re-initiation of IH. Repeat angiography revealed improved vasospasm except for persistent narrowing of the right MCA. Although her hemiparesis improved at higher MAP (110-130 mm Hg), she developed (new) aphasia. Head CT showed hypodensity in the left temporal and parietal-occipital regions. MRI confirmed left hemispheric FLAIR signal abnormality without DWI change. Given these findings, consistent with edema not infarction, vasopressors were weaned and MAP was allowed to return to baseline. Within hours her aphasia resolved and her ischemic (contralateral) motor deficits did not return. Follow-up head CT demonstrated significant improvement in the left-hemispheric abnormalities.

Early recognition of PRES is even more imperative in the setting of vasospasm where management is completely opposite. Critical in this case was recognizing that: 1) deficits worsened when MAPs were raised further; and 2) the new deficits were different (and contralateral) to the prior DINDs and vasospasm. Unilateral PRES has not been previously reported. The hemisphere affected by vasospasm was likely protected from high perfusion pressures while the contralateral hemisphere was exposed to pressures above limits of autoregulation. Greater caution may be required when utilizing IH when vasospasm is significantly asymmetric.

In patients with subarachnoid hemorrhage (SAH), higher hemoglobin (HGB) has been associated with better outcomes, but packed red blood cell (PRBC) transfusions with worse outcomes. We performed a prospective pilot trial of goal HGB after SAH.

Forty-four patients with SAH and high risk for vasospasm were randomized to goal HGB concentration of at least 10 g/dL or 11.5 g/dL. We obtained blinded clinical outcomes at 14 days (NIH Stroke Scale and modified Rankin Scale, mRS), 28 days (mRS) and 3 months (mRS), and blinded interpretation of brain MRI for cerebral infarction at 14 days. This trial is registered at www.stroketrials.org.

Forty-four patients were randomized. Patients with goal HGB 11.5 g/dL received more PRBC units per transfusion ( ) and rates of independence on the mRS at 14 days (65% vs. 44%) and 28 days (80% vs. 67%) were similar, but favored higher goal HGB (P>0.1 for all). In patients with goal HGB of 10 g/dL, all nine cerebral infarctions involved cortex, while in the goal 11.5 g/dL group, only 3 involved cortex (P=0.04).

Higher goal hemoglobin in patients with SAH seems to be safe and feasible, and may reduce infarction of cerebral cortex. A phase III trial of goal HGB after SAH is warranted. 

Early tracheostomy has been shown to reduce length of Neuro ICU stay, incidence of ventilator associated pneumonia, and increase in ventilator free days for a diversity of acute neurological diseases. However, identification of early predictors/ risk factors and ideal timing for tracheostomy in patients suffering from SAH has been defined yet.

We analyzed data from patients enrolled in our prospective Subarachnoid Hemorrhage Outcomes Database between April 1996 and January 2010. We excluded patients who underwent withdrawal of care, were made DNR/DNI, or died within the first week. Independent predictors for tracheostomy were identified using univariate and multivariate logistic regression analyses amongst clinical and radiological variables.

Amongst 1254 SAH patients admitted during the study period 271 met inclusion criteria and 28% (N=73) of these underwent tracheostomy. In a multivariate logistic regression model adjusting for admission glucose, abnormal chest x-ray, and ictal infarcts, admission Hunt and Hess 

In addition to admission clinical presentation, amount of blood on CT scan and seizures we found several in-hospital complications to be associated with the need for tracheostomy. Interventions directly targeted towards these complications might be most promising in reducing the risk for subsequent tracheostomy. Future prospective analysis will be requiring for evaluating validity and clinical impact of the model.

Cerebral vasospasm is a significant cause of morbidity and mortality after aneurysmal subarachnoid hemorrhage (aSAH). Following aneurysm rupture, an ensuing inflammatory response is implicated in the initiation of vasospasm. We determined whether elevations in the concentrations of various inflammatory cytokines were predictive of vasospasm.

A prospective observational cohort study was conducted in which cerebrospinal fluid (CSF) samples were collected from aSAH patients on days 1, 3, 5, and 7 after aneurysm rupture. Using a commercially-available multiplex human cytokine immunoassay, each sample was analyzed for 85 inflammatory cytokines or soluble cytokine receptors. For each cytokine, the mean area under the curve (AUC) was compared between patients with and without vasospasm (Student's t-test). The AUC was calculated up to day 7 or the sample day just prior to vasospasm. Only patients with clinical or radiographic vasospasm severe enough to require systemic or cerebral intra-arterial vasodilator administration were considered to experience vasospasm.

Four of ten aSAH patients (54.5 ± 4.88 years; 80% female) experienced vasospasm between days 6 and 8 posthemorrhage, and all patients survived at least 10 days post-hemorrhage. Eight cytokines that were measured in the CSF had a larger mean AUC among patients with vasospasm compared to those without vasospasm. Differences approached or achieved statistical significance for the following cytokines: IL-11 (p=0.020), IL-15 (p=0.023), SCF (p = 0.035), sTNF RI (p = 0.061), MCP-4 (p= 0.064), MIP-3beta/CCL19 (p=0.074), sIL-6R (p=0.086), and IL-10 (p=0.089), without correction for repeated measures. The mean AUC was similar between both groups for all other cytokines tested.

Concentrations of select inflammatory cytokines in the days following aSAH are higher among patients experiencing vasospasm. Such biomarkers may be used to identify patients who are at increased risk of developing vasospasm and require more aggressive treatment or preventative measures.

The International Subarachnoid Aneurysm Trial reported that patients with ruptured aneurysms had better outcomes when treated with coil embolization. Unfortunately, the reason for this discrepancy is unclear. It is well known that following aneurysmal subarachnoid hemorrhage (SAH), a cytokine mediated inflammatory process ensues. We compared cytokine concentrations in patients with ruptured aneurysms that underwent surgical clipping or coiling.

A prospective observational cohort study was performed in which plasma and cerebrospinal fluid (CSF) samples were collected on days 0, 1, 3, 5, 7, and 10 after aneurysm rupture. Each sample was analyzed simultaneously for circulating cytokines or soluble cytokine receptors using a commercially-available multiplex human cytokine immunoassay. For each cytokine, the mean area under the curve (AUC) following surgical intervention was compared between patients undergoing clipping versus coiling (Student's t-test). Hunt Hess score, Fisher grade, and initial cytokine concentration were used to control for baseline stroke severity.

Twelve patients (51.4 ± 9.3 years; 83% female) underwent surgical clipping (n=6) or coiling (n=6) within 48 hours of aneurysm rupture. The median Hunt Hess score and Fisher grade were both 3. One patient had rheumatoid arthritis; no patient received immunosuppression. Two cytokines and one soluble cytokine receptor measured in the CSF each had a larger mean AUC among patients undergoing clipping compared to coiling. Differences approached or achieved statistical significance for SCF (p=0.008), MIP-3beta/CCL19 (p=0.053), and sTNF-RI (p=0.099), without correction for repeated measures. Three cytokines measured in the plasma each had a larger mean AUC among patients undergoing coiling compared to clipping, which included IL-11 (p=0.035), MCP-2 (p=0.055), and 6Ckine/CCL2 (p=0.068).

The method of treating a ruptured intracranial aneurysm appears to influence the inflammatory process that is associated with SAH. Local and systemic differences in cytokine expression following surgical clipping or coiling may account for divergent outcomes.

Systemic inflammatory response syndrome (SIRS) is characterized by changes in temperature, heart rate, respiratory rate, or white blood cell count, often heralding infection. SIRS can occur without infection in subarachnoid hemorrhage (SAH). There is evidence that procalcitonin (PCT) may serve as a diagnostic marker to distinguish between infection and SIRS but its role in SAH has not been determined.

A prospective, observational study including all patients with aneurysmal SAH admitted to the intensive care unit (ICU) at an academic tertiary medical center from 08/2009-06/2010. Baseline PCT values were obtained on all subjects at the admission and PCT values were repeated with each episode of SIRS or clinical suspicion of infection. Clinicians were blinded to PCT values. Infection was defined by clinical, microbiologic and radiographic data. Patients were followed until discharge from the ICU or death. Based on available PCT literature, a threshold of 0.5 ng/ml was used for upper limit of normal.

Thirty-nine patients were enrolled. Twenty-three patients had more than one PCT level drawn during their admission, as clinically indicated. Six out of 23 patients fulfilled clinical criteria for infection. Three of those 6 with infection had elevation in the PCT levels, 6.5, 1.2, 1.8 respectively. Seventeen out of 23 patients had no elevation of PCT levels and no evidence of infection despite fulfilling SIRS criteria (negative predictive value 85%). Fisher's exact test of the 2x2 table yielded a p-value of 0.011.

The results of our prospective, observational study suggest that PCT may be a useful adjunct to exclude infection among SAH patients with SIRS. A larger prospective study will be needed to confirm this observation, as well as to help guide antibiotic usage in these patients.

Many SAH patients have previous medical conditions requiring treatment with antiplatelet agents; which may lead to worse hemorrhages in the setting of aneurysm rupture. Our objective was to determine whether prior antiplatelet use affects severity of the bleed, measured by SAH and IVH sum scores, in a population of patients with spontaneous aneurismal SAH.

We used retrospective analysis of prospectively collected data of patients with SAH admitted to our NICU between April 2004 and April 2010. Antiplatelet use was defined as taking antiplatelets within 7 days of the aneurysmal rupture.

The main outcome was severity of the SAH measured by the admission SAH and IVH sum scores. We conducted linear regression to identify if antiplatelet use was independently associated with SAH and IVH sum scores after controlling of other predictors. As a secondary outcome we analysed the association with aneurysmal rebleed.

Among 644 patients, 84 (13%) were on antiplatelets prior to SAH. There was no significant association between SAH sum score (p=0.11) and IVH sum score (p=0.09) in unadjusted analysis. We also did not find association in our secondary outcome of aneurysmal rebleed (p=0.20).

We concluded that the use of antiplatelet medications was not associated with worse SAH, measured by SAH and IVH sum scores on the admission to the hospital.

Cerebral vasospasm (VS) is a frequently encountered complication of SAH that is associated with delayed brain ischemia and poor neurological outcome. Approximately 60-70% of SAH patients have angiographic evidence of VS, half of them being symptomatic (sVS). The pathogenesis of VS is multifactorial and not completely understood. Soluble endoglin (sEng) is an antiangiogenic protein that has been associated with pre-eclampsia which, as is the case of SAH, is characterized by endothelial dysfunction and VS. Here we examined whether the cerebrospinal fluid (CSF) and serum levels of sEng in patients with SAH are associated with VS in SAH patients.

The study included adult patients with Fisher-3 SAH (n=8). Demographics, medical history, and history of exposure to elicit drugs were documented. Patients were followed prospectively and samples of CSF and serum were obtained between days 5 and 7 post-bleed. sVS was defined as a neurological deterioration in the setting of angiographically proven VS and in the absence of other active conditions that could explain the occurrence of neurological decline. sEng levels were determined using commercially available ELISA test (R&D). Protein levels in the CSF and serum were measured and utilized to calculate CSF sEng index. Patients were subcategorized into those with sVS and those who did not develop VS (noVS). The levels of sEng in both groups were compared using the nonparametric Mann-Whitney U test.

Age, history of hypertension, exposure to elicit drugs, RBC count in the CSF and proteinorrachia were not statistically different within groups. The mean CSF level of sEng was statistically higher in the sVS group (n=4; 0.244±0.089 pg/ml) than in the noVS group (n=4; 0.1087±0.0495 pg/ml; p=0.025). A similar trend was observed in the serum (3.711±0.286 pg/ml in the sVS group vs. 2.989±0.766 in the noVS group); these differences, however, did not reach statistical significance (p=0.297). CSF sEng in both groups was >2 indicating this mediator was produced intratechally.

Elevated levels of intratecally synthesized sEng are observed in SAH patients with sVS. This suggests sEng may participate in the pathogenesis of VS.

Due to the presence a complex anatomical and a hemodynamic profile at the anterior communication artery (AComA), especially when both A2 segments originate from a single A1 segment of the anterior cerebral artery (ACA), a successful surgical or endovascular repair of AComA aneurysm does not always guarantee good outcome. Objective of our study is to report our experiences of endovascular repair of ruptured AComA aneurysms including wide neck and complex aneurysms

Consecutive patients with the diagnosis of ruptured AComA aneurysms who underwent with endovascular coiling from July 2006 to July 2008 were enrolled. Patients' demographics including Hunt and Hess grade, fisher scale, procedure related complication and outcome were collected.

52 patients with mean age of 52 ± 14 years underwent successful coiling of AComA ruptured aneurysm of which 31/52 (65%) wide neck and complex. Procedure related morbidity was observed in 3/52 without mortality or permanent disability. Intraoperative rupture of aneurysm as manifested by the extravasations on the angiogram was observed in two wide neck cases which resolved with subsequent coils placement. Both of these patients achieved good outcome. The MCA was completely revascularized using 2 mg TPA and MERCI retrieval device, and achieved GOS 5 in 30 days. Complete and near complete obliteration of aneurysms was observed in 96% and subtotal in 2 cases. 30 days good outcome was observed in 70.5% of cases (GOS 5 in 49%, GOS 4 in 21.5%), disability in 15.6% (GOS 3 in 15.6%) and poor outcome in 13.7% (GOS 1 (dead) in 13.7%). Poor outcome and disabilities was associated with high H&H grade.

Endovascular treatment is feasible in most ruptured AComA aneurysms including those with wide neck and complex in nature. The most common but rare challenges are intraoperative rupture of aneurysm and thromboembolic event, which could be successfully treated with good outcome Financial Support: None

The benefits of monitoring cerebral blood flow (CBF) in brain injured patients are apparent. New techniques combining near infrared extinction (NIRE) and indocyanine green (ICG) dye dilution to estimate CBF are available. However, transcutaneous NIRE with optodes applied over the skin is controversially discussed, because the signal is contaminated by extracerebral tissues (skin, skull, cerebrospinal fluid layer). Recently a new brain tissue probe for combined monitoring of intracranial pressure (ICP), CBF and oximetry with NIRE has been developed.

For NIRE 2 approaches are applied. 1. A conventional intraparenchymal probes for ICP monitoring is supplied with optical fibers (NeMo Probe®). The light is coupled out into the brain tissue and collected after absorption and scattering with a light detector. 2. A plaster based patch carrying optodes is placed over skin (NeMo Patch®). Central venous injections of 0.2mg/kgbw ICG are performed. Measurement values are collected in parallel, with the NeMo Probe® and NeMo Patch® (NeMo System®, NeMoDevices, Zuerich, Switzerland).. Regional values for the mean transit time of ICG (mttICG), CBF and CBV are calculated.

With a prototype of the probe in a first patient with subarachnoid hemorrhage 6 pairs of repetitive measurements were performed. Mean values were for mttICG 5.6 + 0.2 sec, CBF 22.3 + 2.8 ml/100g/min and CBV 2.1 + 0.3 ml/100g.

NIRE allows the synchronous determination of multiple parameters with one single device. By measurements in parallel with the NeMo Probe and NIRS optodes placed over the skin, furthermore, new algorithms can be developed to subtract the contribution from extracerebral tissues from the signal obtained by transcranial NIRE.

There are no known reliable indicators for endotracheal intubation and mechanical ventilation in patients presenting acutely with a seizure. Clinical practice varies widely, and long term implications of mechanical ventilation are not known. The objective of this study was to examine the predictors of endotracheal intubation when presenting to the emergency department with seizures.

We retrospectively collected data on all patients who presented to a county hospital ED with seizures over a 12-month period and were admitted to the hospital. Hospital records were reviewed and demographics, clinical presentation, acute treatments, complications during hospital stay and discharge information were recorded. Intubated patients were compared with non-intubated patients.

We identified 112 subjects. Intubated patients (N=26) were similar to non-intubated patients (N=86) in age and gender. The two groups were similar in seizure types, number of seizures prior to arrival, mode of arrival to the ED, occurrence of seizures in the ED and seizure etiology. There were significant differences in reported agitation (46% versus 17%, p=0.007), respiratory distress (23% versus 2%, p= 0.002) and hypoxia (27% versus 5%, p=0.03). Benzodiazepines were used more frequently (77% versus 42%, p=0.03) and brain imaging were obtained more frequently (100% versus 65%, p<0.001) with intubated patients than with non-intubated patients. Pneumonia was diagnosed in the hospital more frequently (31% versus 0%, p<0.001) and antibiotics used more often (50% versus 0%, p<0.001) among intubated patients. Hospital stay was significantly longer for intubated patients than for nonintubated patients (average 6.54 days versus 1.95 days, p<0.001). There were no differences in heart failure, vasopressor use, or discharge destination.

Agitation, respiratory distress, hypoxia and benzodiazepine use were significant predictors of intubation which was associated with increased rate of pneumonia and longer hospital stay. Development and testing of specific postseizure intubation algorithm is needed.

Financial Support: None

De l'hemorrhagie dans le corps vitre au cours de l'hémorrhagie cérébrale

Terson hemorrhage in patients suffering aneurysmal subarachnoid hemorrhage: predisposing factors and prognostic significance

Intraocular and optic nerve sheath hemorrhage in cases of sudden intracranial hypertension

Optic nerve sheath hemorrhage. The Jackson memorial Lecture

Magnetic Resonance imaging findings of Terson's syndrome suggesting a possible vitreous hemorrhage mechanism

Knowledge of stroke risk, signs of stroke, and the need for stroke education among children with sickle cell disease and their caregivers

Stroke prevention and treatment in sickle cell disease

Sickle cell diseases: current therapeutic options and potential pitfalls in preventive therapy for transcranial Doppler abnormalities

The management and outcome of patients with Myasthenia Gravis treated acutely in a neurological intensive care unit

Myasthenic crisis

Anemia is associated with poor outcomes following aneurysmal subarachnoid hemorrhage (SAH). It remains unclear whether this association can be modified with more aggressive use of red blood cell (RBC) transfusions.

We performed a survey of North American academic neurointensivists, vascular neurosurgeons and multi-disciplinary intensivists who regularly care for SAH patients, in order to determine the hemoglobin (Hb) concentrations which commonly trigger a decision to transfuse. We also assessed how decision-making is guided by advanced neurological monitoring, and what clinician or patient-specific factors may influence practices.

The survey was sent to 531 clinicians, of which 282 (53%) responded. In a hypothetical patient with high-grade SAH, the mean Hb concentration at which clinicians would choose to administer RBCs was 8.2 (8.1-8.3) g/dl. Transfusion practices were comparatively more restrictive with low-grade SAH [7.9 (7.7-8.0) g/dl] (p<0.0001) and more liberal with delayed cerebral ischemia (DCI) [8.6 (8.5-8.7) g/dl] (p<0.0001). In each setting, there was a wide range of opinions ( Figure 1 ). In multivariable analysis, neurosurgeons had more liberal Hb goals than intensivists (p=0.003), and were more likely to administer two, rather than one unit of RBCs (56% vs. 19%, p<0.0001). Use of protocols was predictive of more restrictive practices (p=0.03 for low-grade SAH) and multi-modal neurological monitoring was associated with a more liberal approach (p=0.04 for DCI). The proportion of clinicians more likely to transfuse based on P bt O 2 values of <15, 15-20 and 20-25 mmHg was 65%, 22%, and 1%, respectively ( Figure 2) . The corresponding proportions for lactate: pyruvate ratios of >40, 35-40, 30-35 and 25-30 were 36%, 18%, 5% and 2%, respectively.

There is widespread variation in transfusion thresholds for SAH patients across North America. Practices are heavily influenced by the specific, dynamic clinical characteristics of patients and may be further modified by clinician specialty, use of protocols, and advanced neurological monitoring. 

Rebleeding of ruptured aneurysms is a leading cause of morbidity and mortality after subarachnoid hemorrhage (SAH). Studies have identified risk factors for rebleeding and its impact on outcome; however, little is known about the neurological and medical complications of rebleeding, including its effect on the development of vasospasm and delayed cerebral ischemia (DCI).

A randomized, nested case-control study was performed on patients in a SAH prospective outcomes database. Ninety-six cases of rebleeding and 207 controls matched for Hunt Hess Grade and presence of giant aneurysm (>10mm) were identified between the years of 1996 and 2009. Poor outcome was defined as modified Rankin Score of

There were no differences in the baseline demographic, clinical, or radiological characteristics. There was no association between rebleeding and the development of infectious complications such as pneumonia, urinary tract infections, sepsis, or meningitis. In a multivariable logistic regression of medical complications, rebleeding was associated with myocardial ischemia (OR: 5.1, 95% CI: 1.6 -16.2, P = 0.005), congestive heart failure (OR: 4.0, 95% CI: 1.4 -11.5, P = 0.009), and the development of fever (OR: 2.8, 95% CI: 1.2 -6.3, P = 0.02)). There was no statistically significant difference between groups in the incidence of DCI (23% vs. 23%, p=1.0) or in the mean postbleed day of vasospasm onset (6 +/-3 vs. 6 +/-3, p=0.6). Patients who rebled were more likely to have poor outcomes at 3 months (79% vs 50%, p=0.001).

Rebleeding disposes patients with SAH to higher risk of fever and cardiac complications. However, there is neither an increase in the incidence of DCI, nor acceleration in the time to DCI. Future studies interventions aimed at improving outcomes by reducing the risk of rebleeding should also focus on treatment and prevention of secondary medical complications. 

Out of hospital cardiac arrest occurs at a rate of 300,000 a year in the U.S and most of the victims are killed within minutes. Only 15% are successfully resuscitated but 3% of total victims will survive with good neurological outcome. Therapeutic hypothermia after cardiac arrest reduces cascades of cytotoxic reactions in the brain and it improves neurological outcomes by 15-25%. Our hospital is a 350 bed community hospital in suburban Los Angeles, California and we began using the state of the art computerized endovascular cooling catheter system, Thermogard IVTM to treat the cardiac arrest victims in April of 2009. We now have treated 14 patients and the results are so remarkable with 50% good survivals that we would like to share our experiences with others.

The treatment was initiated in the ED and continued at the ICU. Patients were eligible for inclusion when brought by EMS being unconscious after out of hospital cardiac arrest. They were intubated, intravenously sedated and given paralytics for mechanical ventilation. Cooling catheter was inserted femorally or subclavian if patient is within the window of 6 hours from the collapse and GCS less than 8.

There were 14 consecutive patient treatments from 4/13/09 to 5/10/10. Patient survivals to a good neurological recovery was astounding 50% (7 out of 14) as compared to a historical recovery range of 3-15%. Patients tolerated the mild hypothermia very well with remarkably low rates of shivering, coagulopathies and thrombocytopenia.

Our experience suggests that mild hypothermia therapy using endovascular cooling catheter within 6 hours after the cardiac arrest is safe and improve neurological outcome dramatically.Community hospitals need to consider developing a protocol to treat the cardiac arrest victims who have ROSC but remains unconscious.Financial Support: Primary author is a member of speaker's bureau of Zoll Circulation since 3/12/2010.