key: cord-0943753-lvkdbqrz authors: Navapurkar, V.; Bartholdson-Scott, J.; Maes, M.; Higginson, E.; Forrest, S.; Pereira Dias, J.; Parmar, S.; Heasman-Hunt, E.; Polgarova, P.; Brown, J.; Titti, L.; Routledge, M.; Sapsford, D.; Torok, E.; Enoch, D.; Wong, V.; Curran, M. D.; Brown, N.; Herre, J.; Dougan, G.; Conway Morris, A. title: Development and implementation of a customised rapid syndromic diagnostic test for severe pneumonia date: 2020-06-05 journal: nan DOI: 10.1101/2020.06.02.20118489 sha: 049277f1abca0fc5e46e5309994cc51f3064cf91 doc_id: 943753 cord_uid: lvkdbqrz Background The diagnosis of infectious diseases has been hampered by reliance on microbial culture. Cultures take several days to return a result and organisms frequently fail to grow. In critically ill patients this leads to the use of empiric, broad-spectrum antimicrobials and mitigates against stewardship. Methods Single ICU observational cohort study with contemporaneous comparator group. We developed and implemented a TaqMan array card (TAC) covering 52 respiratory pathogens in ventilated patients undergoing bronchoscopic investigation for suspected pneumonia. The time to result was compared against conventional culture, and sensitivity compared to conventional microbiology and metagenomic sequencing. We observed the clinician decisions in response to array results, comparing antibiotic free days (AFD) between the study cohort and comparator group. Findings 95 patients were enrolled with 71 forming the comparator group. TAC returned results 61 hours (IQR 42-90) faster than culture. The test had an overall sensitivity of 93% (95% CI 88-97%) compared to a combined standard of conventional culture and metagenomic sequencing, with 100% sensitivity for most individual organisms. In 54% of cases the TAC results altered clinical management, with 62% of changes leading to de-escalation, 30% to an increase in spectrum, and investigations for alternative diagnoses in 9%. There was a significant difference in the distribution of AFDs with more AFDs in the TAC group (p=0.02). Interpretation Implementation of a customised syndromic diagnostic for pneumonia led to faster results, with high sensitivity and measurable impact on clinical decision making. Funding Addenbrookes Charitable Trust, Wellcome Trust and Cambridge NIHR BRC The diagnosis of infectious diseases has been hampered by reliance on microbial culture. Cultures take several days to return a result and organisms frequently fail to grow. In critically ill patients this leads to the use of empiric, broad-spectrum antimicrobials and mitigates against stewardship. Single ICU observational cohort study with contemporaneous comparator group. We developed and implemented a TaqMan array card (TAC) covering 52 respiratory pathogens in ventilated patients undergoing bronchoscopic investigation for suspected pneumonia. The time to result was compared against conventional culture, and sensitivity compared to conventional microbiology and metagenomic sequencing. We observed the clinician decisions in response to array results, comparing antibiotic free days (AFD) between the study cohort and comparator group. Findings 95 patients were enrolled with 71 forming the comparator group. TAC returned results 61 hours (IQR 42-90) faster than culture. The test had an overall sensitivity of 93% (95% CI 88-97%) compared to a combined standard of conventional culture and metagenomic sequencing, with 100% sensitivity for most individual organisms. In 54% of cases the TAC results altered clinical management, with 62% of changes leading to deescalation, 30% to an increase in spectrum, and investigations for alternative diagnoses in 9%. There was a significant difference in the distribution of AFDs with more AFDs in the TAC group (p=0.02). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint For many decades the diagnosis of infectious diseases has relied on a combination of clinical assessment and the culture of microorganisms. Various factors mitigate against ex vivo growth, including the relative fastidiousness of organisms and the presence of factors which impair division or kill the organism -most notably intercurrent antimicrobial chemotherapy. As a result, it is common for patients with clear clinical evidence of infection to have negative microbiological cultures. 1,2 Furthermore, the time taken to obtain a result by conventional culture can be several days. 3 Optimising antibiotic therapy can, therefore, be challenging, especially in patients who are at risk of multidrug resistant organisms. 2 In critically ill patients, this frequently results in the empiric use of several broad-spectrum agents, often simultaneously, with predictable consequences for antimicrobial resistance and other forms of antimicrobial-related harm. 4 Conversely, failure to identify the causative organism can lead to failure to select the correct antimicrobial, which is associated with poor outcomes. 5 Pneumonia amongst ventilated, critically ill patients can be especially difficult to diagnose. 6 Most critically ill patients are systemically inflamed, 7 clinical examination is unreliable 8 and there are multiple causes of radiographic lung infiltrates most of which are non-infectious. 9 In the case of ventilator-associated pneumonia (VAP) it is estimated that only 20-40% of patients suspected of having infection actually have it confirmed, 10 although this confirmation relies on the imperfect diagnostic standard of microbiological culture. The development of host-based biomarkers for infection, such as C-reactive protein, 11 procalcitonin, 12 and alveolar cytokine concentrations 13 have been advanced as useful measures to help rationalise antimicrobial use. However, their utility in the diagnosis 11, 12 and antibiotic stewardship 14, 15 of pneumonia has been challenged. There is, therefore, a pressing need for rapid, sensitive, multi-pathogen focussed diagnostic tests for pneumonia which can aid effective and efficient antimicrobial optimisation. The use of TaqMan array cards (TAC), enabling single-plex real-time polymerase chain reaction (RT-PCR) to simultaneously target multiple respiratory pathogens, has previously shown promising results compared to conventional culture methods. 16 However, previous experience had demonstrated that multiplex PCR with restricted coverage of common respiratory pathogens had a limited impact on clinical decision making. 17 Most respiratory molecular diagnostics focus largely on viral pathogen detection which will miss important opportunistic bacterial pathogens that cause infections amongst critically ill patients. In this study we set out to develop and evaluate the clinical utility of a RT-PCR based syndromic diagnostic for severe pneumonia, using a customised TAC loaded with primers for 52 known respiratory pathogens covering 17 viral, 30 bacterial, and 5 fungal targets. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint Detailed methods are included in the Supplementary appendix. Patients were recruited from a single, 20 bedded teaching hospital Intensive Care Unit (ICU). The unit is a mixed general medical-surgical unit which supports transplantation and haematology-oncology services. The unit does not admit cardiothoracic or neurotrauma patients. Prospectively identified patients were eligible for inclusion if they were receiving invasive mechanical ventilation, the treating clinician suspected pneumonia and was planning to perform diagnostic bronchoscopy. Patients who were not included in the study due to lack of study team availability had their data reviewed retrospectively and acted as a comparator group. Additional comparator patients were included from the month prior to and the month following the study. Organism coverage for the TAC was selected from the known microbial flora encountered in the ICU, supplemented by review of the literature concerning causative organisms reported in severe pneumonia 6 and the authors' previous experience of molecular diagnostics in pneumonia. 17, 18 The species covered by the card are shown in figure 1. The card also included primers to target the endogenous control RNase P and the internal control MS2, all primers were validated prior to clinical use (supplemental section). Bronchoscopy was undertaken in accordance with existing unit protocols and similar to that used in previous studies. 10 An ethylenediaminetetraacetic acid (EDTA) blood sample and blood culture (BACTEC, Beckton Dickinson, Winnersh, UK) were taken at the time of bronchoscopy. Bronchoalveolar lavage (BAL) samples were processed according to the UK Standards for Microbiology Investigations (SMI). 19 Further details of conventional culture and in-house multi-plex PCR assays are contained in the supplemental section. Following review by a consultant microbiologist, results were returned to the ICU team. Clinical microbiology advice was available 24 hours/day, and patients underwent daily combined ICU-Microbiology multi-disciplinary review in keeping with existing unit practice. The study did not mandate any course of action by the treating clinical team, however the notes were reviewed to assess whether the information from the TAC had influenced patient management. Conventional microbiology results were returned to clinicians via the electronic health record, however in practice these were returned some time after the TAC results. Data were collected on the use of antibiotics amongst patients in the study and from the comparator group, with the number of days alive and free of antibiotics in the seven days following lavage compared (antibiotic-free days, AFDs) 15 . The TAC was compared to a composite reference standard of conventional microbial culture and PCR, and shotgun metagenomics, as it is recognised that culture alone is an imperfect measure of the microbial composition of a sample. Details of the metagenomic sequencing are in the supplemental methods. A flow model for establishing diagnostic tests and decision points for antibiotic prescribing was created (figure S1). Costs for conventional microbiological tests were obtained from Public Health England, whilst costs for the TAC were determined by the cost of the array and staff time to run. The protocol, agreed prior to study initiation, set out the primary and secondary endpoints. The primary endpoints were agreement between conventional microbiology, with sequencing added to form a composite reference during the study, and the TAC, assessed by sensitivity and median time to result for conventional culture and TAC assessed by Wilcoxon matched-pairs test. Where a reference standard was missing this was treated as 'negative', in accordance with pessimistic assumptions about index test performance. Indeterminate results were excluded from analysis. Analysis was conducted using Graphpad prism v5.0 (Graphpad inc, City, CA, USA). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint As the clinical performance characteristics of TAC were not known beforehand, and the incidence of infections with specific organisms could only be inferred from conventional culture which is known to be insensitive, a formal power calculation was not performed. A pragmatic study size of 100 patients was selected to balance cost against including sufficient numbers to be able to make a judgement on card utility. The study was approved by the Leeds East Research Ethics Committee (17/YH/0286) and registered with clinicaltrials.gov (NCT03996330). The retrospective assessment of routinely collected data from the comparator group received a consent waiver and was conducted under a protocol approved by the institutional review board (A095506). The study funders had no role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; nor in the decision to submit the paper for publication The sequences/assays on the card were validated. To ensure cards were spotted correctly with the primers and probes in their assigned pods, the nine synthetic plasmids constructs containing all the target sequences were processed to quality-check each new batch of TAC cards in a checker-board fashion (see supplemental section) All five microorganisms from the Quality Control for Molecular Diagnostics 2018 Sepsis EQA Pilot Study were successfully detected (table S1). Recruitment 100 patients were recruited between February 2018 and August 2019. The recruitment diagram is shown in figure 2 . 64% of eligible patients were recruited, with lack of the study team availability the only reason for nonrecruitment. Table 1 shows demographic and clinical features of the study population and comparator group. Five patients were recruited twice, having suffered a subsequent respiratory deterioration following the initial episode. These episodes occurred a median of 11 days apart (range 4-42 days). In one further patient samples were taken from two separate segments of the lung during the same bronchoscopic procedure. As these were taken at the same time only the primary sample corresponding to the site of maximal radiographic change has been included in the main results. The results of both samples are presented in table S2 for completeness. Although transitory reduction in oxygenation was common during bronchoscopy, no patient in the study suffered any protocol-defined serious adverse events. The co-primary outcome for this study was time to result for the microarray compared to conventional microbial culture. Median difference in time to result between TAC and conventional culture was 61 hours (IQR 42-90) (figure S2). The minimum TAC time to return was 4 hours, with median time to result 22 hours (IQR 7-24), most of the delays in results arising from the inability to run the card outside routine working hours. The total number of positive targets on the TAC across the 100 patients was 248 (table S3) , covering 112 bacteria, 18 fungi and 50 viruses (table 2). The total number of negative targets was 4952 with 21 patients having completely negative TAC results (table S3) . To compare the card, Influenza A sub-types, HSV glyD and Enterococcus faecium ddl gene targets were excluded if they were a duplicate to Flu A, HSV and E. faecium, respectively, as was the mecA resistance marker as this gene was not looked for by conventional microbiology or by sequencing analysis bringing the total to 226 'hits'. Conventional microbial culture and conventional PCR were undertaken on all 100 samples, whilst metagenomic sequencing was performed on residual BAL samples from 98 (two patients with suspected containment level 3 infection could not be sequenced). Conventional microbial culture was positive for 25 organisms in 22 patients, with additional 41 organisms detected by conventional PCR or antigen testing in a further 30 patients (tables 2 and S3). A combination of shotgun and 16S amplicon sequencing approaches revealed 116 organisms in 72 patients (tables 2 and S3), but did not yield any significant reads for 26 patients, of which 18 matched the . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint negative patients identified by TAC and conventional microbiology. Out of the 80 samples with pathogens detected, only 29 presented with a single organism. In the majority of cases (51 samples), we observed mixed infections (table S3 and figure S3) . Taking the combined standard of conventional microbiology and sequencing, a total of 146 organisms were detected, with 136 of these detected by TAC (overall sensitivity 93% 95% CI 88-97%) (table 2) . 90 organisms were detected by TAC only. Looking at all TAC targets across the 100 patient samples, 4952 tests were negative by TAC, giving a nominal specificity of 98% (95% CI 98-99%). In terms of the 10 'missed organisms', one organism, that was positive by both culture and sequencing albeit in different patients, was Citrobacter freundii, for which we did not have a sequence on the card. A further 5 pathogens were detected by sequencing (Staphylococcus aureus, Legionella sp, and Staphylococcus epidermidis) or both culture and sequencing (E. faecium in two patients), and although detected by TAC, these did not pass the internal quality control standards required for reporting and were considered 'negative' results. The remaining three organisms, two Rhinovirus by conventional PCR and one Staphylococcus sp by sequencing, were not detected by TAC at all. In one case the conventional culture was reported as 'Mixed respiratory tract flora at >10 4 CFU/ml' with TAC detection of E. faecium, C. albicans, and Staph. epidermidis which was deemed an indeterminate result for conventional culture, however sequencing was successful on this sample. One case of Aspergillus fumigatus was detected on the TAC, and although no Aspergilli were grown on culture the lavage galactomannan antibody test was highly positive (5.92 Units, reference range <0.5units) and the patient had radiographic evidence of fungal pneumonia (ground glass halo sign on computed tomography of the chest) alongside risk factors for fungal infection (neutropaenia, mechanical ventilation, sepsis, and broadspectrum antibiotic use). However, two further patients met similar criteria for suspected fungal infection but were negative for aspergillus by TAC. Whilst these counted as 'true negatives' as we had set out in our original protocol, we suggest caution in the use of this card for ruling out aspergillus infection. Notably, in the 10 samples from the 5 patients enrolled twice, the organisms detected demonstrated remarkable homogeneity between samples taken within the same ICU admission (table S2) , as did samples taken from two separate areas of the lung on the same day (table S2) . In total, 25 organisms were grown on conventional culture at >/=10 4 CFU/ml, the conventional cut off for quantitative culture of lavage. 20 The median cycles to threshold (Ct) for these organisms that were also detected by TAC was 27 (IQR 24-29, range 20-33). A similar range of Ct values was noted in the QCMD pilot study samples (table S1). Nuclear material extracted from the EDTA blood sample was run on the array card. This was compared with blood culture taken at the time of bronchoscopy. The yield from both conventional culture and TAC was limited, with 5 positive organisms on conventional testing and 20 organisms detected on the array card (table S4) . Streptococcus pneumoniae was the only bacteria detected from blood on the array, being detected 5 times. 14 of the organisms were detected in both lavage and blood, including all the cases of Streptococcus pneumoniae in blood. In 54 (54%) of cases the results from the TAC led to a change in management of the patients (table 3) . 34 (63%) of these decisions involved decrease in antimicrobial intensity, whilst 17 (31%) involved broadening of the spectrum of coverage and in five cases the result prompted a search for an alternative diagnosis (two also involved de-escalation of antimicrobials). In some patients several changes occurred. In a further five cases the research team's opinion was that management could have been changed (all five cases would have been deescalation). Patients in the comparator group demonstrated fewer days alive without antibiotics (distribution of AFDs shown in figure 3 ). The costs of conventional microbiological testing were £222 (US$ 286), whilst the costs of TAC and bacteriological culture for sensitivity testing was £171 (US$ 220), a difference of £51 (US$65). Figure S1 shows workflow with and without TAC. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint In this study we demonstrate that a customised molecular diagnostic, designed to meet the needs of a specific clinical setting can be developed, produce accurate results in a clinically important timeframe and influence patient management decisions. This was within the context of a well-established antimicrobial stewardship programme. Results were returned to clinicians 52 hours faster than conventional cultures, and the accuracy of the results was confirmed by culture, sequencing or both. Notably, many previous studies in respiratory infection specifically exclude critically ill patients, who are those who may suffer the most from inappropriate antimicrobial therapy 5 but are also at greatest risk of antimicrobial-associated harm. 4 Importantly this can be achieved at no additional cost to the laboratory. Molecular diagnostic platforms for respiratory infection syndromes have, until recently, largely focussed on viral pathogens. Whilst distinguishing viral from bacterial infection can be helpful in managing uncomplicated infection, the frequent occurrence of bacterial co-infection in hospitalised patients 18 and the need to optimise antimicrobial therapy whilst limiting the over-use of these precious drugs has led to repeated calls for effective bacterial-focussed diagnostics. 21 Patients who develop critical illness as a result of pneumonia present a particular challenge, as these infections may involve a variety of bacterial, fungal and viral pathogens. 2 Although the source of acquisition of severe pneumonia may indicate the risk of particular organisms, 6, 22, 23 this is an imperfect method of predicting the causative organism. 24 The difficulty in predicting the causative organism, coupled with the insensitivity of growth-based culture detection leads to frequent use of multiple, broad-spectrum antimicrobials. Even when culture results in growth of an organism, the time-frame on which these results are available mitigates against antimicrobial stewardship, and in practice de-escalation is rare 25 as there remains a constant concern about organisms which remain undetected. Although there is now widespread acceptance of the presence of a respiratory microbiome, 26, 27 the lungs of ventilated patients present a particular challenge to highly sensitive molecular diagnostics. The proximal respiratory tract of ventilated patients becomes rapidly colonised with predominantly Gram negative organisms 28 as the microbiome is disrupted by incompletely understood mechanisms including systemic inflammation, translocation of enteric bacteria and intercurrent antimicrobial therapy. 29, 30 This can occur in the absence of infection, and there is a risk that highly sensitive techniques will be more susceptible to the detection of pathogenically irrelevant colonising organisms, driving an unintended increase in antimicrobial use. The use of quantitative cultures and protected lower airways specimens has been advocated as ways of distinguishing infection from colonisation. 20, 31 We adapted this approach in this study, using the quasi-quantitative Ct value provided by RT-PCR and testing protected broncho-alveolar samples. This ensured that the dominant change in practice seen was the de-escalation of antimicrobials. However, it should be noted that although almost all studies of molecular diagnostics focus on reduction in antimicrobial use 15, 32 we found a substantial proportion of changes involved broadening the coverage through agent escalation or addition to cover organisms that had not been anticipated. Had our study been judged by conventional 'reduction in antimicrobial use' the modest reduction in antibiotic-free days seen may have been considered irrelevant. We suggest that consensus measures of antimicrobial optimization be developed to assess the clinical utility of novel microbial diagnostics. One of the problems that has beset microbial diagnostics studies has been the absence of a 'gold standard' against which the candidate can be assessed, 32 as it is acknowledged that conventional culture is imperfect. For this study we developed a new composite 'confirmation standard', using both conventional culture and metagenomic sequencing. As can be seen in table 2 and table S3 , the TAC hits were confirmed by one or both of these techniques in nearly all cases whilst there were very few 'missed' organisms. The card was, overall, more sensitive than either of the confirmatory techniques. We do not believe these additional TAC hits to be true 'false positives' but rather reflect the enhanced sensitivity of targeted PCR. Notably, all of these TAC-only detections occurred at high Ct values, indicating limited and almost certainly clinically irrelevant numbers of organisms. We therefore believe that the sequencing and culture results can give clinicians considerable confidence in the results provided. The selection of organisms targeted on the card was crucial, and informed by our previous experience where omission of key organisms of concern significantly limited the impact of a similar card. 17 Given the case mix of our unit, with a high proportion of immunosuppressed patients, we opted to include a number of low pathogenicity organisms, such a coagulase-negative Staphylococci, Enterococci and Candida albicans on the card. The detection of these organisms can be challenging to interpret, and indeed given the knowledge that . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint many critically ill patients may be immunoparietic even if not classically immunosuppressed 6 their significance remains uncertain. Their presence, and the frequency of polymicrobial detection, indicates a dysbiotic response in the lungs of critically ill patients, 29, 30, 33 and knowing how to respond to this dysbiosis remains an area of uncertainty. However, in seeking to develop a card which could be applied to all our patients with pneumonia, rather than only subsets, we opted for broad coverage and close consultation with clinical microbiologists to support interpretation and implementation. The organisms covered by our card were shown to be appropriate in the context of critically ill ventilated patients, with only one organism cultured that was not included on the card. One of the advantages of the TAC as a platform is its customizability, and single targets can be easily added or modified without having to re-optimise the entire panel, which would also allow it to be adapted to other units or rapidly modified to deal with an emerging infectious threat. It is also possible to include antimicrobial resistance markers, which might further help guide antimicrobial management. We only included one resistance gene, MecA, and because of the low prevalence of MRSA in our unit this did not impact management. This study had a number of strengths. We developed a test which specifically met the needs of a particular unit, and implemented it in the context of a well-established stewardship programme. We managed to include a significant proportion of the patients undergoing bronchoscopic investigation for pneumonia, and thus our results are likely to be representative of our unit's workload. Whilst not a randomised trial, our comparator group was similar, and by being drawn from the same time period of the study and having been managed by the same clinicians, we believe the differences seen are due to the diagnostic approach used. Even when clinicians are presented with a validated, accurate test, achieving a change in prescribing practice is challenging 15 and it is encouraging to see the changes reported here. We set out our schema for managing the approach to results of the TAC in the supplemental results ( figure S4 ). In conclusion, we have developed a customised respiratory diagnostic which covers a broad range of pathogenic organisms. In a pragmatic, observational study of the implementation of this tool we were able to see significant impact on clinical decision making, leading to antimicrobial optimization in a majority of patients tested and reflected in changes relative to a comparator group. The TAC was validated against two distinct standards, and can be adapted for use in units or centres with differing flora. We believe this represents a significant advance in the diagnostics of severe pneumonia, and have adopted this technology across our hospital for the investigation of such patients. Authorship (CReDIT) VN -Conceptualisation, resources, investigation, writing-review and editing, project administration, funding acquisition, supervision. JBS -resources, investigation, data curation, formal analysis, writing-original draft. MM-resources, investigation, writing-review and editing. EH-investigation, writing-review and editing. SF-investigation, writing-review and editing. JD-investigation, writing-review and editing. SP-investigation, writing-review and editing. EHH-investigation, writing-review and editing. PP-investigation, data curation writing-review and editing. JB-investigation, data curation writing-review and editing. LT-investigation, data curation writing-review and editing. MR-data curation writing-review and editing. DS-data curation writing-review and editing. ET-conceptualisation, investigation, data curation writing-review and editing DE-investigation, writing-review and editing. VW-formal analysis,investigation, data curation writing-review and editing. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. (7):594-600. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint Table 2 : Organisms detected by conventional microbiological testing (left hand column), by TAC (middle column), and by microbial sequencing (right hand column). * Legionella urinary antigen test positive. ** Positive BAL galactomannan enzyme immunoassay (>0.5 units) with CT consistent with fungal pneumonia and known risk factors but fungal cultures were not positive. ***One hit not found in same patient. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 5, 2020. Table 3 : changes in treatment made as a result of findings from the TAC (some patients had more than one change within a given type -e.g. both reduction spectrum and cessation of antimicrobials) . *in two cases antimicrobials were reduced to prophylactic dose as patients remained at risk of opportunistic infections . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint Figure 1 : layout of TAC . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint Figure 3 : Distribution of days alive and free of antibiotics in the seven days following bronchoscopy and lavage in the TAC and comparator cohorts. Following first lavage only for patients who had more than one BAL during ICU admission. Numbers in each category and percentage shown below graph, p value by Mann-Whitney U test. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.02.20118489 doi: medRxiv preprint Microbial aetiology of community-acquired pneumonia and its relation to severity Prevalence and Outcomes of Infection Among Patients in Intensive Care Units in 2017 Comparison of Unyvero P55 Pneumonia Cartridge, in-house PCR and culture for the identification of respiratory pathogens and antibiotic resistance in bronchoalveolar lavage fluids in the critical care setting Antimicrobial-associated harm in critical care: a narrative review Clinical Importance of Delays in the Initiation of Appropriate Antibiotic Treatment for Ventilator-Associated Pneumonia Management of pneumonia in intensive care The Natural History of the Systemic Inflammatory Response Syndrome (SIRS) Comparative diagnostic performances of auscultation, chest radiography, and lung ultrasonography in acute respiratory distress syndrome Causes of fever and pulmonary densities in patients with clinical manifestations of ventilator-associated pneumonia Diagnostic importance of pulmonary interleukin-1{beta} and interleukin-8 in ventilator-associated pneumonia C-reactive protein as a marker of ventilator-associated pneumonia resolution: a pilot study Procalcitonin for diagnosis of bacterial pneumonia in critically ill patients during 2009 H1N1 influenza pandemic: a prospective cohort study, systematic review and individual patient data meta-analysis Diagnostic accuracy of pulmonary host inflammatory mediators in the exclusion of ventilator-acquired pneumonia Procalcitonin-Guided Use of Antibiotics for Lower Respiratory Tract Infection Biomarker-guided antibiotic stewardship in suspected ventilator-associated pneumonia (VAPrapid2): a randomised controlled trial and process evaluation Clinical evaluation of a multi-parameter customized respiratory TaqMan® array card compared to conventional methods in immunocompromised patients Evaluating the use of a 22-pathogen TaqMan array card for rapid diagnosis of respiratory pathogens in Intensive Care Public Health England. Investigation of bronchoalveolar lavage, sputum and associated specimens UK SMI B 57 issue 3 Ventilator-associated pneumonia Is the Present Definition of Health Care-Associated Pneumonia the Best Way to Define Risk of Infection with Antibiotic-Resistant Pathogens? Clinical Characteristics and Treatment Patterns Among Patients With Ventilator-Associated Pneumonia Integrating host response and unbiased microbe detection for lower respiratory tract infection diagnosis in critically ill adults The dynamics of the pulmonary microbiome during mechanical ventilation in the intensive care unit and the association with occurrence of pneumonia Bacterial colonization patterns in mechanically ventilated patients with traumatic and medical head injury. Incidence, risk factors, and association with ventilator-associated pneumonia The dynamics of the pulmonary microbiome during mechanical ventilation in the intensive care unit and the association with occurrence of pneumonia Pathophysiological role of respiratory dysbiosis in hospital-acquired pneumonia Evaluation of the effect of diagnostic methodology on the reported incidence of ventilator-associated pneumonia Use and Evaluation of Molecular Diagnostics for Pneumonia Etiology Studies