Summary
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 Addenbrooke’s Charitable Trust, Wellcome Trust and Cambridge NIHR BRC
Evidence before this study The problems of microbial culture-negative pneumonia are well described in the literature. We conducted Medline searches from 1996 to December 2020 using the MeSH terms ‘polymerase chain reaction’, ‘pneumonia’ and ‘critical care’. Most reported studies concerned viral PCR only or single organism PCR and the great majority of critical care-based studies were retrospective studies on stored samples. We identified one study which retrospectively examined the impact on antimicrobial prescribing of a non-quantitative assay designed for blood cultures applied to respiratory samples. This analysis showed a reduction in antibiotic use. A small randomised trial examining blood-based PCR in patients with abdominal and pulmonary sepsis demonstrated faster time to result and no difference in time to therapy change or rate of therapy change. Notably several studies recorded that the narrow spectrum of organisms covered by molecular testing may lead to inappropriate withholding of antibiotics if the tests were used to guide therapy. Clinical trials largely excluded the most severe cases of pneumonia, including those requiring mechanical ventilation. Recent studies have demonstrated the difficulty in using novel tests to change entrenched antimicrobial prescribing patterns amongst critical care physicians. Our group has previously investigated TaqMan array card (TAC) technology for syndromic diagnosis, but found that limited pathogen coverage led to limited impact on clinical decision making.
Added value of this study This observational study demonstrated that a customised, broad-spectrum syndromic test for severe pneumonia could lead to substantial change in treatment. This is in the context of an Intensive Care Unit (ICU) with an established antimicrobial stewardship programme and daily input from microbiology and molecular specialists. The assay was proven to be accurate, with almost all organisms detected confirmed by culture and/or metagenomic sequencing, as well as excellent performance against a panel of organisms used for national reference standards.
Implications of all available evidence The use of multiplex and multi-organism molecular approaches offers rapid, sensitive testing which detects organisms in excess of those identified by conventional culture. Thus, a suitably broad range of targets, ideally including antibiotic resistance genes, is required to safely and effectively optimise antimicrobial therapy, and rapid diagnostics. Tests with a narrower range, or which cannot be customised to local microbial flora, may have limited impact on antimicrobial prescribing. Further work is required on determining cut-offs for distinguishing colonisation from infection, alongside integration with host-markers of inflammation and clinical impact.
Introduction
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 unreliable8 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 concentrations13 have been advanced as useful measures to help rationalise antimicrobial use. However, their utility in the diagnosis11,12 and antibiotic stewardship14,15 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.
Methods
Detailed methods are included in the Supplementary appendix.
Setting
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.
Recruitment
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.
Test development
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 pneumonia6 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).
layout of TAC
Sampling procedure
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.
Conventional microbiological testing
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.
Metagenomic sequencing
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.
Economic analysis
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.
Statistical analysis
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).
Study size
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.
Ethical and regulatory approvals and funding
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).
Role of the funders
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
Results
Initial validation
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 non-recruitment. 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.
APACHE II, acute physiology and chronic health evaluation II, Fi02, fraction of inspired oxygen.
Time to result
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.
Comparison of organisms detected by TAC compared to conventional microbiology and sequencing
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’.
* 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.
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 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 >104 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 broad-spectrum 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).
Quantitation
In total, 25 organisms were grown on conventional culture at >/=104 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).
Comparison of conventional blood culture and TAC
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.
Impact on clinical decision making
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 de-escalation). Patients in the comparator group demonstrated fewer days alive without antibiotics (distribution of AFDs shown in figure 3).
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.
*in two cases antimicrobials were reduced to prophylactic dose as patients remained at risk of opportunistic infections
Economic analysis
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.
Discussion and conclusions
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 therapy5 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 patients18 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 rare25 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 organisms28 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 use15,32 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 many critically ill patients may be immunoparietic even if not classically immunosuppressed6 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 challenging15 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.
MDC-Conceptualisation, resources, investigation, writing-review and editing, project administration, funding acquisition, supervision.
NB-Conceptualisation, resources, investigation, writing-review and editing, project administration, supervision.
JH-Conceptualisation, resources, investigation, writing-review and editing.
GD-Conceptualisation, resources, investigation, writing-review and editing, project administration, funding acquisition, supervision.
ACM-Conceptualisation, methodology, resources, investigation, writing-original draft, project administration, formal analysis, funding acquisition, supervision.
Declaration of interest
MDC is the inventor on a patent held by the Secretary of State for Health (UK government) EP2788503, which covers some of the genetic sequences used in this study. All other authors declare no conflict of interest.
Acknowledgements
The authors thank the consultant intensivists of the John V Farman Intensive Care Unit, Drs P Bradley, P Featherstone, S Ford, M Georgieva, A Johnston, R Mahroof, J Martin, J Preller, K Patel, C Summers M Trivedi, J Varley, Pharmacist L Radford and the nursing and physiotherapy teams who managed the patients, and the patients and their families who consented to the study. We also thank Torsten Seemann for access to his Kraken database. The study was funded by Addenbrooke’s Charitable Trust and the NIHR Cambridge Biomedical Resource Centre (grant held by Professor Dougan), Dr Conway Morris is supported by a Clinical Research Career Development Fellowship from the Wellcome Trust (WT 2055214/Z/16/Z). The metagenomic sequencing was funded by the Wellcome Trust.
References
Subject Area
Reviews and Context
0
Comment
0
TRIP Peer Reviews
0
Community Reviews
0
Automated Services
4
Blogs/Media
Author Videos