Clinical interpretation of enteric molecular diagnostic tests
2019; Elsevier BV; Volume: 25; Issue: 12 Linguagem: Inglês
10.1016/j.cmi.2019.08.018
ISSN1469-0691
AutoresGillian A.M. Tarr, Phillip I. Tarr, Stephen B. Freedman,
Tópico(s)Clostridium difficile and Clostridium perfringens research
ResumoIn this issue of Clinical Microbiology and Infection, Tilmanne et al. [[1]Tilmanne A. Martiny D. Quach C. Wautier M. Vandenberg O. Lepage P. et al.Enteropathogens in pediatric gastroenteritis: comparison of routine diagnostic methods to molecular ones.Clin Microbiol Infect. 2019; 25: 1519-1524Abstract Full Text Full Text PDF Scopus (15) Google Scholar] add to the growing body of literature comparing commercial molecular diagnostic testing platforms to conventional laboratory approaches for children with acute gastroenteritis. We commend the authors for highlighting the important issues of co-detection of definite or plausible pathogens in children with gastroenteritis and the high prevalence of such agents in controls without diarrhoea. They report, as have others, that molecular diagnostics detect enteric pathogens in more patients than do older methods (i.e. culture, microscopy and enzyme immunoassay (EIA)) [2Harrington S.M. Buchan B.W. Doern C. Fader R. Ferraro M.J. Pillai D.R. et al.Multicenter evaluation of the BD max enteric bacterial panel PCR assay for rapid detection of Salmonella spp., Shigella spp., Campylobacter spp. (C. jejuni and C. coli), and Shiga toxin 1 and 2 genes.J Clin Microbiol. 2015; 53: 1639-1647Crossref PubMed Scopus (66) Google Scholar, 3Buchan B.W. Olson W.J. Pezewski M. Marcon M.J. Novicki T. Uphoff T.S. et al.Clinical evaluation of a real-time PCR assay for identification of Salmonella, Shigella, Campylobacter (Campylobacter jejuni and C. coli), and Shiga toxin-producing Escherichia coli isolates in stool specimens.J Clin Microbiol. 2013; 51: 4001-4007Crossref PubMed Scopus (56) Google Scholar, 4Deng J. Luo X. Wang R. Jiang L. Ding X. Hao W. et al.A comparison of Luminex xTAG(R) Gastrointestinal Pathogen Panel (xTAG GPP) and routine tests for the detection of enteropathogens circulating in Southern China.Diagn Microbiol Infect Dis. 2015; 83: 325-330Crossref PubMed Scopus (33) Google Scholar]. While increased sensitivity is intuitively appealing because it allows expanded attribution of illness to specific agents, thereby potentially providing diagnostic clarity, such expanded detection also poses important challenges. To advance our understanding of these issues, we believe clinical interpretation and professional guidance should focus on three questions. Laboratory approaches to acute gastroenteritis lack true reference standards (i.e. even culture does not fill this role), obligating other approaches to understand the diagnostic accuracy of molecular panel results. Polymerase chain reaction (PCR) multi-analyte assays are increasingly found to have inaccuracies [2Harrington S.M. Buchan B.W. Doern C. Fader R. Ferraro M.J. Pillai D.R. et al.Multicenter evaluation of the BD max enteric bacterial panel PCR assay for rapid detection of Salmonella spp., Shigella spp., Campylobacter spp. (C. jejuni and C. coli), and Shiga toxin 1 and 2 genes.J Clin Microbiol. 2015; 53: 1639-1647Crossref PubMed Scopus (66) Google Scholar, 5Duong V.T. Phat V.V. Tuyen H.T. Dung T.T. Trung P.D. Minh P.V. et al.Evaluation of Luminex xTAG gastrointestinal pathogen panel assay for detection of multiple diarrheal pathogens in fecal samples in Vietnam.J Clin Microbiol. 2016; 54: 1094-1100Crossref PubMed Scopus (41) Google Scholar, 6Buss S.N. Leber A. Chapin K. Fey P.D. Bankowski M.J. Jones M.K. et al.Multicenter evaluation of the BioFire FilmArray gastrointestinal panel for etiologic diagnosis of infectious gastroenteritis.J Clin Microbiol. 2015; 53: 915-925Crossref PubMed Scopus (313) Google Scholar]. For the Luminex xTAG gastrointestinal pathogen panel (GPP) specifically, Duong et al. found suboptimal sensitivity for every pathogen examined [[5]Duong V.T. Phat V.V. Tuyen H.T. Dung T.T. Trung P.D. Minh P.V. et al.Evaluation of Luminex xTAG gastrointestinal pathogen panel assay for detection of multiple diarrheal pathogens in fecal samples in Vietnam.J Clin Microbiol. 2016; 54: 1094-1100Crossref PubMed Scopus (41) Google Scholar], we described missed Salmonella and norovirus infections [7Zhuo R. Cho J. Qiu Y. Parsons B.D. Lee B.E. Chui L. et al.Alberta Provincial Pediatric EnTeric Infection T. High genetic variability of norovirus leads to diagnostic test challenges.J Clin Virol. 2017; 96: 94-98Crossref PubMed Scopus (9) Google Scholar, 8Kellner T. Parsons B. Chui L. Berenger B.M. Xie J. Burnham C.A. et al.Comparative evaluation of enteric bacterial culture and a molecular multiplex syndromic panel in children with acute gastroenteritis.J Clin Microbiol. 2019; 57Crossref PubMed Scopus (20) Google Scholar], and Tilmanne et al. now report that Luminex did not detect six of 28 Campylobacter spp [[1]Tilmanne A. Martiny D. Quach C. Wautier M. Vandenberg O. Lepage P. et al.Enteropathogens in pediatric gastroenteritis: comparison of routine diagnostic methods to molecular ones.Clin Microbiol Infect. 2019; 25: 1519-1524Abstract Full Text Full Text PDF Scopus (15) Google Scholar]. Specificity is almost always reported as >98% [2Harrington S.M. Buchan B.W. Doern C. Fader R. Ferraro M.J. Pillai D.R. et al.Multicenter evaluation of the BD max enteric bacterial panel PCR assay for rapid detection of Salmonella spp., Shigella spp., Campylobacter spp. (C. jejuni and C. coli), and Shiga toxin 1 and 2 genes.J Clin Microbiol. 2015; 53: 1639-1647Crossref PubMed Scopus (66) Google Scholar, 3Buchan B.W. Olson W.J. Pezewski M. Marcon M.J. Novicki T. Uphoff T.S. et al.Clinical evaluation of a real-time PCR assay for identification of Salmonella, Shigella, Campylobacter (Campylobacter jejuni and C. coli), and Shiga toxin-producing Escherichia coli isolates in stool specimens.J Clin Microbiol. 2013; 51: 4001-4007Crossref PubMed Scopus (56) Google Scholar, 4Deng J. Luo X. Wang R. Jiang L. Ding X. Hao W. et al.A comparison of Luminex xTAG(R) Gastrointestinal Pathogen Panel (xTAG GPP) and routine tests for the detection of enteropathogens circulating in Southern China.Diagn Microbiol Infect Dis. 2015; 83: 325-330Crossref PubMed Scopus (33) Google Scholar], but PCR panels do yield false positive results [[8]Kellner T. Parsons B. Chui L. Berenger B.M. Xie J. Burnham C.A. et al.Comparative evaluation of enteric bacterial culture and a molecular multiplex syndromic panel in children with acute gastroenteritis.J Clin Microbiol. 2019; 57Crossref PubMed Scopus (20) Google Scholar], which are dwarfed by the number of true negatives. How can a clinician know if a result is correct? We have combined quantitative in-house and commercial confirmatory PCR assays and clinical phenotype [[8]Kellner T. Parsons B. Chui L. Berenger B.M. Xie J. Burnham C.A. et al.Comparative evaluation of enteric bacterial culture and a molecular multiplex syndromic panel in children with acute gastroenteritis.J Clin Microbiol. 2019; 57Crossref PubMed Scopus (20) Google Scholar] to try to confirm results, but this degree of verification is impracticable outside research settings. Clinical laboratory workflows may entail confirmatory culture or other testing for positive results, but this delays result reporting and increases costs. Clinicians can also refer to the positive and negative predictive values (PPV, NPV) to understand the likelihood that a test result is true. For relatively rare pathogens, the NPV is often quite high even in the presence of low sensitivity while the PPV can be low and is heavily influenced by even small losses in specificity. For more common pathogens, PPV can be much higher with only modest degradations to NPV. Importantly, the specific testing context may imply higher pre-test probability of a positive, such as when testing a child with bloody diarrhoea for Shigella, or lower pre-test probability, such as when testing a child hospitalized for several weeks for enteric parasites. Decision support tools (i.e. testing guidelines or clinical prediction rules) can theoretically guide clinicians to restrict testing to situations where the probability of a positive is high, one of the pillars of diagnostic stewardship [[9]Morgan D.J. Malani P. Diekema D.J. Diagnostic stewardship-leveraging the laboratory to improve antimicrobial use.JAMA. 2017; 318: 607-608Crossref PubMed Scopus (121) Google Scholar]. However, in the field of paediatric gastroenteritis, current guidelines are inadequate [[10]Tarr G.A.M. Chui L. Lee B.E. Pang X.L. Ali S. Nettel-Aguirre A. et al.Performance of stool testing recommendations for acute gastroenteritis when used to identify children with nine potential bacterial enteropathogens.Clin Infect Dis. 2019; 69: 1173-1182Crossref PubMed Scopus (14) Google Scholar], highlighting the need for improved decision support tools. Detection does not imply that a viable, disease-causing pathogen is present. Molecular panels can test for >20 pathogens simultaneously, including many that rarely, if ever, cause disease in high-income countries (e.g. enteropathogenic Escherichia coli). Detected agents may be non-pathogenic colonizers, and shedding for extended periods after illness resolution may result in detection of pathogens not responsible for the current illness. As Tilmanne et al. [[1]Tilmanne A. Martiny D. Quach C. Wautier M. Vandenberg O. Lepage P. et al.Enteropathogens in pediatric gastroenteritis: comparison of routine diagnostic methods to molecular ones.Clin Microbiol Infect. 2019; 25: 1519-1524Abstract Full Text Full Text PDF Scopus (15) Google Scholar] highlight, quantitative test results might distinguish pathogens causing active disease, such as the fluorescence-based method developed by Leal et al. for C. difficile [[11]Leal Jr, S.M. Popowitch E.B. Levinson K.J. John T.M. Lehman B. Rios M.B. et al.Quantitative thresholds enable accurate identification of Clostridium difficile infection by the Luminex xTAG gastrointestinal pathogen panel.J Clin Microbiol. 2018; 56 (e01885-17)Crossref PubMed Scopus (6) Google Scholar], but establishing cut-points for actionability requires considerable population-based data. Detection of multiple pathogens is also common but highly variable [2Harrington S.M. Buchan B.W. Doern C. Fader R. Ferraro M.J. Pillai D.R. et al.Multicenter evaluation of the BD max enteric bacterial panel PCR assay for rapid detection of Salmonella spp., Shigella spp., Campylobacter spp. (C. jejuni and C. coli), and Shiga toxin 1 and 2 genes.J Clin Microbiol. 2015; 53: 1639-1647Crossref PubMed Scopus (66) Google Scholar, 4Deng J. Luo X. Wang R. Jiang L. Ding X. Hao W. et al.A comparison of Luminex xTAG(R) Gastrointestinal Pathogen Panel (xTAG GPP) and routine tests for the detection of enteropathogens circulating in Southern China.Diagn Microbiol Infect Dis. 2015; 83: 325-330Crossref PubMed Scopus (33) Google Scholar, 5Duong V.T. Phat V.V. Tuyen H.T. Dung T.T. Trung P.D. Minh P.V. et al.Evaluation of Luminex xTAG gastrointestinal pathogen panel assay for detection of multiple diarrheal pathogens in fecal samples in Vietnam.J Clin Microbiol. 2016; 54: 1094-1100Crossref PubMed Scopus (41) Google Scholar, 6Buss S.N. Leber A. Chapin K. Fey P.D. Bankowski M.J. Jones M.K. et al.Multicenter evaluation of the BioFire FilmArray gastrointestinal panel for etiologic diagnosis of infectious gastroenteritis.J Clin Microbiol. 2015; 53: 915-925Crossref PubMed Scopus (313) Google Scholar]. Tilmanne et al. found that the Luminex detected pathogens in up to four times the number of patients compared with conventional diagnostics, but the proportion with co-detected pathogens also increased substantially [[1]Tilmanne A. Martiny D. Quach C. Wautier M. Vandenberg O. Lepage P. et al.Enteropathogens in pediatric gastroenteritis: comparison of routine diagnostic methods to molecular ones.Clin Microbiol Infect. 2019; 25: 1519-1524Abstract Full Text Full Text PDF Scopus (15) Google Scholar]. In such situations, the clinician must decide which is the more (or most) actionable agent. For example, antibiotics that are appropriate for one co-detected organism (i.e. Shigella) might be contraindicated for another (i.e. Shiga toxin-producing E. coli and Salmonella). The condition and history of the host should be considered when assessing the likelihood that a detected pathogen is causative. Clinical phenotype can help distinguish causative from non-causative agents; however, striking clinical differences have been difficult to identify [[12]Nicholson M.R. Van Horn G.T. Tang Y.W. Vinje J. Payne D.C. Edwards K.M. et al.Using multiplex molecular testing to determine the etiology of acute gastroenteritis in children.J Pediatr. 2016; 176: 50-56 e2Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar]. Exposure history may help resolve aetiology. For example, recent antibiotic use may shift the likelihood that detected C. difficile are more than colonizers, and illnesses in international travellers and immigrants may be due to agents not typically considered pathogens. Medical history (e.g. being immunocompromised) should also be considered when interpreting test results. At the population level, we can tease out a pathogen's role by simultaneously evaluating controls without gastroenteritis symptoms. The attributable fraction can be used to estimate the proportion of clinical detections that reflect aetiology as opposed to carriage. Tilmanne et al. report striking differences in detection between cases and controls for some pathogens (e.g. Campylobacter spp.) and equivalence for others (e.g. Entamoeba histolytica) [[1]Tilmanne A. Martiny D. Quach C. Wautier M. Vandenberg O. Lepage P. et al.Enteropathogens in pediatric gastroenteritis: comparison of routine diagnostic methods to molecular ones.Clin Microbiol Infect. 2019; 25: 1519-1524Abstract Full Text Full Text PDF Scopus (15) Google Scholar]. However, they did not determine the attributable fraction, which would have entailed adjusting for confounders of the pathogen–gastroenteritis relationship, such as age and co-detections. The Global Enteric Multicenter Study (GEMS) and Malnutrition and Enteric Disease Study (MAL-ED) [13Kotloff K.L. Nataro J.P. Blackwelder W.C. Nasrin D. Farag T.H. Panchalingam S. et al.Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control study.Lancet. 2013; 382: 209-222Abstract Full Text Full Text PDF PubMed Scopus (2290) Google Scholar, 14Platts-Mills J.A. Babji S. Bodhidatta L. Gratz J. Haque R. Havt A. et al.Investigators M-ENPathogen-specific burdens of community diarrhoea in developing countries: a multisite birth cohort study (MAL-ED).Lancet Glob Health. 2015; 3: e564-e575Abstract Full Text Full Text PDF PubMed Scopus (564) Google Scholar] both employed population attributable fractions, a closely related measure. Studies conducted in a variety of settings that take this next step are urgently needed to determine whether estimates of the attributable fraction are constant across geographic regions. As demonstrated by GEMS and MAL-ED in research contexts, attributable fraction calculations can also be informed by PCR cycle threshold results [14Platts-Mills J.A. Babji S. Bodhidatta L. Gratz J. Haque R. Havt A. et al.Investigators M-ENPathogen-specific burdens of community diarrhoea in developing countries: a multisite birth cohort study (MAL-ED).Lancet Glob Health. 2015; 3: e564-e575Abstract Full Text Full Text PDF PubMed Scopus (564) Google Scholar, 15Liu J. Platts-Mills J.A. Juma J. Kabir F. Nkeze J. Okoi C. et al.Use of quantitative molecular diagnostic methods to identify causes of diarrhoea in children: a reanalysis of the GEMS case-control study.Lancet. 2016; 388: 1291-1301Abstract Full Text Full Text PDF PubMed Scopus (472) Google Scholar]. If clinical quantitative PCR assays were calibrated for this purpose, more nuanced interpretations could be made. We need evidence beyond accuracy and attribution that costly enteric panels improve outcomes. Enteric syndromic panels with rapid turnaround times may avoid delayed or incorrect diagnoses, inappropriate treatment and unnecessary additional testing (e.g. microbiological, radiological, serological) [[16]Grisaru S. Berenger B. Freedman S. Case 19-2018: a 15-year-old girl with acute kidney injury.N Engl J Med. 2018; 379: e34Crossref PubMed Scopus (3) Google Scholar]. They might also stem empiric and inappropriate uses of antibiotics for gastroenteritis. Tilmanne et al. do not indicate if the 12 cases who received antibiotics were appropriately treated but do report that four additional Salmonella cases might have received antimicrobials had the Luminex results been available in real time [[1]Tilmanne A. Martiny D. Quach C. Wautier M. Vandenberg O. Lepage P. et al.Enteropathogens in pediatric gastroenteritis: comparison of routine diagnostic methods to molecular ones.Clin Microbiol Infect. 2019; 25: 1519-1524Abstract Full Text Full Text PDF Scopus (15) Google Scholar]. However, irrelevant results might invite reflexive use of antibiotics when, for example, an enteropathogenic E. coli is identified. Conversely, a potential benefit would be the ability to guide clinicians to look for alternative aetiologies (e.g. appendicitis, intussusception, inflammatory bowel disease) when a panel detects no pathogens. Enteric panels may also prompt targeted, pathogen-specific anticipatory guidance, which might reassure caregivers thereby stemming unnecessary care-seeking behaviours. Studies focused on understanding clinical outcomes, health system policy and economics are needed to gauge the value of enteric molecular diagnostics. Potential outcomes of interest include return outpatient visits, inpatient hospitalization, length of stay, antimicrobial treatment, intravenous fluid administration, invasive disease (i.e. bacteraemia), cost of care, days missed from work and school, long-term sequelae and post-infectious complications. Observational studies in this area are likely to take the form of pre-post designs, or comparisons of medical centres using conventional diagnostics to those using molecular diagnostics or single centres that offer both conventional and molecular tests. Confounding looms large in each of these designs, whether it be from secular trends, inter-centre variation in patient populations and clinical practice, or the factors driving the choice of testing modality for a given patient, and would need to be carefully controlled. Clinical trials to estimate the effect of molecular diagnostics on patient outcomes are also challenging. They cannot be blinded, because the breadth of targets tested and rapid turnaround time are hallmarks of the molecular gastrointestinal pathogen panels and the means through which they are likely to impact clinical care. Study cohorts must be large, to achieve balance in pathogens and because outcomes of greatest interest, such as invasive disease and hospitalization, are relatively rare. Clinicians' testing practices must also be considered, as should patients without diarrhoea (i.e. isolated vomiting). Another crucial aspect to consider is the impact of culture-independent diagnostic testing on bacterial enteropathogen surveillance and antimicrobial susceptibility testing. Without an isolate, public health entities cannot type or 'fingerprint' strains, crucial for identifying outbreaks, or monitor antibiotic resistance trends, and susceptibility testing is necessary both in the clinical context and for public health surveillance. Multiplex molecular tests can also cost substantially more than conventional diagnostics, although sequential or directed testing using specific panels (i.e. limited bacterial, extended bacterial, viral, parasitic) may be more efficient. Cost–benefit analyses have suggested that assay expenses are offset by savings in care in some settings [[17]Goldenberg S.D. Bacelar M. Brazier P. Bisnauthsing K. Edgeworth J.D. A cost benefit analysis of the Luminex xTAG Gastrointestinal Pathogen Panel for detection of infectious gastroenteritis in hospitalised patients.J Infect. 2015; 70: 504-511Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar]. As more studies elucidate the impact of molecular testing on health outcomes, it will be important to update cost analyses with new information. Tilmanne et al. [[1]Tilmanne A. Martiny D. Quach C. Wautier M. Vandenberg O. Lepage P. et al.Enteropathogens in pediatric gastroenteritis: comparison of routine diagnostic methods to molecular ones.Clin Microbiol Infect. 2019; 25: 1519-1524Abstract Full Text Full Text PDF Scopus (15) Google Scholar] and studies like it are important, because they demonstrate the potential advantages and pitfalls of molecular enteric syndromic panels. However, this is just the tip of the iceberg, and large, multicentre studies are needed to clarify diagnostic test performance and disease attribution. Most importantly we need to understand how these tests impact care and outcomes. Although molecular panels for gastroenteritis have entered into widespread use, we need to close the gaps in our current knowledge to understand best practices for their implementation and interpretation. While these data are accruing, we urge diagnostic microbiologists, possibly in combination with professional societies, to offer annotated reports to providers based on current (though fluid) knowledge, indicating the confidence with which the agents identified are likely to cause human disease. Outside the current work, P.I.T. reports personal fees as a consultant and membership on the scientific advisory board for MediBeacon Inc.; he also holds equity in the company. P.I.T. and S.B.F. report personal fees as consultants for Takeda Pharmaceuticals, and P.I.T. reports consulting fees for the Bill & Melinda Gates Foundation on infant infections. S.B.F. reports non-financial support from Luminex Corporation, BioMérieux and Copan Diagnostics to support research activities. In addition, PIT has a patent 20190142976 pending to MediBeacon Inc., and is an unpaid consultant to Immunova relating to development of a therapeutic agent for Shiga toxin-producing E. coli. S.B.F. is supported by the Alberta Children's Hospital Foundation Professorship in Child Health and Wellness. Enteropathogens in paediatric gastroenteritis: comparison of routine diagnostic and molecular methodsClinical Microbiology and InfectionVol. 25Issue 12PreviewStudies of acute gastroenteritis (AGE) are hampered by the lack of routine diagnostic methods with good sensitivity and specificity. Molecular methods are increasingly used for clinical purposes, but the clinical significance of a positive result remains a challenge. In this study we aimed to compare results of routine diagnostic methods and molecular methods in symptomatic children and asymptomatic controls. Full-Text PDF Open Archive
Referência(s)