Cholera rapid diagnostic tests at the host-microbe interface: Key Considerations for Global Deployments

BackgroundEffective cholera outbreak response requires accurate bedside rapid diagnostic tests (RDTs) because access to laboratories is often limited. Formative studies suggest cholera diagnostics have multiple vulnerabilities, including antibiotics and predation by bacteriophage ( phage) specific to Vibrio cholerae (Vc). MethodsWe conducted a prospective nationwide study in Bangladesh among over 2000 patients with diarrhoeal disease to characterize how these vulnerabilities impact RDT performance. Assays included culture, qPCR and mass spectrometry. FindingsWith the current gold standard of culture or qPCR Vc positivity, we found no effect of phage on RDT performance. When the diagnostic criteria were expanded to include phage, there was a small decrease in RDT sensitivity. In contrast, large increases in sensitivity and specificity were observed among patients with moderate and severe dehydration. Using the expanded definition, the odds of RDT positivity decreased among cholera patients with phage exposure. The effect was most robust among patients with severe dehydration. Antibiotic were detected in over 80% of samples by LC-MS/MS which limited testing for effects on RDTs. Applying these findings, we estimated that restricting RDT use to severe patients with no reported antibiotic exposure increases sensitivity by 50% compared to unrestricted use. If phage were a diagnostic proxy for Vc, we estimate RDT would miss an additional 17% of cholera cases. InterpretationCholera RDTs have critical limitations that require consideration in global deployments. Inclusion of phage detection in diagnostic criteria may improve case detection which requires further study. The impact of these findings likely extends to other diseases where diagnostics share similar vulnerabilities. RESEARCH IN CONTEXTO_ST_ABSEvidence before this studyC_ST_ABSWe conducted two Pubmed searches for reports published after Jan 1, 2000, in all languages. The first search terms were [bacteriophage OR phage] AND [pathogen] AND [diagnostic test]. The primary search criteria identified 57 publications. A secondary criterion excluded reviews and papers on phage display technology. Among twenty remaining articles, bacteriophage were used as proxies for pathogens within diverse genera (Yersinia, Mycobacterium, Burkholderia, Staphylococcus, Salmonella, Bacillus, Escherichia, Acinetobacter, Vibrio); phage profiles were also used as biomarkers for infection. The second search terms were [rapid diagnostic test OR RDT] AND [Vibrio]. After excluding reviews, 101 papers were identified that covered immunologic (antibody-based lateral flow assays), molecular (PCR, qPCR, nl-qPCR) and mass spectrometry-based assays; only 5 articles related to vibriophage detection of which three were permutations of the same initiative. Added value of this studyTo our knowledge, this is the first study to investigate the vulnerabilities of RDT performance (e.g., sensitivity and specificity) to virulent bacteriophage and antibiotics in a large multi-site study. The hypothesis that antimicrobials and disease severity might impact RDT performance by decreasing or increasing the target number, respectively, is not novel. However, there is a lack of literature from large clinical studies that rigorously tests this hypothesis. Therefore, the added value of this study is a pragmatic evaluation of this hypothesis and a proof of concept for how to test these complex questions for other less tractable diseases. Implications of all the available evidenceInside the cholera field, we applied the findings by estimating that changing to a mode of restricted RDT use in which RDTs are reserved for severely dehydrated patients who report no antibiotic exposure, significantly improves the sensitivity. Furthermore, we estimate that an additional 17% of cases would be missed by RDT if the case definition were expanded to include phage detection as a proxy for the pathogen. In this context, the RDT remains with limitations that require consideration when optimizing global RDT deployments. Implications likely extend to other diseases where diagnostics share similar vulnerabilities.