Accuracy of upper respiratory tract samples to diagnose Mycobacterium tuberculosis: a systematic review and meta-analysis

Abstract

Structured summaryBackgroundPulmonary tuberculosis (PTB) due toMycobacterium tuberculosis(Mtb) can be challenging to diagnose because of difficulty obtaining samples, and suboptimal sensitivity of existing tests. We investigated the performance characteristics and diagnostic accuracy of upper respiratory tract tests for diagnosing PTB and hypothesised they would have sufficient accuracy and utility to improve PTB diagnosis.MethodsA systematic review and meta-analysis was conducted by searching MEDLINE, Cinahl, Web of Science, Global Health, and Global Health Archive databases up to 31/01/2021, a second search was conducted for the period 1/1/2021 - 27/5/2022 (subsequently extended to 6/12/2022) to identify studies that reported on the accuracy of upper respiratory tract sampling for TB diagnosis compared to microbiological reference standards. We used a random-effects meta-analysis with a bivariate hierarchical model to estimate pooled sensitivity and specificity, stratified by sampling method. Bias was assessed using QUADAS- 2 criteria. Study registered with PROSPERO (CRD42021262392).Findings10,159 titles were screened for inclusion, 274 studies were assessed for full text review, and 71, comprising 119 test comparisons published between 1933 and 2022 were included in the systematic review (53 in meta-analysis). For laryngeal swabs, pooled sensitivity was 57.8% (95% CI 50.5-65.0%), specificity was 93.8% (95% CI 88.4-96.8%) and diagnostic odds ratio (DOR) was 20.7 (95% CI 11.1-38.8). Nasopharyngeal aspirate sensitivity was 65.2% (95% CI 52.0-76.4%), specificity was 97.9% (95% CI 96.0-99.0%) and DOR was 91.0 (95% CI 37.8-218.8). Oral swabs sensitivity was 56.7% (95% CI 44.3-68.2%), specificity was 91.3% (95% CI 81.0-96.3%), and DOR was 13.8 (95% CI 5.6-34.0).InterpretationUpper respiratory tract sampling holds promise to expand access to TB diagnosis, including for people who can’t produce sputum. Exploring historical methods using modern microbiological techniques may further increase the options for alternative sample types.Prospective studies are needed to optimise accuracy and utility of sampling methods in clinical practice.FundingHRS is funded by the MRC through the MRC DTP programme at LSTM [Grant number MR/N013514/1].Research in contextEvidence before this studyGlobally in 2021, an estimated 4.2 million of 10.6 million people with incident tuberculosis (TB) disease went undiagnosed, emphasising the urgent need for new diagnostic methodologies. Most TB diagnostics are performed on sputum samples, but people who need TB tests are often unable to produce sputum. Upper respiratory tract sampling for TB diagnosis was widely used historically and holds promise to expand non-sputum-based diagnosis.Added value of this studyWe systematically reviewed and synthesised through meta-analysis diagnostic accuracy evaluations of upper respiratory tract sampling for TB. Historically, upper respiratory tract sampling for TB diagnosis was commonly used, with 39/71 studies conducted before 1970, although in recent years there has been a resurgence of interest in oral sampling. We show that upper respiratory tract samples have acceptable sensitivity and specificity compared to sputum culture, and, if testing is optimised using newer molecular and culture-based methods, may be capable of meeting WHO target produce profiles.Implications of all the available evidenceUpper respiratory tract sampling methodologies for TB (oral sampling, and sampling from the larynx and nasopharynx) may hold promise to expand access to TB diagnosis, including for people who can’t produce sputum. These sampling strategies can be optimised using modern microbiological techniques to increase access to diagnostics for TB.