Diagnostic accuracy of an exhaled breath test for TB in hospitalized patients with cough or risk.

<sec><title>INTRODUCTION</title>Closing the TB diagnostic gap is an urgent priority, for which non-sputum-based tests are needed. We evaluated the diagnostic accuracy of Aeonose, an exhaled breath test (EBT), as a TB triage test.</sec><sec><title>METHODS</title>Patients with cough or TB risk factors admitted to a tertiary hospital in Lima, Peru, were prospectively enrolled and underwent EBT. We evaluated EBT sensitivity and specificity for diagnosing pulmonary TB using culture and Xpert as primary and secondary reference standards and conducted stratified analyses based on risk factors.</sec><sec><title>RESULTS</title>EBT sensitivity was 85% (95% CI 72.9-93.4), and specificity was 51% (95% CI 46.0-56.6) in the training cohort (n = 417). EBT sensitivity was 70% (95% CI 47.1-86.8), and specificity was 54% (95% CI 44.8-63.6) in the validation cohort (n = 139) using the culture reference standard, with higher sensitivity (78%) when using the Xpert reference standard (n = 156). Sensitivity (60%) and specificity (48%) were lower when patients with prior TB were included. In a subset of participants randomly selected for interviews, 94% (15/16) preferred EBT to sputum-based testing.</sec><sec><title>CONCLUSIONS</title>EBT had moderate sensitivity and low specificity as a TB triage test in this hospitalised cohort with cough or risk factors. Diagnostic accuracy was lower in people with prior TB.</sec>.