INTRODUCTION: Testing for anti-TB drugs in small hair samples may serve as a non-invasive tool to measure cumulative drug exposure and/or adherence, as these determine treatment success. We aimed to assess how well hair assays of TB drugs predict TB treatment outcomes.METHODS: A small thatch of hair, ~30 strands, was cut from the occipital region in adults and children from a prospective TB cohort in India. Isoniazid (INH), acetyl-INH and pyrazinamide (PZA) were extracted from the hair samples and quantified using liquid-chromatography-tandem mass spectrometry. The relationship between drug concentrations in hair and time to unfavourable outcomes was assessed using Cox-proportional hazards regression models.RESULTS: A two-fold increase in hair acetyl-INH concentrations in the 264 participants in our cohort with hair assays for TB drugs indicated a lower hazard of unfavourable TB treatment outcomes (aHR 0.67, 95%CI 0.44-1.02) and TB treatment failure (aHR 0.65, 95%CI 0.42-1.01). Higher summed concentrations (a summed measure of INH and acetyl-INH) indicated a lower hazard of treatment failure (aHR 0.69, 95%CI 0.45-1.05)CONCLUSION: Hair levels of INH and its metabolite may predict TB treatment outcomes, indicating the potential utility of this measure to assess and optimise TB treatment outcomes.