Abstract

Background Bedaquiline-based oral short-course regimens (SCR) for rifampicin-resistant tuberculosis (RR-TB) are highly effective in clinical trials but outcomes in programmatic settings may be more modest. We evaluated clinical and bacteriological outcomes with a seven-drug, linezolid-containing SCR in a high-burden programmatic setting. Methods This prospective cohort study enrolled adults with newly diagnosed RR-TB who were started on the oral SCR in the Eastern Cape Province, South Africa. The primary outcome was World Health Organization-defined end-of-treatment success. Secondary outcomes were TB-free survival (composite of alive, absence of a positive Mycobacterium tuberculosis culture, and treatment completed or in care) at 18 months and time to sputum culture conversion (SCC). Results In total, 248 participants were included, 173 (69.8%) of whom were human immunodeficiency virus (HIV) positive. Culture conversion by 90 days was 96.8% (median time to SCC: 29 days, 95% confidence interval [CI]: 27–31). Treatment success was 37.5% (93/248). Reasons for unsuccessful treatment included switching to individualised regimens (35.1%, 87/248), loss to follow-up (19.4%, 48/348), and death (8.1%, 20/248). At 18 months, 157 (63.3%) participants achieved TB-free survival, with a cumulative mortality of 21.6% (95% CI: 16.1–29.0). Baseline 3+ smear (adjusted odds ratio [aOR]: 3.38, 95% CI: 1.28–8.95), higher age (aOR: 1.05, 1.01–1.08), and lower albumin (aOR: 0.94, 0.88–0.99), but not HIV status, were associated with unfavourable outcome at 18 months. Conclusions The oral SCR performed poorly in a high-burden TB programme. Strategies to support the implementation of effective new regimens for RR-TB are needed to translate outcomes from clinical trials into practice.