Abstract

Background: Rising prevalence of bedaquiline resistance undermines benefits from this life-saving drug for rifampicin-resistant tuberculosis (RR tuberculosis). Despite increasing awareness, patient-level outcomes for bedaquiline-resistant tuberculosis have not been well characterised and case management has been poorly defined. Methods: We did a retrospective cohort study of bedaquiline-resistant tuberculosis with matched RR tuberculosis controls at a tuberculosis referral hospital in East London, South Africa. Cases included patients aged 13 years or older with a phenotypic bedaquiline-resistant Mycobacterium tuberculosis isolate identified between Jan 1, 2018 and June 30, 2023. Controls with confirmed bedaquiline-susceptible tuberculosis, matched 1:1 by baseline culture status, age, and HIV status, were selected from a prospective observational study conducted during an overlapping period at the same facility. Primary outcomes included time to sputum culture conversion (SCC), a modified WHO-defined unfavourable outcome, and tuberculosis-free survival (alive, with SCC, and in care or treatment completed) up until 18 months. Adjusted analyses used Cox proportional hazards and logistic regression models. Findings: 82 patients with bedaquiline-resistant tuberculosis were included, 57 (70%) of whom were HIV positive. Bedaquiline was prescribed for 72 (88%) of 82 patients and meropenem (plus amoxicillin–clavulanate) for 32 (39%) of 82. Together with bedaquiline, the most frequently prescribed drugs included clofazimine, linezolid, and terizidone. Median time to SCC after treatment initiation was 175 days (IQR 100–254) in the bedaquiline-resistant cohort and 32 days (30–42) in matched controls. In the analysis of the combined cohorts, bedaquiline resistance (adjusted hazard ratio 0·03, 95% CI 0·0023–0·29, p=0·003) was associated with longer time to SCC when adjusted for baseline microscopy grade and baseline fluoroquinolone resistance. WHO treatment outcomes in those with bedaquiline-resistant tuberculosis were unfavourable in 54 (67%) of 81 patients, driven by treatment failure in 35 (43%) of 81. At 18 months, 43 (52%) of 82 patients had reached tuberculosis-free survival, 19 (23%) of 82 had died, and 50 (79%) of 63 survivors were still on treatment. Interpretation: Current treatment options for bedaquiline-resistant tuberculosis result in prolonged therapy, delayed microbiological responses, and poor clinical outcomes. Implementation of more rapid resistance testing, including targeted next-generation sequencing, and access to novel treatment options within randomised controlled trials for bedaquiline-resistant tuberculosis, are priorities for tuberculosis programmes.