Effect of bedaquiline on mortality in South African patients with drug-resistant tuberculosis: a retrospective cohort study

BACKGROUND:Addition of bedaquiline to treatment for multidrug-resistant tuberculosis was associated with an increased risk of death in a phase 2b clinical trial, resulting in caution from WHO. Following a compassionate access programme and local regulatory approval, the South African National Tuberculosis Programme began widespread use of bedaquiline in March, 2015, especially among patients with extensively drug resistant tuberculosis for whom no other effective treatment options were available. We aimed to compare mortality in patients on standard regimens with that of patients on regimens including bedaquiline. METHODS:In this retrospective cohort study, we analysed patient data from the South African rifampicin-resistant tuberculosis case register (EDRweb), and identified additional mortality using the national vital statistics register. We excluded patients who started treatment before July 1, 2014, or after March 31, 2016; patients younger than 15 years or older than 75 years; patients without documented rifampicin resistance, and patients with pre-extensively drug-resistant tuberculosis (multidrug-resistant tuberculosis with further resistance to a second-line injectable or fluoroquinolone). We compared all-cause mortality between patients who received bedaquiline in treatment regimens and those who did not. Patients who did not receive bedaquiline had kanamycin or capreomycin and moxifloxacin as core medicines in their regimen. We estimated hazard ratios for mortality separately for multidrug-resistant or rifampicin-resistant tuberculosis and extensively drug-resistant tuberculosis and adjusted using propensity score quintile strata for the potential confounders of sex, age, HIV and antiretroviral therapy status, history of prior tuberculosis, valid identification number, and year and province of treatment. FINDINGS:24 014 tuberculosis cases were registered in the EDRweb between July 1, 2014, and March 31, 2016. Of these, 19 617 patients initiated treatment and met our analysis eligibility criteria. A bedaquiline-containing regimen was given to 743 (4·0%) of 18 542 patients with multidrug-resistant or rifampicin-resistant tuberculosis and 273 (25·4%) of 1075 patients with extensively drug-resistant tuberculosis. Among 1016 patients who received bedaquiline, 128 deaths (12·6%) were reported, and there were 4612 deaths (24·8%) among 18 601 patients on the standard regimens. Bedaquiline was associated with a reduction in the risk of all-cause mortality for patients with multidrug-resistant or rifampicin-resistant tuberculosis (hazard ratio [HR] 0·35, 95% CI 0·28-0·46) and extensively drug-resistant tuberculosis (0·26, 0·18-0·38) compared with standard regimens. INTERPRETATION:Our retrospective cohort analysis of routinely reported data in the context of high HIV and extensively drug-resistant tuberculosis prevalence showed that bedaquiline-based treatment regimens were associated with a large reduction in mortality in patients with drug-resistant tuberculosis, compared with the standard regimen. FUNDING:None.