In 2016, there were 600 000 cases of rifampicin-resistant tuberculosis globally, and an estimated 190 000 people died from multidrug-resistant tuberculosis.1 Rifampicin-resistant tuberculosis can be rifampicin mono-resistant, multidrug-resistant (that is, resistant to both rifampicin and isoniazid), extensively drug-resistant (that is, multidrug-resistant plus at least resistance to fluoroquinolones and second-line injectable drugs), or pre-extensively drug-resistant (that is, multidrug-resistant plus resistance to either a fluoroquinolone or a second-line injectable drug).2 Rifampicin-resistant tuberculosis requires 9–24 months of treatment using second-line antituberculosis drugs and is associated with high mortality.3, 4 Across all countries reporting to WHO, among patients initiating tuberculosis treatment only 54% of patients with multidrug-resistant or rifampicin-resistant tuberculosis and 30% of patients with extensively drug-resistant tuberculosis (2014 cohort) were successfully treated.1
Bedaquiline is a diarylquinoline that inhibits mycobacterial ATP synthase.5 As of June, 2017, at least 89 countries reported using bedaquiline for treatment of rifampicin-resistant tuberculosis.1 Stage 1 phase 2b clinical trial results showed an increase, from 9 to 48%, in the proportion of patients with multidrug-resistant tuberculosis who converted to a negative sputum culture at 8 weeks when bedaquiline was added to a standard multidrug-resistant tuberculosis regimen, with no significant increase in the frequency or severity of adverse drug reactions.6 However, the final outcomes at 120 weeks for stage 2 of the same phase 2b trial showed a statistically significant imbalance in mortality in the two treatment arms, with ten deaths occurring in the 79 patients exposed to bedaquiline (12·7%) and two deaths (2·5%) in the 79 patients in the placebo arm (p=0·02).7 None of the deaths in the bedaquiline arm were attributed to bedaquiline by the investigators. Nonetheless, interim WHO guidelines recommended use of bedaquiline in rifampicin-resistant tuberculosis only when there is second-line drug resistance, the patient is not eligible for the standard treatment for rifampicin-resistant tuberculosis, or when there are no other treatment options.8, 9 Similarly, because of the reported increased risk of mortality,7 regulatory approval in the USA included a black box warning10 to only use bedaquiline when an effective treatment regimen cannot otherwise be provided.11
Research in context
Evidence before this study
In 2014, only 54% of patients initiating treatment for multidrug-resistant or rifampicin-resistant tuberculosis had successful outcomes. In South Africa, about 20% of such patients die during the standard long-course 18–24 months of second-line treatment. Patients with multidrug-resistant or rifampicin-resistant tuberculosis who are co-infected with HIV and those with resistance to second-line injectable drugs and fluoroquinolones, including those with extensively drug-resistant tuberculosis, are 2–3 times more likely to die compared with people who are HIV-negative and those without second-line injectible or fluoroquinolone resistance. Results from phase 2b clinical trials of bedaquiline (TMC207-C208) showed significant benefit in terms of the proportion of patients who culture converted, time to culture conversion, and proportion of patients achieving cure when bedaquiline was added to the standard multidrug-resistant tuberculosis treatment. However, more deaths occurred in the bedaquiline plus background regimen arm compared with the standard multidrug-resistant tuberculosis treatment arm (10 of 79 patients vs 2 of 81 patients). The deaths in the bedaquiline arm were not attributed to bedaquiline by the investigators, but the significant increased risk of mortality led to a black box warning attached to the 2012 US Food and Drug Administration bedaquiline approval. WHO bedaquiline guidelines (2013 and 2017 revision) were also cautious, recommending bedaquiline only when an effective treatment regimen could not be constructed with other WHO-recommended drugs.
Added value of this study
Since March 2015, bedaquiline has been used within the South African National Tuberculosis Programme for all patients for whom an effective regimen could not be constructed (ie, those with second-line drug resistance or toxicity to the standard regimen). In our study, we analysed the South African drug-resistant tuberculosis case register from July 1, 2014, and March 31, 2016, and 18 601 patients who initiated drug-resistant tuberculosis regimens without bedaquiline were compared to 1016 patients who initiated bedaquiline-containing regimens. We used propensity score strata to adjust for potential confounders. In this cohort, bedaquiline was associated with a 3 times reduction in the adjusted hazard ratio for mortality. Results from our large cohort treated under the state tuberculosis programme and in a population with high prevalence of HIV and second-line drug resistance provided evidence that bedaquiline is associated with reduced rather than increased mortality in drug-resistant tuberculosis.
Implications of all the available evidence
Initial recommendations for bedaquiline weighed the benefits of the increased rate of culture conversion and cure against the unexplained higher risk of mortality observed in randomised controlled trials. However, given the evidence from our study that patients receiving bedaquiline are not at increased mortality risk, this risk evaluation could change. As bedaquiline was associated with increased culture conversion and cure in clinical trials and with decreased mortality in our large observational cohort, clinicians and policy makers should re-evaluate the practice and guidance of only using bedaquiline in drug-resistant tuberculosis treatment when there are no other options.
WHO categorised South Africa as a country with a high burden of tuberculosis, HIV tuberculosis, and multidrug-resistant tuberculosis in its 2017 global report.1 19 073 laboratory-confirmed multidrug-resistant or rifampicin-resistant tuberculosis cases were diagnosed and 967 extensively drug-resistant cases were diagnosed in 2016. Similar to global reporting, 54% (n=11 111) of the 2014 multidrug-resistant or rifampicin-resistant tuberculosis cohort in South Africa were successfully treated. Short-term mortality is highest in patients with extensively drug-resistant tuberculosis, compared with other forms of tuberculosis—42% of the 2014 South African cohort died during extensively drug-resistant tuberculosis treatment.1 Long-term survival for extensively drug-resistant tuberculosis is even poorer—in one South African cohort study, 73% had died within 5 years of initiating treatment.12
In January, 2013, the South African National Tuberculosis Programme, with the support of non-governmental and academic partners, established the Bedaquiline Clinical Access Programme so that patients with pre-extensively drug-resistant tuberculosis and extensively drug-resistant tuberculosis at specialised public health facilities in South Africa could receive bedaquiline.13 Early results of the South African Bedaquiline Clinical Access Programme cohort showed high amounts of culture conversion and encouraging outcomes, including survival.14 In October, 2014, the South African Medicines Control Council regulatory authority approved the use of bedaquiline for treatment of multidrug-resistant and rifampicin-resistant tuberculosis. Starting in March, 2015, the South African National Tuberculosis Programme began the process of rolling out bedaquiline as an additional drug to strengthen the existing regimens for rifampicin-resistant tuberculosis.
In our study, we analysed survival from routinely reported data from the South African National Tuberculosis Programme to establish the hazard ratio (HR) for mortality for patients who received bedaquiline compared with patients who received regimens that did not include bedaquiline in public sector health-care facilities in South Africa.