Treatment outcomes and drug-related adverse events across four short-course RR-TB regimens in South Africa: A retrospective analysis

BACKGROUND: Shorter all-oral regimens for the treatment of multidrug-resistant tuberculosis or rifampicin-resistant tuberculosis (MDR/RR-TB) have improved efficacy, but their performance under programmatic conditions remains insufficiently characterized. METHODS: We retrospectively evaluated treatment outcomes and drug-related adverse events (AEs) among 396 randomly selected patients from four facilities in three provinces across South Africa. Patients received four different short-course regimens sequentially introduced into national programme: i) injectable 9-11-month (injectable-regimen), ii) bedaquiline-ethionamide-containing (ethionamide-regimen), iii) bedaquiline-linezolid-containing (linezolid-regimen), and iv) bedaquiline-pretomanid-linezolid-levofloxacin regimen (BPaL-L). RESULTS: Compared with the injectable regimen, the ethionamide-regimen (adjusted odds ratio [aOR] 1.98, 95% CI: 1.07-3.65) and BPaL-L (aOR 2.31; 95% CI: 1.23-4.33) had higher odds of treatment success; the linezolid-regimen did not (aOR 1.68; 95% CI: 0.92-3.06). The linezolid and BPaL-L regimens had similar odds of grade 3 or higher AEs, and comparable AE-related drug discontinuation. Toxicity patterns differed, from hearing loss with injectables to anaemia and neurotoxicity with BPaL-L. CONCLUSION: Under programmatic conditions, the BPaL-L regimen achieved the highest treatment success while maintaining manageable safety profiles. These findings provide valuable evidence to guide patient management and support implementation of novel MDR/RR-TB regimens in high-burden settings.