Body mass index (BMI) at RR-TB diagnosis as an independent predictor of treatment outcomes: a retrospective analysis

Abstract Background Rifampicin-resistant tuberculosis (RR-TB) remains a major public health concern, particularly in African populations where undernutrition is common. Many participants with RR-TB initiate treatment with a low body mass index (BMI), which has been linked to higher mortality and treatment failure. Methods This retrospective study analysed data from 895 participants diagnosed with rifampicin-resistant tuberculosis (RR-TB) enrolled across five clinical trials conducted between 2009 and 2024 in KwaZulu Natal, South Africa. Participants were classified by BMI (normal weight, overweight, and underweight). Cox proportional hazards regression models assessed the relationship between baseline BMI and treatment outcomes, adjusting for potential confounders. Results At baseline, 51% (542/894) of the participants had a normal weight, 16% were overweight, and 34% were underweight. The majority were people living with HIV (PLWHA) (85%) (761/895); 89% (675/761) were on ART, and 51% (240/475) had a suppressed viral load. At day 400, underweight participants had a higher mortality rate per 100 person-years (14.86, 95% CI: 8.13–24.94) compared to those with normal weight (6.27, 95% CI: 3.00-11.52) and overweight individuals (2.06, 95% CI: 0.05–11.50). Underweight participants had an increased risk of unfavourable outcomes (HR = 1.85, 95% CI = 1.35–2.53, p < 0.001), while overweight participants had a reduced risk (HR = 0.39, 95% CI = 0.31–0.97, p = 0.040). Unsuppressed viral load (HR = 1.77, 95% CI = 1.14–2.73, p = 0.011) and absence of ART (HR = 3.00, 95% CI = 1.62–5.55, p < 0.001) were significant risk factors. Conclusion Low BMI is a strong predictor of poor treatment outcomes for RR-TB. Nutritional assessment and interventions should be integrated into RR-TB management procedures to improve patient outcomes. Trial registration Not applicable. Graphical Abstract