Abstract
The impact of HIV viral suppression on multidrug resistant tuberculosis (MDR-TB) treatment outcomes among people with HIV (PWH) has not been clearly established. Using secondary data from a cluster-randomized clinical trial among people with MDR-TB in South Africa, we examined the effects of HIV viral suppression at MDR-TB treatment initiation and throughout treatment on MDR-TB outcomes among PWH using multinomial regression. This analysis included 1479 PWH. Viral suppression (457, 30.9%), detectable viral load (524, 35.4%), or unknown viral load (498, 33.7%) at MDR-TB treatment initiation were almost evenly distributed. Having a detectable HIV viral load at MDR-TB treatment initiation significantly increased risk of death compared to those virally suppressed (relative risk ratio [RRR] 2.12, 95% CI 1.11-4.07). Among 673 (45.5%) PWH with a known viral load at MDR-TB outcome, 194 (28.8%) maintained suppression, 267 (39.7%) became suppressed, 94 (14.0%) became detectable, and 118 (17.5%) were never suppressed. Those who became detectable (RRR 11.50, 95% CI 1.98-66.65) or were never suppressed (RRR 9.28, 95% CI 1.53-56.61) were at significantly increased risk of death (RRR 6.37, 95% CI 1.58-25.70), treatment failure (RRR 4.54, 95% CI 1.35-15.24), and loss to follow-up (RRR 7.00, 95% CI 2.83-17.31; RRR 2.97, 95% CI 1.02-8.61) compared to those who maintained viral suppression. Lack of viral suppression at MDR-TB treatment initiation and failure to achieve or maintain viral suppression during MDR-TB treatment drives differences in MDR-TB outcomes. Early intervention to support access and adherence to antiretroviral therapy among PWH should be prioritized to improve MDR-TB treatment outcomes.