Acquired resistance during short-course treatment for rifampicin-resistant tuberculosis

Objectives Shorter regimens represent a significant advancement for rifampicin-resistant tuberculosis (RR-TB) treatment. However, data on acquired drug resistance (ADR) remain limited. Methods This study was nested within TB-TRUST serial trials for shorter treatment for RR-TB in China. Participants without resistance to fluoroquinolone and second-line injectable drugs received either a bedaquiline-free oral regimen or the WHO-recommended injectable-containing regimen. Participants with fluoroquinolone resistance were treated with a bedaquiline-based oral regimen. All participants with two or more isolates successfully sequenced by whole-genome sequencing were included in this study. ADR was determined using whole-genome sequencing data by identifying mutations in a predefined panel of resistance-associated genes. Results Among 114 participants included, 16 (14.0%; 95% CI, 8.8-21.6%) experienced at least one ADR event (17 events in total), with a median onset of 17 (range, 14-605) days from treatment initiation. ADR was most common for pyrazinamide (6/70, 8.6%; 95% CI, 4.0-17.5%), followed by bedaquiline (5/111, 4.5%; 95% CI, 1.9-10.1%), ethambutol (2/48, 4.2%; 95% CI, 1.2-14.0%), fluoroquinolones (4/100, 4.0%; 95% CI, 1.6-9.8%), and clofazimine (4/111, 3.6%; 95% CI, 1.4-8.9%). No ADR was detected for linezolid or cycloserine. ADR was more frequent in participants with poor treatment adherence (31.1% (5/16) vs. 11.2% (11/98), p 0.048). Among 13 participants with bacteriological failure, ADR was identified in two cases. Conclusions Shorter treatment for RR-TB carries a non-negligible risk of ADR. Poor adherence might increase the likelihood of ADR, and early emergence of ADR may indicate suboptimal regimen potency. Continued surveillance is warranted, and further studies are needed to evaluate the clinical association between ADR and treatment outcomes.