Acquired Resistance in Different Treatment Cohorts for Rifampicin-resistant Tuberculosis: Results from the TB-TRUST and TB-TRUST-plus Trials

Abstract Introduction All-oral shorter treatment regimens represent a landmark advance for patients with rifampicin resistant tuberculosis (RR-TB). However, the promise of shorter treatment is challenged by our limited knowledge of the risk of acquired drug resistance during the treatment. Methods A multicenter randomized controlled trial, TB-TRUST (NCT03867136), was initiated in China, to evaluate a 6-9 month all-oral regimen among patients with RR-TB after excluding resistance to fluroquinolones and injectable agents. The investigator regimen (bedaquiline-free regimen) included levofloxacin, linezolid, cycloserine, pyrazinamide, and (or) clofazimine, while the control was an injectable-containing regimen of seven drugs. Those with fluoroquinolone resistance were enrolled in the TB-TRUST-plus study (NCT04717908) and were treated with a 6-9 month bedaquiline-containing regimen that companion drugs included linezolid, cycloserine, pyrazinamide, and (or) clofazimine. Baseline and serial strains with intervals exceeding 8 weeks were included in evolutionary analysis. Whole-genome sequencing (WGS) was conducted on all isolates using NovaSeq 6000 with at least 100-fold coverage, analyzing all single nucleotide polymorphisms (SNPs) with >10% mutation frequency to detect subpopulations and microevolution. Fixed mutations were identified at mutation frequencies >80%. Mutations associated with resistance were identified for 27 resistance-related genes across 11 anti-TB drugs. Results Interpretable WGS results were obtained from 347 participants including 136 in the control group, 144 in the bedaquiline-free group and 67 in the bedaquiline-containing group. Among patients in the TB-TRUST study, bedaquiline, clofazimine, cycloserine, linezolid and pyrazinamide resistance-conferring mutations were identified in 0 (0.0%), 19 (6.8%), 38 (13.6%), 4 (1.4%), and 69 (24.6%) of baseline isolates, respectively. In the TB-TRUST plus cohort, the mutations to those drugs were detected in 0 (0.0%), 4 (6.0%), 12 (17.9%), 0 (0.0%), and 41 (61.1%), respectively. A total of 114 patients were included in the evolutionary analysis (48 in the control group, 46 in the bedaquiline-free group, and 20 in the bedaquiline-containing group). The highest acquired resistance rates were observed for ethambutol, pyrazinamide, and bedaquiline in the respective groups. In the bedaquiline-free group, pyrazinamide acquired resistance was identified in four patients, featuring mixed mutations including pncA Y103 and pncA F58S, pncA C14R and pncA A134V, as well as pncA V139A and pncA T142A. In the bedaquiline-containing group, three patients developed mutations related to bedaquiline resistance including rv0678 S68N (n=2) and atpE E61D (n=1). No fixed mutation to bedaquiline was identified. Conclusion Ensuring at least four effective drugs is beneficial to prevent the acquisition of drug resistance during shorter treatments for RR-TB.