Continuous surveillance of drug-resistant TB burden in Rwanda: a retrospective cross-sectional study.

BACKGROUND: Since the roll-out of the Xpert MTB/RIF assay, continuous surveillance can provide an estimate of rifampicin-resistant TB (RR-TB) prevalence, provided high drug susceptibility testing (DST) coverage is achieved. We use national data from Rwanda to describe rifampicin DST coverage, estimate the prevalence of RR-TB and assess its predictors. METHODS: Routinely collected DST data were entered into an electronic TB case-based surveillance system. DST coverage was calculated among all bacteriologically confirmed pulmonary TB patients notified from 1 July 2019 to 30 June 2020 in Rwanda. The prevalence of RR-TB was estimated among those with DST results. Univariable and multivariable analysis was performed to explore predictors for RR TB. RESULTS: Among 4066 patients with bacteriologically confirmed pulmonary TB, rifampicin DST coverage was 95.6% (4066/4251). RR-TB was diagnosed in 73 patients. The prevalence of RR-TB was 1.4% (53/3659; 95% CI 1.09 to 1.89%) and 4.9% (20/406; 95% CI 3.03 to 7.51%) in new and previously treated TB cases, respectively. Predictors of RR-TB were: (1) living in Kigali City (adjusted OR [aOR] 1.65, 95% CI 1.03 to 2.65); (2) previous TB treatment (aOR 3.64, 95% CI 2.14 to 6.19); and (3) close contact with a known RR-TB patient (aOR 11.37, 95% CI 4.19 to 30.82). CONCLUSIONS: High rifampicin DST coverage for routine reporting allowed Rwanda to estimate the RR-TB prevalence among new and previously treated patients.

Rifampicin resistance conferring mutations among Mycobacterium tuberculosis strains in Rwanda.

Background: The World Health Organization-endorsed phenotypic and genotypic drug-susceptibility testing (gDST/pDST) assays for the detection of rifampicin-resistant (RR) tuberculosis (TB), may miss some clinically relevant rpoB mutants, including borderline mutations and mutations outside the gDST-targeted hotspot region. Sequencing of the full rpoB gene is considered the reference standard for rifampicin DST but is rarely available in RR-TB endemic settings and when done indirectly on cultured isolates may not represent the full spectrum of mutations. Hence, in most such settings, the diversity and trends of rpoB mutations remain largely unknown. Methods: This retrospective study included rpoB sequence data from a longitudinal collection of RR-TB isolates in Rwanda across 30 years (1991-2021). Results: Of 540 successfully sequenced isolates initially reported as RR-TB, 419 (77.6%) had a confirmed RR conferring mutation. The Ser450 Leu mutation was predominant throughout the study period. The Val170Phe mutation, not covered by rapid gDST assays, was observed in only four patients, three of whom were diagnosed by pDST. Along with the transition from pDST to rapid gDST, borderline RR-associated mutations, particularly Asp435Tyr, were detected more frequently. Borderline mutants were not associated with HIV status but presented lower odds of having rpoA-C compensatory mutations than other resistance-conferring mutations. Conclusion: Our analysis showed changes in the diversity of RR-TB conferring mutations throughout the study period that coincided with the switch of diagnostic tools to rapid gDST. The study highlights the importance of rapid molecular diagnostics reducing phenotypic bias in the detection of borderline rpoB mutations while vigilance for non-rifampicin resistance determinant region mutations is justified in any setting.