Drug-resistance profiles, population structure, genomic clustering, and temporal trends in drug resistance among Mycobacterium tuberculosis complex isolates in Ukraine, 2019–2023: a multicentre cohort study

Background Ukraine has a high burden of rifampicin-resistant (RR) tuberculosis (TB). We characterised whole-genome sequencing (WGS)-inferred drug-resistance (DR) profiles, population structure, genomic clustering, and temporal trends in DR among Mycobacterium tuberculosis complex (Mtbc) isolates during 2019–2023, amid the COVID-19 pandemic, large-scale displacement, health-system disruptions, and changes in DR-TB-policy. Methods In this multicentre cohort study, we analysed WGS data from pretreatment Mtbc isolates collected across 18 of 24 Ukrainian regions. WGS enabled phylogenetic classification, resistance prediction, clustering. Clinical data were prospectively collected. Findings WGS was completed for 4162 Mtbc-isolates, of which 3112/4162 (74.8%) were at least RR. Among 3040 multidrug-resistant (MDR) isolates, 1101/3040 (36.2%) had WGS-inferred fluoroquinolone (FQ) resistance and 78/3040 (2.6%) were classified as extensively DR-TB (XDR-TB); 57 had bedaquiline resistance, 32 had linezolid resistance, and 11 had resistance to both drugs. Lineage 2 (L2) accounted for 1608/1939 (82.9%) of MDR, 891/1023 (87.1%) of pre-XDR, and 70/78 (89.7%) of XDR isolates. Overall, 1712/3112 (55.0%) of DR-isolates formed genomic clusters; the three largest comprised 671/3112 (21.6%) of DR-TB cases and consisted exclusively of L2-isolates. WGS-inferred FQ resistance declined from 276/675 (40.9%) to 216/767 (28.2%) during 2019–2023 (absolute difference −12.7 percentage points [95% CI −17.5 to −7.8]; FDR-adjusted q=4.1 × 10 −7 ). Interpretation DR-TB in this cohort was characterised by the predominance of genomically clustered L2 strains and frequent FQ resistance. Although WGS-inferred bedaquiline resistance and XDR-TB were uncommon, combined FQ, bedaquiline, and linezolid resistance warrants continued genomic surveillance. Temporal trends may reflect concurrent epidemiological, diagnostic, and health-system changes. Funding NIAID/USCRDF; BMBF; Deutsche Forschungsgemeinschaft; EvoLUNG.