Discordance Between Phenotypic and WGS-Based Drug Susceptibility Testing Results for Some Anti-Tuberculosis Drugs: A Snapshot Study of Paired Mycobacterium tuberculosis Isolates with Small Genetic Distance

Background: Current tuberculosis treatment regimens primarily rely on phenotypic drug susceptibility testing and rapid molecular assays. Although whole-genome sequencing (WGS) offers a promising alternative, disagreements between phenotypic and molecular testing methods remain. In this retrospective study, we compared the phenotypic and WGS-predicted drug resistance profiles of paired Mycobacterium tuberculosis isolates with small genetic distances (<= 10 single nucleotide variants) obtained from patients with longitudinal single-episode or recurrent tuberculosis. Additionally, we investigated the distribution of drug-resistance-conferring variants among the identified M. tuberculosis genotypes. Methods: Paired M. tuberculosis isolates from 46 patients with pulmonary tuberculosis (2002-2019) were analyzed. Spoligotyping was performed for all the isolates. WGS data were processed using TB-Profiler software to genotype the strains and detect variants in M. tuberculosis genes associated with drug resistance. The significance of these variants was evaluated using the M. tuberculosis variant catalog developed by the World Health Organization. Phenotypic drug susceptibility test results were obtained from patients' medical records. Results: Among the 46 isolate pairs, 25 (54.3%) harbored drug-resistance-associated variants, with 20 demonstrating identical WGSpredicted drug resistance profiles. Drug-resistant isolate pairs belonged to Lineages 2 and 4, with the most common sub-lineages being 2.2.1 (SIT1 and SIT190 spoligotypes), and 4.3.3 (SIT42). Agreement between phenotypic and WGS-based drug susceptibility testing was highest (>90%) for rifampicin, isoniazid, ethambutol, fluoroquinolones, streptomycin, and amikacin when calculated for M. tuberculosis isolates or isolate pairs. In most discordant cases, isolate pairs harbored variants that could cause low- or moderate- level resistance or were previously associated with variable minimum inhibitory concentrations. Notably, such discrepancies mostly occurred in one isolate from the pair. In addition, differences in resistance-related variant distributions among M. tuberculosis genotypes were observed for most of the analyzed drugs. Conclusion: The simultaneous performance of phenotypic and WGS-based drug susceptibility testing creates the most accurate drug resistance profile for M. tuberculosis isolates and eliminates important limitations of each method.