Résumé
Introduction. This study underscores the critical role of identifying heteroresistant infections of Mycobacterium tuberculosis (Mtb) in enhancing the diagnostics of tuberculosis (TB). These conditions complicate diagnostics and treatment, underlining the need for advanced techniques to detect and characterize resistant populations effectively. Hypothesis/Gap statement. Current diagnostics may fail to identify heteroresistance and mixed infections, limiting the understanding of their impact on treatment outcomes. Aim. This pilot study aimed to phenotypically and genotypically characterize rifampicin-heteroresistant clinical isolates and assess their genetic diversity and resistance patterns. Methodology. A retrospective analysis of 2,917 Mtb genomes from Peru (1999–2020) was conducted using MTBseq and TB-Profiler. Techniques included indirect microscopic observation drug susceptibility, MIC determination via tetrazolium microplate assay, agar proportion method and sequencing. From each clinical isolate, three colonies were isolated from both rifampicin-supplemented (1 µg mL −1 ) and drug-free media for subsequent phenotypic and genotypic characterization, including rpoB sequencing. Results. Of the 2,917 genomes analysed, 14.6% were classified as mixed infections, 3.8% exhibited heteroresistance to at least 1 drug between 21 antibiotics analysed and 0.79% were rifampicin-heteroresistant. Colonies from rifampicin-supplemented media displayed high resistance (MIC >1 µg mL −1 ) with mutations such as S450L in the RpoB protein. In contrast, those from drug-free media exhibited sensitivity to rifampicin (MIC <1 µg ml −1 ), harbouring other RpoB mutations including D435Y, L452P and L430P. Notably, some colonies retained WT RpoB sequences, suggesting a diversity of subpopulations within isolates. Conclusion. Whole-genome sequencing and phenotypic analysis confirmed the coexistence of rifampicin-susceptible and rifampicin-resistant Mtb populations within single clinical isolates. Subculturing in drug-free media favoured the selection of sensitive strains, emphasizing the critical need for advanced diagnostic tools to accurately detect and characterize heteroresistant and mixed infections. These findings pave the way for more targeted treatment strategies to combat antimicrobial resistance in TB.