P27

Timely detection of fluoroquinolone resistance is essential for the effective treatment of rifampicin-resistant tuberculosis. Hetero-resistance, the co-occurrence of drug-susceptible and drug-resistant populations, can be difficult to detect by molecular methods when the drug-resistant bacilli are a minority population. We aimed to determine the accuracy of the novel GeneXpert MTB/XDR assay to detect fluoroquinolone hetero-resistance caused by the gyrA D94G variant, a common variant conferring high-level fluoroquinolone resistance. We used a gyrA and gyrB wild-type strain and its isogenic gyrA D94G daughter strain to generate quality-controlled hetero-resistant mixtures. The isolates were plated for colony counting and mixed at different ratios (0.5%-100% resistant). The ratio of resistant to susceptible populations was then determined genotypically by Deeplex-MycTB. The mixtures were assessed phenotypically by minimum inhibitory concentration testing using EUCAST broth microdilution. The mixtures are available as thermolysates in the Belgian Coordinated Collections of Microorganisms (www.bccm.belspo.be). Deeplex-MycTB demonstrated the accurate proportion of intended susceptible and resistant populations of all mixtures above 1% (Deeplex-MycTB cut-off). All mixtures were phenotypically resistant to moxifloxacin and levofloxacin, including the mixture containing only 0.5% of gyrA D94G bacilli. Our results demonstrate that the Xpert MTB/XDR assay detects resistance to fluoroquinolones when gyrA D94G variants are present at ≥20%, highlighting its value as a rapid molecular diagnostic. Future studies should assess the assay's detection limit for other common gyrA and gyrB mutants, including variants that confer low-level fluoroquinolone resistance.