Objectives: Heteroresistance, described both in terms of various point mutations resulting in different levels of resistance and in terms of a mixture of mutant and WT bacilli, is identified in up to one-third of fluoroquinolone (FQ)-resistant Mycobacterium tuberculosis isolates. Heteroresistance is a challenge for current phenotypic and genotypic susceptibility testing (DST) regimes. We aimed to compare the performances of different phenotypic and genotypic DST in the context of FQ heteroresistance by mimicking, in a murine model, the course of selection of FQ resistance during treatment.
Methods: The capacity of different phenotypic DST [Lowenstein-Jensen (LJ) medium containing either 2 mg/L ofloxacin or 0.5, 1 or 2 mg/L moxifloxacin] and genotypic DST (gyrA/B Sanger sequencing) to detect FQ resistance was analysed.
Results: Ninety-seven percent of mice harboured a heterogeneous population. The proportion of mice in which FQ resistance was detected varied according to the medium used (97% for 0.5 mg/L moxifloxacin, 80% for 2 mg/L ofloxacin, 47% for 1 mg/L moxifloxacin and 25% for 2 mg/L moxifloxacin). Compared with phenotypic DST, genotypic DST had a low sensitivity for detection of resistance (33%).
Conclusions: Our study shows the in vivo complexity of FQ resistance emergence and the poor sensitivity of Sanger DNA sequencing for detection of heteroresistance. Our data support the use of 0.5 mg/L moxifloxacin in LJ for detection of FQ resistance, but not the recent increase in the ofloxacin critical concentration from 2 to 4 mg/L given in the WHO recommendations.
© The Author 2016. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.