Differential drug susceptibility patterns of Mycobacterium chimaera and other members of the Mycobacterium avium-intracellulare complex.

OBJECTIVES: To determine MIC distributions for Mycobacterium chimaera, Mycobacterium intracellulare, Mycobacterium colombiense and Mycobacterium avium, and to derive tentative epidemiological cut-off (ECOFF) values.

METHODS: A total of 683 bacterial isolates (M. chimaera, n = 203; M. intracellulare, n = 77; M. colombiense, n = 68; M. avium, n = 335) from 627 patients were tested by broth microdilution according to CLSI protocol M24-A2 on Sensititre RAPMYCOI plates. MICs were interpreted based on CLSI breakpoints for clarithromycin, and tentative breakpoints for amikacin, moxifloxacin and linezolid. Tentative ECOFFs were determined by visual approximation and the ECOFFinder algorithm.

RESULTS: Modal MIC, MIC50 and MIC90 values were within ± one dilution step from the respective aggregated data set for 47/48 (97.9%), 48/48 (100%) and 48/48 (100%) species-drug combinations. Clarithromycin wild-type populations were mostly classified as susceptible (MIC90 4-8 mg/L; S ≤8 mg/L). Rifabutin MICs were lower than those of rifampicin. Tentative moxifloxacin, linezolid and amikacin breakpoints split wild-type populations. No ECOFFs could be set for rifampicin, ethambutol, ciprofloxacin, isoniazid, trimethoprim/sulfamethoxazole and doxycycline because of truncation of MIC distributions. Agreement between the visually determined and the modelled 97.5% ECOFFs was 90.9%. All 99.0% ECOFFs were one titre step higher than by visual approximation.

CONCLUSIONS: Drug susceptibility patterns of M. chimaera are comparable to those of closely related species. Except for clarithromycin, breakpoints for Mycobacterium avium-intracellulare complex should be re-evaluated. Statistical determination of the 99.0% ECOFF may be superior to visual approximation.