Multicenter evaluation of the MB/BACT system for susceptibility testing of Mycobacterium tuberculosis.

The reliability of the MB/BACT system for susceptibility testing of Mycobacterium tuberculosis to pyrazinamide, rifampin, isoniazid, streptomycin, and ethambutol was compared to the BACTEC 460TB system. The proportion method was used to resolve discrepant results by an independent arbiter. Two interpretative methods were used, with an undiluted control (direct control) and a diluted control (10(-1) control). As no significant difference was observed between the two controls, the method with the direct control was adopted as the most accurate one. One hundred sixty-six strains were tested, with an overall agreement of 98.3%. After resolution of the 18 discrepant results by the proportion method, the sensitivity and specificity of the MB/BACT system were 100% for rifampin, isoniazid, and pyrazinamide. For ethambutol, sensitivity was 92.3% at the critical concentration and 33% at the high concentration, and specificity was 100% at both concentrations. For streptomycin, sensitivity was 100% at the critical concentration and 80% at the high concentration, and specificity was 98.6% at the critical concentration and 100% at the high concentration. The rifampin, isoniazid, streptomycin, and ethambutol susceptibility test results were obtained in 6.6 days with the MB/BACT versus 5 days with the BACTEC 460TB. The pyrazinamide susceptibility test results were obtained in 7.8 days with the MB/BACT, versus 6.7 days with the BACTEC 460TB. These data demonstrate that the fully automated MB/BACT system is a very reliable method for rapid susceptibility testing of M. tuberculosis against rifampin, isoniazid, and pyrazinamide. Sensitivity results have to be improved for ethambutol and streptomycin, especially at the high concentration.