The control of multi-drug-resistant tuberculosis has been hampered by the lack of simple, rapid and sensitive methods for assessing bacterial growth and antimicrobial susceptibility. Due to the increasing incidence and high frequency of mutations, it is unlikely that culture methods will disappear in the foreseeable future. Therefore, the need to modernize methods for rapid detection of viable clinical isolates, at a minimum as a gold standard, will persist. Previously, we confirmed the feasibility of using the Gel Microdrop (GMD) Growth Assay for identifying sub-populations of resistant Mycobacteria by testing different laboratory strains. Briefly, this assay format relies on encapsulating single bacterium in agarose microspheres and identifying clonogenic growth using flow cytometry and fluorescent staining. In this study, we modified the GMD Growth Assay to make it suitable for clinical applications. We demonstrated the effectiveness and safety of this novel approach for detecting drug susceptibility in clinically relevant laboratory strains as well as clinical isolates of Mycobacterium tuberculosis. Correlation between results using the GMD Growth Assay format and results using two well characterized methods (Broth Microdilution MIC and BACTEC 460TB) was 87.5% and 90%, respectively. However, due to the inherent sensitivity of flow cytometry and the ability to detect small (< 1%) sub-populations of resistant mycobacteria, the GMD Growth Assay identified more cases of drug resistance. Using 4 clinically relevant mycobacterial strains, we assessed susceptibility to primary anti-tuberculosis drugs using both the Broth Microdilution MIC method and the GMD Growth Assay. We performed 24 tests on isoniazid-resistant BCG, Mycobacterium tuberculosis H37Ra and Mycobacterium avium strains. The Broth Microdilution MIC method identified 7 cases (29.1%) of resistance to INH and EMB compared to the GMD Growth Assay which identified resistance in 10 cases (41.6%); in 3 cases (12.5%), resistance to INH and EMB was detected only with the GMD Growth Assay. In addition, using 20 Mycobacterium tuberculosis clinical isolates, we compared results using BACTEC 460TB method performed by collaborators and the GMD Growth Assay. Eight of 20 (40%) clinical isolates, which were not identified as drug-resistant using the conventional BACTEC 460TB method, were resistant to 1, 2, or 3 different concentrations of drugs using the GMD Growth Assay (13 cases of 140 experiments). In one case (isolate 1879), resistance to 10.0 mu g/ml of STR detected using BACTEC 460TB method was not confirmed by the GMD Growth Assay. Thus, the overall agreement between these methods was 90% (14 discrepant results of 140 experiments). These data demonstrate that the GMD Growth Assay is an accurate and sensitive method for rapid susceptibility testing of Mycobacterium tuberculosis for use in clinical reference laboratory settings. (c) 2005 Elsevier B.V. All rights reserved.
