We conducted a cross-sectional, hospital-based study between January 2006-March 2008 to estimate the resistance of Mycobacterium tuberculosis
to first-line drugs in patients with tuberculosis at a Brazilian hospital. We evaluated the performance of the [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide] (MTT) microplate assay compared with the Bactec-MGIT 960™ system for mycobacteria testing. The prevalence of resistance in M. tuberculosis
was 6.7%. Multidrug-resistance [resistance to rifampicin (RMP) and isoniazid (INH)], INH-resistance and streptomycin (SM)-resistance accounted for 1%, 3.8% and 3.8% of all resistance, respectively, and all isolates were susceptible to ethambutol (EM). The resistance was primary in four cases and acquired in three cases and previous treatment was associated with resistance (p = 0.0129). Among the 119 M. tuberculosis
isolates, complete concordance of the results for INH and EM was observed between the MTT microplate and Bactec-MGIT 960™ methods. The observed agreement for RMP was 99% (sensitivity: 90%) and 95.8% for SM (sensitivity 90.9%), lower than those for other drugs. The MTT colourimetric method is an accurate, simple and low-cost alternative in settings with limited resources.