The MIDITECH colorimetric susceptibility test with automated reading is a modification of the standard broth microdilution method that uses a 3-(4,5-dimetylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) dye for detecting viable bacteria. The method can be applied to non-fastidious aerobic Gram-negative bacteria, staphylococci and Enterococcus faecalis. To assess the reliability of this method, we compared susceptibility data obtained by this test with standard NCCLS microdilution assay results. For this purpose, 15 antibiotics and a well characterized set of 527 Gram-negative and Gram-positive bacterial isolates collected and stored at the Division of Infectious Diseases and Chemotherapy (Vienna General Hospital, Austria), yielding 5751 organism-antibiotic combinations, were analysed in duplicate. The overall essential agreement (+/-1 log(2) dilution) between the MIDITECH and NCCLS methods was 96.18 +/- 0.67%. The colorimetric assay compared with the reference method produced MICs < or = 2 log(2) dilutions and > or = 2 log(2) dilutions in 2.34% and 1.48% comparisons, respectively. For 326 Gram-negative bacteria, the absolute interpretative agreement of both methods ranged from 87.12% for ampicillin-sulbactam to 99.85% for meropenem (mean 94.86%); 417 (4.92%) minor, three (0.05%) major and 15 (0.63%) very major errors were found. For 127 staphylococci and 74 E. faecalis isolates, the absolute interpretative agreement ranged from 90.80% for ciprofloxacin to 100% for vancomycin and linezolid (mean 96.96%); 81 (2.77%) minor, three (0.15%) major and eight (0.83%) very major errors were found. For most of the clinically important aerobically growing pathogens, the MIDITECH colorimetric test provided reliable quantitative susceptibility data. The main advantage of this method is simple performance, automated reading and data processing without expensive investments.