We compared the E-test to the broth microdilution method for testing the susceptibility of 115 clinical isolates of Acinetobacter baumannii to colistin. Twenty-two (19.1%) strains were resistant to colistin and 93 (80.8%) strains were susceptible according to the reference broth microdilution method. A categorical agreement of 98.2% was found, with only two (1.7%) very major errors. Agreement within 1 twofold dilution between the E-test and the broth microdilution was 16.5%. Complete agreement was found for the strains for which MICs fell within the range of 0.25 to 1 microg of colistin/ml. However, there was poor concordance, particularly in extreme dilutions with higher MICs by the E-test method.