Eighteen rapidly growing mycobacteria were tested for susceptibility to amikacin by six different antibiotic susceptibility procedures to assess method variability and factor variation within a single method. Using amikacin MICs determined by the microdilution method as the reference standard, results for Mycobacterium chelonae were on average eight-fold higher by the macrodilution method and two-fold higher by the BACTEC, 1% standard proportion, and agar dilution methods. For Mycobacterium fortuitum, macrodilution MICs were on average four-fold higher than microdilution results; however, for this species, agar dilution, the 1% standard proportion method, and the BACTEC method showed good correlation with microdilution testing. The use of different test media and incubation in increased CO2 tension increased amikacin MICs for Mycobacterium chelonae. An inoculum effect was observed with both species, especially when the organism concentration increased from 10(5) cfu/mL to 10(6) cfu/mL for broth testing and 10(5) to 10(6) cfu per spot for agar dilution. These results indicate that different antibiotic susceptibility methods and test conditions markedly influence MICs of amikacin for these rapidly growing mycobacteria.