Comparison of visual and spectrophotometric methods of MIC endpoint determinations by using broth microdilution...

A study to compare three different methods for reading MIC endpoints tested by the broth microdilution modification of the National Committee for Clinical Laboratory Standards (Villanova, Pa.) reference method was conducted. MICs of amphotericin B, flucytosine, fluconazole, itraconazole, and a new triazole, D0870, were determined for five reference yeast strains and 100 clinical isolates of Candida spp. MICs were read visually according to National Committee for Clinical Laboratory Standards guidelines from microdilution trays that had been (VS) and had not been (V) shaken. MICs were also determined spectrophotometrically (SP) at 492 nm. SP endpoints were determined as the concentrations resulting in a > or = 50% inhibition of growth (flucytosine and azoles) and a > or = 90% inhibition of growth (amphotericin B) relative to control growth. The five reference strains were tested nine times each against all five antifungal agents, and the MIC results for each reading method were compared with a 3-log2 dilution reference range determined by the macrodilution (M27-P) method. Overall, 84 to 100% of the MICs determined by V, 93 to 100% of those determined by VS, and 89 to 100% of those determined by SP fell within the 3-log2 dilution reference range for each reference strain and antifungal agent. Reproducibility was 99% for V and SP and 98% for VS. Agreement among the three methods of reading ranged from 97 to 99%. Excellent agreement among reading methods was also observed for all antifungal agents when tested against 100 clinical isolates. Agreement between the standard V method (no agitation) and VS ranged from 99 to 100%, and that between V and SP ranged from 89 to 99%. The VS and SP reading methods provided more definitive endpoints than the V method, which does not involve shaking.