Oropharyngeal candidiasis (OPC) is a common opportunistic infection in human immunodeficiency virus (HIV)-infected patients and other immunocompromised hosts. Clotrimazole troches are widely used in the treatment of mucosal candidiasis. However, little is known about the potential contribution of clotrimazole resistance to the development of refractory mucosal candidiasis. We therefore investigated the potential emergence of resistance to clotrimazole in a prospectively monitored HIV-infected pediatric population receiving this azole. Adapting the National Committee for Clinical Laboratory Standards M27-A reference method for broth antifungal susceptibility testing of yeasts to clotrimazole, we compared MICs in macrodilution and microdilution assays. We further analyzed the correlation between these in vitro findings and the clinical response to antifungal therapy. One isolate from each of 87 HIV-infected children was studied by the macrodilution and microdilution methods. Two inoculum sizes were tested by the macrodilution method (10(3) and 10(4) CFU/ml) in order to assess the effect of inoculum size on clotrimazole MICs. The same isolates also were tested using a noncolorimetric microdilution method. Clotrimazole concentrations ranged from 0.03 to 16 microg/ml. Readings were performed after incubation for 24 and 48 h at 35 degrees C. For 62 (71.2%) of 87 clinical isolates, the MICs were low (< or =0.06 microg/ml). The MIC for 90% of the strains tested was 0.5 microg/ml, and the highest MIC was 8 microg/ml. There was no significant difference between MICs at the two inoculum sizes. There was 89% agreement (+/-1 tube) between the microdilution method at 24 h and the macrodilution method at 48 h. If the MIC of clotrimazole for an isolate of C. albicans was > or =0.5 microg/ml, there was a significant risk (P < 0.001) of cross-resistance to other azoles: fluconazole, > or = 8 microg/ml (relative risk [RR] = 8.9); itraconazole, > or =1 microg/ml (RR = 10). Resistance to clotrimazole was highly associated with clinically overt failure of antifungal azole therapy. Six (40%) of 15 patients for whom the clotrimazole MIC was > or =0.5 microg/ml required amphotericin B for refractory mucosal candidiasis versus 4 (5.5%) of 72 for whom the MIC was <0.5 microg/ml (P = 0.001; 95% confidence interval = 2.3 to 22; RR = 7.2). These findings suggest that an interpretive breakpoint of 0.5 microg/ml may be useful in defining clotrimazole resistance in C. albicans. The clinical laboratory's ability to determine MICs of clotrimazole may help to distinguish microbiologic resistance from the other causes of refractory OPC, possibly reducing the usage of systemic antifungal agents. We conclude that resistance to clotrimazole develops in isolates of C. albicans from HIV-infected children, that cross-resistance to other azoles may develop concomitantly, and that this resistance correlates with refractory mucosal candidiasis.