The prediction of clinical outcome during antifungal therapy is an issue of paramount importance in clinical research, because no consistently reliable parameters are available. Minimum inhibitory concentration (MIC) values and breakpoint interpretation may serve as surrogate criteria until standardized in vitro antifungal susceptibility testing is developed, especially for fluconazole. With reproducible susceptibility testing methods for Candida species now available, tentative fluconazole interpretive breakpoints derived from MIC values determined by the National Committee on Clinical Laboratory Standards (NCCLS) M27-T broth macrodilution method are open for public comment. Besides the in vitro susceptibility of the fungus, clinical response to antifungal therapy with fluconazole depends to a great extent on the daily dose and corresponding plasma and tissue concentrations. Promising results from a few dose-finding studies in non-neutropenic patients show that fluconazole doses of up to 1000 mg day-1 result in higher clinical response rates than lower dosages. Therapeutic success depends substantially on achieving fluconazole plasma and tissue levels that are sufficiently higher than MIC values indicated by in vitro testing. However, this simplified concept must be related to the clinical situation and it is essential to consider the immunological status and underlying disease of the patient. Misinterpretation of MIC values may result in selection of an antifungal agent for life-threatening infections that does not provide optimal efficacy or toleration.