We conducted a prospective, randomized, double-blind, placebo-controlled study to determine whether or not ketoconazole could prevent yeast colonization or invasion in critically ill adult surgical patients. Fifty-seven patients in a surgical intensive care unit (SICU) with three or more clinical risk factors for Candida infection were randomized to receive ketoconazole, 200 mg via the gastrointestinal tract daily (27 patients), or placebo (30 patients). Patients with hepatic dysfunction were excluded. The study was continued for 21 days or until one week after discharge from the SICU, whichever was longer. Stool cultures were obtained every three days and other cultures as indicated clinically. Patients were observed for yeast colonization (sputum, urine, stool, or wound) and invasion (fungemia or deep tissue focus). The incidence of Candida colonization was significantly lower in the ketoconazole group than the placebo group. Invasive yeast sepsis developed in five (17%) of the placebo-treated patients and in no patient in the ketoconazole group, a significant difference. Length of stay in the SICU was significantly lower in the ketoconazole group, as were the basic SICU patient charges. Sixty percent of the patients with invasive fungal sepsis died.