From 16 to 30 April 1990, four of 364 (1%) postsurgical patients at one hospital developed Candida albicans fungemia or endophthalmitis. The case patients' surgeries were clustered on two days. To identify risk factors for C. albicans infections, we conducted a cohort study comparing these 4 patients with 67 control patients who had surgeries on the same days but did not acquire C. albicans infections. The participation of anesthesiologist 9 (relative risk [RR], undefined; P < 0.001) and receipt of intravenous propofol, an anesthetic agent without preservative, which was administered by an infusion pump (RR, 8.8; P = 0.048) were identified as risk factors for C. albicans infections. The anesthetic had been recently introduced in the hospital. Hand cultures of 8 of 14 (57%) anesthesiologists were positive for Candida species; one yielded C. albicans. Anesthesiologist 9 was the only one to use stored syringes of propofol in the infusion pump and to reuse propofol syringes. DNA fingerprinting with a digoxigenin-labeled C. albicans repetitive element 2 probe and electrophoretic karyotyping showed two distinct banding patterns among patient isolates. We hypothesize that extrinsic contamination of propofol by anesthesiologist 9 likely resulted in C. albicans infections. These data suggest that strict aseptic techniques must be used when preparing and administering propofol.