A randomized trial was conducted to compare the efficacy and safety of fluconazole versus that of amphotericin B in the treatment of candidemia in non-neutropenic adults. Enrollment was stratified by disease severity (APACHE II score). Patients were randomized (1:1) to receive amphotericin B 0.6 mg/kg/day (cumulative dose 8 mg/kg) or fluconazole 800 mg intravenous loading dose, then 400 mg daily for four weeks (intravenous for at least 10 days). Patients were monitored for six months. A total of 106 patients were enrolled. A protocol amendment implemented midway through the trial required patients to be removed from the study and treated with amphotericin B if species identification indicated candidemia due to Candida glabrata or Candida krusei. Baseline characteristics were similar for the two groups; 103 patients (fluconazole, 50; amphotericin B, 53) met the major enrollment criteria. The intention-to-treat analysis indicated successful therapy in 50% of fluconazole recipients compared to 58% of the amphotericin B group (p = 0.39; one-sided 95% CI, -8 to 24%). The efficacy analysis included 84 patients (fluconazole, 42; amphotericin B, 42); successful outcomes were observed in 57% and 62% of cases in the fluconazole and amphotericin B groups, respectively (p = 0.66: one-sided 95% CI, -12 to 22%). The mortality at day 14 for the fluconazole group was 26% and for the amphotericin B group 21% (p = 0.52; chi-square test) and remained similar throughout the course of follow-up, Drug-related adverse events were more frequent with amphotericin B than with fluconazole and prompted switching of therapy for two (4%) and zero cases, respectively. Fluconazole and amphotericin B were associated with similar clinical response rates and survival in the treatment of candidemia among non-neutropenic patients; however, drug-related adverse events were more frequent with amphotericin B.