OBJECTIVE We conducted a comparison of itraconazole versus amphotericin B plus flucytosine in the initial treatment of cryptococcal meningitis in patients with AIDS and established the efficacy of itraconazole as maintenance treatment. METHODS The trial was a prospective, randomized, and non-blinded study. METHODS The study was performed at an academic centre for AIDS, Amsterdam, The Netherlands. METHODS Twenty-eight HIV-1-seropositive men with a presumptive diagnosis of cryptococcal meningitis, randomized between 5 February 1987 and 1 January 1990, were included for analysis. METHODS Oral itraconazole (200 mg twice daily), versus amphotericin B (0.3 mg/kg daily) intravenously plus oral flucytosine (150 mg/kg daily) was administered for 6 weeks followed by maintenance therapy with oral itraconazole (200 mg daily) to all patients. METHODS Outcome measures were a complete or partial response, recrudescence and relapse. RESULTS A complete response was observed in five out of the 12 patients who completed 6 weeks of initial treatment with itraconazole versus all 10 patients who completed treatment with amphotericin B plus flucytosine (P = 0.009). A partial response was observed in seven out of the 14 patients assigned to itraconazole. During maintenance therapy, recrudescence (n = 6) or relapse (n = 1) occurred in seven out of the 12 patients initially assigned to itraconazole, whereas two relapses occurred among nine patients initially treated with amphotericin B plus flucytosine (P = 0.22); recurrence of clinical symptoms was significantly related to a positive cerebrospinal fluid culture at 6 weeks (P = 0.003). CONCLUSIONS Itraconazole is less effective compared with amphotericin B plus flucytosine in achieving a complete response in initial therapy in AIDS patients with cryptococcal meningitis.