To evaluate whether oral fluconazole reduces the risk of a first episode of cryptococcal meningitis in HIV-infected patients, we conducted a case-control study of patients cared for in a university teaching hospital and two urban HIV-outpatient clinics. Cases consisted of HIV-infected patients with CD4 cell counts less than 250/microL who developed a first episode of cryptococcal meningitis between July 1, 1990, and June 30, 1993. For each case (n = 18), 4 control subjects were chosen from HIV-infected patients (CD4 count < 250/microL) whose cerebrospinal fluid was negative for cryptococcal antigen and culture, and who were matched by age, sex, and time of lumbar puncture. There were no significant differences between cases and controls in age, sex, insurance status, mean CD4 count, history of oral candidosis, presence of a previous AIDS-defining illness, the number of visits to the HIV-outpatient clinic, or use of antiretroviral therapy. In the 6 months before lumbar puncture, 2 of 18 cases (11%) and 26 of 72 controls (36%) were exposed to fluconazole, a finding that gives a matched odds ratio (adjusted for race, route of HIV infection, and CD4 count) of 0.08 (95% CI 0.01-0.84; p = 0.035) and indicates a 92% protective efficacy. We conclude that fluconazole reduces the risk of a first episode of cryptococcal meningitis in HIV-infected patients with a CD4 count less than 250/microL. Although the optimum dose and duration of fluconazole could not be determined, our results suggest that less than daily use was effective in the prevention of cryptococcal meningitis.