OBJECTIVE: To define more rapidly effective initial antifungal regimens sustainable in resource-constrained settings. METHODS: Cohort study in SW Uganda: Thirty HIV-seropositive, antiretroviral therapy-naïve, patients with first episode cryptococcal meningitis were treated with high dose fluconazole (1200 mg/d for 2 weeks, then 800 mg/d until ART started) plus amphotericin B (AmB, 1 mg/kg/d), with routine normal saline and potassium supplementation, for the initial 5 days. Outcome measures were early fungicidal activity (EFA), determined by serial quantitative CSF cultures, safety, and mortality. RESULTS: EFA was -0.30 ± 0.11 log CFU/day calculated over the first 2 weeks of treatment, with no reduction in the rate of clearance between days 5 and 14. There was no grade IV hypokalemia or elevated creatinine, and no grade III or IV anemia or elevation of ALT. AmB or high dose fluconazole were not stopped early in any patient. Mortality was 23% at 2, and 28% at 10 weeks. CONCLUSIONS: Short course AmB was associated with rapid clearance of infection and was well-tolerated, suggesting it could be used safely in many centres currently relying on fluconazole monotherapy. Phase III trials are needed in African centres to compare short course with the standard 2-week course of AmB.