Résumé
BACKGROUND:Mortality from cryptococcal meningitis remains very high in Africa. In the ACTA trial, 2 weeks of fluconazole (FLU) plus flucytosine (5FC) was as effective and less costly than 2-week amphotericin-based regimens. However, many African settings treat with FLU monotherapy and the cost effectiveness of adding 5FC to FLU is uncertain.METHODS:Effectiveness and costs of FLU+5FC were taken from ACTA, which included costing analysis at the Zambian site. Effectiveness of FLU was derived from cohorts of consecutively enrolled patients, managed in respects other than drug therapy, as were participants in ACTA. FLU costs were derived from costs of FLU+5FC in ACTA, by subtraction of 5FC drug and monitoring costs.Cost-effectiveness of FLU+5FC vs FLU alone was measured as the incremental cost-effectiveness ratio (ICER). Probabilistic sensitivity analysis assessed uncertainties, and a bivariate deterministic sensitivity analysis examined the impact of varying mortality and 5FC drug costs on the ICER.RESULTS:Mean costs per patient were US$847 (95%CI:776-927) for FLU+5FC, and US$628 (95%CI:557-709) for FLU. 10 week mortality was 35.1% (95%CI 28.9-41.7) with FLU+5FC and 53.8% (95%CI: 43.1-64.1) with FLU. At the current 5FC price of $US1.30 per 500mg tablet, the ICER of 5FC+FLU versus FLU alone was US$65 (95%CI: 28-208) per life year saved. Reducing 5FC cost to between US$0.80 and US$0.40 per 500mg resulted in an ICER between US$44 and US$28 per life year saved.CONCLUSIONS:Addition of 5FC to FLU is cost-effective for cryptococcal meningitis treatment in Africa and if made available widely could substantially reduce mortality rates among HIV-infected persons in Africa.