Background: Access to antiretroviral therapy in resource-poor countries is extremely limited by high costs and lack of infrastructures. Aim of this project, coordinated by the Community of Sant’Egidio, is then the evaluation of feasibility and impact of free-of-charge combination antiretroviral therapy (CART) in a poor semi-rural area of Mozambique.

Methods: Prospective, observational cohort study of HIV-patients (pts) recruited from Nov 2001 to Aug 2002, followed by outpatient and home-care service in Matola, Mozambique. Since Feb 2002, CD4 and HIV-RNA are routinely measured; generic antiretrovirals (AZT or d4T+3TC+Nevirapine [NVP] in combination tabs) are provided according to given protocols. Pts are monitored for medication adherence, toxicity, and efficacy. Other than CART, all pts received nutritional supplementation, prophylaxis with cotrimoxazole (CTX-proph), and treatment of TB and malaria, if needed. All consecutive pts with baseline CD4 < 200 were analysed by an intent-to-treat approach. Follow-up was censored on Aug 31, 2002.

Results: Of 148 pts included in the study, 73% were females, median age 30 yrs (range 15–58), CD4 109 (0–199), VL 5.26 log (1.69–6.70). All pts underwent VCT and accepted treatment when prescribed. Within 1 month from the first visit, 82 pts (55%) received CART, while 66 received only nutritional supplementation, CTX-proph and, if needed, treatment of TB and malaria. Adherence to CART was very high; WHO grade 3 toxicity occurred in 3 pts (who temporarily interrupted CART). There were no significant differences in terms of baseline clinical or laboratory parameters, age, VL, CD4, and weight, between those receiving or not CART. After 8 weeks of CART, mean change in VL was -2.6 log (71% < 500 c/ml), and mean CD4 increase was 164 cells/μl. Kaplan-Meier curves showed a cumulative probability of 24-weeks survival of 91% in CART and 43% in no CART (log rank p < 0.001). At Cox regression, CART was significantly associated with time-to-death (HR 0.10; 95%, CI 0.04–0.28), while CD4 showed a trend (100 cells higher, HR 0.28;0.08–1.03). In a multivariate Cox model CART remained independently associated with longer survival (HR 0.13; 0.01–0.49).

Conclusions: Use of CART based upon generic-manufactured antiretrovirals produced an adjusted mean reduction of 87% (95% CI 51%–99%) of death hazard as compared with no CART but same background care in HIV pts with CD4 < 200 in Mozambique. Recruitment and follow-up in the DREAM cohort are continuing.