One year of HAART in Mozambique: survival, virological and immunological results of DREAM project in Adults and Children

Background. Since February 2002 the DREAM programme, run in Mozambique by the Community of Sant’Egidio, provided free-of-charge to the patients both HAART and immunological and virological monitoring. Methods: DREAM is working in the Mozambican public health sector. HAART is offered to eligible HIV positive patients in both day hospital and home care service. Health education as well as nutritional assessment and supplementation are included in the protocol. As on 15th August 2003, 802 HIV+ individuals (510 female) are included in the community and home care programme run in the Machava center (Maputo province): close to 50% had a CD4 count below 200/mm3 and more than one third was 3-4 WHO clinical stage. Generic ARV are provided to the patients. Results: More than 50% of patients met the criteria to start ARV treatment. The overall rate of lost to follow up was 9.3%, higher in the not treated group (12.7% vs 7.2%). Patient receiving HAART had more likely a lower CD4 count (median value 186 vs 396; p<0.001) lower Haemoglobin (mean value 10.1 vs 10.6; p=0.004), lower weigth (mean value 55.3 vs 58.5 Kg; p=0.001). On the sub-sample of patient receiving HAART the percentage of people with CD4 count below 200 decreased from 68.2% to 22.5% after one year of treatment. Close to 75% reach an undetectable viral load (<50 copies/ml – bDNA 3.0) within one year of treatment. The percentage of death was 12.5%, with a median time of survival of 70 days in the not treated patients. The follow up time was higher in the treated sub-sample (290 days vs 140). Univariate Cox’ model shows a significant correlation with the following variables: Haemoglobin (RR: 2.65 IC95%: 2.01-3.49), Viral Load (RR: 2.29; IC95% 1.57-3.35), CD4 count (RR:1.55 IC95% 1.37-1.75), HAART (RR: 1.54; IC95% 1.04-2.29), age (RR: 1.03 IC95% 1.01-1.05) and weight (RR: =0.93 IC95% 0.91-0.95). Multivariate Cox ‘model shows the independent role as predictor of longer survival of haemoglobin, CD4 count, Viral load and HAART with a risk of 2.34, 1.41, 1.74 and 2.98 respectively (p≤ 0.001). Conclusion: One year follow up shows the durable effect of generic ARV in patients living in LRS. The lost to follow up rate was below 10% and is lower in patients starting therapy, directly confirming an overall good adherence in HAART subsample. The survival analysis of the 2 groups suggests an emerging role of haemoglobin and body weight as predictors of mortality and markers of the general health status. The high mortality levels in the non treated patients addressees to reassess the criteria to start HAART.