The purpose of this study was to assess the one-year efficacy of highly active antiretroviral therapy (HAART) administered by general practitioners in a primary care community clinic in rural South Africa. We performed an observational cohort study of 675 treatment-naive human immunodeficiency virus (HIV)-infected patients (including 66 children) who began HAART at least 12 months prior to the data analyses. Throughout treatment, the CD4+ T-cell count (percentage of CD4+ T-cells in children) and plasma HIV-RNA level were determined and the patient's weight was recorded. The primary outcome was mortality. Secondary outcomes were viral suppression, immunological response, and weight gain. One year after the start of HAART, 100 of the 675 (15%) patients were lost to follow-up and 119 patients (18%), including six children, died. Mortality was highest during the first few months of treatment. Based on an on-treatment analysis at one year after the start of therapy, 83% of adults and 71% of children had a viral load <400 copies/ml; the viral load was <50 copies/ml in 70% of adults and 61% of children. At one year, the mean CD4+ T-cell count in adults had increased by 236/mm(3), and the mean body mass index (BMI) had increased by 3.5 kg/m(2). In children, the mean CD4% had increased by 17.6. A low Karnofsky score and a low baseline CD4+ T-cell count were independently associated with death. In addition to these factors, a low baseline BMI and gender were predictive of a poor immunological outcome. Our study shows that adequately monitored HIV/acquired immunodeficiency syndrome (AIDS) care administered by general practitioners and their staff is feasible and leads to good results in a rural, primary care center in sub-Saharan Africa. In order to achieve even better results, early mortality should be reduced and efforts should be made to start HAART earlier.