OBJECTIVE Few studies have investigated the efficacy of antiretroviral therapy among HIV-infected children in resource-poor settings. This observational, retrospective analysis describes the clinical, immunologic, and virologic effects of highly active antiretroviral therapy in treatment-naive, HIV-infected children in Mombasa, Kenya. In keeping with a public health approach, all children were treated by using a simplified, nationally approved, triple-drug regimen. METHODS Clinical data and stored plasma samples from 29 children who were followed prospectively between April 2003 and October 2004 were analyzed. All children received generic formulations of nevirapine, zidovudine, and lamivudine and were evaluated at baseline and at 3, 6, 9, 12, and 15 months. At each visit, weight and CD4 lymphocyte counts were measured and plasma samples were stored for analysis. HIV RNA load was determined retrospectively at baseline and 9 months after initiation of therapy. RESULTS The mean age of the children was 8.5 years (range: 2-16 years). At baseline, the mean CD4 count (+/-SD) was 182.3 x 10(6) cells per microL (+/-145.6). On treatment, CD4 counts increased step-wise by a mean of 187 x 10(6) cells per microL at 3 months, 293 cells per microL at 6 months, 308 cells per microL at 9 months, 334 cells per microL at 12 months, and 363 cells per microL at 15 months. The mean plasma viral load decreased from a baseline level of 622,712 to 35,369 copies per mL, and at 9 months was undetectable in 55% of the patients. Mean z scores for weight for age increased from a baseline of -1.61 to -1.12 at 12 months into therapy. CONCLUSIONS A public health approach using 1 treatment regimen in generic form showed excellent efficacy among treatment-naive, HIV-infected children in a resource-limited country. Clinical and immunologic improvement occurred in all patients, but 9 months after the start of therapy, only 55% of the children had an undetectable viral load.