OBJECTIVE Determine outcome differences between orphaned and non-orphaned children receiving antiretroviral therapy (ART). METHODS Retrospective review of prospectively recorded electronic data. METHODS Nine HIV clinics in western Kenya. METHODS 279 children on ART enrolled between August 2002 and February 2005. METHODS Orphan status, CD4%, sex- and age-adjusted height (HAZ) and weight (WAZ) z scores, ART adherence, mortality. RESULTS Median follow-up was 34 months. Cohort included 51% males and 54% orphans. At ART initiation (baseline), 71% of children had CDC clinical stage B or C disease. Median CD4% was 9% and increased dramatically the first 30 weeks of therapy, then leveled off. Parents and guardians reported perfect adherence at every visit for 75% of children. Adherence and orphan status were not significantly associated with CD4% response. Adjusted for baseline age, follow-up was significantly shorter among orphaned children (median 33 vs. 41 weeks, P = 0.096). One-year mortality was 7.1% for orphaned and 6.6% for non-orphaned children (P = 0.836). HAZ and WAZ were significantly below norm in both groups. With ART, HAZ remained stable, while WAZ tended to increase toward the norm, especially among non-orphans. Orphans showed identical weight gains as non-orphans the first 70 weeks after start of ART but experienced reductions afterwards. CONCLUSIONS Good ART adherence is possible in western rural Kenya. ART for HIV-infected children produced substantial and sustainable CD4% improvement. Orphan status was not associated with worse short-term outcomes but may be a factor for long-term therapy response. ART alone may not be sufficient to reverse significant developmental lags in the HIV-positive pediatric population.