BACKGROUND: Substitution of stavudine with zidovudine may lead to recovery from lipodystrophy (LD) in HIV-infected children. METHODS: We prospectively followed HIV-infected children enrolled in an earlier LD study conducted between 2002 and 2004 at Chiang Mai University Hospital in northern Thailand. In 2006, stavudine was substituted with zidovudine. All children were evaluated by a clinical LD checklist modified from that of the European Pediatric LD study group together with waist/hip measurement at baseline and 24, 48, 72, and 96 weeks after substitution. The waist-to-hip ratios were converted to age- and sex-adjusted z scores based on normal ranges in healthy Thai children. RESULTS: Forty-five lipodystrophic children with 36 episodes of lipohypertrophy and 22 episodes of lipoatrophy were enrolled. By weeks 48 and 96 after substitution, 40% and 47% of lipohypertrophy resolved, whereas 59% and 73% of lipoatrophy resolved, respectively. The rate of resolution of lipoatrophy was higher than that of lipohypertrophy at 48 weeks after substitution and thereafter. Ninety-six weeks after changing to zidovudine therapy, 8 children still had LD (1 with both lipoatrophy and lipohypertrophy, 7 with lipohypertrophy). No clinically significant hematologic adverse event was observed. CONCLUSIONS: Substitution of stavudine with zidovudine resulted in decreased severity or resolution of LD among HIV-infected children and adolescents.