OBJECTIVE: Highly active antiretroviral therapy (HAART) has been associated with lipodystrophy (LD) in adults but data are more limited for children. The purpose of this study was to determine the prevalence of and risk factors for LD in Tanzanian children receiving HAART by clinical assessment and to compare the results with anthropometric data. DESIGN AND METHODS: A cross-sectional study was performed in a cohort of HIV-infected children aged 1-18 years receiving HAART in a single center in Moshi, Tanzania. Age, gender, past and current medication regimens and anthropometric measurements were recorded. A clinical scoring method was used to assess LD. Backward binary multivariate logistic regression was used to determine relationships between anthropometric measurements and the presence of clinical LD. RESULTS: Among 210 HIV-infected children, the prevalence of LD was 30% (95% confidence interval [CI]: 23.8-36.2) overall, 19% (95% CI: 13.7-24.3) for lipoatrophy only, 3.8% (95% CI: 1.2-6.4) for lipohypertrophy only and 7.1% (95% CI: 3.6-10.6) for the mixed type. Most cases were mild. Older age and use of stavudine increased the risk of LD. Overall, the study population was stunted but not underweight. In children with relatively lower weight-for-height (<1), only the mid-upper arm circumference was found to be associated with lipoatrophy, while nearly all anthropometric measurements were associated with lipoatrophy in the well-nourished (weight-for-height ≥1) children. CONCLUSIONS: Our findings demonstrate that LD is a significant problem among Tanzanian HIV-infected children receiving HAART. Anthropometric measurements predicted LD in well-nourished children but generally failed to do so in relatively wasted children. Our findings support current efforts to avoid stavudine use in children.