Stavudine use is a contributing factor for lipoatrophy, whereas use of abacavir or zidovudine is less likely to cause this complication. The TARHEEL study was a 48-week, open-label study that assessed changes in lipoatrophy after abacavir (86 patients [73%]) or zidovudine (32 patients [27%]), 300 mg twice daily, was substituted for stavudine for 118 human immunodeficiency virus (HIV)-infected patients (HIV type 1 RNA level, <400 copies/mL) with virological suppression who had developed lipoatrophy after > or =6 months of stavudine-based treatment. At week 48, full-body dual-energy x-ray absorptiometry demonstrated a median increase in arm fat of 35%, leg fat of 12%, and trunk fat of 18%, compared with the baseline level. These improvements coincided with fat gain in lipoatrophic areas that was documented by computerized tomography. Results of a "body image" questionnaire showed that a substantial percentage of patients reported some or a lot of fat gain in the arms (22%), legs (18%), buttocks (19%), and face (27%). HIV suppression was maintained over the study period. In conclusion, replacing stavudine with abacavir or zidovudine resulted in improvement in stavudine-induced lipoatrophy.