Background:  Thymidine NRTI (tNRTI) are strong inhibitors of PPAR-g. Conflicting data exist on the efficacy of the PPAR-g agonists glitazones on lipoatrophy in the setting of tNRTI. We report the first study of a glitazone in lipoatrophy in HIV-infected subjects treated with tNRTI-sparing regimens.

Methods:  This randomized, double-blind, placebo-controlled study evaluated the effect of rosiglitazone (ROSI) on limb fat in HIV-infected subjects with lipoatrophy who discontinued their tNTRI at least 24 weeks prior to enrollment. They were randomized to ROSI (4 mg twice daily) vs placebo for 48 weeks. Dual energy X-ray absortiometry (DEXA)-scans and fasting metabolic assessments were serially performed. Distributionally appropriate paired tests were used to determine significant changes from baseline within groups, and 2-sample tests were used to determine differences at baseline and for changes from baseline between groups.

Results:  We enrolled 71 subjects:  17% were female and 51% white. At baseline, the mean ±SD limb fat (grams) was 6533±4381 and 6413±3272; p = 0.70, in ROSI and placebo arms, respectively. At entry, 26 (37%) had insulin levels >15 μU/mL, and 92% of subjects had HIV-1 RNA <400 copies/mL. Baseline characteristics were similar between groups except for higher mean total cholesterol levels (mg/dL) in the placebo arm (180 vs 200; p = 0.045). Overall, 9 subjects were lost to follow-up (4 on ROSI). Of these, only 1 (in ROSI group) had discontinued the study for a possible adverse event (possible exacerbation of pre-diagnosed coronary artery disease). No subjects modified their entry NRTI during the study. At week 48, 98% of subjects had HIV-1 RNA <400 copies/ mL. Limb fat (grams) increased significantly more in the rosi group than in the placebo group (mean±sd of 911±1215 vs 253±1039; p = 0.018). The mean HOMA and insulin levels (μU/mL) decreased from baseline only in the ROSI arm –1.9±6.8 vs 0.5±2.6 and –7.4±27.3 vs 2.7±12.2; p = 0.03 and 0.02, respectively). Triglycerides, HDL- and non-HDL cholesterol did not change significantly within or between groups. Six subjects (3 on ROSI) started statin or fibrate during the study. Total bone mineral density, CD4⁺ cell count and body mass index did not change significantly between groups.

Conclusions:  In the absence of thymidine NRTI, ROSI significantly improves lipoatrophy and insulin resistance, without adversely affecting lipids or bone density.