Background:  Peripheral lipoatrophy is associated with nucleoside analogue ART, but the contribution of protease inhibitors alone to lipodystrophy remains unclear. We assessed body fat changes and potential mediators in a randomized, controlled trial of lopinavir/ritonavir (LPV/r)+zidovudine/lamivudine (ZDV/3TC) induction followed by LPV/r monotherapy, vs efavirenz (EFV)+ZDV/3TC for 96 weeks of treatment.

Methods:  We randomized 155 ART-naive HIV⁺ subjects to LPV/r+ZDV/3TC induction for 24 to 48 weeks followed by LPV/r monotherapy (LPV/r arm, n = 104), or EFV+ZDV/3TC (EFV arm, n = 51). Subjects were followed for 96 weeks with DEXA scans every 24 weeks. Lipoatrophy (>20% limb fat loss) and lipohypertrophy (>20% trunk fat gain) were assessed.  Baseline serum lipids and other metabolic parameters were evaluated for association with percent changes in limb or trunk fat among subjects completing 96 weeks of treatment. 

Results:  In the LPV/r arm, 74 subjects (71%) and, in the EFV arm, 32 (63%) had serial DEXA scans available to 96 weeks. No baseline differences in DEXA measurements were observed (mean trunk fat:  9.3 kg LPV/r, 8.3 kg EFV, p = 0.46; mean limb fat 8.2 kg LPV/r, 7.4 kg EFV, p = 0.45). A significant difference in limb fat change from baseline was observed (median +18% LPV/r, –9% EFV, p <0.001 between groups) at week 96, while trunk fat changes were similar (+14% LPV/r, +15% EFV). Lipoatrophy was observed in 5% in the LPV/r arm and 34% in the EFV arm (p <0.001) and lipohypertrophy was observed in 45% and 44%, respectively (p >0.99); 0% (LPV/r arm) and 16% (EFV arm) had both lipoatrophy and lipohypertrophy. Baseline and changes from baseline in lipids and metabolic parameters were comparable between treatment groups, except for a higher median triglyceride increase in the LPV/r group (+0.67 vs +0.47 mM/L, p = 0.02).  Subjects with low baseline CD4 counts were more likely to have limb fat increases. Age, other baseline demographics, and levels of blood lipids and TNF soluble receptors 1 and 2 were not associated with limb or trunk fat changes. The difference between treatment groups in limb fat changes remained significant after adjusting for baseline CD4 count (p <0.001). There was no statistically significant change in peripheral blood mononuclear cell mtDNA in either group.

Conclusions:  Treatment with LPV/r monotherapy (compared with EFV+ZDV/3TC) was significantly and independently associated with sparing of peripheral lipoatrophy.