Recent studies suggest body shape changes, fat redistribution, and fat lipoatrophy may be related to the NRTI component of patients' HAART and not to the protease inhibitor (PI) component. The hypothesis of this study is that thymidine analogues such as stavudine (d4T) and zidovudine (ZDV) cause lipoatrophy more so than non-thymidine analogues and that removal of thymidine analogues from HAART in patients with defined lipoatrophy will reverse this process.

In Step 1, patients will undergo axial mid-thigh and abdomen computer tomography (CT) scans. If the CT scans are readable, patients are restrictively and randomly assigned to 1 of 2 treatment arms in Step 2. Patients in Arm A-1 will replace the thymidine analogue component (stavudine [d4T] or zidovudine [ZDV]) of their HAART with abacavir (ABC). Patients in Arm B-1 will discontinue their current HAART and will receive a PI and a nonnucleoside reverse transcriptase inhibitor (NNRTI), either lopinavir/ritonavir (LPV/r) and nevirapine (NVP) or atazanavir, ritonavir, and NVP. Patients currently on efavirenz (EFV) not provided by the study may choose to continue with EFV instead of switching to NVP. Comparisons will be made to the baseline values of subcutaneous fat measured by mid-thigh and abdominal CT. Patients in Arms A-1 and B-1 remain on study for a total of 48 weeks and do not advance to Step 3.

Two additional groups (Arms A-2 and B-2) made no changes to HAART for 28 weeks to evaluate the natural history of change in lipoatrophy over time; accrual into these groups and into Step 3 has been discontinued. At Week 28, patients in Arms A-2 and B-2 were registered to Step 3 and switched from HAART to a designated new treatment. Arm A-2 patients will replace d4T or ZDV with ABC for 48 weeks. Arm B-2 patients replace their HAART with LPV/r plus NVP for 48 weeks. If patients in Arms A-2 and B-2 have not completed the 28-week delay and have not switched regimens, they will enter Step 4 and be reregistered into Arms A-1 and B-1, respectively, remaining on their treatment assignment for 48 weeks. If patients in Arms A-2 and B-2 have already switched regimens, then they will continue on their new regimens until Week 76.