There are limited data about dyslipidemia in Asian patients treated with combination antiretroviral therapy. To assess the relative association of different protease-inhibitor-containing regimens with the degree of dyslipidemia, fasting lipid levels were compared during 110 weeks in 250 nucleoside-experienced but protease-inhibitor-naïve Thai patients beginning treatment with 5 protease-inhibitor-containing regimens. Regimens were (1) stavudine, didanosine, and saquinavir; (2) zidovudine, lamivudine, and saquinavir; (3) zidovudine, lamivudine, and indinavir; (4) zidovudine, lamivudine, and ritonavir-boosted indinavir; and (5) efavirenz and ritonavir-boosted indinavir. Triglyceride levels were available for all patients; total cholesterol and high-densitylipoprotein cholesterol levels were available for patients receiving indinavir. The strongest predictors of dyslipidemia after beginning protease-inhibitor-based therapy were treatment regimen and baseline dyslipidemia. Triglycerides, total cholesterol, and high-density-lipoprotein cholesterol changes from baseline to week 110 were significant in patients taking ritonavir-boosted indinavir. Efavirenz and ritonavir-boosted indinavir were associated with significant high-density-lipoprotein cholesterol increases compared with other regimens. Non-stavudine-containing non-boosted protease-inhibitor-based highly active antiretroviral treatment regimens had the least association with dyslipidemia.