The aim of this study was to quantify the change in saquinavir and amprenavir exposure when combined and used with low-dose ritonavir; to evaluate 24-week safety and immunologic and virologic response. It was a randomized, nonblinded, prospective study. There were 11 HIV-1-infected, antiretroviral-experienced, male and female subjects > or = 18 years old, median HIV-1 RNA and CD4(+) T-cell count of 4.86 log copies/mL and 10(6) cells/mm(3), respectively. Subjects were randomly assigned to receive saquinavir 1000 mg/ritonavir 100 mg every 12 hours or amprenavir 600 mg/ritonavir 100 mg every 12 hours for 7 days. After 12-hour pharmacokinetic sampling, the third protease inhibitor (PI) was added, and pharmacokinetics sampling was repeated 14 days later. Subsequent PI dosage adjustments were based on real-time pharmacokinetic assessment. Saquinavir did not affect amprenavir or ritonavir pharmacokinetics. Amprenavir decreased area under the concentration-time curve (AUC(0-12h)) and C(12h) for saquinavir 82 and 61%, and 74 and 75% for ritonavir. An adjusted PI regimen of amprenavir 600 mg/saquinavir 1400 mg/ritonavir 200 mg every 12 hours returned saquinavir exposure to baseline. At 24 weeks, HIV RNA declined a median of 1.55 log copies/mL and CD4(+) T-cell counts increased a median of 52 cells/mm(3). Gastrointestinal events predominated and were mild to moderate. These data suggest that amprenavir/saquinavir/ritonavir may be a viable salvage regimen in heavily PI-experienced individuals. New formulations of amprenavir and saquinavir may simplify this regimen.