The safety, efficacy, and mutual interactions of combination amprenavir with lopinavir/ritonavir as deep salvage treatment were investigated in a prospective 24-week pilot study. HIV-infected patients (n = 22) with virologic failure to nucleoside reverse transcriptase inhibitors (NRTIs), non-NRTIs, and at least 2 protease inhibitors received 400/100 mg of lopinavir/ritonavir with 600 mg of amprenavir every 12 hours combined with NRTIs. Patients receiving the same doses of lopinavir/ritonavir (n = 10) or amprenavir with ritonavir (n = 8) were chosen as controls for pharmacokinetic analyses. Mean changes from baseline HIV RNA levels and CD4 counts after 24 weeks were -1.13 log10 copies/mL and +88 x 10 cells/L, respectively. The mean plasma trough concentration (Cmin)and peak concentration (Cmax) of amprenavir were 104% and 228% lower and the Cmin of lopinavir was 46% lower in patients in whom the drugs were coadministered than in controls. There were 4 permanent drug discontinuations because of toxicity. An inhibitory quotient (IQ) of amprenavir higher than 0.8 was the best predictor of virologic outcome at 24 weeks, even after adjusting for amprenavir Cmin or phenotypic susceptibility. Deep salvage therapy using lopinavir/ritonavir with amprenavir is sufficiently safe and shows partial efficacy. When these drugs are coadministered, therapeutic drug monitoring should be employed and the IQ can be used to determine target drug levels.