Adult AIDS Clinical Trials Group 5043 examined pharmacokinetic (PK) interactions between amprenavir (APV) and efavirenz (EFV) both by themselves and when nelfinavir (NFV), indinavir (IDV), ritonavir (RTV), or saquinavir (SQV) is added. A PK study was conducted after the administration of single doses of APV (day 0). Subjects (n = 56) received 600 mg of EFV every 24 h (q24h) for 10 days and restarted APV with EFV for days 11 to 13 with a PK study on day 14. A second protease inhibitor (PI) (NFV, 1,250 mg, q12h; IDV, 1,200 mg, q12h; RTV, 100 mg, q12h; or SQV, 1,600 mg, q12h) was added to APV and EFV on day 15, and a PK study was conducted on day 21. Controls continued APV and EFV without a second PI. Among subjects, the APV areas under the curve (AUCs) on days 0, 14, and 21 were compared using the Wilcoxon signed-rank test. Ninety-percent confidence intervals around the geometric mean ratios (GMR) were calculated. APV AUCs were 46% to 61% lower (median percentage of AUC) with EFV (day 14 versus day 0; P values of <0.05). In the NFV, IDV, and RTV groups, day 21 APV AUCs with EFV were higher than AUCs for EFV alone. Ninety-percent confidence intervals around the GMR were 3.5 to 5.3 for NFV (P < 0.001), 2.8 to 4.5 for IDV (P < 0.001), and 7.8 to 11.5 for RTV (P = 0.004). Saquinavir modestly increased the APV AUCs (GMR, 1.0 to 1.4; P = 0.106). Control group AUCs were lower on day 21 compared to those on day 14 (GMR, 0.7 to 1.0; P = 0.042). African-American non-Hispanics had higher day 14 efavirenz AUCs than white non-Hispanics. We conclude that EFV lowered APV AUCs, but nelfinavir, indinavir, or ritonavir compensated for EFV induction.