Most protease inhibitors available for the treatment of human immunodeficiency virus (HIV) infection are highly bound to plasma proteins, mainly alpha-1 acid glycoprotein. Therapeutic drug monitoring (TDM) of total protease inhibitor (PI) concentrations has been increasing in the past few years; however, the pharmacological activity of the PIs is dependent on unbound drug entering cells harboring HIV. There is little information available on unbound drug concentrations of these drugs in vivo. The aim of the study was to measure unbound plasma concentrations of lopinavir (LPV) and to relate them to the total plasma concentrations to establish the unbound percentage in vivo during a full dosage interval. A pharmacokinetic study was performed in HIV-infected subjects (n = 23; median CD4 cell count = 290 x 10(6) cells x L(-1); viral load < 50 copies x mL(-1)) treated with a LPV/ritonavir (RTV)-containing regimen. Ultrafiltration was used to separate unbound LPV in all plasma samples (n = 115). Equilibrium dialysis was also used to compare with ultrafiltration measurements in 10/23 patients at baseline and 2 hours after drug intake. Total and unbound LPV concentrations were measured by a fully validated method using high-performance liquid chromatography-mass spectometry (HPLC-MS/MS). Based on a comparison of AUC(unbound)AUC(total), the mean (+/- SD) unbound percentage of LPV from all the samples studied (n = 115) was 0.92% (+/- 0.22) when measured with ultrafiltration and 1.32% (+/- 0.44) when equilibrium dialysis was used (n = 20), showing a higher drug recovery (P = 0.048). The unbound percentage of LPV was found to be significantly higher after 2 h than at baseline (P < 0.05 with both methods), suggesting a concentration-dependent binding of LPV that has not been observed in vitro. However, the clinical significance of such phenomena is still unclear.