Background: Published lopinavir (LPV) pharmacokinetic data, generated in treatment naïve patients (pts), may not adequately reflect LPV pharmacokinetics and its clinical variability. Our objective was to describe the pharmacokinetics of LPV at steady-state in treatment-experienced pts with population methods, and to relate LPV exposure measures to response.

Methods: CCTG 578 is an ongoing, randomized, 3x2 factorial study of 3 adherence interventions crossed with therapeutic drug monitoring of approved PIs and NNRTIs vs standard care. Drug levels drawn pre-, 2  and 4 hrs post-witnessed dose (wk 2) and randomly (wks 4, 6) were analyzed by HPLC for LPV and ritonavir (RTV) among LPV-treated subjects. Thirty-three (33) pts with 174 LPV plasma levels were used in the NONMEM program to develop a population pharmacokinetic model. The impacts of co-variates were evaluated by changes in the objective function (model goodness of fit). Post-hoc Bayesian estimates of individual subject’s LPV exposure measures were obtained. C₁₂/IC₅₀ ratios based on screening phenotypes (ViroLogic), RNA and D RNA (wks 1, 2, 4, and 6) were compared to predicted LPV exposure measures using two-tailed Pearson correlation co-efficients.

Results: Concomitant NNRTI use and RTV levels were associated with higher LPV oral clearance (CL/F) and bioavailability, respectively. Inclusion of RTV in the model decreased variability in apparent volume of distribution (V_d/F) and CL/F estimates by 48% and 11%, respectively (mean RTV level = 0.5 mg/L). A one-compartment model estimated the V_d/F to be 129 L (97% CV) and CL/F to be 8.1 L/hr (9.3 L/hr if on NNRTI) (29% CV). Residual error was 22%. Paradoxically, higher LPV clearance (lower LPV levels) overall (n = 27, wk 6) was associated with greater RNA reduction (r = -0.3 to -0.5, wks 1 to 6). This was driven by the group of pts (n = 13, wk 6) who had been off therapy for ³ 4 months at initiation of the LPV regimen (r = -0.4 to -0.7). C₁₂/IC₅₀ ratios were not associated with D RNA for wks 1 to 6, overall. However, for the group who had been on therapy at screening (n = 14, wk 6), higher C₁₂/IC₅₀ ratios were associated with greater RNA reduction at wks 1 to 6 (r = -0.3 to -0.7).

Conclusions: LPV oral clearance was higher than literature values reported for treatment naïve pts (~ 5 L/hr) despite comparable RTV levels. The C₁₂/IC₅₀ ratio was predictive of RNA changes only for pts who were on therapy when phenotyped.