Limitations of fexofenadine limited sampling strategy using plasma concentrations and partial area under the concentration-time curve to estimate transporter activity in healthy adults.

Abstract

Fexofenadine is a probe drug used to phenotype P-glycoprotein (P-gp) and organic anion transporting polypeptide (OATP) 1B1/3 activities. This study evaluated a limited sampling strategy using plasma concentrations and/or partial area under the concentration versus time curves (AUCs) to estimate systemic exposure and, potentially, P-gp and OATP1B1/3 activities. Plasma concentration versus time data were obtained from 53 healthy adult participants (22 females) from four published studies. Participants were administered a single oral dose (120 mg) of fexofenadine during constitutive P-gp and OATP1B1/3 conditions. Concentration-time data were divided into a training (n=18) and validation (n=35) set. Backwards stepwise linear regression generated single-, 2-timepoint, and partial AUC limited sampling models (LSMs). Noncompartmental analysis methods were used to determine total AUC (AUC0-lNF) from intensive sampling. Coefficient of determination (r2) and bias and precision were assessed via relative percent mean prediction error (%MPE), relative percent mean absolute error (%MAE), and relative percent root mean square error (%RMSE). The geometric mean observed AUC0-INF was 1,680 ng×h/mL. The 2-, 5-, and 2- plus 5-hour LSMs met backwards stepwise linear regression significance (p<0.15) to remain in the model but had unacceptable %RMSE (17-29%). The majority of partial AUC LSMs had unacceptable r2 (0.21-0.83), with all models having unacceptable %MAE (12-35%). Fexofenadine limited sampling strategy using single-timepoint, 2-timepoint, and partial AUCs were unable to accurately estimate AUC0-lNF and thus constitutive P-gp and OATB1B1/3 activities in healthy adults. Timepoints that were not measured or selected may have improved LSM performance.