Population pharmacokinetics and limited sampling strategy for therapeutic drug monitoring of mycophenolate mofetil in Japanese patients with lupus nephritis

Abstract

BackgroundMycophenolate mofetil (MMF), a prodrug of the immunosuppressive agent mycophenolic acid (MPA), is difficult to administer because of the pharmacokinetic complexity of MPA. Although dosage adjustment according to the 12-h area under the concentration–time curve (AUC0-12) is thought to be desirable, multiple blood samplings for AUC calculation may pose a clinical challenge. A limited sampling strategy (LSS) would provide a solution; however, little is known about MPA pharmacokinetics in lupus nephritis patients, especially in those with Asian backgrounds, or few, if any, LSSs are reported for them.MethodsThirty-four adult Japanese patients receiving MMF for lupus nephritis were examined retrospectively. MPA pharmacokinetics were investigated, and a PPK model was developed using Phoenix® NLME™ software. Single and double blood sampling strategies from Bayesian estimation using the PPK model and from multiple linear regression were compared. Tolerability was also evaluated.ResultsIn the pharmacokinetic analysis, renal function and serum albumin had significant effects on dose-normalized AUC0-12; and serum albumin, concomitant proton pump inhibitor (PPI) and iron/magnesium oxide did on dose-normalized maximum concentration. As a PPK model, a two-compartment model was developed with a transit absorption model and first-order elimination, in which creatinine clearance and serum albumin were covariates for MPA clearance. The double sampling strategy at 1 and 4 h by multiple linear regression showed the best agreement with the observed AUC0-12 (r2 = 0.885). Of the single sampling strategies, the one at 6 h by Bayesian estimation performed best (r2 = 0.769). The tolerability evaluation showed that correlations were suggested for gastrointestinal involvement.ConclusionsThe present study developed the first PPK model of MPA for Japanese lupus nephritis patients. As for LSSs, a double sampling strategy at 1 and 4 h by multiple linear regression would work best; when only a single blood sampling is allowed, a strategy at 6 h by Bayesian estimation using the PPK model developed in this study would be best. The LSSs good enough for clinical use may facilitate safer, more effective, and individualized therapy.