Pharmacokinetic-pharmacodynamic modeling of the miotic effects of dihydrocodeine in humans

The purpose of this study was to evaluate the pharmacokinetic-pharmacodynamic interrelations of pupillary effects of dihydrocodeine by two different analytic approaches.

Dihydrocodeine plasma concentrations and miotic effects were available from a previous study with 24-h measurements after administration of 60 mg dihydrocodeine to nine healthy young men. Plasma concentration versus time course was described either by a one-compartment model or by linear splines using NONMEM. Dihydrocodeine concentrations at the effect site were obtained by convolution of a first-order transfer function with the function describing the plasma concentration versus time courses, and miotic effects were related to effect-site concentrations by a sigmoidal pharmacodynamic model.

Bayesian individual fits of miotic effects were only slightly better with the spline approach than with the compartmental approach (median individual absolute weighted residuals 0.046 versus 0.058, respectively, Wilcoxon test p = 0.008; residual errors of an additive error model 0.0979 versus 0.184, respectively). Both approaches provided similar pharmacokinetic-pharmacodynamic population parameter values. The transfer half-life between plasma and effect site was 21.1 min (95% CI 11.1-34.7 min) and 19.8 min (95% CI 11.9-34 min) with spline and compartmental approaches, respectively, and miosis occurred with EC50 of 207 or 230 ng/ml, respectively.

Two modeling approaches to the miotic effects of dihydrocodeine provided similar transfer half-lives between plasma and effect site, which also agreed with previous independently estimated values obtained from analgesic effects, suggesting that pupil size is a valid biomarker to estimate the value of ke0 for opioid central nervous system (CNS) effects.