AbstractWe previously analysed the serum concentrations of dihydrocodeine in a 1‐month‐old infant with respiratory depression after being prescribed dihydrocodeine phosphate 2.0 mg/day divided t.i.d. for 2 days. The purpose was to develop a full physiologically based pharmacokinetic (PBPK) model that could account for these and other drug monitoring results. Based on experiments in Caco‐2 cell monolayers, the effective permeability of dihydrocodeine in human jejunum was established as 1.28 × 10−4 cm/s. The in vitro Vmax/Km values for dihydrocodeine demethylation mediated by recombinant cytochrome P450 2D6 and 3A4 were 0.19 and 0.066 μl/min/pmol, respectively, and for dihydrocodeine 6‐O‐glucuronidation mediated by recombinant UGT2B4 and 2B7, the Vmax/Km values were 0.14 and 0.22 μl/min/mg protein, respectively. Renal clearance was calculated as 5.37 L/h on the total clearance value multiplied by the fraction recovered in urine. The reported plasma concentration–time profiles of dihydrocodeine after intravenous administration in healthy volunteers were used to adjust the tissue partitioning ratios. The developed model simulated the pharmacokinetic profiles of dihydrocodeine after single and multiple oral administrations reasonably well in the same population. Subsequently, the validated model was used to simulate pharmacokinetic profiles for five pediatric cases, including the 1‐month‐old Japanese boy and a 14‐year‐old Japanese girl who took an overdose of dihydrocodeine phosphate (37 mg). The simulated pharmacokinetic profiles for five virtual pediatric subjects matching the age, gender, and P450 2D6 phenotype of each case approximately reflected the observed values. These results suggested that our dihydrocodeine PBPK model reproduced the results of clinical cases reasonably well for subjects.