The steady-state plasma concentrations of imipramine and desipramine were measured after a more than 2-week treatment with 0.39 to 1.39 mg/kg/day of imipramine hydrochloride in 28 Japanese patients with major depression who had been phenotyped simultaneously with mephenytoin (for CYP2C19-related status) and with metoprolol (for CYP2D6-related status) before initiating the antidepressant therapy. Patients consisted of five poor metabolizers (PMs) of CYP2C19 with an extensive metabolizer (EM) phenotype of CYP2D6, whereas the remainder were EMs for both of the phenotypes. The mean respective concentrations (corrected by mg/kg) of imipramine and the sum of imipramine plus desipramine were 2.4 and 1.8 times greater in the CYP2C19-related PM than in the EM group, and these two variables correlated with the log10 urinary excretion of 4'-hydroxymephenytoin (rs = -0.73 and -0.64, both p < 0.01, respectively), but not with the metabolic ratio (MR) of metoprolol/alpha-hydroxymetoprolol. The mean N-demethylation index (MR of desipramine/imipramine) was significantly (p < 0.01) less in the PM than in the EM group. This index correlated with the 4'-hydroxylation of S-mephenytoin (rs = -0.51, p < 0.01), but not with the alpha-hydroxylation of metoprolol, implying that imipramine N-demethylation is under a coregulatory pharmacogenetic control of CYP2C19, but not of CYP2D6. In conclusion, by taking into account that the incidence of the PMs of CYP2C19 is much greater (18-23%) than that of CYP2D6 (< 1%) in Japanese population, the individually predetermined assessment of the CYP2C19-mediated metabolic capacity of imipramine would be more valuable than that of the CYP2D6-mediated capacity for forecasting the steady-state concentrations of imipramine and desipramine in Japanese depressive patients, thereby attaining an individualized optimization of imipramine therapy. Obviously, a pharmacodynamic assessment study conducted simultaneously with predetermined CYP2C19 status is required for supporting this contention.