After an initial placebo week, 37 depressed in-patients were treated with the fixed dose of 75 mg clomipramine b. d. A sparteine test was carried out during the placebo period and again during the second week of active therapy. Blood for drug assay was collected at the end of the inter-dose interval in the (morning) at weekly intervals. Clomipramine and four metabolites (desmethylclomipramine, didesmethylclomipramine, 8-hydroxyclomipramine, and 8-hydroxydesmethylclomipramine) in plasma were assayed by reversed phase HPLC. The clomipramine and desmethylclomipramine steady-state plasma levels varied by factors of 11 and 9, respectively, and the clomipramine/8-hydroxyclomipramine and desmethylclomipramine/8-hydroxydesmethylclomipramine ratios both varied by 7-fold. During the placebo week, 36 patients were phenotyped as extensive metabolizers (EM) (metabolic ratio, MR, 0.1-2.0), and one patient was phenotyped as a poor metabolizer (PM) (MR > 300). During clomipramine treatment, one patient changed phenotype from EM to PM (MR = 140). In the EM, the median of the MR increased from 0.4 to 2.3. There was a statistically significant correlation between the MR before and during clomipramine treatment, even when the PM was excluded. Neither the steady-state plasma clomipramine levels nor the clomipramine/desmethylclomipramine ratios showed a significant correlation with the MR. In contrast, the desmethylclomipramine and didesmethylclomipramine steady-state levels and the desmethylclomIpramine/8-hydroxydesmethylclomipramine and clomipramine/8-hydroxyclomipramine ratios showed a significant positive correlation with the MR. The PM had the highest steady-state plasma desmethylclomipramine level and the highest desmethylclomipramine/8-hydroxydesmethylclomipramine ratio. These correlation coefficients (rs) were generally increased when the correlation analyses were based on the MR obtained during clomipramine treatment. The results suggest that the 8-hydroxylation of clomipramine and of desmethylclomipramine are catalyzed by the same isozyme that oxidises sparteine, CYP2D6. The N-demethylation of clomipramine appears to be less clearly related to the activity of CYP2D6. Clomipramine appeared to cause more potent inhibition of sparteine oxidation than that seen previously with other tricyclic antidepressants.