1. Sparteine and mephenytoin phenotyping tests were carried out in 327 healthy Danish subjects. Two weeks later each subject took 25 mg imipramine followed by urine collection for 24 h. The urinary content of imipramine, desipramine, 2-hydroxy-imipramine and 2-hydroxy-desipramine was assayed by h.p.l.c. 2. The medians of the hydroxylation ratios (i.e. 2-hydroxy-metabolite over parent compound) were 6 to 14 times higher in 300 extensive metabolizers of sparteine (EMs) as compared with 27 poor metabolizers (PMs), but none of the ratios separated the two phenotypes completely. 3. There were 324 EM of mephenytoin (EMM) and three PM (PMM) in the sample. The demethylation ratios between desipramine, 2-hydroxy-desipramine and their corresponding tertiary amines showed statistically significant correlations with the mephenytoin S/R isomer ratio (Spearman's rs: -0.20 and -0.27, P < 0.05). 4. The demethylation ratios were higher in 80 smokers than in 245 non-smokers. This indicates that CYP1A2, which is induced by cigarette smoking, also catalyzes the N-demethylation of imipramine. 5. CYP2D6 genotyping was carried out by PCR in 325 of the subjects, and the D6-wt allele was amplified in 298 EMs, meaning that they were genotyped correctly. One PMs was D6-wt/D6-B, another PMs had the genotype D6-wt/ and hence both were misclassified as EMs. The remaining 25 PMs were D6-A/D6-B (n = 5), D6-B/ (n = 18) or D6-D/D6-D (no PCR amplification, n = 2).(ABSTRACT TRUNCATED AT 250 WORDS)