The relationship between genetic polymorphism and imipramine metabolism has never been studied in a family study. A sparteine/mephenytoin test was carried out in 31 parents and 20 siblings of 18 Danish poor metabolizers of sparteine (PMs). One week 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 HPLC. There were 10 PMs (20%; 9.8-33%, 95% confidence interval) and 41 extensive metabolizers of sparteine (EMs) among parents and siblings. In 26 of the 28 PMs among probands and relatives, there were concordance between phenotype and genotype: D6-A/D6-D (n = 2), D6-A/D6-B (n = 5), D6-B/(n = 15) or D6-B/D6-D (n = 4). Two PMs were apparently heterozygous (EMs), D6-wt/D6-B. Accordingly, based on the present sample of 28 PMs the specificity of the genotype test was 100% and the sensitivity was 92.9%. Two EMs were homozygous dominant D6-wt/and 39 were heterozygous EMs; D6-wt/D6-D (n = 5), D6-wt/D6-B (n = 27), D6-wt/D6-A (n = 6), D6-wt/D6-wt* (unknown mutation) (n = 1). As previously reported in a population study the hydroxylation ratios (i.e. 2-hydroxymetabolite over parent compound) of imipramine were much lower in PMs than in EMs. This and the pedigrees confirmed the co-segregation of sparteine oxidation, imipramine 2-hydroxylation and the CYP2D6 genotype. None of the hydroxylation ratios could separate EMs and PMs completely, mainly because the 2-hydroxylation of imipramine also depends on P450s other than CYP2D6.