We investigated the impact of the genotype of CYP2D6 on the hydroxylation of desipramine in eighteen patients who were administered desipramine hydrochloride per os. Significantly higher plasma concentration of desipramine/daily dose of desipramine/body weight was observed in the subjects with two mutated alleles than in the subjects with either no mutated alleles or one mutated allele (two mutated alleles versus no mutated alleles=530.4+/-215.2 versus 118.1+/-63.9 ng/ml/mg/kg, t=5.68, P<0.001; two mutated alleles versus one mutated allele=530.4+/-215.2 versus 176.2+/-62.3 ng/ml/mg/kg, P<0.001; One-way analysis of variance followed by Bonferroni's multiple comparison test, respectively). Significantly higher ratio of desipramine/2-hydroxy-desipramine was observed in the subjects with two mutated alleles compared to subjects with no mutated alleles or the subjects with one mutated allele (two mutated alleles versus one mutated allele=4.39+/-0.36 versus 2.00+/-0.64, t=5.12, P<0.001; two mutated alleles versus no mutated alleles=4.39+/-0.36 versus 2.02+/-0.59, t=4.42, P<0.01). The genotyping of CYP2D6 only grossly predicts the steady state concentration of desipramine, mainly predicts the risk of getting very high plasma levels. Within each genotype there is marked interindividual variability.