Therapeutic drug monitoring data for amitriptyline (AT) and nortriptyline (NT) collected during 10 years (total of 4,278 analyses in 2,937 patients) were evaluated to study how other drugs affect the kinetics at steady state. The distribution of the ratio concentration/daily dose (C/D) in patients treated with the antidepressant only was compared with that in patients on different concomitant drugs. Patients on phenothiazines or dextro-propoxyphene had a significantly higher mean CID of NT than controls, both when AT and when NT had been given. The highest values were seen with levomepromazine and thioridazine. On the contrary, the mean CID of both AT and NT in patients on carbamazepine was about 50% lower than in those treated with the antidepressant only. Benzodiazepines did not affect the steady-state kinetics of AT or NT. Intraindividual comparisons of the ratio CID in subjects with analyses performed when off and on concomitant drugs corroborate previous results showing that drugs metabolized by the debrisoquine hydroxylase (CYP2D6) inhibit the metabolism of NT and that carbamazepine induces the metabolism of both AT and NT. Modeling of the dose dependency of the NT interactions with levomepromazine, perphenazine, and thioridazine revealed that the ratio CID was most affected at low doses of the antidepressant and at high doses of the phenothiazine. The distribution of the doses given was the same in patients on monotherapy as in patients with interacting drugs, which means that many patients treated with phenothiazines had concentrations above the therapeutic range and that most patients treated with carbamazepine had subtherapeutic levels. The present study shows that therapeutic drug monitoring may serve as a valuable tool to discover and quantify drug interactions.