Olanzapine (Zyprexa) was approved for general prescription in Sweden in November 1996, and an HPLC-based therapeutic drug monitoring (TDM) routine for serum olanzapine (OLA) and its major metabolite, N-demethylolanzapine (DMO) was established in February 1997. During 1997 to 1999, a total of 753 TDM requests for a total of 545 Swedish patients was analyzed. Additional patient information on certain clinical variables was collected on specifically designed TDM request forms. After the exclusion process, samples from 194 patients were found to be eligible for further scrutiny. The concentration-to-dose (C/D) ratio for OLA varied 25-fold and that of DMO 22-fold. Women had a higher (P < 0.01) median C/D ratio for OLA than men (median, 7.2 nmol/L/mg vs 5.2 nmol/L/mg). Nonsmokers had a higher (P < 0.001) C/D ratio for OLA than smokers (median, 9.2 nmol/L/mg vs 4.0 nmol/L/mg). Smokers got higher prescribed (P < 0.05) doses of OLA than nonsmokers did. In the group with reported side effects, the median serum OLA concentration was 22% higher (P < 0.05) than in the group without side effects. Patients co-medicated with carbamazepine had a 71% lower median C/D ratio for OLA than patients on OLA monotherapy. The present TDM-based follow-up suggests that the influence of gender, smoking habits, and certain drug interactions may need to be considered for optimal dosage of OLA. TDM may be used for this purpose more readily in the future.