This study investigated whether the smokinginducible cytochrome P450 (CYP) 1A2 and the polymorphic CYP2D6 play significant roles in the metabolism of olanzapine and its clinical effects at steady-state treatment. Caffeine and debrisoquine were used as measures of CYP1A2 and CYP2D6, respectively. After drug therapy for 15 days, the effect of olanzapine on the activities of CYP1A2 and CYP2D6 was also examined. Seventeen psychiatric patients (9 men and 8 women) were orally administered olanzapine, at a mean +/- standard deviation (SD) dosage of 10 mg/d for all smokers (n = 8) and 7.5 +/- 2.5 mg/d (range, 5-10 mg) for nonsmokers (n = 9;p <0.01). The plasma concentration-to-dose (C:D) ratio was closely correlated to the CYP1A2 activity ( s = -0.89;p <0.0001). The mean urinary caffeine indexes of nonsmokers and smokers were 17 +/- 8 and 101 +/- 44, respectively, indicating that smoking had induced a sixfold higher CYP1A2 activity (p <0.0001). Likewise, the olanzapine plasma C:D ratio (ng.mL.mg) was about fivefold lower in smokers (7.9 +/- 2.6) than in nonsmokers (1.56 +/- 1.1;p <0.0001). On day 15 of the antipsychotic therapy, the percentage decrease in Brief Psychiatric Rating Scale (BPRS) total score relative to the predosing score (in the drug-free period) was higher for nonsmokers than for smokers (30.4 +/- 10% vs. 12.5 +/- 14%;p <0.01). Six nonsmokers and three smokers experienced side effects with olanzapine. After 15 days of drug treatment, olanzapine had caused significant (p <0.0001) and substantial CYP1A2 inhibition (by 50%) in comparison with predosing values, and such inhibition can contribute to adverse drug interactions. In conclusion, smoking-induced increased CYP1A2 activity significantly diminished plasma olanzapine concentrations and the antipsychotic effect of the drug. The performance of a simple caffeine test may assist in individualization of the olanzapine dosage.