Olanzapine is an atypical antipsychotic effective in the treatment of schizophrenia. The present study has examined the potential use of target concentration monitoring of olanzapine in plasma as a marker of clinical response and an aid in patient management. Fifty-three patients (mean age 32 years; 40 M, 13 F) with a DSM-IVR diagnosis of schizophrenia completed a 6-week trial of oral olanzapine. Participants received once-daily olanzapine, and their psychotic symptoms were measured with the PANSS (Positive and Negative Symptom Scale) on admission and again after 6 weeks. Responders were classified as having a >/=20% decrease in PANSS scores. Plasma olanzapine was quantified by high-performance liquid chromatography. Receiver operator characteristic (ROC) curve analysis was used to identify a break point in plasma olanzapine that might serve as a surrogate for PANSS classification, and the two methods were compared using the McNemar chi2 test. After 6 weeks the median olanzapine dose was 15 mg/d (range 5-30 mg/d), and the mean plasma olanzapine was 32 micrograms/L at a mean of 13.5 hours after dose. With the PANSS (total), there were 42 responders and 11 nonresponders. ROC curve analysis for total PANSS identified a break point at 23 micrograms/L plasma olanzapine, with the proportions of responders and nonresponders identified by PANSS and the plasma break point being similar. Similar break points were found for the positive, negative, and global PANSS subscores. Nevertheless, these relationships were very modest, and at best the target plasma olanzapine concentration identified only 20% more responders than nonresponders. We suggest that plasma olanzapine monitoring can be used for dose-response optimization, but only to complement the normal clinical evaluation process.