Objective: Clozapine is a unique atypical antipsychotic with superior efficacy in treatment-resistant schizophrenia. Plasma concentration of clozapine and its major metabolite N-desmethylclozapine (NDMC) as well as the ratio of NDMC to clozapine have been reported to be predictors of clozapine response. Here we evaluate these as well as other measures in an effort to find predictors of response to clozapine in our early-onset treatment-refractory population. Method: Fifty-four children and adolescents participated in double-blind (n = 22) or open-label (n = 32) clozapine trials. Clinical evaluations took place at baseline, week 6 on clozapine, and at 2- to 6-year follow-up. The data were analyzed in relation to demographics, age at onset, IQ, clozapine dose, and plasma concentrations of prolactin, clozapine, NDMC, and NDMC/clozapine ratio. Stepwise regression and correlation analyses were performed to find predictors of treatment response. Results: Clinical improvement after 6 weeks of clozapine treatment, as measured by the percentage of improvement on the Brief Psychiatric Rating Scale and the Scale for the Assessment of Positive Symptoms, was strongly associated with the NDMC/clozapine ratio at the 6-week time point (Pearson correlation coefficient: r = 0.41; p less than 0.01 for Brief Psychiatric Rating Scale and r = 0.43; p less than 0.01 for Scale for the Assessment of Positive Symptoms). Although the rate of side effects was higher than that typically found in the adult population, it did not appear to be related to clozapine dose, clozapine or NDMC plasma concentrations, or NDMC/clozapine ratio. Outcome at long-term follow-up, as measured by Children's Global Assessment Scale, was associated with lesser illness severity at baseline and with greater improvement during the initial 6 weeks of clozapine treatment. Conclusions: The NDMC/clozapine ratio may be a valuable predictor of response to clozapine and may suggest new approaches to clozapine treatment. (Contains 5 tables.)