Schizophrenic outpatients (n=102) whose condition had stabilized with clozapine (CLZ) therapy and were being maintained on CLZ were followed for 1 year. Clinical status and concentrations of serum clozapine (CLZ) and its metabolite norclozapine (NCLZ) were evaluated periodically or when relapse occurred. Relapse was defined as a significant exacerbation of psychotic symptoms or hospitalization. Thirty-three patients relapsed and 69 did not. Relapse patients displayed significantly lower serum concentrations of CLZ and a sum of CLZ and NCLZ at endpoint than non-relapses (CLZ: 162 ng/ml vs. 237 ng/ml, p<0.001; CLZ+NCLZ: 225 ng/ml vs. 301 ng/ml, p<0.001). When all subjects were pooled together, a significant inverse correlation was observed between percent increase in the total score on the Brief Psychiatric Rating Scale (BPRS) from baseline and serum levels of CLZ alone (r=0.404, p<0.001) and the sum of CLZ and NCLZ (r=0.364, p<0.001). Relapses and non-relapses were well separated by a threshold CLZ serum concentration of 200 ng/ml with a sensitivity of 73% and a specificity of 80%. The threshold value represented about a 40% lower serum CLZ level than concentration achieved in acute treatment. Survival analysis showed a similarity of the relapse risk over time defined by the CLZ serum threshold and by symptomatic criteria. These results suggest that effective relapse prevention may require maintenance of patients at CLZ serum concentrations above 200 ng/ml and above 60% of the acute-phase level during long-term maintenance treatment of schizophrenia.