The authors describe a clinical trial of 170 patients who received clozapine over a ten year period between September 1989 and September 1999. It is a retrospective study, describing individual responses. Each patient was his own control before and with treatment. The study also compared individuals within the group of patients whose treatment was stopped and those whose treatment was continuing at the time of the study. Data was collected by analysing all patients' records and by direct enquiry of prescribers. Diagnosis was according to DSM IV criteria. Assessment included: socio-epidemiological data (sex, age, marital status and family situation, education, military and professional status, level of benefits and social support); data related to the illness (age of onset, age at first contact with a psychiatrist, diagnosis, level of hospital contact); data concerning prescriptions of drugs (indications, average dose, duration of treatment, side effects, reason for stopping and other drugs taken at the same time); 170 patients were prescribed clozapine: 96 of them were continuing to take clozapine at the time of the study while 74 patients had stopped. The characteristics of the two groups are described. They show the severity of the illnesses concerned: early onset of illness and early psychiatric care, the absence in many patients of a partner or family, their low level of employment, high dependence on social assistance. Concerning diagnostic criteria, the range of diagnoses included mostly paranoid schizophrenia, then unclassified schizophrenia then schizoaffective disorders. The indication of clozapine prescription was in the majority of the cases (87%) an inefficiency of classical neuroleptic therapy. The average dose was 401 mg per day: 388 mg for the group continuing treatment; 417 for the group which had stopped their treatment. For the patients who continued taking clozapine, the average time of treatment was just over 4 years, with a maximum of 110 months. The tolerance of clozapine was good, with 35% not suffering any side effects. Neutropenia was the commonest side effect (4.1% - a higher incidence than previously reported with one case only of agranulocytosis (0.59%). The other adverse effects were in accordance with known data: sedation affected 22.4% of patients; hypersalivation 13.5%; postural hypotension 7.6%; malocclusion 7.6%; weight gain (>5 kg) 7.1%. Treatment was stopped for side effects in 17.1% of patients; for ineffectiveness in 14.7% and 3% of patients died during treatment (their death attributed to clozapine) from seizures, intestinal obstruction or agranulocytosis. Clozapine significantly reduced the need for other associated psychotropic drugs. 25.3% of all patients were on monotherapy when on clozapine compared with 6.5% before (31.2% compared with 3.1% for those patients continuing treatment). The need for supplementary medication to reduce side effects was much less. However 22% of patients taking clozapine at the time of the study are still on an anticholinergic drug. On the basis of the analysis of 5 successive terms of treatment lasting 12 months, we have shown that for each patient: clozapine significantly reduces the length of hospitalisation compared with standard neuroleptics; it allows for out patient management and continuing integration in the community; the critical length of treatment for the group of patients studied with regard to the need for hospitalisation is 18 months. For patients whose treatment with clozapine was stopped, we noted that with the continued input from the team of carers even after clozapine was stopped, patients who had been seriously ill for long period of time continued to improve.