The aim of this study was to investigate possible clinical predictors of the outcome of first-episode schizophrenia. The clinical charts of the first episode of psychosis and the follow-up period over at least 4 years for 67 DSM-III-R schizophrenic patients were reviewed. According to the number of psychotic relapses observed during the follow-up period, patients were sub-divided into two groups: mono- and multi-episode patients. The main demographic and clinical variables recorded at the first episode were compared between the two groups. A logistic regression analysis was performed to test a model for the possible predictors of the two different patterns of outcome.Multi-episode patients had an earlier onset of the illness and a longer "duration of untreated psychosis" (DUP), defined as the interval between the onset of the first psychotic symptoms and the first antipsychotic treatment. The Brief Psychiatric Rating Scale (BPRS) total scores were lower and the "avolition/apathy" scores of the scale for the assessment of negative symptoms (SANS) were higher in multi-episode patients. The logistic regression analysis results confirmed the DUP and the pre-treatment BPRS scores to be significant predictors of the outcome. These findings confirm that the timing in recognizing and treating the early symptoms of schizophrenia, even when subtle, is a core issue for the clinical management of the disorder.