This prospective, double-blind, parallel-group study evaluated ziprasidone in the prevention of psychotic relapse and in the long-term treatment of the negative symptoms of schizophrenia. Inpatients with chronic schizophrenia ceased neuroleptic treatment, after written informed consent was given, and were randomized to either ziprasidone 40 mg/day (n = 76), ziprasidone 80 mg/day (n = 72), ziprasidone 160 mg/day (n = 71) or placebo (n = 75) for 1 year. Assessments included the Positive and Negative Symptom Scale (PANSS) and the Global Assessment of Functioning (GAF). Relapse was defined as Clinical Global Impression (Improvement) score of ≥6 and/or a score ≥6 on PANSS items P7 (hostility) or G8 (uncooperativeness) on 2 successive days. Patients who relapsed were immediately discontinued and received appropriate effective treatment. Group mean ages and mean durations of current hospitalization were 48.8-50.8 years and 62.1-74.6 months, respectively. Group mean baseline PANSS scores were 85.1-88.9 and mean baseline PANSS-derived depression factor was low (11.6-12.5). Group mean baseline GAF scores were 47.3-48.4. Survival analysis demonstrated that ziprasidone significantly reduced the risk of psychotic relapse. The probability of relapse at 1 year was significantly lower in the ziprasidone 40, 80 and 160 mg/day groups (40.5%, 34.6% and 35.8% respectively) compared with placebo (70.8%; P = 0.02, = 0.001 and <0.001, respectively). The intention-to-treat (ITT) analysis of the PANSS negative symptom subscale showed statistically significant improvement with ziprasidone compared with placebo (P ≤ 0.05) from Week 16 (third assessment on treatment) with further improvement observed through to 1 year. A similar pattern of improvement was observed in the ITT analysis of those patients who had predominantly negative symptoms at baseline who were treated with ziprasidone. The ITT analysis of GAF scores also showed a statistically significant difference between groups in favor of ziprasidone (P < 0.001) at 1 year. The tolerability of ziprasidone was excellent. Both the overall incidence of adverse events and discontinuations due to adverse events were lower in all three ziprasidone groups compared with the placebo group. Weight gain, EPS and orthostatic hypotension were rarely associated with ziprasidone. Mean Simpson-Angus, Abnormal Involuntary Movements and Barnes Akathisia scores improved in all groups over the course of the study and were similar to placebo in all ziprasidone groups. Concomitant anticholinergic use and beta-blocker use was very low with ziprasidone and placebo. Mean body weight decreased slightly in all four groups over the course of the study. No pattern of laboratory abnormalities emerged with ziprasidone and the frequency of transaminase elevations was similar to placebo. Median prolactin levels decreased from baseline in all groups. This study demonstrates that ziprasidone is effective and very well tolerated in the prevention of relapse and in the long-term treatment of negative symptoms of schizophrenia.