Previous reports in referral populations have emphasized the poor prognosis of dilated cardiomyopathy. This study evaluated mortality and morbidity in patients presenting at a referral center between 1989 and 1993. One hundred seventy-two consecutive patients were studied. At presentation, 82 were in New York Heart Association functional class III/IV. Mean (+/- SD) left ventricular end-diastolic dimension was 69 +/- 11 mm, ejection fraction was 25 +/- 10%, VO2 max was 21 +/- 9 mL/min/kg, and sodium was 136 +/- 9 mM. Treatments included vasodilators (n = 157, 92%), anticoagulation (n = 50, 29%), amiodarone (n = 52, 30%), and cardiac defibrillator (n = 5, 3%). During the follow-up period (mean, 26 +/- 29 months), 16 patients died and 60 developed progressive heart failure; 46 (27%) required cardiac transplantation. The majority of the patients (102, 59%) were stable or improved. Established prognostic determinants (left ventricular end-diastolic dimension, ejection fraction, sodium, and arrhythmia) were of low predictive value for the development of progressive heart failure or sudden death. The 1- and 2-year probabilities of death or transplantation was 16 and 21%, respectively (death only 6 and 7%, respectively). These observations are subject to referral bias, but suggest that the majority of patients can remain stable. Any improvement in survival compared to earlier experience can be due to earlier diagnosis, availability of transplantation, and new heart failure management strategies.