OBJECTIVE To analyse the changes in mortality in dilated cardiomyopathy over the past 15 years and to identify the factors that might have influenced survival. METHODS Follow up study of 235 patients (aged 16-70) systematically enrolled on a register from 1 January 1978 to 31 December 1992. METHODS Hospital department of cardiology. METHODS Three groups corresponding to three periods of 5 years: group 1 (diagnosis between 1 January 1978 and 31 December 1982) 26 patients; group 2 (diagnosis between 1 January 1983 and 31 December 1987) 65 patients; and group 3 (diagnosis between 1 January 1988 and 31 December 1992) 144 patients. METHODS Death or heart transplantation. RESULTS Two and four year survival was 73.8% and 53.8% in group 1, 87.7% and 72.3% in group 2, and 90.3% and 82.9% in group 3 (P = 0.02). During the 15 years of the study period the number of cases increased progressively and the baseline clinical characteristics changed (that is, patients were younger and less severely affected), partly explaining the improvement in survival. None the less, the three mortality curves tended to diverge progressively and the improvement in survival in the different groups was still significant after stratification for the severity of the disease, suggesting that treatment had a sustained effect. A progressively higher proportion of patients were treated with angiotensin converting enzyme (ACE) inhibitors and more recently with beta blockers. In group 2, after stratification for the severity of heart failure, patients who were treated with ACE inhibitors showed a better survival than patients who were not. Furthermore, analysis of group 3 showed that beta blockers had a significant additive effect with conventional therapy both by intention to treat and actual treatment. Four year survival in patients with mild and moderate to severe heart failure treated with beta blockers, and usually digitalis and ACE inhibitors, was respectively 90% and 87.5%. CONCLUSIONS The improvement in the survival of patients with dilated cardiomyopathy over the past 15 years may be explained by earlier diagnosis, new treatments, and a change in the clinical characteristics of the patients at enrolment.