AIMS: This study analyses the association between cardiovascular diseases with and without drug treatment, self-rated health, and all-cause mortality during a 12-year period. METHODS: Genetic and familial environmental effects were controlled for in subsets of the analyses. Data for these analyses were collected in 1984, 1987, 1990, and 1993 as part of the Swedish Adoption/Twin Study of Aging (SATSA), and through record linkage to mortality data from the National Cause of Death Register through 1996. RESULTS: In multivariate analyses, both cardiovascular disease and low self-rated health were associated with higher mortality. Age, sex, lifestyle, diabetes, respiratory problems, cancer, depression, marital status, and social network were controlled for in the analyses. Furthermore, development of cardiovascular disease and/or drug use is associated with a decline in perceived health and higher mortality rates. The co-twin control analyses suggest that drug-treated cardiovascular disease has a marginal effect on survival beyond the effects of bad self-rated health and genetic liability to a certain survival time. CONCLUSIONS: This study shows that subjective health ratings are important predictors of mortality for persons with cardiovascular disease.