OBJECTIVES: To analyse short- and long-term case-fatality trends following admission to hospital with a first acute myocardial infarction, in men and women between 1986 and 1995, after adjusting for risk factors known to influence survival. DESIGN: A Scottish-wide retrospective cohort study. SETTING: The Linked Scottish Morbidity Record Database was analysed. This contains accurate data on all hospital admissions since 1981, for the Scottish population of 5.1 million. It is linked to the Registrar General's death certificate data. SUBJECTS: All 117 718 patients admitted to Scottish hospitals with a principal diagnosis of first acute myocardial infarction (ICD-9 code 410) between 1986 and 1995. MAIN OUTCOME MEASURES: The outcome was death, both in and out of hospital, from any cause, at 30 days, 1 year, 5 and 10 years. RESULTS: Overall case-fatality following hospital admission with acute myocardial infarction was 22. 2%, 31.4%, 51.1% and 64.0% at 1 month, 1 year, 5 and 10 years, respectively. Multivariate analyses identified statistically significant independent prognostic factors. Thirty day mortality increased twofold for each decade of increasing age, and increased with any prior admission to hospital. When comparing the most deprived category to that of the most affluent, men had a 10% increased mortality (P<0.01), whilst women had an increased mortality of 4% (not significant). After adjustment for age, sex, deprivation and prior admission to hospital, case-fatality rates fell significantly between 1986 and 1995. Short-term case-fatality fell by 46% in men (27% in women) and long-term by 34% in men (30% in women) (both P<0.001). CONCLUSIONS: Population-based case-fatality rates in Scotland have fallen dramatically since 1986, particularly in men. The increasing survival in patients admitted to hospital suggests that the trial-based efficacy of modern therapies is now translating into population-based effectiveness. However, an individual's life expectancy still halves after a diagnosis of acute myocardial infarction. Of the variables that we could examine, age was the most powerful predictor of prognosis.