OBJECTIVES: Temporal trends in the incidence of ST-elevation myocardial infarction (STEMI) have been declining in many countries, while the incidence of non-ST elevation myocardial infarction (NSTEMI) has reached a plateau or even increased. The reasons for these changing trends have yet to be explained. We analysed these trends and short-term mortality from acute coronary syndromes in a nationwide cohort study over 35 years in Iceland. DESIGN: Retrospective cohort study using a national MI registry. SETTING: Iceland. PARTICIPANTS: All cases of myocardial infarction in individuals aged 25-74 years in Iceland 1981-2015. METHODS: Each case was classified as STEMI, NSTEMI or no ECG taken. ECG recordings were classified according to Minnesota criteria. OUTCOME MEASURES: Trends of STEMI and NSTEMI incidence and 1-day and 28-day mortality were obtained from the National Death Registry. RESULTS: A total of 10 348 cases were identified (mean age 61 years, 76.4% male). These were categorised as STEMI (32.7%), NSTEMI (45.8%) and no ECG taken (21.5%). We detected a significant 3.7% annual decline in the incidence of first MI. The age-adjusted incidence of STEMI showed an 83.2% decline, most pronounced after 1994, while for NSTEMI the decline was 66.5%, reaching a plateau from the year 1989 onwards. In Iceland, the uptake of highly sensitive biomarkers was initiated in 1997 (cardiac troponin T) and 2012 (high-sensitive troponin T), respectively. CONCLUSIONS: The different temporal trends in the incidence of STEMI and NSTEMI cannot be explained only by the uptake of highly sensitive biomarkers in 1997 and 2012. The change in population-level risk factor exposure is likely to have influenced atherosclerotic plaque burden and thrombotic mechanisms. Finally, increasing uptake of cardioprotective pharmacological and interventional therapy may have resulted in a primary preventive effect on plaque rupture and thrombosis and thus on the rates of STEMI and NSTEMI disproportionally.