Background: The characteristics of patients with acute myocardial infarction (MI) admitted to 37 Chilean hospitals (GEMI Registry Group), have been analyzed in the periods 1993-1995 and 1997-1998. Aim: To report the changes in hospital mortality between these 2 periods, with a particular emphasis on the impact of treatment. Patients and methods: Between 1993-1995 we collected information from 2,957 patients and between 1997-1998 we registered 1,981 patients with MI. Analysis of the changes in mortality between periods was adjusted by demographic variables, coronary risk factors, MI location, Killip class on admission and the different therapeutic strategies utilized. The effects of different treatments on hospital mortality were adjusted by the previously determined mortality risk variables. Results: Hospital mortality decreased from 13.3% to 10.8% between both periods (Odds Ratio (OR) 0.78, confidence intervals (95%) (CI) 0.65-0.93). A significant reduction in mortality was observed among patients below 60 years of age, in men, in diabetics and in subjects with an infarction classified as Killip class over II. The use of beta blockers (OR 0.65, CI 0.42-0.99) and intravenous nitrates (OR 0.78, CI 0.61-0.99) and the lower use of calcium channel blockers (OR 0.72, CI 0.60-0.87) were significantly associated with a lower mortality. The administration of angiotensin converting enzyme inhibitors was associated with a 29.3% mortality reduction (OR 0.69, CI 0.47-1.02). Conclusions: There has been a significant reduction in the mortality rate for MI in Chilean hospitals during the 2 registry periods analyzed, which was significant among some high risk patients and was related to treatment changes, according to evidence based guidelines (Rev Méd Chile 2002; 130: 368-378)