The veteran and the rookie.

Clinical registries have systematically shown that routine clinical practice deals with sicker patients and faces higher morbidity and mortality than the artificial environment of strictly defined randomized-controlled trials might otherwise suggest. The series of Terkelsen et al., in the present issue of the European Heart Journal, focuses on a cohort of 654 patients, representing all patients admitted to hospital for acute myocardial infarction (AMI) from a catchments area of 139 000 inhabitants in Denmark. Their data are of particular interest because (i) they constitute an attempt to describe ‘real life’ in a comprehensive manner by targeting and checking all hospital admissions, rather than relying on voluntary reports from selected participating centres; and (ii) because angiographic and survival status were obtained up to 1 year. Their analysis confirms the persistent high, in-hospital mortality of AMI in the general population (14%), and also reveals three main surprising facts: the median age of admitted patients was advanced (73 years), the incidence of non-ST-elevation myocardial infarction (NSTEMI) was high (54% of all those diagnosed as AMI) and the prognosis of NSTEMI (31% 1 year mortality) was worse than that of STEMI (21% 1 year mortality). The median age is clearly an important parameter to keep in mind when assessing the significance of a 14% early mortality. In AMIS, GRACE and the EHS, the mean age was under 70 in all three registries, and hospital mortality varied between 2.4 and 11.8% depending on the subgroups considered. Older patients are often excluded from randomized-controlled trials, and they may also be more likely not to be entered into open registries because of late arrival at hospital, death in the emergency room or failure to be admitted to either the ICU or the catheterization laboratory. The present series reminds us that the elderly constitute a large— and growing—subset of patients admitted to hospital for AMI. Some of the other multivariate predictors of hospital mortality also correspond to typical exclusion criteria used in randomized trials: excessive delay in reaching the hospital and azotaemia are good examples. Thus, for registry data to be adequately assessed, investigators need to remember that acquisition of sufficient information concerning essential cardiac and non-cardiac parameters is of major importance: the very factors that exclude patients from randomized trials are often those with the greatest impact on clinical outcome. Both acute management and discharge treatment reported in the series by Terkelsen et al. are probably a good reflection of the average European approach. Only 48% of patients with NSTEMI benefited from an early invasive strategy, however, and only 35–51% of patients were discharged with lipid-lowering therapy. Both of these can be seen as less than optimal in view of current knowledge and may have also contributed to the high observed mortality.