The effect of recurrent events on register-based estimates of level and trends in incidence of acute myocardial infarction.

Although changes in incidence of first acute myocardial infarctions (AMI) are of primary interest in the evaluation of preventive efforts, only few studies have used this measure. In the present study, the risk of recurrence over time in subjects with first AMI is analyzed, and the effect of inclusion of recurrent cases on the estimation of level and trend in the incidence of AMI is evaluated. The National Patient Register of Hospital Discharges and the Causes-of-Death Register were linked, and all cases of admission for AMI and fatal manifestation of the disease since 1977 and until 1992 in the Danish population were identified. New events occurred during the following 12 years in 46% of men and 42% of women with their first AMI in 1980. Ninety percent of the recurrent events occurred during the first 5 years. Using absence of events during only 1 preceding year as the inclusion criteria, the incidence rate would be overestimated by about 20%-30%. However, if the preceding event-free period was of the same duration throughout the study period, the trends in AMI rates were not altered by expanding the event-free period up to 14 years before the index event. Although rates of AMI based on the total number of affected persons without AMI in the previous year overestimate the true incidence by 20%-30%, trends in these rates reflect trends in rates of first events with reasonable accuracy.