Surveillance of Sudden Death After Myocardial Infarction

To the Editor: We have some concerns about the design of the surveillance study of sudden death after myocardial infarction (MI) by Dr Adabag and colleagues.¹ First, the Mayo Clinic provides half of the primary care and nearly all specialty care for the community.² A high proportion of the working population was employed in the health care industry. In addition to race and ethnicity, local diagnosis and treatment characteristics and the degree of awareness of disease may influence the outcome and prognosis of MI. These characteristics of this study population were not representative of a general population and therefore may not be valid for accurate comparisons to be made across geographic regions.

Second, from 1979 to 2005 diagnosis and understanding of MI evolved enormously. The target International Classification of Diseases, Ninth Revision (ICD-9) codes should have included 402 (hypertensive heart disease), 427 (cardiac dysrhythmia), 428 (heart failure), 518.4 (acute edema of the lung), 429.2 (cardiovascular disease, unspecified), and 440 (atherosclerosis), which have also been used in other studies.^{3 - 4} This would have allowed a full assessment of MI and minimized the difference in disease diagnosis between 1979 and 2005.

Third, heart failure that occurred after hospitalization for incident MI but before discharge was included as an intercurrent event, but recurrent ischemia was defined only as hospitalization for recurrent MI or unstable angina using the physicians' diagnoses. The difference will tend to augment an association with heart failure and undervalue recurrent ischemia. Before adjusting for baseline characteristics, recurrent ischemia was associated with an increased risk of sudden cardiac death (hazard ratio, 1.31; 95% confidence interval, 1.01-1.71; P = .04); after adjustment the association became nonsignificant (hazard ratio, 1.26; 95% confidence interval, 0.96-1.65; P = .09). The lower limits of these 2 confidence intervals are very close to 1 (indicating no association), so a little change in classification of recurrent ischemia might change the results. With regard to sudden cardiac death, Jokhadar et al⁵ reported that its declining trend parallels that of recurrent ischemic events.

Fourth, sudden cardiac death was defined as out-of-hospital deaths. But in some studies^{3 - 4} it includes in-hospital (during rehospitalization) events with rapid fatal course, for which no or few data on history, enzymes, or electrocardiogram were available.