AN ECHOCARDIOGRAPHIC METHOD FOR SELECTING HIGH-RISK PATIENTS SHORTLY AFTER ACUTE MYOCARDIAL-INFARCTION, FOR INCLUSION IN MULTICENTER STUDIES (AS USED IN THE TRACE STUDY)

The aim of our study was to examine if echocardiography can reproducib ly be used in a multicentre study to select high risk patients with re duced left ventricular function early after an acute myocardial infarc tion (MI). In the TRAndolapril Cardiac Evaluation Study (TRACE) patien ts with reduced left ventricular systolic function were randomized 3-7 days post MI to receive either the ACE inhibitor trandolapril, or pla cebo. Twenty-seven Danish centres participated and 7001 consecutive MI patients were screened for entry. Local doctors and technicians who h ad received a brief but thorough training course recorded a two-dimens ional echocardiographic examination on videotape 2-6 days after MI. Wi thin 24 h, wall motion index (WMI) was visually assessed by one of two cardiologists (examiners) with considerable experience in echocardiog raphy. A WMI of less-than-or-equal-to 1.2 (corresponding to a left ven tricular ejection fraction (LVEF) less-than-or-equal-to0.35) meant tha t the patient was eligible for randomization in the TRACE study. Two o ther experienced cardiologists with substantial experience in echocard iography (controllers) performed blind reassessment of 155 randomly ch osen videotapes. We showed that 93% of the 7001 screened MIs had an as sessable echocardiogram. WMI was less-than-or-equal-to1.2 in 37% of pa tients. The one-year mortality was inversely related to WMI, being 60% , 30%, 14% and 11% in patients with a WMI<0.8, 0.8-1.2, 1.3-1.6 and >1 .6, respectively. In the random sample of 155 videorecordings that wer e reevaluated, 97% were found to be technically adequate for analysis both by the examiners and the controllers. Comparing the examiners wit h the controllers, the reproducibility analysis showed 95% confidence limits for a single estimate of LVEF of +/-0.13. Comparison between th e two examiners showed corresponding confidence limits of +/-0.10. Usi ng WMI of 1.2 (LVEF approximately 0.35) as a discriminative value the concordance between examiners and controllers was 80%. Thus, evaluatio n by experienced cardiologists of videotaped echocardiographic examina tions recorded by briefly but thoroughly trained investigators appears to be a reliable and reproducible method for the selection of high ri sk patients shortly after MI in multicentre studies.