Pathophysiology and Natural History Coronary Artery Disease

We examined the relationships among ventricular arrhythmias, left ventricular dysfunction, and mortality after the occurrence of myocardial infarction in 766 patients who enrolled in a nine-hospital study and underwent two special tests. Frequency and repetitiveness of ventricular premature depolarizations (VPDs) were determined by computer analysis of predischarge 24 hr electrocardiographic recordings. The left ventricular ejection fraction (LVEF) was determined by radionuclide ventriculography and dichotomized at its optimal value of 30%. Frequency of VPDs was divided into three categories: (1) less than one per hour, (2) one to 2.9 per hour, and (3) three or more per hour. Repetitiveness of VPDs was also divided into three categories: (1) no repetitive VPDs, (2) paired VPDs, and (3) VPD runs. These variables were related, one at a time and jointly, to total mortality and to deaths caused by arrhythmias. The hazard ratios for dying in the higher or highest risk stratum vs the lower or lowest stratum for each variable (adjusted for the effects of the others) were: LVEF below 30%, 3.5; VPD runs, 1.9; and VPD frequency of three or more per hour, 2.0. There were no significant interactions among the three variables with respect to effects on the risk of mortality. There was a suggestion of an interaction between each risk variable and time after infarction. LVEF below 30% was a better predictor of early mortality (less than 6 months) and the presence of ventricular arrhythmias was a better predictor of late mortality (after 6 months). The results of this large multicenter study support the conclusion that ventricular arrhythmias and left ventricular dysfunction are independently related to mortality risk. Circulation 69, No. 2, 250-258, 1984. PREVIOUS STUDIES have suggested that ventricular arrhythmias occurring in patients who have suffered myocardial infarction are associated with increased mortality in 1 to 4 years of follow-up. 1-9 This association has been found in nearly every study, although the strength of the association has varied considerably. From the Division of Cardiology, Department of Medicine, and Division of Biostatistics, School of Public Health, Columbia University, New York; and the Departments of Medicine and Biostatistics, Washington University School of Medicine, St. Louis. Supported in part by NIH grants HL-22982 and HL-70204 from the National Heart, Lung and Blood Institute; by grant RR-00645 from the Research Resources Administration, Bethesda; and by funds from the Gebbie Foundation, Jamestown, NY; Merck, Sharp and Dohme Research Laboratories, West Point, PA; and the Winthrop and Chemow Foundations, New York, NY. Address for correspondence: J. T. Bigger, M.D., Arrhythmia Control Unit, Columbia-Presbyterian Medical Center, 630 West 168th St., New York, NY 10032. Received April 4, 1983; revision accepted Sept. 19, 1983. *All editorial decisions for this article, including selection of reviewers and the final deposition were made by a guest editor. This procedure applies to all manuscripts with authors from the Washington University School of Medicine. **Committee members and enrolling hospitals appear in an appendix to this article. High frequency and various "complex" forms of ventricular arrhythmias have been proposed as the best predictors of subsequent mortality.'-' Of the complex forms, pairs and runs of three or more consecutive ventricular premature depolarizations (VPDs) have the strongest association with subsequent mortality.8 9 The relationship between ventricular arrhythmias and left ventricular dysfunction after myocardial infarction is a controversial issue. Two large studies have shown an independent effect of ventricular arrhythmias on mortality after adjustment is made for left ventricular dysfunction.3' 4 However, both studies used a short electrocardiogram (ECG) and clinical heart failure as the indicators of left ventricular dysfunction. Other workers have concluded that ventricular arrhythmias are so strongly associated with left ventricular dysfunction that this association accounts for the relationship between ventricular arrhythmias and death. '° The initial report from the Multicenter Post-Infarction Program presented our primary a priori analysis. 1 1 An average VPD frequency of greater than 10 per hour CIRCULATION 250 by gest on July 6, 2017 http://ciajournals.org/ D ow nladed from PATHOPHYSIOLOGY AND NATURAL HISTORY-CORONARY ARTERY DISEASE was used as the index of arrhythmia and ejection fraction less than 40% (determined by radionuclide ventriculography) was used as the index of left ventricular dysfunction. The a priori analysis showed a significant association between frequency of VPDs and mortality when adjustment was made for left ventricular dysfunction. Since our primary analysis dichotomized the variables at values selected before we examined the data, the relationships among ventricular arrhythmias, left ventricular dysfunction, and mortality were not fully explored. The purpose of this study is to define the interrelationships among frequency of VPDs, repetitiveness of VPDs, left ventricular dysfunction, and mortality in a large multicenter study. This information should help to establish strategies for detection of high-risk patients and to determine the feasibility of selected drug treatment (e.g., antiarrhythmic drugs) for secondary prevention. Methods Patients. Patients were selected from nine participating hospitals located in New York City, Rochester, NY, St. Louis, and Tucson. The diagnosis of a definite acute myocardial infarction required the presence of two of the following: (1) a clinical history of central chest pressure, pain, or tightness lasting 30 min or more, (2) ECG Q waves that fulfilled one of the major criteria for abnormal Q waves of the Minnesota ECG code2 with evolutionary ST and T wave changes on serial tracings, or (3) MB isoenzyme fraction greater than 4% of the total creatine kinase, or elevation of creatine kinase or aspartine transaminase levels greater than the upper limit of normal in the hospital laboratory for 2 days or more during the coronary care unit phase without other clinical reasons for the enzyme elevation. During the 2 year recruitment period (January 1, 1979, to December 31, 1980), 4090 patients were screened and 1417 patients were eligible (i.e., had myocardial infarction and were under 70 years of age). Eligible patients were excluded from enrollment if they had life-threatening comorbidity or if they lived too far away for follow-up. Of the 1417 eligible patients, 866 lacked exclusion criteria and were enrolled after they signed written informed consent statements. Follow-up. Enrolled patients had follow-up contact either by a clinical visit or by phone at 3, 6, and 12 months after infarction and then annually. The study design called for a common termination date (December 31, 1981). Therefore follow-up ranged from 12 to 36 months; the average duration of follow-up was 22 months. Thirty-six patients (4.2%) were lost to follow-up during the study: 13 in the first year, 19 in the second year, and four in the third year. Data acquisition. The 452 clinical variables collected for each patient included demographics, medical history, course of treatment in the coronary care unit, 12 lead ECG findings, left ventricular ejection fraction (LVEF), 24 hr ECG recording, a low-level predischarge treadmill exercise test,13 medical/cardiac status at discharge, and follow-up data on rehospitalization and mortality events through December 31, 1981. Missing values in the routine clinical data were in the range of 0.5% or less. For the special tests analyzed in this study, 47 (5%) did not undergo a 24 hr ECG recording and 56 (6%) did not undergo measurement of LVEF for a variety of logistical and technical reasons. Thus, of the 866 patients enrolled, 819 had a 24 hr Vol. 69, No. 2, February 1984 ECG recording and 766 patients had both a 24 hr ECG recording and a radionuclide ventriculogram. Primary risk variables. Two-channel (leads V, and V,) 24 hr ambulatory ECGs were analyzed by computer with manual overread.'415 Holter ECG tapes were randomly assigned for computer analysis to either the Columbia University Arrhythmia Control Unit or to the Washington University Arrhythmia Analysis Laboratory.'4 5 Before the program began, the two laboratories demonstrated excellent agreement in their quantitative identification of VPD frequency and repetitiveness of VPDs. As part of ongoing quality-control procedures, 140 24 hr ECGs were selected at random by the data coordinating center for a second computer analysis. There were 77'crossreadings (Columbia University vs Washington University) and 63 repeat readings (by 'the same institution) over the duration of the study. A two-way analysis of variance was used to assess the degree of bias between and within institutions for the computed frequency of VPDs. To measure degree of agreement for this quantitative variable, the intraclass correlation coefficient (ICC) was used.'6 An exact binomial version of McNemar's test for matched samples was used to test for bias in the detection of repetitive VPDs (defined as two or more VPDs in a row).'7 The kappa statistic was used as the measure of agreement for this dichotomous variable. 17 Kappa is interpreted in the same way as the ICC for quantitative data; however, values of kappa are lower for similar degrees of correlation.17 For the repeat readings from each of the two centers and the crossreadings, there was no significant bias present for VPD frequency. The ICCs revealed excellent agreement for the intracenter and intercenter repeat readings for VPD frequency (ICC > .90 in each case). For VPD runs, the exact version of the McNemar test revealed no evidence of bias in the detection of repetitive VPDs. The kappa statistics showed very good agreement for the repeat reading's and crossreadings (kappa = .79 for the repeat readings on either system and kappa = .55 for the crossreadings). LVEF was measured by radionuclide ventriculography with patients in the supine position. The first-pass technique used a 20 to 30 degree right anterior oblique projection during acquisition of data, whereas the multiple-gated technique (MUGA) was performed in a 45 to 60 degree left anterior oblique projection. Quality-control procedures for the LVEF measurements have been described previously. " Mortality data. Detailed information was collected on the clinical circumstances surrounding each of the enrolled patients who died on or before December 31, 1981. A four-member mortality committee classified each death as to geographic location, underlying cause, elapsed time between onset of acute cardiac symptoms and death, and the suspected mechanism of death in patients dying of atherosclerotic heart disease. Each death was classified by the method of Hinkle (sed table 1) as follows: deaths primarily due to arrhythmia, Hinkle class I; deaths primarily due to nonarrhythmic causes, Hinkle class II (circulatory failure); and deaths not classifiable, Hinkle class 111.18 The file containing the mortality data was closed to the investigators until the time of the primary analysis. For the present analysis, two mortality end points were used: death from any cause and potentially avoidable death due to arrhythmia (from the Hinkle classification). Because judgment has to be applied to classify deaths and because all the information pertinent to classification is almost never available, we used total mortality as the primary end point for analysis. Statistical methods. The relationships among frequency of VPDs, repetitiveness of VPDs, and left ventricular dysfunction on one hand and total or arrhythmic (Hinkle class I) mortality on the other were initially explored by means of the Kaplan-Meier 251 by gest on July 6, 2017 http://ciajournals.org/ D ow nladed from