Cardiac Arrest Complicating Acute Myocardial Infarction: Predictability and Prognosis
MARTIN J. CONLEY, MD J. FREDERICK McNEER, MD KERRY L. LEE, PhD GALEN S. WAGNER, MD ROBERT A. ROSATI, MD Durham, North Carolina
From the Division of Cardiology, Department of Medicine, Department of Surgery and the Department of Community Health Sciences, Duke University Medical Center, Durham, N. C. This work was supportedby ResearchGrant HL-17670 from the National Heart,Lung,and Blood Institute, Bethesda, Md. and Grant HS-01613from the National Center for Health Services Research, Rockville, Md. and TrainingGrantsLM-07003from the National Library of Medicine and HL-07101 from the NationalHeart, Lung, and Blood Institute, Bethesda,Md. ManuscriptreceivedMay 12, 1976; revised manuscript received July 6, 1976, accepted July 7, 1976. Address for reprints: RobertA. Rosati,MD, P.O. Box 3337, Duke University Medical Center, Durham, N. C. 27710.
Eleven percent of 905 consecutive patients with acute myocardial infarction admitted to the coronary care unit at Duke University Medical Center experienced cardiac arrest. Subgroups of patients at high and low risk for cardiac arrest were identified. Cardiac arrest was experienced by 17 percent of patients with signs of heart failure on admission but by only 3 percent of patients without diabetes mellitus, prior myocardial infarction or heart failure by history or on admission. Only 59 percent Of patients with cardiac arrest survived hospitalization compared with 88 percent of those without cardiac arrest. Long-term survival for the 7 6 5 hospital survivors was significantly greater in the group without than in the group with arrest at each yearly Interval from 1 through 5 years; the 2 year survival rate was 50 and 77 percent, respecUvely, in these two groups. Many of the deaths among the hospital survivors occurred in patients with signs Of heart failure during hospitalization. Among 668 hospital survivors who had mild or no heart failure during hospitalization, cardiac arrest continued to be a significant predictor of mortality. The mode of death among hospital survivors did not differ in the groups with and without cardiac arrest; for example, the incidence rate of sudden death in the two groups was 44 and 37 percent, respectively. In light of recent reports suggesting that the prophylactic use of antiarrhythmic agents can virtually eliminate Ventricular fibrillation during the hospital phase of acute myocardial infarction, we contend that such use may substantially reduce both long-term and hospital mortality after acute myocardial infarction.
The prognosis of patients who survive hospitalization for acute myocardial infarction complicated by cardiac arrest has been reported to be similar to that of patients whose infarction was not thus complicated, i-7 With the exception of one report s suggesting an excess mortality at 12 months in patients with acute infarction complicated by ventricular fibrillation, previous studies have either included unmonitored cardiac arrests or have had no concurrent control group of patients without cardiac arrest for comparison. The purpose of this report was to define subgroups of patients at high and low risk for cardiac arrest during hospitalization for acute myocardial infarction and to determine the effect of such arrest on hospital and long-term survival. Recent studies 9,i° have suggested that vigorous prophylactic treatment for ventricular arrhythmias can substantially reduce the incidence of ventricular fibrillation complicating acute myocardial infarction. We believe that if cardiac arrest can be shown to be related to mortality in hospital survivors of acute myocardial infarction, vigorous efforts to eliminate ventricular arrhythmias should become an integral part of the management of infarction. This report presents our experience with 905 consecutive patients admitted to the cardiac care unit at Duke University Medical Center for acute myocardial infarction.
January 1977
The American Journal of CARDIOLOGY
Volume 39
7
CARDIAC ARREST AFTER MYOCARDIAL INFARCTION--CONLEY El" AL.
Methods Between December 15, 1965 and September 1, 1974 a total of 939 patients were admitted to the cardiac care unit at Duke University Medical Center with acute myocardial infarction. Information on these patients was collected prospectively beginning in July 1967 and retrospectively before July 1967. - All data were stored in a computerized information system, n Table I lists the 25 historical characteristics used in this analysis. Thirty-four patients who had an unmonitored arrest before admission to the cardiac care unit and 22 patients whose cardiac arrest was secondary to ventricular asystole were eliminated from the study. The remaining 905 patients were assigned to one of two groups depending on whether their hospital course was or was not complicated by a monitored cardiac arrest. Acute myocardial infarction was defined as the presence of at least two of the following three factors: (1) a clinical history of chest pain suggesting acute myocardial infarction; (2) typical QRS changes in serial electrocardiograms; (3) characteristic changes in serum enzymes. The enzyme changes included transiently elevated lactic dehydrogenase (LDH) or creatinine phosphokinase (CPK) levels, or both, without abnormal serum glutamic pyruvic transaminase (SGPT) levels or other possible causes for elevation. More recently the isoenzymes of LDH and CPK have been used for this determination. Cardiac arrest was defined as a primary episode of ventricular tachycardia or ventricular fibrillation requiring resuscitation (external cardiac massage or direct current cardioversion, or both). Episodes of ventricular asystole and
TABLE I Prevalence Rate of Historical Characteristics in Pat;ents With and Without Cardiac Arrest Group Without Arrest
(%)
Male sex Age > 50 years White
Group With Arrest
8
January 1977
T A B L E II Prevalence Rate of Hospitalization Characteristics in Patients With and W i t h o u t Cardiac Arrest
(%)
69 84 76
69 81 73
60 51 43 7 1 33 20 11 29 7 10 3 7 7 29 17 21 7 45 58 47
66 57 51 6 0 44 25 11 39 4 13 4 10 11 26 28 26 9 49 56 50
60 26 8 5
44 41 15 0
Past history of: Coronary artery disease Angina Typical angina Atypical angina Nonanginal chest pain Previous myocardial infarction (MI) Old anterior MI Old posterior MI Old diaphragmatic MI Old lateral MI Cerebrovascular disease Transient ischemic attacks Stroke Peripheral vascular disease Lipid disorder Congestive heart failure Diabetes mellitus Renal disease Hypertension Smoking Family history of ischemic heart disease Admission class for congestive heart failure (Killip) Class I Class II Class III Class IV
death occurring as the terminal event of a gradually deteriorating condition were not considered cardiac arrest. No patient had cardiogenic shock (KiUip class IV) at the time of cardiac arrest. FoUow-up data on life or death status was obtained for the hospital survivors at yearly intervals from I through 5 years. The cause of death was determined for all patients who died. Information was obtained by a staff cardiologist on a clinical visit or by a research associate in direct telephone contact. The follow-up data were 99 percent complete; that is, follow-up information at each interval for all patients was known in 99 percent of instances in which the patient had reached the appropriate interval. S t a t i s t i c a l analysis: The chi-square test was used to determine the significance of each of the 25 historical characteristics as a predictor of cardiac arrest after hospitalization for acute myocardial infarction. A stepwise chi-square analysis using combinations of these same 25 characteristics was performed to identify subgroups of patients who were at high and low risk for cardiac arrest. The hospital survival rates in the groups with and without cardiac arrest were compared using the chi-square test. Long-term survival rates of hospital survivors in the two groups were calculated using the life-table method. 12 Differences in survival rates through the 5 year follow-up period were tested for statistical significance using the chi-square statistic within the categorical linear models framework. 13The cause of death among hospital survivors in the two groups was compared using the chi-square test to determine if there were differences in the proportion of deaths in each group that were sudden, related to protracted heart failure or of other cardiovascular or noncardiovascular origin.
The American Journal of CARDIOLOGY
Group Without Arrest
Group With Arrest
46 33 10 12 18 8
24 33 9 35 40 19
33 15 38 18 32 47
36 11 46 29 19 62
4 5 10
8 12 20
14 8 1 7
22 12 1 12
3 5 5 3
6 7 10 4
(%)
Worst class for congestive heart fadure (Killip) Class I Class II Class III Class IV Sustained hypotension Extension of myocardial infarction Site of acute infarction Anterior Posterior Diaphragmatic Lateral Persistent sinus bradycardia Persistent sinus tachycardia Rapid atrial arrhythmias Paroxysmal atrial tachycardia Atrial flutter Atrial fibrillation Heart block First degree Second degree, type I Second degree, type II Third degree Conduction disturbances Left bundle branch block Right bundle branch block Left anterior hemiblock Left posterior hemiblock
Volume 39
(%)
CARDIAC ARREST AFTER MYOCARDIAL INFARCTION~ONLEY ET AL.
Long-term survival analysis: To determine whether cardiac arrest was independently related to long-term survival after hospitalization for acute myocardial infarction, the 25 historical and 20 hospitalization characteristics (Table II) were examined in hospital survivors to identify significant risk factors relative to long-term survival. The distribution of these significant risk factors in the two patient groups was compared to identify baseline inequalities. Patients described by characteristics that were significant risk factors relative to longterm survival and that were unequally distributed between the two groups were eliminated in stepwise fashion starting with that characteristic that was most significant individually as a risk factor. Life-table survival rates were then recalculated for these corrected population data in which characteristics that were significant risk factors relative to survival were equally distributed between the groups with and without cardiac arrest. Differences in long-term survival between the two groups were again compared using the chi-square statistic within the categorical linear models framework. Results
Functional class for heart failure: The hospital course of 102 of the 905 patients (11 percent) was complicated by a monitored cardiac arrest. The yearly cardiac arrest rate ranged from 7 to 18 percent (Table III), but there was no consistent change in the rate of cardiac arrest from 1966 through 1974. Forty-five patients who subsequently had a cardiac arrest were in Killip class I for heart failure on admission, 42 were in class II and 15 were in class III. Few patients changed their functional class for heart failure between admission and the episode of cardiac arrest. Accordingly, at the time of cardiac arrest 46 patients were in class I, 40 were in class II and 16 were in class III. Consistent with our definition of cardiac arrest as a primary event is the fact that no patient was in class IV at the time of cardiac arrest. Predictors of cardiac arrest: Examination~)f the 25 Historical characteristics (Table I) revealed that only admission class for heart failure, history of congestive heart failure and history of myocardial infarction were significant individual predictors of cardiac arrest after hospitalization for acute myocardial infarction. Admission class for heart failure was the most significant predictor. Of 336 patients with signs of heart failure on admission (Killip class II or III), 57 patients (17 percent) subsequently had a cardiac arrest during their hospi-
talization. Forty-two patients who were in class IV for heart failure on admission were not included since by definition they could not experience cardiac arrest. The incidence of cardiac arrest among patients without signs of heart failure on admission (Killip class I) was less than 9 percent (45 of 527 patients). This difference was statistically significant (X 2 = 13.98, DF = 1, P
MARTIN J. CONLEY, MD J. FREDERICK McNEER, MD KERRY L. LEE, PhD GALEN S. WAGNER, MD ROBERT A. ROSATI, MD Durham, North Carolina
From the Division of Cardiology, Department of Medicine, Department of Surgery and the Department of Community Health Sciences, Duke University Medical Center, Durham, N. C. This work was supportedby ResearchGrant HL-17670 from the National Heart,Lung,and Blood Institute, Bethesda, Md. and Grant HS-01613from the National Center for Health Services Research, Rockville, Md. and TrainingGrantsLM-07003from the National Library of Medicine and HL-07101 from the NationalHeart, Lung, and Blood Institute, Bethesda,Md. ManuscriptreceivedMay 12, 1976; revised manuscript received July 6, 1976, accepted July 7, 1976. Address for reprints: RobertA. Rosati,MD, P.O. Box 3337, Duke University Medical Center, Durham, N. C. 27710.
Eleven percent of 905 consecutive patients with acute myocardial infarction admitted to the coronary care unit at Duke University Medical Center experienced cardiac arrest. Subgroups of patients at high and low risk for cardiac arrest were identified. Cardiac arrest was experienced by 17 percent of patients with signs of heart failure on admission but by only 3 percent of patients without diabetes mellitus, prior myocardial infarction or heart failure by history or on admission. Only 59 percent Of patients with cardiac arrest survived hospitalization compared with 88 percent of those without cardiac arrest. Long-term survival for the 7 6 5 hospital survivors was significantly greater in the group without than in the group with arrest at each yearly Interval from 1 through 5 years; the 2 year survival rate was 50 and 77 percent, respecUvely, in these two groups. Many of the deaths among the hospital survivors occurred in patients with signs Of heart failure during hospitalization. Among 668 hospital survivors who had mild or no heart failure during hospitalization, cardiac arrest continued to be a significant predictor of mortality. The mode of death among hospital survivors did not differ in the groups with and without cardiac arrest; for example, the incidence rate of sudden death in the two groups was 44 and 37 percent, respectively. In light of recent reports suggesting that the prophylactic use of antiarrhythmic agents can virtually eliminate Ventricular fibrillation during the hospital phase of acute myocardial infarction, we contend that such use may substantially reduce both long-term and hospital mortality after acute myocardial infarction.
The prognosis of patients who survive hospitalization for acute myocardial infarction complicated by cardiac arrest has been reported to be similar to that of patients whose infarction was not thus complicated, i-7 With the exception of one report s suggesting an excess mortality at 12 months in patients with acute infarction complicated by ventricular fibrillation, previous studies have either included unmonitored cardiac arrests or have had no concurrent control group of patients without cardiac arrest for comparison. The purpose of this report was to define subgroups of patients at high and low risk for cardiac arrest during hospitalization for acute myocardial infarction and to determine the effect of such arrest on hospital and long-term survival. Recent studies 9,i° have suggested that vigorous prophylactic treatment for ventricular arrhythmias can substantially reduce the incidence of ventricular fibrillation complicating acute myocardial infarction. We believe that if cardiac arrest can be shown to be related to mortality in hospital survivors of acute myocardial infarction, vigorous efforts to eliminate ventricular arrhythmias should become an integral part of the management of infarction. This report presents our experience with 905 consecutive patients admitted to the cardiac care unit at Duke University Medical Center for acute myocardial infarction.
January 1977
The American Journal of CARDIOLOGY
Volume 39
7
CARDIAC ARREST AFTER MYOCARDIAL INFARCTION--CONLEY El" AL.
Methods Between December 15, 1965 and September 1, 1974 a total of 939 patients were admitted to the cardiac care unit at Duke University Medical Center with acute myocardial infarction. Information on these patients was collected prospectively beginning in July 1967 and retrospectively before July 1967. - All data were stored in a computerized information system, n Table I lists the 25 historical characteristics used in this analysis. Thirty-four patients who had an unmonitored arrest before admission to the cardiac care unit and 22 patients whose cardiac arrest was secondary to ventricular asystole were eliminated from the study. The remaining 905 patients were assigned to one of two groups depending on whether their hospital course was or was not complicated by a monitored cardiac arrest. Acute myocardial infarction was defined as the presence of at least two of the following three factors: (1) a clinical history of chest pain suggesting acute myocardial infarction; (2) typical QRS changes in serial electrocardiograms; (3) characteristic changes in serum enzymes. The enzyme changes included transiently elevated lactic dehydrogenase (LDH) or creatinine phosphokinase (CPK) levels, or both, without abnormal serum glutamic pyruvic transaminase (SGPT) levels or other possible causes for elevation. More recently the isoenzymes of LDH and CPK have been used for this determination. Cardiac arrest was defined as a primary episode of ventricular tachycardia or ventricular fibrillation requiring resuscitation (external cardiac massage or direct current cardioversion, or both). Episodes of ventricular asystole and
TABLE I Prevalence Rate of Historical Characteristics in Pat;ents With and Without Cardiac Arrest Group Without Arrest
(%)
Male sex Age > 50 years White
Group With Arrest
8
January 1977
T A B L E II Prevalence Rate of Hospitalization Characteristics in Patients With and W i t h o u t Cardiac Arrest
(%)
69 84 76
69 81 73
60 51 43 7 1 33 20 11 29 7 10 3 7 7 29 17 21 7 45 58 47
66 57 51 6 0 44 25 11 39 4 13 4 10 11 26 28 26 9 49 56 50
60 26 8 5
44 41 15 0
Past history of: Coronary artery disease Angina Typical angina Atypical angina Nonanginal chest pain Previous myocardial infarction (MI) Old anterior MI Old posterior MI Old diaphragmatic MI Old lateral MI Cerebrovascular disease Transient ischemic attacks Stroke Peripheral vascular disease Lipid disorder Congestive heart failure Diabetes mellitus Renal disease Hypertension Smoking Family history of ischemic heart disease Admission class for congestive heart failure (Killip) Class I Class II Class III Class IV
death occurring as the terminal event of a gradually deteriorating condition were not considered cardiac arrest. No patient had cardiogenic shock (KiUip class IV) at the time of cardiac arrest. FoUow-up data on life or death status was obtained for the hospital survivors at yearly intervals from I through 5 years. The cause of death was determined for all patients who died. Information was obtained by a staff cardiologist on a clinical visit or by a research associate in direct telephone contact. The follow-up data were 99 percent complete; that is, follow-up information at each interval for all patients was known in 99 percent of instances in which the patient had reached the appropriate interval. S t a t i s t i c a l analysis: The chi-square test was used to determine the significance of each of the 25 historical characteristics as a predictor of cardiac arrest after hospitalization for acute myocardial infarction. A stepwise chi-square analysis using combinations of these same 25 characteristics was performed to identify subgroups of patients who were at high and low risk for cardiac arrest. The hospital survival rates in the groups with and without cardiac arrest were compared using the chi-square test. Long-term survival rates of hospital survivors in the two groups were calculated using the life-table method. 12 Differences in survival rates through the 5 year follow-up period were tested for statistical significance using the chi-square statistic within the categorical linear models framework. 13The cause of death among hospital survivors in the two groups was compared using the chi-square test to determine if there were differences in the proportion of deaths in each group that were sudden, related to protracted heart failure or of other cardiovascular or noncardiovascular origin.
The American Journal of CARDIOLOGY
Group Without Arrest
Group With Arrest
46 33 10 12 18 8
24 33 9 35 40 19
33 15 38 18 32 47
36 11 46 29 19 62
4 5 10
8 12 20
14 8 1 7
22 12 1 12
3 5 5 3
6 7 10 4
(%)
Worst class for congestive heart fadure (Killip) Class I Class II Class III Class IV Sustained hypotension Extension of myocardial infarction Site of acute infarction Anterior Posterior Diaphragmatic Lateral Persistent sinus bradycardia Persistent sinus tachycardia Rapid atrial arrhythmias Paroxysmal atrial tachycardia Atrial flutter Atrial fibrillation Heart block First degree Second degree, type I Second degree, type II Third degree Conduction disturbances Left bundle branch block Right bundle branch block Left anterior hemiblock Left posterior hemiblock
Volume 39
(%)
CARDIAC ARREST AFTER MYOCARDIAL INFARCTION~ONLEY ET AL.
Long-term survival analysis: To determine whether cardiac arrest was independently related to long-term survival after hospitalization for acute myocardial infarction, the 25 historical and 20 hospitalization characteristics (Table II) were examined in hospital survivors to identify significant risk factors relative to long-term survival. The distribution of these significant risk factors in the two patient groups was compared to identify baseline inequalities. Patients described by characteristics that were significant risk factors relative to longterm survival and that were unequally distributed between the two groups were eliminated in stepwise fashion starting with that characteristic that was most significant individually as a risk factor. Life-table survival rates were then recalculated for these corrected population data in which characteristics that were significant risk factors relative to survival were equally distributed between the groups with and without cardiac arrest. Differences in long-term survival between the two groups were again compared using the chi-square statistic within the categorical linear models framework. Results
Functional class for heart failure: The hospital course of 102 of the 905 patients (11 percent) was complicated by a monitored cardiac arrest. The yearly cardiac arrest rate ranged from 7 to 18 percent (Table III), but there was no consistent change in the rate of cardiac arrest from 1966 through 1974. Forty-five patients who subsequently had a cardiac arrest were in Killip class I for heart failure on admission, 42 were in class II and 15 were in class III. Few patients changed their functional class for heart failure between admission and the episode of cardiac arrest. Accordingly, at the time of cardiac arrest 46 patients were in class I, 40 were in class II and 16 were in class III. Consistent with our definition of cardiac arrest as a primary event is the fact that no patient was in class IV at the time of cardiac arrest. Predictors of cardiac arrest: Examination~)f the 25 Historical characteristics (Table I) revealed that only admission class for heart failure, history of congestive heart failure and history of myocardial infarction were significant individual predictors of cardiac arrest after hospitalization for acute myocardial infarction. Admission class for heart failure was the most significant predictor. Of 336 patients with signs of heart failure on admission (Killip class II or III), 57 patients (17 percent) subsequently had a cardiac arrest during their hospi-
talization. Forty-two patients who were in class IV for heart failure on admission were not included since by definition they could not experience cardiac arrest. The incidence of cardiac arrest among patients without signs of heart failure on admission (Killip class I) was less than 9 percent (45 of 527 patients). This difference was statistically significant (X 2 = 13.98, DF = 1, P
