Avoiding interpretive pitfalls when assessing arrhythmia suppression after myocardial infarction: insights from the long-term observations of the placebo-treated patients in the Cardiac Arrhythmia Pilot Study (CAPS)

The Cardiac Arrhythmia Pilot Study (CAPS) was a 1 year trial that analyzed the safety and effectiveness of arrhythmia suppression in 502 patients surviving acute myocardial infarction who had greater than or equal to 10 ventricular premature depolarizations/h or greater than or equal to 5 runs of ventricular tachycardia on a Holter recording obtained 6 to 60 days after the acute infarction. Because 100 of these patients received placebo in a double-blind fashion for 1 year, a comprehensive objective analysis was performed of spontaneous arrhythmia changes based on real data rather than statistical estimates. In the CAPS placebo group, 19% developed some serious clinical event in 1 year (death, heart failure, proarrhythmia) that could likely be attributable to antiarrhythmic drug toxicity. A significant reduction in the frequency of ventricular premature depolarizations (p = 0.004) occurred in the first few weeks of "therapy" with a further significant (p less than 0.04) decrease between 3 to 12 months. After initiation of placebo antiarrhythmic therapy, 27% had "apparent ventricular premature depolarization suppression" (greater than or equal to 70% reduction) after one Holter recording evaluation and nearly half (48%) after six Holter recordings to assess suppression were performed.(ABSTRACT TRUNCATED AT 250 WORDS)

Statistical interim monitoring of the Cardiac Arrhythmia Suppression Trial

The Cardiac Arrythmia Suppression Trial was stopped much earlier than planned. Statistical considerations played a very important role in the decision. Flexible group sequential testing was developed for the trial by implementing a Lan and DeMets procedure with use of the permutation test. We compute P-values from the joint permutation distribution of the test statistics, so we do not need to estimate the sampling distribution which in general is rather difficult to do without strict assumptions. The method also gives an exact test for small samples and allows us to use more complicated or non-Gaussian statistics. We also utilized stochastic curtailment ideas to evaluate various scenarios that might occur during the course of the trial, which assisted the Data and Safety Monitoring Board in making appropriate decisions.

Biobehavioral variables and mortality or cardiac arrest in the Cardiac Arrhythmia Pilot Study (CAPS)

The frequency of ventricular premature complexes and the degree of impairment of left ventricular ejection fraction are major predictors of cardiac mortality and sudden death in the year after acute myocardial infarction. Recent studies have implicated psychosocial factors, including depression, the interaction of social isolation and life stress, and type A-B behavior pattern, as predictors of cardiac events, controlling for known parameters of disease severity. However, results tend not to be consistent and are sometimes contradictory. The present investigation was designed to test the predictive association between biobehavioral factors and clinical cardiac events. This evaluation occurred in the context of a prospective clinical trial, the Cardiac Arrhythmia Pilot Study (CAPS). Five-hundred two patients were recruited with greater than or equal to 10 ventricular premature complexes/hour or greater than or equal to 5 episodes of nonsustained ventricular tachycardia, recorded 6 to 60 days after a myocardial infarction. Baseline behavioral studies, conducted in approximately 66% of patients, included psychosocial questionnaires of anxiety, depression, social desirability and support, and type A-B behavior pattern. In addition, blood pressure and pulse rate reactivity to a portable videogame was assessed. The primary outcome was scored on the basis of mortality or cardiac arrest. Results indicated that the type B behavior pattern, higher levels of depression and lower pulse rate reactivity to challenge were significant risk factors for death or cardiac arrest, after adjusting statistically for a set of known clinical predictors of disease severity. The implication of these results for future research relating behavioral factors to cardiac endpoints is discussed.

Imbalance in treatment assignments in stratified blocked randomization

Blocking and stratification are used in preparing randomization assignments to ensure that there will be nearly equal numbers of patients in each treatment group and that the groups will be similar with respect to important covariates. Stratified blocked randomization will create near balance within strata, but imbalance for the total trial may still occur. The variance for the total trial imbalance D is derived and examples from clinical trials are given. Under reasonable assumptions, if the blocking factor is size B in each of K strata, then max D = KB/2 and var D = K(B + 1)/6. These results may be used in planning a trial to estimate the overall imbalance expected for various choices of B and K. A conditional variance is given that allows the probability of an observed imbalance at the completion of a trial to be evaluated. Overall imbalance is about as likely with stratified blocked randomization as with simple randomization unless the total sample size N is appreciably larger than K X B. So long as the blinding is maintained, the block sizes should be chosen to be as small as possible.

Sudden cardiac arrest in Israel: factors associated with successful resuscitation

Out-of-hospital cardiac arrests were studied in Israel from 1984 to 1985. More than 3,500 patients in cardiac arrest received paramedic care. Eighty-three percent of cases were caused by underlying heart disease. Overall, 17% of patients with arrest caused by heart disease were admitted and 7% were discharged from the hospital. There was a wide variation in the percent discharged among the 15 paramedic service areas, ranging from 0% to 13%. Factors associated with successful resuscitation included witnessed collapse, rhythm of ventricular fibrillation, short interval from collapse to cardiopulmonary resuscitation (CPR) and delivery of advanced cardiac life support, collapse at public location, and bystander initiation of CPR. Improvements in survival are likely to result if CPR is more frequently and promptly initiated and the time to arrival of definitive paramedic care can be improved.

Lay person use of automatic external defibrillation

The ability of lay persons to learn and retain defibrillation skills using an automatic external defibrillator (AED) was assessed. Thirty-four family members of cardiac arrest survivors were trained in CPR techniques and defibrillation, and evaluated for performance of skills immediately following training and at six-week and three-month follow-ups. All but two were successfully trained to complete three cycles of CPR interspersed with three defibrillatory shocks in an average of four minutes with the first shock delivered in an average of two minutes. Although there were decrements in the speed and quality of performance at each follow-up period (P less than .01), we conclude that most lay persons can learn to operate an AED safely and under simulated conditions provide defibrillatory shocks an average of eight minutes faster than typical response times of emergency medical technicians. These results suggest that AEDs can be placed in many homes of patients at high risk for cardiac arrest.

Spontaneous variability of ventricular arrhythmias in patients at increased risk for sudden death after acute myocardial infarction: consecutive ambulatory electrocardiographic recordings of 88 patients

The Cardiac Arrhythmia Pilot Study, sponsored by the National Heart, Lung, and Blood Institute, is a multicenter, prospective, randomized, double-blind trial designed to identify patients having 10 or more ventricular premature complexes (VPCs) per hour within 6 to 60 days of acute myocardial infarction. The present investigation selected patients after acute myocardial infarction who had ambulatory electrocardiographic qualifying arrhythmia for CAPS. An additional baseline electrocardiogram was recorded before enrollment in the study to assess baseline spontaneous variability of VPCs. A total of 88 patients (15 women, 73 men, aged 57 +/- 10 years) were studied. The 43 patients (49%) receiving beta-blocking drugs were included because the dose was not altered between the 2 consecutive electrocardiographic recordings. This investigation shows that a 95% reduction in VPCs is required to document a significant drug effect rather than variability alone if 1 day of control and 1 day of treatment electrocardiographic recording are compared. Similarly, based on 1 day of electrocardiographic recording before and after antiarrhythmic therapy, 1,780% increase in VPC frequency is required to establish "arrhythmia aggravation" from an antiarrhythmic drug rather than from variability alone based on a 95% confidence interval. Variability of ventricular arrhythmias is independent of left ventricular function, whereas patients taking beta-blocking therapy tend to have greater VPC variability (p = 0.052), even though VPC frequencies were lower (59 +/- 19 vs 138 +/- 31 VPCs/hour, p less than 0.006) than those not taking beta-blocking drugs.

Home placement of automatic external defibrillators among survivors of ventricular fibrillation

We performed a study to determine how many patients at highest risk for an episode of ventricular fibrillation (those who have already survived such an event) are potential candidates for automatic external defibrillator placement and accepting of such a device in their homes. All VF survivors in King County, Washington, during 1984 were screened for possible enrollment in the study. Of 95 survivors of out-of-hospital VF, 63 (66%) were eligible and of 47 patients approached, 38 (81%) agreed to participate in the study. These findings suggest that approximately half (product of 66% and 81% acceptance rate equals 53%) of VF survivors potentially could use automatic external defibrillators.

Identification of cardiac arrest by emergency dispatchers

Prompt identification of cardiac arrest by emergency dispatchers can save valuable time and increase the likelihood of successful resuscitation. The authors reviewed 516 cardiac and 146 non-cardiac calls to identify features of a probable cardiac arrest call. The results indicate that information about sex, location, and activity is of little use in the identification of cardiac arrest. When the patient is over 50 years old and the caller is emotional, the possibility of cardiac arrest is high, suggesting that questions about consciousness and breathing should be asked immediately. Additional information can be obtained or telephone cardiopulmonary resuscitation (CPR) instructions can be given after dispatch of an emergency vehicle.

The potential use of automatic defibrillators in the home for management of cardiac arrest

Ventricular fibrillation, an abnormal cardiac rhythm, occurs in at least two-thirds of the 400,000 people who die out of the hospital from sudden cardiac arrest. This rhythm can be treated successfully by electric countershock, a procedure known as defibrillation. The survival rate following such cardiac arrest is directly related to the rapidity of response; the shorter the time from collapse to defibrillation, the more patients will survive. There are two basic options to shorten the time from collapse to defibrillatory shock. The first is to upgrade the emergency medical system. The second is to provide spouses and family members of potential cardiac arrest patients with automatic home defibrillators. This article considers the effectiveness of the second option, home defibrillation, compared with that of an equally costly upgrade in existing emergency medical service systems. The comparisons depend on the existing level of emergency medical service system, the cost of the home defibrillator, and the rate at which a home defibrillator would be used appropriately. The comparisons suggest that in many circumstances home defibrillation is an appropriate option to be considered.

Automatic external defibrillation: evaluations of its role in the home and in emergency medical services

Many recent efforts to improve emergency medical services (EMS) and increase survival rates are simply efforts to get defibrillation to patients as rapidly as possible. In the 1960s physicians traveled in mobile coronary care units to bring the defibrillator to cardiac arrest patients. Later, paramedics, rather than physicians, were used. During the late 1970s the concept of early out-of-hospital defibrillation expanded as emergency medical technicians (EMTs) learned to defibrillate. Researchers in several settings confirmed the effectiveness of early defibrillation by EMTs. The automatic detection of ventricular fibrillation (VF) creates new opportunities for the early defibrillation concept. This includes both automatic implantable defibrillators and automatic external defibrillators (AED). The King County, Washington, EMS is conducting two projects to evaluate AEDs. One is a randomized, controlled crossover study in which EMTs use either an AED or a standard manual defibrillator. Outcome measurements include time to countershock, conversion rates, and survival rates. In the second project family members of patients who have survived out-of-hospital VF randomly receive an AED and cardiopulmonary resuscitation (CPR) instruction, or CPR instruction alone. This study was designed to determine whether family members can be trained adequately to use the device effectively. Psychological tests measure the effect of learning about, living with, and using such technology. These studies may help define the role of AEDs in the future management of out-of-hospital VF.

Survivors of out-of-hospital cardiac arrest: morbidity and long-term survival

Sudden cardiac death accounts for two thirds of death due to coronary artery disease. Advanced cardiac life support can now be brought directly to patients with out-of-hospital cardiac arrest, and in this country, as many as 30% of such patients can be discharged from the hospital annually. Certain clinical and resuscitation-related factors are predictive of mortality and morbidity. The best clinical predictors of long-term survival are absence of previous history of myocardial infarction, lack of congestive heart failure during hospitalization, and age less than 60 years. Resuscitation-related predictors of long-term survival are a short time collapse to cardiopulmonary resuscitation (CPR), and a short time from collapse to CPR combined with a short time to provision of definitive care. The majority of cardiac arrest survivors are able to resume previous levels of function.

Epidemiology of cardiac arrest and resuscitation in children

A surveillance system identified all out-of-hospital cardiac patients under the age of 18 who received emergency care in suburban King County, Washington. The etiology, cardiac rhythm, and outcome were identified for each case. During a 6 1/2-year period, 119 cardiac arrests occurred (annual incidence, 12.7/100,000 among individuals less than 18). Sudden infant death was the most common etiology (32%), and drowning was the second most common (22%). The most common rhythm was asystole, accounting for 66% of all rhythms. Six percent of patients treated with basic EMT care were discharged, compared with 7% of patients treated with EMT and paramedic care. In contrast to resuscitation from cardiac arrest in adults, the likelihood of successful resuscitation in children is very poor. This is due to different etiologies and the higher proportion of asystole seen in pediatric cardiac arrest as compared with adults.

Evaluation of noninvasive tests of cardiac damage in suspected cardiac contusion

Nonpenetrating trauma to the chest can result in cardiac damage that may be overlooked because of associated injuries and the lack of obvious thoracic injury. The clinical diagnosis of important cardiac damage in this setting is difficult. We evaluated noninvasive tests for detecting myocardial damage in 100 patients with severe, nonpenetrating chest trauma. The noninvasive tests included serial ECG, serial total CPK and CPK-MB enzymes, continuous Holter monitor recording to detect dysrhythmia, and technetium-99m pyrophosphate scintigraphy. Peak CPK-MB elevations occurred in 72 patients. ECG abnormalities were noted in 70 patients, and 27 patients had Lown grade 3 or greater dysrhythmias. Fifteen patients died and all had autopsies. The noninvasive abnormalities were nonspecific and did not reflect myocardial contusion that led to clinically important cardiac complications.

Community awareness of emergency phone numbers

Knowledge of numbers to call for medical emergencies was compared among communities with three different call numbers: 1) 911; 2) regional seven-digit numbers; and 3) local seven-digit numbers. Correct responses were 85 per cent in the 911 communities; 47 per cent in areas with regional systems, and 36 per cent in areas with local systems. Persons living adjacent to a 911 area were more likely to believe 911 was the emergency number (28 per cent) than persons not adjacent to 911 areas (12 per cent).

The ACLS score. Predicting survival from out-of-hospital cardiac arrest

We developed a score predictive of survival following out-of-hospital cardiac arrest from an analysis of factors associated with 611 cases. The score is calculated from four pieces of information readily obtainable by emergency personnel directly at the scene. The four items are as follow: A, arrest witnessed; C, cardiac rhythm; L, lay bystander cardiopulmonary resuscitation (CPR); S, speed (response time of paramedic unit). Among 22 patients with favorable findings on all four predictive variables (witnessed arrest, ventricular fibrillation, bystander CPR, paramedic response time less than four minutes), 15 (70%) were discharged alive. The ACLS score for this group of patients was 70%. Among 97 patients with the most unfavorable findings (whose ACLS score was 0), one (1%) was discharged. We believe the score can provide emergency personnel with a realistic appraisal of the likelihood of successful resuscitation.

Out-of-hospital cardiac arrest: improved survival with paramedic services

Survival after out-of-hospital cardiac arrest was studied in a suburban community (population 304000) before and after addition of paramedic services. During period 1 emergency medical technicians provided basic emergency care (cardiopulmonary resuscitation at the scene of collapse and during the journey to hospital). In period 2 additional care was given at the scene of collapse by paramedics capable of advanced emergency care (defibrillation, endotracheal intubation, drugs). During the 3-yr study 585 patients with cardiac arrest caused by heart disease received prehospital emergency resuscitation. Paramedic services improved the rate of live admission to the coronary-care or intensive-care unit from 19% to 34% (p less than 0.001) and the rate of discharge from 7% to 17% (p less than 0.01). The mean time from collapse to delivery of advanced emergency care was 27.5 min during period 1 with technician services, and 7.7 min during period 2 with paramedic services. Ventricular fibrillation caused cardiac arrest in nearly all patients who survived; it occurred in 91 of the 160 (57%) patients during period 1 whose rhythms were determined and in 192 of the 343 (56%) patients during period 2. The decreased time from collapse to delivery of advanced emergency care accounted for the improved survival with paramedic services.

Management of out-of-hospital cardiac arrest. Failure of basic emergency medical technician services

Survival after out-of-hospital cardiac arrest treated by emergency medical technicians (EMTs) with basic life support was studied in four communities with a combined population of 380,000. During a two-year period, 18 (6%) of 321 patients with cardiac arrest were resuscitated and ultimately discharged from the hospital. This figure is compared with 55 (22%) of 253 discharged in adjacent suburban communities with paramedic services. The evident factor accounting for the difference in survival rates was the time from collapse to receiving definitive care (advanced cardiac life support)--26 minutes in the EMT area compared to 7.8 minutes in the paramedic area.

Evaluation of paramedic programs using outcomes of prehospital resuscitation for cardiac arrest

Two evaluation methods, one statistical and one comparative, were developed to assess the effectiveness of paramedic programs in King County, Washington. The outcome of hospital admission following prehospital cardiac arrest was used as a measure of effectiveness. In the statistical method, actual outcomes were compared with predicted outcomes. Predictive variables for admission were time from collapse to initiation of cardiopulmonary resuscitation and time from collapse to definitive care. Given knowledge of the predictive variables, the statistical evaluation enabled us to determine the probability of the outcome following cardiac arrest. In the comparative method, outcomes were compared with a standard in an adjacent community. Using this method, we identified program elements that could lead to improved outcome. Both evaluation methods are easily implemented.

Cardiac resuscitation in the community. Importance of rapid provision and implications for program planning

Several time-related variables involving resuscitation from out-of-hospital cardiac arrest were studied. Short time intervals from collapse to initiation of cardiopulmonary resuscitation (CPR) and to provision of definitive care were significantly associated with survival from cardiac arrest. The two times were jointly related, and one short time without the other was unlikely to result in survival. If CPR was initiated within four minutes and if definitive care was provided within eight minutes, 43% of patients survived. If either time was exceeded, the changes of survival fell dramatically. The time to initiation of CPR and definitive care are factors directly influenced by emergency medical service program decisions. A realistic option to improve time to initiation of CPR is widespread citizen CPR training. A possible option to improve the time to definitive care is the training of emergency medical technicians in defibrillation.

Evaluating emergency medical services: Quasi-experimental outcome studies

The need to evaluate expensive, dramatic, and politically sensitive emergency medical services programs when classical controlled trials are neither ethically nor practically possible can be satisfied by quasi-experimental designs. The sequential implementation of paramedic services in several suburban areas provided a natural experimental situation in which to evaluate whether addition of the service could significantly alter the outcome of cardiac emergencies compared to the basic emergency medical technician program previously available. Before measurements and after measurements were made in a study area plus two control areas: one with paramedic services in both time periods and the other with emergency medical technician service throughout. Preliminary results indicate successful resuscitation increased from 20% to 32% (p less than .05) and discharge from the hospital went from 8% to 18% (p less than .01). The implications for program and policy decisions are noted. Development of studies that evolved from this work are outlined.