|
|
| |||||||||||||||||||||||||||||||||||||||
Background Transcutaneous cardiac pacemakers generate electrical stimuli that pace the heart through external electrodes that adhere to the chest wall. Transcutaneous pacing has been useful in some patients with bradycardia, but its efficacy in patients with asystole and full cardiac arrest has been limited, possibly because of delays in the initiation of pacing. We studied the efficacy of early transcutaneous pacing in patients with out-of-hospital asystolic cardiac arrest.
Methods For three years we provided transcutaneous pacemakers to about half the fire districts in a large emergency-medical-services system (the intervention group). In these districts, we authorized emergency medical technicians (EMTs) to begin transcutaneous pacing in patients with cardiac arrest and primary asystole or post-defibrillation asystole. Pacing was done as early as possible, before endotracheal intubation or intravenous medication. EMTs in the other fire districts (the control group) treated similar patients with basic cardiopulmonary resuscitation but without transcutaneous pacing.
Results The EMTs in the intervention group initiated transcutaneous pacing in 112 of the 278 patients with primary asystole. Of these patients, 22 (8 percent) were admitted to the hospital, and 11 (4 percent) were discharged. Among the 259 patients treated by the EMTs in the control group, 21 (8 percent) were admitted to the hospital, and 5 (2 percent) were discharged. The two groups did not differ significantly with respect to the rate of hospital admission or survival. Survival after early pacing for post-defibrillation asystole was no better than survival after pacing for primary asystole.
Conclusions Transcutaneous pacing appears to offer no benefit in patients with asystolic cardiac arrest, even when it is performed as early as possible by EMTs in the field. Our data suggest that the widespread implementation of early transcutaneous pacing for out-of-hospital asystolic cardiac arrest would be ineffective.
This study attempted to counteract both possible reasons for the failure of transcutaneous pacing in patients with full cardiac arrest. To solve the "late arrival" problem, we equipped emergency medical technicians (EMTs) with transcutaneous pacemakers. These rescuers are usually the first emergency personnel to assist patients with cardiac arrest. Pacing by EMTs also solved the "late initiation" problem, since we authorized the EMTs to begin transcutaneous pacing before the arrival of paramedics. This meant that patients underwent pacing before endotracheal intubation and before the administration of intravenous medications. The discouraging results we observed have implications for the use of external pacing in emergency-medical-services systems.
Methods
The King County Emergency Medical Services System
Our Emergency Medical Services System serves approximately 920,000 people in a 2028-square-mile area adjacent to but not including the city of Seattle. The program is organized as a two-tiered response system. EMTs respond as the first tier. More than 1500 EMTs are stationed in 39 fire districts, and all are trained and equipped to use either conventional or automated external defibrillators. A total of 86 paramedics stationed in nine paramedic units respond as the second tier, which provides advanced life support. A cardiac-arrest surveillance system gathers data on patients and records the outcome of efforts at resuscitation28,29,30,31,32.
Study Design
This evaluation was designed as a controlled, prospective study. Transcutaneous external pacemakers were distributed selectively to the fire districts, so that half the anticipated cardiac arrests would occur in fire districts where EMTs were equipped with transcutaneous pacemakers, and half would occur in fire districts without transcutaneous pacemakers. Our cardiac-arrest surveillance system provided the average number of cardiac arrests per year in each fire district. We used a similar design in a trial of the use of automated external defibrillators by EMTs33.
The intervention group included all patients with cardiac arrest whose initial cardiac rhythm was found to be asystole by the emergency personnel in the 16 fire districts where the EMTs were equipped with pacemakers. The control group included all such patients in the 23 fire districts where the EMTs were not equipped with pacemakers. Patients in both groups received basic cardiopulmonary resuscitation, positive-pressure ventilation with supplemental oxygen, defibrillation with either automated external defibrillators or conventional defibrillators, and intubation and intravenous medications from paramedics on their arrival. In addition, personnel in units equipped with pacemakers were instructed to institute pacing in patients who converted from ventricular fibrillation to asystole after defibrillation (post-defibrillation asystole). Patients were included in the analysis if emergency personnel initiated resuscitation efforts or continued the efforts initiated by bystanders.
It was not possible to obtain informed consent from the patients in a study of this nature. The study design was reviewed and approved by the Human Subjects Review Board of the Seattle-King County Department of Public Health.
Pacing Devices
Emergency personnel used three different pacemakers during this study. The Zoll NTP transcutaneous pacemaker (Zoll Medical, Cambridge, Mass.) is a free-standing device that does not perform defibrillation. Two adhesive pacing electrodes are applied, one on the surface of the chest and one on the posterior aspect. The device delivers operator-adjusted currents up to 140 milliamperes (measured against a 50-ohm test load), at rates up to 180 beats per minute. In this study, for patients in cardiac arrest, the devices were set to deliver the pacing stimulus at a fixed rate, without reference to any electrocardiographic signal (asynchronous pacing).
The other pacemakers were the QuickPace (Physio-Control, Redmond, Wash.) and HeartAid 97 (Cardiac Resuscitator, Lake Oswego, Oreg.). The HeartAid 97 differed from the other devices in that it delivered the pacing stimulus through the same adhesive pads used to monitor the patient's heart rate and to deliver defibrillatory shocks to the sternal apex. The HeartAid delivered pacing impulses of 100 milliamperes at a rate of 80 beats per minute. The pulse duration of the Zoll NTP was 40 msec, that of the QuickPace 20 msec, and that of the HeartAid 97 10 msec. During the analysis of the data, we searched for any differences in outcome related to the use of the various devices.
Treatment Protocols
The detailed treatment protocols that the EMTs followed for cardiac arrest have been published previously30,33,34,35. We trained 397 EMTs to perform transcutaneous pacing in two-to-three-hour classes. We authorized the EMTs to institute pacing only in patients with asystole, which we defined as the absence of electrical activity greater than 1 mm (calibrated to 10 mm per millivolt) for at least 15 seconds36.
The EMTs started pacing by attaching the pacing electrodes, making two cable connections, and pressing two to three device controls. The transcutaneous pacemakers were set at 80 beats per minute, with the current at 80 to 100 milliamperes. The EMTs increased the current level until the monitor screen displayed a response to the pacing stimulus, or until they reached the maximal current setting (360 milliamperes). The EMTs continued pacing for one minute, after which they checked for a spontaneous pulse and assessed the heart rhythm. Chest compressions were performed simultaneously with the transcutaneous pacing. EMTs timed the chest compressions with the pectoral-muscle contractions from the pacing stimulus. If there was no cardiac rhythm and no pulse, these cycles were repeated until the paramedics assumed the care of the patients. The paramedics performed endotracheal intubation and administered intravenous medications to all patients, whether or not they had undergone pacing.
Collection of Data and Clinical Outcomes
The paramedics and EMTs completed a medical incident report on all patients with cardiac arrest whom they treated. In this form they noted clinical variables such as age and sex, whether the arrest had been witnessed, and whether bystanders had initiated cardiopulmonary resuscitation, as well as variables related to the emergency-medical-services system, such as the initial recorded rhythm, the time of collapse, and event-to-event time intervals (from the collapse or call to the start of cardiopulmonary resuscitation or the arrival of rescue personnel).
In all resuscitation attempts, the EMTs were required to attach their defibrillators, which contain dual-channel tape recorders to record voices and the patient's heart rhythms. We reviewed these recordings for details related to transcutaneous pacing, including the response of the cardiac rhythm and the time of initiation and duration of pacing. Research assistants obtained additional information on every patient through telephone or personal interviews with the emergency-medical-services personnel.
We measured two outcome variables: admission to the hospital and survival to discharge. Admission to the hospital required that the patient leave the emergency department with a spontaneous circulation and no continuing cardiopulmonary resuscitation and that he or she survive in the intensive care unit or coronary care unit for at least six hours. Survival to discharge meant the patient was discharged from the hospital either to his or her home or to an extended-care facility.
Statistical Analysis
All analyses were performed according to the principle of intention to treat. We calculated the odds ratios and the 95 percent confidence intervals for the two outcome variables. We analyzed differences in discrete demographic or system-related variables using two-way chi-square tests, and differences in continuous variables using the Mann-Whitney rank test.
Results
Overall Frequency of Cardiac Rhythms and Clinical Outcomes
During the study period, 1056 cardiac arrests occurred, 506 in the intervention districts and 550 in the control districts (Table 1); 39 percent of the patients in the intervention districts had ventricular fibrillation as the initial identified rhythm, as compared with 43 percent in the control districts. Asystole was identified with similar frequency in the intervention districts (55 percent) and the control districts (47 percent). Among patients whose initial rhythm was ventricular fibrillation, more in the control districts were admitted to the hospital than in the intervention districts (Table 1).
|
During the study period, 278 patients with asystole as the first identified rhythm were treated in the intervention districts (Table 2). EMTs initiated pacing for 112 of these patients. EMTs did not initiate pacing for the remaining 166 of these patients for several reasons, including a staggered training schedule, a hesitancy to initiate pacing when paramedics were near, and the steady accumulation of negative outcomes. The lack of pacing was unrelated to technical problems with the devices or to selection of particular patients for pacing on clinical grounds. Twenty-two of the 278 patients (8 percent) were admitted to the hospital; 11 (4 percent) were discharged alive. In the control districts, 259 cardiac arrests occurred in which the initial rhythm was asystole; 21 of the 259 (8 percent) were admitted, and 5 (2 percent) were discharged alive. The odds ratio of either admission (0.97; 95 percent confidence interval, 0.52 to 1.8) or survival to discharge (2.05; 95 percent confidence interval, 0.72 to 5.82) was not significantly different from 1.0 for patients treated by the intervention departments as compared with patients treated by the control departments.
|
In the intervention districts, 196 patients had ventricular fibrillation as their first identified rhythm; 134 of these patients received defibrillation, after which the cardiac rhythm was asystole. EMTs then initiated pacing for 46 of these patients (Table 1 and Table 2). Of the 46 patients who underwent pacing, 7 (15 percent) were admitted, and 2 (4 percent) survived to discharge. In the intervention districts, 88 patients had asystole after defibrillation but did not undergo pacing by the EMTs. Of these 88 patients, 11 (12 percent) were admitted, and 1 (1 percent) was discharged.
In the control group 109 patients were initially in ventricular fibrillation but post-defibrillation asystole developed. Of this group, 11 (10 percent) were admitted to the hospital, and 2 (2 percent) were discharged alive. The odds ratios for admission (1.4; 95 percent confidence interval, 0.62 to 3.1) and discharge (1.2; 95 percent confidence interval, 0.2 to 7.5) were not significantly different from 1.0, indicating that the intervention and control groups did not differ significantly. However, the confidence intervals were wide, reflecting our small samples.
The data in Table 2 allow us to test whether asystole of brief duration (post-defibrillation asystole) responds better to pacing than primary asystole, which has been present for many more minutes. The small sample prevents adjustment for covariates, but the data suggest that asystole of brief duration after defibrillation carries the same grim prognosis as primary asystole that is present for much longer periods; survival was 4 percent (2 of 46) among patients who received pacing for post-defibrillation asystole and 2 percent (2 of 112) among those who underwent pacing with an initial rhythm of asystole (Table 2).
Comparability of the Study Groups
Table 3 compares the study groups in terms of the characteristics of the patients and the emergency-medical-services systems. There were no significant differences between the two study groups in any of these variables.
|
The overall median time from collapse to the start of pacing by EMTs was 9.0 minutes. We compared the outcomes for patients who received pacing earlier (<9 minutes; n = 78) and those who received it later ( 
9 minutes; n = 80). Admission rates for patients who received pacing earlier (8 of 78 [10 percent]) were not significantly different from those for patients who received it later (5 of 80 [6 percent]), nor were discharge rates different for early pacing (2 of 78 [2.6 percent]) and late pacing (2 of 80 [2.5 percent]).
Lack of Anecdotal Success
Of the 158 patients paced by the EMTs, only 13 (8 percent) were admitted to a hospital, and only 4 (3 percent) were discharged alive. We carefully examined the records of these 13 patients for evidence that the pacemaker produced a measurable pacemaker-dependent blood pressure during the resuscitation attempt. This occurred in only 2 (1 percent) of the patients who underwent pacing. In the other 11 survivors, spontaneous circulation developed only after intubation and the administration of intravenous medications. Pacing appeared to be coincidental to survival in this small group. Emergency personnel reported no anecdotal successes with pacemaker therapy during the course of the study.
Discussion
In our study the initiation of transcutaneous pacing by EMTs did not improve survival for patients with out-of-hospital asystolic cardiac arrest. For prehospital transcutaneous pacing to improve outcomes, it would have to be performed very soon after the arrest. EMTs are the first emergency personnel to assist patients with out-of-hospital cardiac arrest. Early defibrillation by EMTs and other early-response emergency personnel is now accepted as standard therapy and is widely practiced30,33,37,38,39,40. The clinical issue we addressed in this study was whether transcutaneous pacing should be added to the armamentarium of early-response emergency personnel. Our data provided no support for the use of pacing by EMTs, despite the current availability of combination defibrillator-pacemaker devices.
Our study also addressed the question of whether transcutaneous pacing would be more effective if performed as one of the first resuscitation interventions, before endotracheal intubation and the administration of intravenous medications. Most clinical studies have reported on the use of transcutaneous pacing as a late intervention, after other efforts have failed. Such an approach would condemn transcutaneous pacing to inevitably poor outcomes. Our data demonstrated that early transcutaneous pacing for asystolic cardiac arrest, performed before intubation and the administration of medications, was associated with the same poor survival as standard therapy. We were particularly disappointed to observe that pacing performed rapidly for post-defibrillation asystole produced no better results than pacing in patients who were initially in asystole.
One could argue that we have not tested the value of very early transcutaneous pacing. The median time from collapse to the arrival of the EMTs was six minutes. The overall median interval from collapse to the start of pacing by EMTs was nine minutes. We think it would be difficult for any emergency-medical-services system to achieve shorter intervals from the time of arrest to the start of pacing.
We observed no differences in outcome among the three different transcutaneous pacemaker devices. Differences in the operation of the devices were addressed by device-specific training. Any preferences for one device over another reflected differences of opinion among the operators rather than important distinctions among types of equipment41,42.
We acknowledge the limitations of this study. The EMTs in the intervention districts failed to initiate pacing for more than half the patients in primary asystole (112 of 278 patients underwent pacing). This raises the question of whether the EMTs may have intentionally selected certain patients to receive either pacing or standard therapy. We do not think the EMTs did this, however. We trained the personnel in the fire districts in staggered groups, rather than all at once. The intervention districts began to contribute to the data collection as soon as we began training. Some cardiac arrests occurred in these departments and were treated by EMTs who had not yet been trained in the use of the pacemakers. We learned also that EMTs would sometimes forgo pacing in two sets of circumstances: toward the end of the study, as they became aware of the discouraging results, and whenever paramedics were close at hand. Statistical propriety required that we analyze survival not among the patients who actually underwent pacing by the EMTs but among all patients treated by EMTs in the intervention departments (the total intervention group, or the intention-to-treat group). Although it is possible that certain patients in the intervention group may have been selected for nontreatment, such selection did not produce differences in survival rates.
A twofold increase in survival with early transcutaneous pacing for patients in asystole would be compatible with our data, but our observed odds ratio of 2.05 did not achieve statistical significance. A much larger study might have confirmed a significant increase in survival, but we believe that the low overall survival rate with pacing does not justify widespread implementation of this technique.
A driving force behind the interest in transcutaneous pacing has been the dismal record of success in treating asystole. Approximately 300,000 to 400,000 people die each year from sudden cardiac arrest outside of hospitals43. Depending on the length of time from collapse to the arrival of emergency personnel, ventricular fibrillation is found in 50 to 60 percent of patients with a witnessed cardiac arrest44,45,46,47. Most of the remaining 120,000 to 200,000 people have asystole, either initially or after deterioration from some other arrest rhythm. Published reports state that patients with asystole, whether primary or secondary, almost never survive46,47,48,49,50,51. Reported rates of resuscitation are only 1 to 4 percent, and even these may be overestimated because of misclassification.
Though there may be rare exceptions, asystole appears to be not a cardiac rhythm but an indicator of a dead heart. An asystolic heart has exhausted all the myocardial energy stores, and all electrical activity vanishes52,53,54,55. The chance that an asystolic heart will respond to any intervention, even direct electrical stimulation, appears to be vanishingly small. In our data, transcutaneous pacing was associated with no statistically significant improvement in rates of admission or survival.
Transcutaneous pacing was first developed in the 1950s in an attempt to solve the rare but tragic problem of Stokes-Adams attacks56,57,58. The first successes occurred when the pacing stimulus was delivered through small needle electrodes inserted beneath the skin56. During the 1960s and 1970s transcutaneous pacemakers were not manufactured, but in the 1980s such devices began to reappear either as an add-on feature with defibrillators or as separate free-standing pacemakers3,17,41,59,60,61. By that time engineers had overcome several technical and clinical problems associated with the first generation of transcutaneous pacing devices1,41,61,62,63.
Despite an increase in the availability of transcutaneous pacers in the 1980s, efforts to treat patients in full cardiac arrest with transcutaneous pacing have not been successful. Clinical studies of pacing have been plagued by the inability to start pacing soon after the heart stops. The shortest average interval from cardiac arrest to the start of pacing in any report has been about 20 minutes. The possibility that a stunned and exhausted myocardium might respond to pacing stimuli more than 20 minutes after the cardiac arrest approaches zero.
The results of this study, plus the experience of many other clinicians and researchers, suggest that out-of-hospital transcutaneous pacing does not materially improve survival for patients with asystolic cardiac arrest, including post-defibrillation asystole. All clinicians and emergency personnel are driven by a desire to give the patients they treat the best possible chance of survival, and transcutaneous pacemakers appeal to this desire. There are more than 50,000 ambulances staffed by paramedics and approximately 450,000 ambulances staffed by EMTs in the United States. Transcutaneous pacemakers cost $4,000 to $7,000. Outfitting every ambulance in the country with a transcutaneous pacemaker would constitute a major health expenditure. Our data suggest that transcutaneous pacing should not be a part of the routine treatment of patients with out-of-hospital asystolic cardiac arrest.
Supported by a grant (HS05740-03) from the Agency for Health Care and Policy Research and by research funds from the Center for Evaluation of Emergency Medical Services, Seattle.
We are indebted to the EMTs and paramedics of King County, Washington, for their important contribution to this study and for their interest in improved patient care; to Mickey Eisenberg, co-director of the Center for Evaluation of Emergency Medical Services, Steve Call, manager of the Emergency Medical Services Division, and Christy Horton, medical director of the King County Paramedics; and to William Marsh, paramedic coordinator, and William Lohr and Jane Scott, Agency for Health Care Policy and Research project coordinators, for their encouragement and support.
Source Information
From the Center for Evaluation of Emergency Medical Services, Department of Medicine, University of Washington Medical Center, Seattle (R.O.C., J.R.G.); the King County Emergency Medical Services Division, Seattle-King County Department of Public Health, Seattle (R.O.C., J.R.G., M.P.L., T.R.H., R.M.N., S.H.); the Department of Biostatistics, University of Washington School of Public Health, Seattle (A.P.H.); and the Overlake Memorial Hospital, Bellevue, Wash. (J.C.). Presented in part at the 64th Scientific Sessions of the American Heart Association, Anaheim, Calif., November 11-14, 1991.
Address reprint requests to Ms. Graves at the King County Emergency Medical Services Division, Seattle-King County Department of Public Health, 110 Prefontaine Pl. S., Suite 500, Seattle, WA 98104.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Transcutaneous Pacing in Patients with Asystolic Cardiac Arrest
Armon C., Cummins R. O.
Extract |
Full Text
N Engl J Med 1993;
329:1277, Oct 21, 1993.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
