- Email: [email protected]
Prolonged resuscitation efforts for cardiac arrest patients who cannot be resuscitated at the scene: Who is likely to benefit?
ORIGINAL CONTRIBUTION
cardiac arrest, prognosis cardiwulmonary resuscitation
Krolonged Resuscitation Efforts for Cardiac Arrest Patients Who Cannot Be Resuscitated at the Scene: Who Is Likely to Benefit?
From the Department of General Internal Medicine, Medical Intensive Care Unit, University Hospital; and Emergency Medical Services System, Leiden, The Netherlands. Receivedfor publication June 19, 1992. Revision received October 16, 1992. Acceptedfor publication December 22, 1992.
Johannes G van der Hoeven, MD
Study objective: To determine who may benefit from
Hendrik Waanders, MD
prolonged resuscitation efforts after therapy by emergency medical services system (EMS) personnel has failed to restore vital signs.
Elizabeth A Compier, MD Pepita KC van der Weyden, MD Arend E Meinders, MD, PhD
Design: Retrospective chart review. Type o f participants: Two hundred
sixteen consecutive adult patients with out-of-hospital cardiac arrest who were admitted to the emergency department without vital signs. Methods: Identification of prehospital resuscitation data, therapy in the ED, hospital course, and final outcome.
Results: Thirty-nine patients (18.1%)were resuscitated successfully. The odds ratio of successful resuscitation in the ED for the patients with ventricular fibrillation at the scene versus those with asystole or electromechanical dissociation was 3.4 (95% confidence interval, 1.5, 7.9). All patients with asystole or electromechanical dissociation, either at the scene or in the ED, died (95% confidence interval, 0, 4.3). Conclusion: Prolonged resuscitation efforts in the ED for patients with asystole or electromechanical dissociation usually are futile after previous efforts by the EMS personnel have failed to restore vital signs. Transportation to the hospital may not be indicated. However, for patients with persistent ventricular fibrillation, transport is indicated. [van der Hoeven JG, Waanders H, Compier EA, van der Weyden PKC, Meinders AE: Prolonged resuscitation efforts for cardiac arrest patients who cannot be resuscitated at the scene: Who is likely to benefit? Ann EmergMed November 1993;22:1659-1663.]
16/1659
ANNALSOFEMERGENCYMEDICINE 22:11 NOVEMBER1993
RESUSCITATION van der Hoeven et al
INTRODUCTION
Patients with an out-of-hospital cardiac arrest have a dismal prognosis, with only 2% to 25% surviving to hospital discharge. 1-5 Several factors have been shown to improve outcome: the presence of a witnessed arrest, immediate bystander CPR, and an initial cardiac rhythm of ventricular fibrillation, followed by early defibrillation. Most survivors of an out-of-hospital cardiac arrest have stable vital signs on arrival in the emergency department. If vital signs have not been restored by the ambulance personnel, prolonged resuscitation efforts usually are futile. Some of these patients may be resuscitated, but cerebral damage usually is severe, and only a few will survive to discharge.6, r Gray and colleagues studied 185 out-of-hospital cardiac arrest victims without vital signs on arrival at the ED.s Sixteen patients were resuscitated successfully, but none survived to hospital discharge. The accompanying editorial stated that if a resuscitation attempt outside the hospital is unsuccessful, further in-hospital procedures are useless and should be abandoned in all but exceptional casesP Oiher important considerations are the financial aspects and the mental burden inflicted on the patient's family during a long ICU stay, with an eventual fatal outcome. To assess who might benefit from continued resuscitation efforts, we reviewed the case histories of 216 consecutive patients with out-of-hospital cardiac arrest who were admitted to our ED without vital signs. MATERIALS AND METHODS
Leiden University Hospital is an 875-bed teaching hospital. It functions as the principal tertiary care referral center for the Leiden area. All cardiac arrest victims preferably are transported to our hospital. The area served is i00 km 2 and has a population of 196,193 (as of January 1990). The male-to-female ratio is 1:1.06. Twelve percent of the population is more than 65 years old. The Leiden area is served by one emergency medical services (~!MS) system, which includes the fire and police department. A four-digit public alarm telephone number (06-11) communication system is used. The median call receipt-to-mobile vehicle interval is less than one minute. Ambulance personnel are nurses trained in intensive or coronary care medicine. When confronted with a patient in cardiac arrest, they institute basic life support (BLS) with manual chest compression and technical airway management (bag-valve-mask). The EMS system authorizes advanced cardiac life support (ACLS), which includes, endotracheal intubation, defibrillation with a conventional
NOVEMBER1993 22:11 ANNALS OF EMERGENCY MEDIC]NE
manual defibrillator, and the use of IV pharmacologic agents such as epinephrine, atropine, lidocaine, and sodium bicarbonate. The protocols in use adhere to those recommended by the American Heart Association. The median time from arrival at the patient side until the first defibrillatory attempt is three minutes. No direct telephone communication is possible between ambulance personnel and the physician in the ED. The EM5 system permits the ambulance personnel to cease resuscitation efforts in the field if extensive resuscitation efforts result in persistent asystole. We reviewed the records of all patients with a persistent out-of-hospital cardiac arrest who were admitted to the ED between January 1, 1989, and January 31, 1992. Patients were excluded from the study if their arrests were due to trauma or drowning or if they were less than 15 years old. Prehospital data concerning location and circumstances of arrest, cardiac rhythm at the scene, presence of bystander CPR, time intervals, and therapy were retrieved from the EMS records. The ElVlS system maintains typed reports of every attempted resuscitation. All time intervals are recorded automatically and none was missing. The callresponse interval is defined as the time between the call receipt and arrival of the ambulance personnel at the patient. The call-hospital interval is defined as the time between the call receipt and arrival of the patient in the ED. The duration of the cardiac arrest before the emergency call receipt usually is difficult to estimate. Therefore, we used the presence of a witnessed arrest and bystander CPR as an estimate of the delay before adequate resuscitation was started. Patients with a witnessed arrest and bystander CPR were thought to have the best prognosis. Hospital records were reviewed for demographic data, cardiac rhythm in the ED, therapy in the t!D, duration of the resuscitation attempt, outcome in the tiD, days in the ICU, duration of hospital stay, and final outcome. The medical history for patients who died in the KD often was incomplete and is not reported. The patients were divided into three categories: patients who died in the ED after a continued resuscitation attempt by the emergency physician, patients who were resuscitated successfully and admitted to the ICU, and hospital survivors. The cause of the cardiac arrest was determined from the circumstances surrounding the event or from postmortem examination. For patients who left the hospital, the Glasgow-Pittsburgh Cerebral Performance and Overall Performance Categories were recorded at the time of discharge and three months or one year later.~O,~
1660/17
RESUSCITATION van der Hoeven et ctI
Statistical analysis was performed using the Number Cruncher Statistical System software package (Version 5.0, 1987, by Dr,jerry L Hintze). To assess differences between patient groups, we used Fisher's exact test or the Z 2 test for categorical data and Student's t-test for continuous data. All differences were considered statistically significant at P < .05. If P < .05, 95% confidence intervals (CIs)
are provided. The odds ratio of hospital discharge is given for the presenting cardiac rhythm in the ED. Summary data are presented as numeric counts, percentages, and mean + SD values, unless noted otherwise. RESULTS
From,january 1, 1989, through January 31, 1992, 216 consecutive patients (158 male and 58 female) with outof-hospital cardiac arrest were admitted to the ED. The mean + SD age was 61.7 + 15.3 years (range, 15 to 88 years). In all patients, advanced resuscitation protocols by the ambulance personnel had failed to restore vital signs. During the sLudy period, 41 patients who had CPR in the field were not transported to the hospital. Seventy-two patients were field resuscitated; 35 of these patients (48.6%) survived to hospital discharge. The cause of arrest was cardiac in 197 patients (91.2%) and did not differ among patient categories. Ventricular fibrillation was the predominant cardiac rhythm at the scene (57.9%), followed by asystole (25.9%) and electromechanical dissociation (16.2%). On arrival at the ED, Figure 1.
Final outcome o~persistent out-of-hospital cardiac arrest in 216 patients
this changed to 23.6%, 48.2%, and 28.2%, respectively. In the ED, resuscitation efforts were continued in all 216 patients following the 1986 ACLS guidelines. 12 All patients had been or were intubated endotracheally, 97.7% received IV medications, and 46.3%,received defibrillatory shocks. In 13 patients (6%) with asystole, transthoracic or transvenous pacing was attempted, but pulses were established in none. One hundred seventyseven patients died in the ED after mean in-hospital resuscitation attempts of 24.1 + 12.4 minutes (range, one to 65 minutes). Thirty-nine patients were transferred to the medical ICU. Thirty-three patients, including all patients with an initial cardiac rhythm of asystole or electromechanical dissociation, died in the hospital after a median stay of two days (range, zero to 34 days). The cause of death in these 33 patient s was severe postanoxic encephalopathy followed by withdrawal of life-sustaining therapy (22), cardiogenic shock (eight), and recurrent cardiac arrest (three). Only six patients survived to hospital discharge (Figure 1). All six had an initial cardiac rhythm of ventricular fibrillation and received prompt bystander CPR after a witnessed cardiac arrest. Five of the six received ACLS by the ambulance personnel, but all failed to respond at any point. The median ICU stay for survivors was two days (range, one to 42 days), with a median hospital stay of i7.5 days (range, 13 to 91 days), Four of six survivors were discharged home with Glasgow-Pittsburgh Cerebral Performance and Overall Performance Categories the same as before the cardiac arrest. Two were discharged to a rehabilitation facility with initially severe neurologic Table 1.
Characteristics of patients and outcomes ~n the ED De
Survived to discharge 6
in ICU
}3
18/1661
Age (mean _+SDyr) Male (%) Witnessed arrest (%) Bystander CPR (%) Call-response interval (rain) Median Mean _+SD Call-hospital interval (rain) Median Mean _+SD Initial cardiac rhythm (%) Ventricular fibrillation Asystole Electromechanical dissociation Prehospital ACLS (%)
Died in ED (N = 177)
Resuscitated (N = 39)
P
61.8 _+16 129 (72,9) 151 (85.3) 116 (65.5)
61.5 _+11.4 29 (74.4) 31 (79.5) 24 (61.5)
.891 .835 .366 .636
5 5.0 +3.0
4 4.4 _+3.1
.261
34 34 _+9.9
30 30.6 +11.7
.097
31 (79.5) 6 (15.4) 2 (5.1) 31 (79.5)
.002 .097 .038 .895
94 (53,1) 50 (28.3) 33 (18.6) 139 (78.5)
ANNALS OF EMERGENCY MEDICINE
22:11
NOVEMBER1993
RESUSCITATION van der Hoeven et al
defects. However, both patients were much improved by one year after the incident and were able to lead acceptably independent lives. Demographic and prehospital data for those who died in the ED and those who were resuscitated successfully are presented (Table 1). Call-response intervals are shown (Figure 2). The call-response interval was defined as zero if cardiac arrest occurred after arrival of the ambulance personnel (31 patients). There was no difference in resuscitation rate for witnessed arrests with bystander CPR (17%), witnessed arrests without bystander CPR (18%), and unwitnessed arrests (24%). The only significant difference between those who died in the ED and those who were resuscitated successfully was the cardiac rhythm at the scene. Successfully resuscitated patients more often had ventricular fibrillation and correspondingly less often had electromechanical dissociation. The odds ratio of successful resuscitation in the ED for the patients who had ventricular fibrillation at the scene versus those with asystole electromechanical dissocation was 3.4 (95% CI, 1.5, 7.9). This difference was even more pronounced for ventricular fibrillation on arrival at the ED versus asystole and electromechanical dissociation (odds ratio of successful resuscitation, 5.7; 95% CI, 2.7, 12.0). Table 2 shows the patients who survived until hospital discharge in relation to the cardiac rhythm on arrival at the ED. Survival was significantly different for patients Figure 2. CalI-response interval for the study population. The callresponse interval was defined as 0 if the cardiac arrest occurred in the presence of the ambulance personnel. No. of patients 30
•
Dead in ED
[ ] Resuscitated 25
with ventricular fibrillation than for patients with asystole or electromechanical dissociation (P = .00014, Fisher's exact test). Comparison of other prehospital resuscitation data (eg, age, sex, witnessed arrest, bystander CPR, cardiac rhythm at the scene, call-response and call-hospital intervals) and medical history (eg, myocardial'infarction, congestive heart failure, chronic obstructive pulmonary disease, malignancy) between paLients who died in the ICU and final survivors showed no significant differences. DISCUSSION
Out-of-hospital cardiac arrest has a dismal prognosis, especially when resuscitation attempts by the ambulance personnel fail to restore vital signs. Continued efforts in the ED usually are futile. Kellermann and colleagues described 240 cases, of which 32 (13.3%) were resuscitated successfully.6 All received ACLS by the ambulance personnel, but other prehospital data were not reported. Patients with ventricular fibrillation in the ED were resuscitated more often than were patients with other rhythms. Only four (1.6%) survived until hospital discharge (two of them with severe neurologic deficits). 6 Bonnin and Swor reported that of 181 patients, ten (6%) were resuscitated successfully.r However, only one patient (0.6%) survived to hospital discharge. The authors could find no subgroup of patients failing to respond to prehospital ACLS efforts who had a better chance of successful resuscitation or hospital discharge, r Another recent retrospective study by Gray and colleagues described 185 cases, of which 16 (9%) were resuscitated, s No patient survived to hospital discharge. Accounts of witnessed versus unwitnessed arrests and of the initiation of bystander CPR were not available. The call-response inLerval and the cardiac rhythm at the scene were reported in only 44% and 62%, respectively. Sixty-two percent received ACLS. Although the authors stated that asystole and electromechanical dissociation probably represent terminal rhythms, they found no difference in the cardiac rhythm between those who died and those who were resuscitated, either at the scene or in the ED. Table 2. Number of discharged patients and cardiac rhythm Resuscitation Outcome Cardiac Rhythm atthe ED
0
1
2
3
4
5
6
7
8
9
10
11
12 13
Call-response interval (min)
NOVEMBER1993
22:11
ANNALS OF EMERGENCY MEDICINE
14
15
Survived to Hospital Discharge
Ventricular fibrillation As~/stole or electromechanical dissociation
Nonsurvivors 45 165
P= .0oo14 by one-sided Fisher'sexacttest.
1662/19
RESUSCITATION van der Hoeven et aI
We studied 216 patients with a persistent out-ofhospital cardiac arrest. In contrast with the three aforementioned studies, we were abIe to retrieve all prehospital resuscitation data. We compared the 177 patients who died in the ED with the 39 who survived to leave the ED. The 18.1% resuscitation rate compares favorably with the study by Gray and colleagues. We found no significant differences in the percentage of witnessed arrests, patients receiving bystander CPR, patients receiving prehospital ACLS, or call-response intervals. This can be explained by selection bias because not all patients who responded to prehospital resuscitation efforts are incIuded in the study population. Furthermore, the long call-hospital interval in both groups tends to diminish the importance of these prognostic factors. A 30-minute resuscitation attempt without restoration of a spontaneous circulation almost uniformly is fatal. 13 Patients with ventricular fibrillation at the scene had the highest chance of successful resuscitation in the ED. If ventricular fibrillation was still present on arrival at the ED, the chance for successful resuscitation was 41.2%. This was significantly'higher than that for asystole or electromechanicaI dissociation. Thirty-three of 39 resuscitated patients died in the hospital. The most frequent cause of death in the ICU was severe postanoxic encephalopathy followedby withdrawal of life-sustaining therapy (67%). None of these patients ever regained consciousness. This also reflects the long period without spontaneous circulation. Only six patients (2.8%) survived to hospital discharge. Two of them initially had severe neurologic deficits, but both were much improved one year after the incident. In these cases, the cardiac rhythm at the scene and in the ED was ventricular fibrillation. Of all patients with ventricular fibrillation in the ED, 12% were discharged alive. All patients with asystole or electromechanicaI dissociation either at the scene or in the ED died. Our results support the view of asystole and electromechanical dissociation as terminal rhythms; they denote a prolonged ischemic period. Although restoration of a stable cardiac rhythm is sometimes possible, severe cerebral damage precludes long-term survivai. New therapies such as the early administration of high-dose norepinephrine may increase the number of patients with a return of spontaneous circulation, but the chances for cerebral recovery are likely Lo be small. > 4 6
unsuccessful, and transport of these patients to the hospital may not be indicated. However, transport for patients with persistent ventricular fibrillation is indicated, as 12% survived to hospital discharge in our study. REFERENCES 1. Eisenberg MS, Hallstrom A, Bergner LB: Long-term survival after out-of-hospital cardiac arrest. N EnglJ Mad 1982;306:1340-1343. 2. Ritter G, Wolfe RA, Goldstein S, et ai: The effect of bystander CPH on survival of out-ofhospitar cardiac arrest victims. Am HeartJ 1985;110:932-937. 3. E[senberg MS, Horwood BT, Cummins RO, et al: Cardiac arrest and resuscitation: A tale of 29 cities. Ann EmergMeal1990;19:179-186. 4. Kentsch M, Stendel M, Berket H, et al: Early prediction of prognosis in out-of-hospital cardiac arrest. IntoneCareMad 1990;16:378-383. 5. Longstreth WT, Cobb LA, Fahranbruch CE, at al: Does age affect outcomes of out-of-hospital car@pulmonary resuscitation? JAMA 1990;264:2109-2110. 6. Kellermann AL, Staves DR, Hackman BB: in-hospital resuscitation following unsuccessful prehospitar advanced cardiac life support: "Heroic efforts" or an exercise in futility? Ann Emarg Marl 1988;17:589-594. 7. Bonnin M J, Swor R~,: Outcomes in unsuccessful field resuscitation attempts. Ann EmergMad 1989;18:507-512. 8. Gray WA, Capone RJ, Most AS: Unsuccessful emergency medical resuscitation--Are continued efforts in the EO justified? N EnglJ Mad 199I ;325:1393-1398. 9. Weaver WD: Resuscitation outside the hospital--What's lacking? N EnglJ Mad 199I;325:1437-1439. 10. Jennett B, Bond M: Assessment of outcome after severe brain damage: A practical scale. Lancet1975;1:480-484. 11. Brain Resuscitation Clinical Trial 1 Study Group: Randomized clinical study of thiopental loading in comatose survivors ef cardiac arrest. N EnglJ Mad 1986;314:397-403. 12. Standards and guidelines for cardiopu[monary resuscitation (CPR)and emergency cardiac care (ECC).JAMA 1986;255:2905-2984. 13. Weaver WD, Cobb LA, Hallstrom AP, et al: Considerations for improving survival from out-e# hospital cardiac arrest. Ann EmergMad 1986;15:1181-1188. 14. Martin D, Werman HA, Brown CG: Four case studies: High-dose epinephrine in cardiac arrest. Ann EmergMad 1990;19:322-326. 15. Barton C, Callaham M: High-dose epinephrine improves the return of spontaneous circulation rates in human victims of cardiac arrest. Ann EmergMad 1991;20:722~725. 16. Cai[aham M, Barton CW, Kayser S: Potential complications of high-dose epinephrine therapy in patients resuscitated from cardiac arrest. JAMA 1991;265:1117-1122.
Address for reprints: Johannes Gvan der Hoeven,MD Medical IntensiveCareUnit, C6-Q University Hospital Leiden PO Box 9600 2300 RCLeiden The Netherlands
CONCLUSION
Prolonged resuscitatio n efforts in the ED for patients with asystole or electromechanical dissociation usually are
20/1663
ANNALS OF EMERGENCY MEDICINE
22:11
NOVEMBER1993
cardiac arrest, prognosis cardiwulmonary resuscitation
Krolonged Resuscitation Efforts for Cardiac Arrest Patients Who Cannot Be Resuscitated at the Scene: Who Is Likely to Benefit?
From the Department of General Internal Medicine, Medical Intensive Care Unit, University Hospital; and Emergency Medical Services System, Leiden, The Netherlands. Receivedfor publication June 19, 1992. Revision received October 16, 1992. Acceptedfor publication December 22, 1992.
Johannes G van der Hoeven, MD
Study objective: To determine who may benefit from
Hendrik Waanders, MD
prolonged resuscitation efforts after therapy by emergency medical services system (EMS) personnel has failed to restore vital signs.
Elizabeth A Compier, MD Pepita KC van der Weyden, MD Arend E Meinders, MD, PhD
Design: Retrospective chart review. Type o f participants: Two hundred
sixteen consecutive adult patients with out-of-hospital cardiac arrest who were admitted to the emergency department without vital signs. Methods: Identification of prehospital resuscitation data, therapy in the ED, hospital course, and final outcome.
Results: Thirty-nine patients (18.1%)were resuscitated successfully. The odds ratio of successful resuscitation in the ED for the patients with ventricular fibrillation at the scene versus those with asystole or electromechanical dissociation was 3.4 (95% confidence interval, 1.5, 7.9). All patients with asystole or electromechanical dissociation, either at the scene or in the ED, died (95% confidence interval, 0, 4.3). Conclusion: Prolonged resuscitation efforts in the ED for patients with asystole or electromechanical dissociation usually are futile after previous efforts by the EMS personnel have failed to restore vital signs. Transportation to the hospital may not be indicated. However, for patients with persistent ventricular fibrillation, transport is indicated. [van der Hoeven JG, Waanders H, Compier EA, van der Weyden PKC, Meinders AE: Prolonged resuscitation efforts for cardiac arrest patients who cannot be resuscitated at the scene: Who is likely to benefit? Ann EmergMed November 1993;22:1659-1663.]
16/1659
ANNALSOFEMERGENCYMEDICINE 22:11 NOVEMBER1993
RESUSCITATION van der Hoeven et al
INTRODUCTION
Patients with an out-of-hospital cardiac arrest have a dismal prognosis, with only 2% to 25% surviving to hospital discharge. 1-5 Several factors have been shown to improve outcome: the presence of a witnessed arrest, immediate bystander CPR, and an initial cardiac rhythm of ventricular fibrillation, followed by early defibrillation. Most survivors of an out-of-hospital cardiac arrest have stable vital signs on arrival in the emergency department. If vital signs have not been restored by the ambulance personnel, prolonged resuscitation efforts usually are futile. Some of these patients may be resuscitated, but cerebral damage usually is severe, and only a few will survive to discharge.6, r Gray and colleagues studied 185 out-of-hospital cardiac arrest victims without vital signs on arrival at the ED.s Sixteen patients were resuscitated successfully, but none survived to hospital discharge. The accompanying editorial stated that if a resuscitation attempt outside the hospital is unsuccessful, further in-hospital procedures are useless and should be abandoned in all but exceptional casesP Oiher important considerations are the financial aspects and the mental burden inflicted on the patient's family during a long ICU stay, with an eventual fatal outcome. To assess who might benefit from continued resuscitation efforts, we reviewed the case histories of 216 consecutive patients with out-of-hospital cardiac arrest who were admitted to our ED without vital signs. MATERIALS AND METHODS
Leiden University Hospital is an 875-bed teaching hospital. It functions as the principal tertiary care referral center for the Leiden area. All cardiac arrest victims preferably are transported to our hospital. The area served is i00 km 2 and has a population of 196,193 (as of January 1990). The male-to-female ratio is 1:1.06. Twelve percent of the population is more than 65 years old. The Leiden area is served by one emergency medical services (~!MS) system, which includes the fire and police department. A four-digit public alarm telephone number (06-11) communication system is used. The median call receipt-to-mobile vehicle interval is less than one minute. Ambulance personnel are nurses trained in intensive or coronary care medicine. When confronted with a patient in cardiac arrest, they institute basic life support (BLS) with manual chest compression and technical airway management (bag-valve-mask). The EMS system authorizes advanced cardiac life support (ACLS), which includes, endotracheal intubation, defibrillation with a conventional
NOVEMBER1993 22:11 ANNALS OF EMERGENCY MEDIC]NE
manual defibrillator, and the use of IV pharmacologic agents such as epinephrine, atropine, lidocaine, and sodium bicarbonate. The protocols in use adhere to those recommended by the American Heart Association. The median time from arrival at the patient side until the first defibrillatory attempt is three minutes. No direct telephone communication is possible between ambulance personnel and the physician in the ED. The EM5 system permits the ambulance personnel to cease resuscitation efforts in the field if extensive resuscitation efforts result in persistent asystole. We reviewed the records of all patients with a persistent out-of-hospital cardiac arrest who were admitted to the ED between January 1, 1989, and January 31, 1992. Patients were excluded from the study if their arrests were due to trauma or drowning or if they were less than 15 years old. Prehospital data concerning location and circumstances of arrest, cardiac rhythm at the scene, presence of bystander CPR, time intervals, and therapy were retrieved from the EMS records. The ElVlS system maintains typed reports of every attempted resuscitation. All time intervals are recorded automatically and none was missing. The callresponse interval is defined as the time between the call receipt and arrival of the ambulance personnel at the patient. The call-hospital interval is defined as the time between the call receipt and arrival of the patient in the ED. The duration of the cardiac arrest before the emergency call receipt usually is difficult to estimate. Therefore, we used the presence of a witnessed arrest and bystander CPR as an estimate of the delay before adequate resuscitation was started. Patients with a witnessed arrest and bystander CPR were thought to have the best prognosis. Hospital records were reviewed for demographic data, cardiac rhythm in the ED, therapy in the t!D, duration of the resuscitation attempt, outcome in the tiD, days in the ICU, duration of hospital stay, and final outcome. The medical history for patients who died in the KD often was incomplete and is not reported. The patients were divided into three categories: patients who died in the ED after a continued resuscitation attempt by the emergency physician, patients who were resuscitated successfully and admitted to the ICU, and hospital survivors. The cause of the cardiac arrest was determined from the circumstances surrounding the event or from postmortem examination. For patients who left the hospital, the Glasgow-Pittsburgh Cerebral Performance and Overall Performance Categories were recorded at the time of discharge and three months or one year later.~O,~
1660/17
RESUSCITATION van der Hoeven et ctI
Statistical analysis was performed using the Number Cruncher Statistical System software package (Version 5.0, 1987, by Dr,jerry L Hintze). To assess differences between patient groups, we used Fisher's exact test or the Z 2 test for categorical data and Student's t-test for continuous data. All differences were considered statistically significant at P < .05. If P < .05, 95% confidence intervals (CIs)
are provided. The odds ratio of hospital discharge is given for the presenting cardiac rhythm in the ED. Summary data are presented as numeric counts, percentages, and mean + SD values, unless noted otherwise. RESULTS
From,january 1, 1989, through January 31, 1992, 216 consecutive patients (158 male and 58 female) with outof-hospital cardiac arrest were admitted to the ED. The mean + SD age was 61.7 + 15.3 years (range, 15 to 88 years). In all patients, advanced resuscitation protocols by the ambulance personnel had failed to restore vital signs. During the sLudy period, 41 patients who had CPR in the field were not transported to the hospital. Seventy-two patients were field resuscitated; 35 of these patients (48.6%) survived to hospital discharge. The cause of arrest was cardiac in 197 patients (91.2%) and did not differ among patient categories. Ventricular fibrillation was the predominant cardiac rhythm at the scene (57.9%), followed by asystole (25.9%) and electromechanical dissociation (16.2%). On arrival at the ED, Figure 1.
Final outcome o~persistent out-of-hospital cardiac arrest in 216 patients
this changed to 23.6%, 48.2%, and 28.2%, respectively. In the ED, resuscitation efforts were continued in all 216 patients following the 1986 ACLS guidelines. 12 All patients had been or were intubated endotracheally, 97.7% received IV medications, and 46.3%,received defibrillatory shocks. In 13 patients (6%) with asystole, transthoracic or transvenous pacing was attempted, but pulses were established in none. One hundred seventyseven patients died in the ED after mean in-hospital resuscitation attempts of 24.1 + 12.4 minutes (range, one to 65 minutes). Thirty-nine patients were transferred to the medical ICU. Thirty-three patients, including all patients with an initial cardiac rhythm of asystole or electromechanical dissociation, died in the hospital after a median stay of two days (range, zero to 34 days). The cause of death in these 33 patient s was severe postanoxic encephalopathy followed by withdrawal of life-sustaining therapy (22), cardiogenic shock (eight), and recurrent cardiac arrest (three). Only six patients survived to hospital discharge (Figure 1). All six had an initial cardiac rhythm of ventricular fibrillation and received prompt bystander CPR after a witnessed cardiac arrest. Five of the six received ACLS by the ambulance personnel, but all failed to respond at any point. The median ICU stay for survivors was two days (range, one to 42 days), with a median hospital stay of i7.5 days (range, 13 to 91 days), Four of six survivors were discharged home with Glasgow-Pittsburgh Cerebral Performance and Overall Performance Categories the same as before the cardiac arrest. Two were discharged to a rehabilitation facility with initially severe neurologic Table 1.
Characteristics of patients and outcomes ~n the ED De
Survived to discharge 6
in ICU
}3
18/1661
Age (mean _+SDyr) Male (%) Witnessed arrest (%) Bystander CPR (%) Call-response interval (rain) Median Mean _+SD Call-hospital interval (rain) Median Mean _+SD Initial cardiac rhythm (%) Ventricular fibrillation Asystole Electromechanical dissociation Prehospital ACLS (%)
Died in ED (N = 177)
Resuscitated (N = 39)
P
61.8 _+16 129 (72,9) 151 (85.3) 116 (65.5)
61.5 _+11.4 29 (74.4) 31 (79.5) 24 (61.5)
.891 .835 .366 .636
5 5.0 +3.0
4 4.4 _+3.1
.261
34 34 _+9.9
30 30.6 +11.7
.097
31 (79.5) 6 (15.4) 2 (5.1) 31 (79.5)
.002 .097 .038 .895
94 (53,1) 50 (28.3) 33 (18.6) 139 (78.5)
ANNALS OF EMERGENCY MEDICINE
22:11
NOVEMBER1993
RESUSCITATION van der Hoeven et al
defects. However, both patients were much improved by one year after the incident and were able to lead acceptably independent lives. Demographic and prehospital data for those who died in the ED and those who were resuscitated successfully are presented (Table 1). Call-response intervals are shown (Figure 2). The call-response interval was defined as zero if cardiac arrest occurred after arrival of the ambulance personnel (31 patients). There was no difference in resuscitation rate for witnessed arrests with bystander CPR (17%), witnessed arrests without bystander CPR (18%), and unwitnessed arrests (24%). The only significant difference between those who died in the ED and those who were resuscitated successfully was the cardiac rhythm at the scene. Successfully resuscitated patients more often had ventricular fibrillation and correspondingly less often had electromechanical dissociation. The odds ratio of successful resuscitation in the ED for the patients who had ventricular fibrillation at the scene versus those with asystole electromechanical dissocation was 3.4 (95% CI, 1.5, 7.9). This difference was even more pronounced for ventricular fibrillation on arrival at the ED versus asystole and electromechanical dissociation (odds ratio of successful resuscitation, 5.7; 95% CI, 2.7, 12.0). Table 2 shows the patients who survived until hospital discharge in relation to the cardiac rhythm on arrival at the ED. Survival was significantly different for patients Figure 2. CalI-response interval for the study population. The callresponse interval was defined as 0 if the cardiac arrest occurred in the presence of the ambulance personnel. No. of patients 30
•
Dead in ED
[ ] Resuscitated 25
with ventricular fibrillation than for patients with asystole or electromechanical dissociation (P = .00014, Fisher's exact test). Comparison of other prehospital resuscitation data (eg, age, sex, witnessed arrest, bystander CPR, cardiac rhythm at the scene, call-response and call-hospital intervals) and medical history (eg, myocardial'infarction, congestive heart failure, chronic obstructive pulmonary disease, malignancy) between paLients who died in the ICU and final survivors showed no significant differences. DISCUSSION
Out-of-hospital cardiac arrest has a dismal prognosis, especially when resuscitation attempts by the ambulance personnel fail to restore vital signs. Continued efforts in the ED usually are futile. Kellermann and colleagues described 240 cases, of which 32 (13.3%) were resuscitated successfully.6 All received ACLS by the ambulance personnel, but other prehospital data were not reported. Patients with ventricular fibrillation in the ED were resuscitated more often than were patients with other rhythms. Only four (1.6%) survived until hospital discharge (two of them with severe neurologic deficits). 6 Bonnin and Swor reported that of 181 patients, ten (6%) were resuscitated successfully.r However, only one patient (0.6%) survived to hospital discharge. The authors could find no subgroup of patients failing to respond to prehospital ACLS efforts who had a better chance of successful resuscitation or hospital discharge, r Another recent retrospective study by Gray and colleagues described 185 cases, of which 16 (9%) were resuscitated, s No patient survived to hospital discharge. Accounts of witnessed versus unwitnessed arrests and of the initiation of bystander CPR were not available. The call-response inLerval and the cardiac rhythm at the scene were reported in only 44% and 62%, respectively. Sixty-two percent received ACLS. Although the authors stated that asystole and electromechanical dissociation probably represent terminal rhythms, they found no difference in the cardiac rhythm between those who died and those who were resuscitated, either at the scene or in the ED. Table 2. Number of discharged patients and cardiac rhythm Resuscitation Outcome Cardiac Rhythm atthe ED
0
1
2
3
4
5
6
7
8
9
10
11
12 13
Call-response interval (min)
NOVEMBER1993
22:11
ANNALS OF EMERGENCY MEDICINE
14
15
Survived to Hospital Discharge
Ventricular fibrillation As~/stole or electromechanical dissociation
Nonsurvivors 45 165
P= .0oo14 by one-sided Fisher'sexacttest.
1662/19
RESUSCITATION van der Hoeven et aI
We studied 216 patients with a persistent out-ofhospital cardiac arrest. In contrast with the three aforementioned studies, we were abIe to retrieve all prehospital resuscitation data. We compared the 177 patients who died in the ED with the 39 who survived to leave the ED. The 18.1% resuscitation rate compares favorably with the study by Gray and colleagues. We found no significant differences in the percentage of witnessed arrests, patients receiving bystander CPR, patients receiving prehospital ACLS, or call-response intervals. This can be explained by selection bias because not all patients who responded to prehospital resuscitation efforts are incIuded in the study population. Furthermore, the long call-hospital interval in both groups tends to diminish the importance of these prognostic factors. A 30-minute resuscitation attempt without restoration of a spontaneous circulation almost uniformly is fatal. 13 Patients with ventricular fibrillation at the scene had the highest chance of successful resuscitation in the ED. If ventricular fibrillation was still present on arrival at the ED, the chance for successful resuscitation was 41.2%. This was significantly'higher than that for asystole or electromechanicaI dissociation. Thirty-three of 39 resuscitated patients died in the hospital. The most frequent cause of death in the ICU was severe postanoxic encephalopathy followedby withdrawal of life-sustaining therapy (67%). None of these patients ever regained consciousness. This also reflects the long period without spontaneous circulation. Only six patients (2.8%) survived to hospital discharge. Two of them initially had severe neurologic deficits, but both were much improved one year after the incident. In these cases, the cardiac rhythm at the scene and in the ED was ventricular fibrillation. Of all patients with ventricular fibrillation in the ED, 12% were discharged alive. All patients with asystole or electromechanicaI dissociation either at the scene or in the ED died. Our results support the view of asystole and electromechanical dissociation as terminal rhythms; they denote a prolonged ischemic period. Although restoration of a stable cardiac rhythm is sometimes possible, severe cerebral damage precludes long-term survivai. New therapies such as the early administration of high-dose norepinephrine may increase the number of patients with a return of spontaneous circulation, but the chances for cerebral recovery are likely Lo be small. > 4 6
unsuccessful, and transport of these patients to the hospital may not be indicated. However, transport for patients with persistent ventricular fibrillation is indicated, as 12% survived to hospital discharge in our study. REFERENCES 1. Eisenberg MS, Hallstrom A, Bergner LB: Long-term survival after out-of-hospital cardiac arrest. N EnglJ Mad 1982;306:1340-1343. 2. Ritter G, Wolfe RA, Goldstein S, et ai: The effect of bystander CPH on survival of out-ofhospitar cardiac arrest victims. Am HeartJ 1985;110:932-937. 3. E[senberg MS, Horwood BT, Cummins RO, et al: Cardiac arrest and resuscitation: A tale of 29 cities. Ann EmergMeal1990;19:179-186. 4. Kentsch M, Stendel M, Berket H, et al: Early prediction of prognosis in out-of-hospital cardiac arrest. IntoneCareMad 1990;16:378-383. 5. Longstreth WT, Cobb LA, Fahranbruch CE, at al: Does age affect outcomes of out-of-hospital car@pulmonary resuscitation? JAMA 1990;264:2109-2110. 6. Kellermann AL, Staves DR, Hackman BB: in-hospital resuscitation following unsuccessful prehospitar advanced cardiac life support: "Heroic efforts" or an exercise in futility? Ann Emarg Marl 1988;17:589-594. 7. Bonnin M J, Swor R~,: Outcomes in unsuccessful field resuscitation attempts. Ann EmergMad 1989;18:507-512. 8. Gray WA, Capone RJ, Most AS: Unsuccessful emergency medical resuscitation--Are continued efforts in the EO justified? N EnglJ Mad 199I ;325:1393-1398. 9. Weaver WD: Resuscitation outside the hospital--What's lacking? N EnglJ Mad 199I;325:1437-1439. 10. Jennett B, Bond M: Assessment of outcome after severe brain damage: A practical scale. Lancet1975;1:480-484. 11. Brain Resuscitation Clinical Trial 1 Study Group: Randomized clinical study of thiopental loading in comatose survivors ef cardiac arrest. N EnglJ Mad 1986;314:397-403. 12. Standards and guidelines for cardiopu[monary resuscitation (CPR)and emergency cardiac care (ECC).JAMA 1986;255:2905-2984. 13. Weaver WD, Cobb LA, Hallstrom AP, et al: Considerations for improving survival from out-e# hospital cardiac arrest. Ann EmergMad 1986;15:1181-1188. 14. Martin D, Werman HA, Brown CG: Four case studies: High-dose epinephrine in cardiac arrest. Ann EmergMad 1990;19:322-326. 15. Barton C, Callaham M: High-dose epinephrine improves the return of spontaneous circulation rates in human victims of cardiac arrest. Ann EmergMad 1991;20:722~725. 16. Cai[aham M, Barton CW, Kayser S: Potential complications of high-dose epinephrine therapy in patients resuscitated from cardiac arrest. JAMA 1991;265:1117-1122.
Address for reprints: Johannes Gvan der Hoeven,MD Medical IntensiveCareUnit, C6-Q University Hospital Leiden PO Box 9600 2300 RCLeiden The Netherlands
CONCLUSION
Prolonged resuscitatio n efforts in the ED for patients with asystole or electromechanical dissociation usually are
20/1663
ANNALS OF EMERGENCY MEDICINE
22:11
NOVEMBER1993