- Email: [email protected]
Arterial pH in out-of-hospital cardiac arrest: Rosponse time as a determinant of acidosis
ArterialpH in Out-of-hospital Cardiac Arrest: Response Time as a Determinant of Acidosis JOSEPH P. ORNATO, MD, EDGAR R. GONZALEZ, PharmD, MARTHA R. COYNE, PharmD, CAROL L. BECK, PharmD, MARK S. COLLINS, BS It is unclear why some victims of out-of-hospital cardiac arrest
are severely acidotic on arrival to the emergency department (ED), whereas others have a pH within normal limits. To explain the difference among patlents, the authors collected data on 119 consecutive out-of-hospltal adult nontraumatic cardiac arrest victims brought to the University of Nebraska Medical Center by paramedic rescue squad between December 1982 and January 1984. Patients who experienced restoration of spontaneous circulation (ROSC) in the field had a normal pH (7.40 f 0.13) as compared with the pH of patlents still recalvlng cardiopulmonary resuscitation (CPR) on arrival at the ED (7.18 f 0.20). A rapid paramedic response time was the best determinant of ROSCand a normal pH on arrival at the ED. Bystander CPR neither signlficantiy increased the number of patients with ROSC in the field nor protected against the development of acidosis, but did improve the neumloglcal outcome of survivors. The presence of acidosis in patients still receiving CPR on arrival in the ED could not be predicted on the basis of paramedic response time, amount of sodium bicarbonate given In the fleid, whether or not the collapse was witnessed, or whether or not bystander CPR had been performed. Patients who were acidotic had a significantly higher (P < 0.001) Pa,,2 (101 + 33 mm Hg) and a lower Pa, (41 f 89 mm Hg) than patients with a normal pH (Pa, 37 f 10 mm Hg, Pao2134 f 107 mm Hg). Adequacy of ventllatlon is the principal determinant of acidosis in patlents who are still receiving CPR on arrlval at the ED. (Am J Emerg Med 1985; 3:498-502) Acidosis occurs frequently during cardiopulmonary resuscitation (CPR) because of the low cardiac output generated by external cardiac compression. It may have both metabolic and respiratory components, because of anaerobic production of lactic acid and hypoventilation. ‘**Regardless of etiology, acidosis is undesirable during resuscitation because it decreases the threshold for ventricular fibrillation, depresses myocardial contractility, impairs normal cardiac automa-
From the University braska.
of Nebraska
Medical
Center,
Manuscript received August 1, 1984; revision 25, 1985; revision accepted April 1, 1985.
Omaha,
received
Ne-
February
Address reprint requests to Dr. Ornato at his present address: Medical College of Virginia, MCV Station Box 525, Richmond, VA 23298. Key Words: Acidosis, cardiac vices, paramedic, resuscitation. 498
arrest,
emergency
medical
ser-
ticity, and blocks the cardio-accelerator and vasopressor action of catecholamines.3-‘0 It is unclear why some patients who have received basic and advanced cardiac life support for pre-hospita1 cardiac arrest have a normal or alkaline pH upon arrival at the hospital emergency department (ED), whereas others have moderate to severe acidosis. Factors likely to affect the initial pH on arrival at the ED include: I) the number of minutes the patient was in cardiac arrest before arrival of the paramedic unit; 2) whether or not bystander CPR was performed; 3) whether or not the cardiac arrest was witnessed; 4) the number of minutes CPR was performed before pH level was obtained; 5) the amount of buffer administered before determination of the pH; 6) the adequacy of ventilation; 7) whether or not the patient has had a return of spontaneous circulation (ROW,) before arrival in the ED; and 8) the effectiveness of perfusion with CPR. This study was undertaken to investigate the determinants of the arterial pH on arrival at the ED in adult patients with pre-hospital cardiac arrest. METHODS Data were collected prospectively on all pre-hospita1 cardiac arrest victims brought to the University of Nebraska Medical Center consecutively from December 1982 to January 1984. The study population included only adult patients (17 years of age or older) with cardiac arrest that was not caused by external physical trauma. All patients were initially treated by both the fire department’s first-responding emergency medical technicians and paramedics employed by the Omaha Fire Division. The city-wide emergency medical services system provides two-tiered response. Basic cardiac life support (CPR) is provided by the fire department’s first-responding emergency medical technicians (mean response time, 3 minutes), and advanced cardiac life support is provided by paramedics (mean response time, 5.5 minutes). All cardiac arrest records were produced by ED personnel using a highly accurate automated electronic flowsheet system.‘i Paramedic response time (defined as the time from dispatch of the paramedic unit by the 911 center to arrival at the patient’s location) was ob-
ORNATO ET AL n ARTERIAL pH IN CARDIAC ARREST
tained from the rescue squad report form. Witnessed arrest was defined as an arrest that occured in the presence of an onlooker (lay person, family, or health professional). Physicians trained and certified in advanced cardiac life support directed all resuscitation efforts using radio-telemetry. The following information was obtained about each patient: age, sex, premonitory symptoms of arrest, paramedic response time, initial arterial blood gas (ABG) results obtained within 10 minutes of the patient’s arrival at the hospital emergency department, whether or not the arrest was witnessed, whether or not bystander CPR was performed, the sequence of rhythms and drugs given during resuscitation, and outcome. Arterial blood gases were drawn percutaneously from the femoral artery by a physician on the cardiac arrest team, If a brisk flashback did not occur during vessel puncture, repeated attempts were made with a different heparinized syringe until the physician was convinced that the blood obtained was of arterial origin. Response times of the first-responding basic life support units were not available for individual patients and could not be entered into our data base (although the mean response for these units during the study period was obtained from the Fire Division). Data were entered into an IBM personal computer text file for processing. The number of ampules of sodium bicarbonate (50 mEq/ampule) administered to each patient before arrival in the ED was counted by the computer. Patients were divided into four groups based on their pH when the initial set of ABG samples were drawn. Severely acidotic patients had pH values below 7.10; moderately acidotic patients had pH values of 7.10 to 7.29; normal patients had pH values from 7.30 to 7.49; alkalotic patients had pH values of 7.50 or greater. All patients were intubated by paramedics in the field with an esophageal obturator airway (EOA) and ventilated with either a bag-valve device or the ventilator portion of a mechanical thumper. The EOA remained in place until arrival at the hospital, where it was removed by an anesthesiologist after insertion of an endotracheal tube. Arterial blood gas samples were drawn before replacement of the esophageal obturator by the endotracheal tube. In those rare cases in which the patient regained consciousness in the field and began to “fight” the airway, the esophageal obturator airway was removed by paramedics immediately. All patients received CPR in the field with a mechanical thumper manufactured by Michigan Instruments, Inc. (Grand Rapids, Michigan). This device delivers 60 compressions per minute and 12 ventilations in a 5:l duty cycle. The ventilatory portion of the thumper delivers an inspired oxygen concentration of approximately 0.80. Neurological outcome of survivors was assessed
from the hospital records taken immediately after resuscitation, at 24-48 hours after resuscitation, and upon hospital discharge. Patients were scored as alert, responsive to verbal stimuli, responsive to painful stimuli, or totally unresponsive. Chi-square analysis with Yate’s correction and student’s t-test were used for statistical analysis. A Pvalue less than or equal to 0.05 was considered statistically significant. RESULTS The study group was composed of 119 adults (82 male, 37 female) with a mean age of 65 + 14 years. The rates of survival to hospital admission and to hospital discharge were 30% and 22%, respectively. Initial cardiac rhythm was ventricular fibrillation in 63 (53%), ventricular tachycardia in four (3%), asystole in 25 (21%), idioventricular in 13 (1 l%), supraventricular in 11 (9%), and miscellaneous in two (2%). Table I compares patients who had ROSC in the field (restart) with patients who were still receiving CPR on arrival to the ED (nonrestart). The restart group had a normal mean pH value. This was significantly different than the mean pH for the nonrestart group, which was moderately acidotic. The mean Pao2 was slightly higher in the restart group than in the nonrestart group, but the difference was not statistically significant. The mean paCo, was significantly lower in the restart group than in the nonrestart group. Paramedic response time was significantly shorter in the restart group. There were more patients in ventricular fibrillation (VF) or ventricular tachycardia (VT) and fewer patients in bradyasystole (asystole or pulseless idioventricular rhythm) in the restart group. The initial rhythm on arrival of the paramedic unit was documented in 82 of the 83 nonrestart patients. The remaining case was treated as missing data. The rate of survival to hospital discharge was significantly higher (P < 0.001) in the restart group (261 36, 72%) than in the nonrestart group (O/83, 0%). Of the six patients found in asystole whose pulses were restarted in the field, none became neurologically alert, and none survived to hospital discharge. There was no significant difference between the groups in the number of patients whose collapse was witnessed or who received bystander CPR before arrival of the first responders. Information as to whether or not the arrest was witnessed was known for only 113 of the 119 patients. The remaining six patients were treated as missing data. Table 2 divides the nonrestart group into subgroups based on the arterial pH value upon the patient’s arrival in the ED. The presence of acidosis in the nonrestart group could not be predicted by the paramedic response time, amount of sodium bicarbonate given in 499
AMERICAN
JOURNAL
OF EMERGENCY
TABLE 1. Comparison of Patients in the Field (Nonrestart*)
MEDICINE
Successfully
n Volume 3, Number 6 n November
Resuscitated
in the Field (Restart)
1985
with Those Not Resuscitated
Nonrestart
Restart Number of patients Age (years; mean 2 SD) Sex (male/female) Pa,, (mm Hg; mean r SD) pH (mean 2 SD) Pa,,, (mm Hg; mean f SD) Response timet (min; mean + SD) No. of witnessed* (%) No. bystander CPR§ (%) Sodium bicarbonate used1 (ampules; Rhythm initially detected (No. [“X,1) Ventricular fibrillation Ventricular tachycardia ldioventricular rhythm Asystole Supraventricular rhythm No. surviving to discharge (%)
36 66 ?I 10 22114 116 2 93 7.40 2 0.13 52 2 53 3.3 + 3.0 20133 (61) 13136 (36) 1.6 2 1.5
mean 2 SD)
26136 2136 O/36 6136 l/36 26136
P-value
-
83 64 2 15 60123 80 2 114 7.18 2 0.20 84 5 89 5.8 ” 5.0 47/80 (59) 34/83 (41) 4.2 2 1.5
(72) (6) (0) (17) (3) (72)
NS NS NS
37182 2182 (2) (45)
I
13/82 (23) 19182 (16) 11/82 (13) 0182 (0)
I
co.004 CO.03 NS
Patients in this group were receiving cardiopulmonary resuscitation at the time of their arrival in the emergency department. T Time between dispatch of the paramedic unit by the 911 service and arrival of the paramedic unit at the scene of the collapse. $ Number of patients whose collapses were witnessed. 5 Number of patients who received cardiopulmonary resuscitation from a bystander at the scene of the collapse. 1’Each ampule contained 50 mEq of sodium bicarbonate. l
DISCUSSION
the field, whether or not the collapse was witnessed, or whether or not bystander CPR had been performed before the arrival of the first responders. Patients who were acidotic had a significantly higher Pacoz and a lower Pao,, than patients with a normal or alkaline pH. Bystander CPR initiated prior to the arrival of the fire engine company significantly (P < 0.05) increased the number of patients who became neurologically fully alert at any time following successful resuscitation. Nine of 47 patients (19%) who received bystander CPR became neurologically alert as compared with four of 67 patients (6%) who did not receive bystander CPR.
TABLE 2.
Analysis
of Cases of Unsuccessful
Resuscitation
Severe Acidosis Bicarbonatet Witnessed+ Bystander CPR§ Response Time (min)” Pao, (mm Hg)ll PH” %o,
(mm
f-W
4.6 +- 1.6 8122 (36) lo/23 (44) 7.5 2 5.1 41 269 6.95 2 0.12 101 & 33
The majority of out-of-hospital cardiac arrests are caused by sudden cardiac dysrhythmia, usually ventricular tachycardia or fibrillation.‘* In such cases, it is reasonable to assume that arterial pH was close to normal before the patient’s collapse. As anaerobic metabolism proceeds, hydrogen ion (in the form of lactic acid) will be produced, and tissue pH will fall.13 Bradyasystole is more common, and VF or VT are less frequent, later in the arrest. 14,15Most survivors of prehospital cardiac arrest will have been in VF or VT on arrival of the paramedic unit.16-18 The survival rate
(Nonrestart
Group)*
Moderate Acidosis 4.0 2 28142 17144 5.1 ? 63 2 7.19 s 70 f
1.5 (67) (39) 5.2 102”” 0.06Tt 36tt
Normal 4.1 r 1.4 6/l 0 (60) 4/10 (40) 6.2 -t 4.8 134 f 107**tt 7.38 f 0.06tt 37 f 1ott
* Patients whose spontaneous circulation was not restarted in the field and who were receiving t Mean (*SD) number of ampules (50 mEq/ampule) received per patient. $ Number (%) of patients whose arrests were witnessed. 5 Number (%) of patients who received CPR from a bystander at the scene of the collapse. 1Mean f SD. ** Difference between these values is statistically significant (P < 0.01). tt Difference between these values is statistically significant (P < 0.001).
500
Alkalotic 3.2 a 1.5 415 (80) 2/5 (40) 3.0 f 2.2 307 t 133j.t 7.63 2 0.76tt 27 c 9tt
CPR at the time of arrival
in the ED.
ORNATO ET AL n ARTERIAL pH IN CARDIAC ARREST
when the initial rhythm zero 14,15,19,20
is bradyasystole
is nearly
In our study, a major discriminant of pH on arrival in the ED was whether the patient had experienced ROSC in the field. Patients whose circulation was restored in the field had a significantly shorter paramedic response time than those whose circulation could not be restored. We believe the more rapid paramedic response time resulted in a higher number of patients in VF or VT and a lower number of patients in bradyasystole on arrival of the paramedics. Neither the witnessing of the collapse nor the performance of bystander CPR improved survival rates in our series. This is in contrast to the finding by Thompson et af“j that bystander CPR significantly improved rates of survival.16 Ritter et UP’ found that bystander CPR only improves survival rate when the paramedic response time is greater than six minutes. The reported paramedic response time in the Seattle system is six minutes,t6 as compared with our times of 3.3 and 5.8 minutes in the restart and nonrestart groups, respectively. We believe our shorter paramedic response times are the most likely reason why bystander CPR could not be shown to affect survival. Bystander CPR resulted in better neurological outcome of survivors, similar to the experience of Thompson et al in Seattle.16 Our findings suggest that CPR produces enough blood flow to measurably protect the brain from anoxic damage, but it produces insufficient flow to prevent acidosis from anaerobic metabolism in less vital areas of the body. The initial arterial pH of patients who did not experience ROSC in the field (nonrestart group) could not be predicted on the basis of paramedic response time, amount of sodium bicarbonate given in the field, witness of the collapse, or whether or not bystander CPR has been performed before the arrival of the firstresponding fire company. The presence of significant hypoxemia and hypercarbia in the acidotic subgroup, as compared with the subgroups with a normal or alkaline pH, suggests that inadequate ventilation and/or perfusion from CPR are the major determinants of the initial pH in this subgroup of out-of-hospital cardiac arrest victims. We suspect that the high incidence of hypoxemia, acidosis, and hypercarbia in our patients who cannot be restarted in the field may be related to the use of the esophageal obturator airway in our system during the time period of the study. The EOA did not appear to provide adequate ventilation to many of our patients, probably because of difftculty in making an adequate seal with the face mask22 in a field environment. This device has been shown to be more effective in a hospital setting using an oxygen-powered breathing device.23
A limitation of the study deserves mention. Although care was taken clinically to insure that blood gas samples were arterial and not venous in origin, the percutaneous method of drawing blood gases can never completely exclude venous origin. We recommend that future studies be conducted using arterial cutdown (radial) and cannulation for repeated sampling . CONCLUSION Victims of out-of-hospital cardiac arrest who are brought to the emergency department by paramedics may have a wide range of values for their initial arterial pH. The single most important determinant of survival and subsequent pH is how quickly the paramedics can respond to the scene, for it is this factor that determines the likelihood that the patient will still be in For those patients who ventricular fibrillation. 17~18,23 do not respond to field resuscitation in our system, the adequacy of ventilation and/or perfusion is the most important determinant of pH. Witness of the collapse and bystander CPR do not significantly affect the initial arterial pH. REFERENCES 1. Redding JS, Pearson JW. Metabolic acidosis: A factor in cardiac resuscitation. South Med J 1967;60:926-932. 2. Chazan JA, Stenson R, Kurland GS. The acidosis of cardiac arrest. N Engl J Med 1968;278:360-364. 3. Gerst PH, Fleming WH, Malm JR. A quantitative evaluation of the effects of acidosis, and alkalosis upon ventricular fibrillation threshold. Surgery 1966;59:1050-1066. 4. Cingolani HE, Mattiazzi AR, Blesa ES, et al. Contractility in isolated mammalian heart muscle after acid-base changes. Circ Res 1970;26:269-278. 5. Williamson JR, Schaffer SW, Ford C, et al. Contribution of tissue acidosis to ischemic injury in the perfused rat heart. Circulation 1976;53:3-14. 6. Stewart JS, Stewart WK, Morgan HG, et al. A clinical and experimental study of the electrocardiographic changes in extreme acidosis and cardiac arrest. Br Heart J 1965;27:490-497. 7. Darby TD, Aldinger EE, Gadsen RH, et al. Effects of metabolic acidosis on ventricular isometric systolic tension and the response to epinephrine and levarterenol. Circ Res 1980;8:1242-1253. 8. Houie 08, Weil MH, Brown EB, et al. Influence of respiratory acidosis on ECG and pressor response to epinephrine, norepinephrine, and metaraminol. Proc Sot Exp Biol Med 1957;94:561-564. 9. Camilion de Hurtado MC, Argel MI, Cingolani HE. Influence of acid-base alterations on myocardial sensitivity to catecholamines. Arch Pharmacol 1981;317:219-224. 10. Burget GE, Visscher MB. Variations of the pH of the blood and the response of the vascular system to adrenalin. Am J Physiol 1927;81:113-123. 11. Ornato JP, Fennigkoh L, Jaeger C. The electronic clipboard: An automated system for accurately recording events during a cardiac arrest. Ann Emerg Med 1981;10:138141.
501
AMERICAN JOURNAL OF EMERGENCY MEDICINE n Volume 3, Number 6 n November 1985
12. Bigger JT. Definition of benign versus malignant ventricular arrhythmias: Targets for treatment. Am J Cardiol 1983;52:47c_54c. 13. Park R. Lactic acidosis. West J Med 1980;133:418-424. 14. lseri LT, Humphrey SB, Siner EJ. Pre-hospital bradyasystolic cardiac arrest. Ann Intern Med 1978;88:741-745. 15. Adgey AAJ, Geddes JS, Mulholland HC, et al. Incidence, significance and management of early bradyarrhythmias complicating acute Ml. Lancet 1969;2:1097-1101. 16. Thompson RG, Hallstrom AP, Cobb LA. Bystander-initiated cardiopulmonary resuscitation in the management of ventricular fibrillation. Ann Intern Med 1979;90:737-740. 17. Roth R, Stewart RD, Cannon RK, et al. Out-of-hospital cardiac arrest: Factors associated with survival. Ann Emerg Med 1984;13:101-103. 18. Eisenberg MS, Bergner L, Hallstrom A. Paramedic programs and out-of-hospital cardiac arrest: Factors associated with successful resuscitation. Am J Public Health 1979;69:30-38.
502
19. Stueven H, Thompson BM, Aprahamian C, et al. Use of calcium in pre-hospital cardiac arrest. Ann Emerg Med 1983;12:136-139. 20. Coon GA, Clinton JE, Ruiz E. Use of atropine for bradyasystolic prehospital cardiac arrest. Ann Emerg Med 1981;10:462-467. 21. Ritter G, Goldstein S, Leighton R, et al. The effect of bystander CPR and arrival time of EMS on successful outof-hospital resuscitation. Am J Emerg Med 1984;2:358. 22. White RD. Controversies in out-of-hospital emergency airway control: Esophageal obstruction or endotracheal intubation? Ann Emerg Med 1984;13:778-781. 23. Kerber RE, Jensen SR, Gascho JA, et al. Determinants of defibrillation: Prospective analysis of 183 patients. Am J Cardiol 1983;52:739-745. 24. Hammargren Y, Clinton JE, Ruiz E. Standardized comparison of the EOA and endotracheal tube in cardiac arrest (abstract). Ann Emerg Med 1984;13:404.
are severely acidotic on arrival to the emergency department (ED), whereas others have a pH within normal limits. To explain the difference among patlents, the authors collected data on 119 consecutive out-of-hospltal adult nontraumatic cardiac arrest victims brought to the University of Nebraska Medical Center by paramedic rescue squad between December 1982 and January 1984. Patients who experienced restoration of spontaneous circulation (ROSC) in the field had a normal pH (7.40 f 0.13) as compared with the pH of patlents still recalvlng cardiopulmonary resuscitation (CPR) on arrival at the ED (7.18 f 0.20). A rapid paramedic response time was the best determinant of ROSCand a normal pH on arrival at the ED. Bystander CPR neither signlficantiy increased the number of patients with ROSC in the field nor protected against the development of acidosis, but did improve the neumloglcal outcome of survivors. The presence of acidosis in patients still receiving CPR on arrival in the ED could not be predicted on the basis of paramedic response time, amount of sodium bicarbonate given In the fleid, whether or not the collapse was witnessed, or whether or not bystander CPR had been performed. Patients who were acidotic had a significantly higher (P < 0.001) Pa,,2 (101 + 33 mm Hg) and a lower Pa, (41 f 89 mm Hg) than patients with a normal pH (Pa, 37 f 10 mm Hg, Pao2134 f 107 mm Hg). Adequacy of ventllatlon is the principal determinant of acidosis in patlents who are still receiving CPR on arrlval at the ED. (Am J Emerg Med 1985; 3:498-502) Acidosis occurs frequently during cardiopulmonary resuscitation (CPR) because of the low cardiac output generated by external cardiac compression. It may have both metabolic and respiratory components, because of anaerobic production of lactic acid and hypoventilation. ‘**Regardless of etiology, acidosis is undesirable during resuscitation because it decreases the threshold for ventricular fibrillation, depresses myocardial contractility, impairs normal cardiac automa-
From the University braska.
of Nebraska
Medical
Center,
Manuscript received August 1, 1984; revision 25, 1985; revision accepted April 1, 1985.
Omaha,
received
Ne-
February
Address reprint requests to Dr. Ornato at his present address: Medical College of Virginia, MCV Station Box 525, Richmond, VA 23298. Key Words: Acidosis, cardiac vices, paramedic, resuscitation. 498
arrest,
emergency
medical
ser-
ticity, and blocks the cardio-accelerator and vasopressor action of catecholamines.3-‘0 It is unclear why some patients who have received basic and advanced cardiac life support for pre-hospita1 cardiac arrest have a normal or alkaline pH upon arrival at the hospital emergency department (ED), whereas others have moderate to severe acidosis. Factors likely to affect the initial pH on arrival at the ED include: I) the number of minutes the patient was in cardiac arrest before arrival of the paramedic unit; 2) whether or not bystander CPR was performed; 3) whether or not the cardiac arrest was witnessed; 4) the number of minutes CPR was performed before pH level was obtained; 5) the amount of buffer administered before determination of the pH; 6) the adequacy of ventilation; 7) whether or not the patient has had a return of spontaneous circulation (ROW,) before arrival in the ED; and 8) the effectiveness of perfusion with CPR. This study was undertaken to investigate the determinants of the arterial pH on arrival at the ED in adult patients with pre-hospital cardiac arrest. METHODS Data were collected prospectively on all pre-hospita1 cardiac arrest victims brought to the University of Nebraska Medical Center consecutively from December 1982 to January 1984. The study population included only adult patients (17 years of age or older) with cardiac arrest that was not caused by external physical trauma. All patients were initially treated by both the fire department’s first-responding emergency medical technicians and paramedics employed by the Omaha Fire Division. The city-wide emergency medical services system provides two-tiered response. Basic cardiac life support (CPR) is provided by the fire department’s first-responding emergency medical technicians (mean response time, 3 minutes), and advanced cardiac life support is provided by paramedics (mean response time, 5.5 minutes). All cardiac arrest records were produced by ED personnel using a highly accurate automated electronic flowsheet system.‘i Paramedic response time (defined as the time from dispatch of the paramedic unit by the 911 center to arrival at the patient’s location) was ob-
ORNATO ET AL n ARTERIAL pH IN CARDIAC ARREST
tained from the rescue squad report form. Witnessed arrest was defined as an arrest that occured in the presence of an onlooker (lay person, family, or health professional). Physicians trained and certified in advanced cardiac life support directed all resuscitation efforts using radio-telemetry. The following information was obtained about each patient: age, sex, premonitory symptoms of arrest, paramedic response time, initial arterial blood gas (ABG) results obtained within 10 minutes of the patient’s arrival at the hospital emergency department, whether or not the arrest was witnessed, whether or not bystander CPR was performed, the sequence of rhythms and drugs given during resuscitation, and outcome. Arterial blood gases were drawn percutaneously from the femoral artery by a physician on the cardiac arrest team, If a brisk flashback did not occur during vessel puncture, repeated attempts were made with a different heparinized syringe until the physician was convinced that the blood obtained was of arterial origin. Response times of the first-responding basic life support units were not available for individual patients and could not be entered into our data base (although the mean response for these units during the study period was obtained from the Fire Division). Data were entered into an IBM personal computer text file for processing. The number of ampules of sodium bicarbonate (50 mEq/ampule) administered to each patient before arrival in the ED was counted by the computer. Patients were divided into four groups based on their pH when the initial set of ABG samples were drawn. Severely acidotic patients had pH values below 7.10; moderately acidotic patients had pH values of 7.10 to 7.29; normal patients had pH values from 7.30 to 7.49; alkalotic patients had pH values of 7.50 or greater. All patients were intubated by paramedics in the field with an esophageal obturator airway (EOA) and ventilated with either a bag-valve device or the ventilator portion of a mechanical thumper. The EOA remained in place until arrival at the hospital, where it was removed by an anesthesiologist after insertion of an endotracheal tube. Arterial blood gas samples were drawn before replacement of the esophageal obturator by the endotracheal tube. In those rare cases in which the patient regained consciousness in the field and began to “fight” the airway, the esophageal obturator airway was removed by paramedics immediately. All patients received CPR in the field with a mechanical thumper manufactured by Michigan Instruments, Inc. (Grand Rapids, Michigan). This device delivers 60 compressions per minute and 12 ventilations in a 5:l duty cycle. The ventilatory portion of the thumper delivers an inspired oxygen concentration of approximately 0.80. Neurological outcome of survivors was assessed
from the hospital records taken immediately after resuscitation, at 24-48 hours after resuscitation, and upon hospital discharge. Patients were scored as alert, responsive to verbal stimuli, responsive to painful stimuli, or totally unresponsive. Chi-square analysis with Yate’s correction and student’s t-test were used for statistical analysis. A Pvalue less than or equal to 0.05 was considered statistically significant. RESULTS The study group was composed of 119 adults (82 male, 37 female) with a mean age of 65 + 14 years. The rates of survival to hospital admission and to hospital discharge were 30% and 22%, respectively. Initial cardiac rhythm was ventricular fibrillation in 63 (53%), ventricular tachycardia in four (3%), asystole in 25 (21%), idioventricular in 13 (1 l%), supraventricular in 11 (9%), and miscellaneous in two (2%). Table I compares patients who had ROSC in the field (restart) with patients who were still receiving CPR on arrival to the ED (nonrestart). The restart group had a normal mean pH value. This was significantly different than the mean pH for the nonrestart group, which was moderately acidotic. The mean Pao2 was slightly higher in the restart group than in the nonrestart group, but the difference was not statistically significant. The mean paCo, was significantly lower in the restart group than in the nonrestart group. Paramedic response time was significantly shorter in the restart group. There were more patients in ventricular fibrillation (VF) or ventricular tachycardia (VT) and fewer patients in bradyasystole (asystole or pulseless idioventricular rhythm) in the restart group. The initial rhythm on arrival of the paramedic unit was documented in 82 of the 83 nonrestart patients. The remaining case was treated as missing data. The rate of survival to hospital discharge was significantly higher (P < 0.001) in the restart group (261 36, 72%) than in the nonrestart group (O/83, 0%). Of the six patients found in asystole whose pulses were restarted in the field, none became neurologically alert, and none survived to hospital discharge. There was no significant difference between the groups in the number of patients whose collapse was witnessed or who received bystander CPR before arrival of the first responders. Information as to whether or not the arrest was witnessed was known for only 113 of the 119 patients. The remaining six patients were treated as missing data. Table 2 divides the nonrestart group into subgroups based on the arterial pH value upon the patient’s arrival in the ED. The presence of acidosis in the nonrestart group could not be predicted by the paramedic response time, amount of sodium bicarbonate given in 499
AMERICAN
JOURNAL
OF EMERGENCY
TABLE 1. Comparison of Patients in the Field (Nonrestart*)
MEDICINE
Successfully
n Volume 3, Number 6 n November
Resuscitated
in the Field (Restart)
1985
with Those Not Resuscitated
Nonrestart
Restart Number of patients Age (years; mean 2 SD) Sex (male/female) Pa,, (mm Hg; mean r SD) pH (mean 2 SD) Pa,,, (mm Hg; mean f SD) Response timet (min; mean + SD) No. of witnessed* (%) No. bystander CPR§ (%) Sodium bicarbonate used1 (ampules; Rhythm initially detected (No. [“X,1) Ventricular fibrillation Ventricular tachycardia ldioventricular rhythm Asystole Supraventricular rhythm No. surviving to discharge (%)
36 66 ?I 10 22114 116 2 93 7.40 2 0.13 52 2 53 3.3 + 3.0 20133 (61) 13136 (36) 1.6 2 1.5
mean 2 SD)
26136 2136 O/36 6136 l/36 26136
P-value
-
83 64 2 15 60123 80 2 114 7.18 2 0.20 84 5 89 5.8 ” 5.0 47/80 (59) 34/83 (41) 4.2 2 1.5
(72) (6) (0) (17) (3) (72)
NS NS NS
37182 2182 (2) (45)
I
13/82 (23) 19182 (16) 11/82 (13) 0182 (0)
I
co.004 CO.03 NS
Patients in this group were receiving cardiopulmonary resuscitation at the time of their arrival in the emergency department. T Time between dispatch of the paramedic unit by the 911 service and arrival of the paramedic unit at the scene of the collapse. $ Number of patients whose collapses were witnessed. 5 Number of patients who received cardiopulmonary resuscitation from a bystander at the scene of the collapse. 1’Each ampule contained 50 mEq of sodium bicarbonate. l
DISCUSSION
the field, whether or not the collapse was witnessed, or whether or not bystander CPR had been performed before the arrival of the first responders. Patients who were acidotic had a significantly higher Pacoz and a lower Pao,, than patients with a normal or alkaline pH. Bystander CPR initiated prior to the arrival of the fire engine company significantly (P < 0.05) increased the number of patients who became neurologically fully alert at any time following successful resuscitation. Nine of 47 patients (19%) who received bystander CPR became neurologically alert as compared with four of 67 patients (6%) who did not receive bystander CPR.
TABLE 2.
Analysis
of Cases of Unsuccessful
Resuscitation
Severe Acidosis Bicarbonatet Witnessed+ Bystander CPR§ Response Time (min)” Pao, (mm Hg)ll PH” %o,
(mm
f-W
4.6 +- 1.6 8122 (36) lo/23 (44) 7.5 2 5.1 41 269 6.95 2 0.12 101 & 33
The majority of out-of-hospital cardiac arrests are caused by sudden cardiac dysrhythmia, usually ventricular tachycardia or fibrillation.‘* In such cases, it is reasonable to assume that arterial pH was close to normal before the patient’s collapse. As anaerobic metabolism proceeds, hydrogen ion (in the form of lactic acid) will be produced, and tissue pH will fall.13 Bradyasystole is more common, and VF or VT are less frequent, later in the arrest. 14,15Most survivors of prehospital cardiac arrest will have been in VF or VT on arrival of the paramedic unit.16-18 The survival rate
(Nonrestart
Group)*
Moderate Acidosis 4.0 2 28142 17144 5.1 ? 63 2 7.19 s 70 f
1.5 (67) (39) 5.2 102”” 0.06Tt 36tt
Normal 4.1 r 1.4 6/l 0 (60) 4/10 (40) 6.2 -t 4.8 134 f 107**tt 7.38 f 0.06tt 37 f 1ott
* Patients whose spontaneous circulation was not restarted in the field and who were receiving t Mean (*SD) number of ampules (50 mEq/ampule) received per patient. $ Number (%) of patients whose arrests were witnessed. 5 Number (%) of patients who received CPR from a bystander at the scene of the collapse. 1Mean f SD. ** Difference between these values is statistically significant (P < 0.01). tt Difference between these values is statistically significant (P < 0.001).
500
Alkalotic 3.2 a 1.5 415 (80) 2/5 (40) 3.0 f 2.2 307 t 133j.t 7.63 2 0.76tt 27 c 9tt
CPR at the time of arrival
in the ED.
ORNATO ET AL n ARTERIAL pH IN CARDIAC ARREST
when the initial rhythm zero 14,15,19,20
is bradyasystole
is nearly
In our study, a major discriminant of pH on arrival in the ED was whether the patient had experienced ROSC in the field. Patients whose circulation was restored in the field had a significantly shorter paramedic response time than those whose circulation could not be restored. We believe the more rapid paramedic response time resulted in a higher number of patients in VF or VT and a lower number of patients in bradyasystole on arrival of the paramedics. Neither the witnessing of the collapse nor the performance of bystander CPR improved survival rates in our series. This is in contrast to the finding by Thompson et af“j that bystander CPR significantly improved rates of survival.16 Ritter et UP’ found that bystander CPR only improves survival rate when the paramedic response time is greater than six minutes. The reported paramedic response time in the Seattle system is six minutes,t6 as compared with our times of 3.3 and 5.8 minutes in the restart and nonrestart groups, respectively. We believe our shorter paramedic response times are the most likely reason why bystander CPR could not be shown to affect survival. Bystander CPR resulted in better neurological outcome of survivors, similar to the experience of Thompson et al in Seattle.16 Our findings suggest that CPR produces enough blood flow to measurably protect the brain from anoxic damage, but it produces insufficient flow to prevent acidosis from anaerobic metabolism in less vital areas of the body. The initial arterial pH of patients who did not experience ROSC in the field (nonrestart group) could not be predicted on the basis of paramedic response time, amount of sodium bicarbonate given in the field, witness of the collapse, or whether or not bystander CPR has been performed before the arrival of the firstresponding fire company. The presence of significant hypoxemia and hypercarbia in the acidotic subgroup, as compared with the subgroups with a normal or alkaline pH, suggests that inadequate ventilation and/or perfusion from CPR are the major determinants of the initial pH in this subgroup of out-of-hospital cardiac arrest victims. We suspect that the high incidence of hypoxemia, acidosis, and hypercarbia in our patients who cannot be restarted in the field may be related to the use of the esophageal obturator airway in our system during the time period of the study. The EOA did not appear to provide adequate ventilation to many of our patients, probably because of difftculty in making an adequate seal with the face mask22 in a field environment. This device has been shown to be more effective in a hospital setting using an oxygen-powered breathing device.23
A limitation of the study deserves mention. Although care was taken clinically to insure that blood gas samples were arterial and not venous in origin, the percutaneous method of drawing blood gases can never completely exclude venous origin. We recommend that future studies be conducted using arterial cutdown (radial) and cannulation for repeated sampling . CONCLUSION Victims of out-of-hospital cardiac arrest who are brought to the emergency department by paramedics may have a wide range of values for their initial arterial pH. The single most important determinant of survival and subsequent pH is how quickly the paramedics can respond to the scene, for it is this factor that determines the likelihood that the patient will still be in For those patients who ventricular fibrillation. 17~18,23 do not respond to field resuscitation in our system, the adequacy of ventilation and/or perfusion is the most important determinant of pH. Witness of the collapse and bystander CPR do not significantly affect the initial arterial pH. REFERENCES 1. Redding JS, Pearson JW. Metabolic acidosis: A factor in cardiac resuscitation. South Med J 1967;60:926-932. 2. Chazan JA, Stenson R, Kurland GS. The acidosis of cardiac arrest. N Engl J Med 1968;278:360-364. 3. Gerst PH, Fleming WH, Malm JR. A quantitative evaluation of the effects of acidosis, and alkalosis upon ventricular fibrillation threshold. Surgery 1966;59:1050-1066. 4. Cingolani HE, Mattiazzi AR, Blesa ES, et al. Contractility in isolated mammalian heart muscle after acid-base changes. Circ Res 1970;26:269-278. 5. Williamson JR, Schaffer SW, Ford C, et al. Contribution of tissue acidosis to ischemic injury in the perfused rat heart. Circulation 1976;53:3-14. 6. Stewart JS, Stewart WK, Morgan HG, et al. A clinical and experimental study of the electrocardiographic changes in extreme acidosis and cardiac arrest. Br Heart J 1965;27:490-497. 7. Darby TD, Aldinger EE, Gadsen RH, et al. Effects of metabolic acidosis on ventricular isometric systolic tension and the response to epinephrine and levarterenol. Circ Res 1980;8:1242-1253. 8. Houie 08, Weil MH, Brown EB, et al. Influence of respiratory acidosis on ECG and pressor response to epinephrine, norepinephrine, and metaraminol. Proc Sot Exp Biol Med 1957;94:561-564. 9. Camilion de Hurtado MC, Argel MI, Cingolani HE. Influence of acid-base alterations on myocardial sensitivity to catecholamines. Arch Pharmacol 1981;317:219-224. 10. Burget GE, Visscher MB. Variations of the pH of the blood and the response of the vascular system to adrenalin. Am J Physiol 1927;81:113-123. 11. Ornato JP, Fennigkoh L, Jaeger C. The electronic clipboard: An automated system for accurately recording events during a cardiac arrest. Ann Emerg Med 1981;10:138141.
501
AMERICAN JOURNAL OF EMERGENCY MEDICINE n Volume 3, Number 6 n November 1985
12. Bigger JT. Definition of benign versus malignant ventricular arrhythmias: Targets for treatment. Am J Cardiol 1983;52:47c_54c. 13. Park R. Lactic acidosis. West J Med 1980;133:418-424. 14. lseri LT, Humphrey SB, Siner EJ. Pre-hospital bradyasystolic cardiac arrest. Ann Intern Med 1978;88:741-745. 15. Adgey AAJ, Geddes JS, Mulholland HC, et al. Incidence, significance and management of early bradyarrhythmias complicating acute Ml. Lancet 1969;2:1097-1101. 16. Thompson RG, Hallstrom AP, Cobb LA. Bystander-initiated cardiopulmonary resuscitation in the management of ventricular fibrillation. Ann Intern Med 1979;90:737-740. 17. Roth R, Stewart RD, Cannon RK, et al. Out-of-hospital cardiac arrest: Factors associated with survival. Ann Emerg Med 1984;13:101-103. 18. Eisenberg MS, Bergner L, Hallstrom A. Paramedic programs and out-of-hospital cardiac arrest: Factors associated with successful resuscitation. Am J Public Health 1979;69:30-38.
502
19. Stueven H, Thompson BM, Aprahamian C, et al. Use of calcium in pre-hospital cardiac arrest. Ann Emerg Med 1983;12:136-139. 20. Coon GA, Clinton JE, Ruiz E. Use of atropine for bradyasystolic prehospital cardiac arrest. Ann Emerg Med 1981;10:462-467. 21. Ritter G, Goldstein S, Leighton R, et al. The effect of bystander CPR and arrival time of EMS on successful outof-hospital resuscitation. Am J Emerg Med 1984;2:358. 22. White RD. Controversies in out-of-hospital emergency airway control: Esophageal obstruction or endotracheal intubation? Ann Emerg Med 1984;13:778-781. 23. Kerber RE, Jensen SR, Gascho JA, et al. Determinants of defibrillation: Prospective analysis of 183 patients. Am J Cardiol 1983;52:739-745. 24. Hammargren Y, Clinton JE, Ruiz E. Standardized comparison of the EOA and endotracheal tube in cardiac arrest (abstract). Ann Emerg Med 1984;13:404.