Does the presence or absence of sonographically identified cardiac activity predict resuscitation outcomes of cardiac arrest patients?

American Journal of Emergency Medicine (2005) 23, 459 – 462

www.elsevier.com/locate/ajem

Does the presence or absence of sonographically identified cardiac activity predict resuscitation outcomes of cardiac arrest patients? Philip Salen MDa,*, Larry Melniker MD, MSb, Carolyn Chooljian MDc, John S. Rose MDd, Janet Alteveer MDe, James Reed PhDa, Michael Heller MDa a

Department of Emergency Medicine, St. Lukes Hospital, St. Lukes Hospital EM Residency, Bethlehem, PA 18015, USA Department of Emergency Medicine, New York Methodist Hospital, NY 11215, USA c Department of Emergency Medicine, University of California, UCSF Fresno, Emergency Medicine Residency, University Medical Center, Fresno, CA 93727, USA d Department of Emergency Medicine, University of California, Davis Medical Center, Sacramento, CA, USA e Department of Emergency Medicine, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, Camden, NJ b

Received 29 October 2004; accepted 18 November 2004

Abstract This study evaluated the ability of cardiac sonography performed by emergency physicians to predict resuscitation outcomes of cardiac arrest patients. A convenience sample of cardiac arrest patients prospectively underwent bedside cardiac sonography at 4 emergency medicine residency–affiliated EDs as part of the Sonography Outcomes Assessment Program. Cardiac arrest patients in pulseless electrical activity (PEA) and asystole underwent transthoracic cardiac ultrasound B-mode examinations during their resuscitations to assess for the presence or absence of cardiac kinetic activity. Several end points were analyzed as potential predictors of resuscitations: presenting cardiac rhythms, the presence of sonographically detected cardiac activity, prehospital resuscitation time intervals, and ED resuscitation time intervals. Of 70 enrolled subjects, 36 were in asystole and 34 in PEA. Patients presenting without evidence of cardiac kinetic activity did not have return of spontaneous circulation (ROSC) regardless of their cardiac rhythm, asystole, or PEA. Of the 34 subjects presenting with PEA, 11 had sonographic evidence of cardiac kinetic activity, 8 had ROSC with subsequent admission to the hospital, and 1 had survived to hospital discharge with scores of 1 on the Glasgow-Pittsburgh Cerebral Performance scale and 1 in the Overall Performance category. The presence of sonographically identified cardiac kinetic motion was associated with ROSC. Time interval durations of cardiac resuscitative efforts in the prehospital environment and in the ED were not accurate predictors of ROSC for this cohort. Cardiac kinetic activity, or lack thereof, identified by transthoracic B-mode ultrasound may aid physicians’ decision making regarding the care of cardiac arrest patients with PEA or asystole. D 2005 Elsevier Inc. All rights reserved.

Presented at the ACEP Scientific Assembly in Boston, MA, October 2003. * Corresponding author. 0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.11.007

460

1. Introduction Cardiac arrest results in the absence of signs of circulation, including the absence of a pulse. The pulse check has been the bgold standardQ usually relied on by clinicians to determine if the heart is or is not beating [1]. Absence of a pulse has been used to indicate the need to place automated external defibrillator leads and initiate chest compressions. Unfortunately, the pulse check is neither a rapid nor a reliable means for detecting the presence or absence of circulation for both laypersons and healthcare providers [1]. Multiple studies have concluded that as a diagnostic test to assess for perfusion during suspected cardiac arrest, the pulse check has serious limitations in specificity, sensitivity, and accuracy [2- 4]. The absence of a palpable pulse does not always reflect cardiac standstill during cardiac resuscitation as inefficient cardiac contractions occur because of many factors, including cardiac tamponade, pulmonary embolism, and pneumothorax [5]. Because some causes of inefficient cardiac contractions are more reversible than others, (ie, hypo/ hyperkalemia, hypovolemia, hypothermia, acidosis, cardiac tamponade), a tool that would allow for early identification of groups of pulseless patients with more or less likelihood of resuscitation would be welcome [1,6]. Ultrasound (US) is well suited to bedside ED use in the care of critically ill patients and provides rapid, noninvasive, potentially useful information regarding the patient’s clinical condition. The importance of transesophageal cardiac US has been recognized in the American Heart Association guidelines for differentiating patients with true electromechanical dissociation (EMD; pulseless patients with cardiac electrical activity but no cardiac contractility) from patients with pseudo-EMD (pulseless patients with cardiac electrical activity and cardiac contractility), a group that should be aggressively treated [1,6]. Recent reports have suggested that transthoracic cardiac sonography might assist in identifying patients with potentially treatable conditions when pulses are not palpable, but its role in the setting of cardiac arrest is unclear [7,8]. This multicenter trial examined the use of cardiac sonography as an adjunctive diagnostic aid and as a predictor of resuscitation outcomes in cardiac arrest patients with pulseless electrical activity (PEA) and asystole.

P. Salen et al. initiated and transport cardiac arrest victims to the nearest ED; no change in transport policy was instituted for this study. The institutional review board granted an exemption of informed consent with the caveat that there were explicit instructions in the study protocol that cardiac US results do not influence the standard American Heart Association decision-making guidelines in the care of the cardiac arrest subjects. Pulseless subjects underwent an initial cardiac US examination on presentation and sequential examinations every 3 to 5 minutes during the resuscitation at the time during which carotid pulse checks were carried out to determine if there was cardiac kinetic activity. Sonographic evidence of cardiac kinetic activity was defined as any detected motion within the heart: atrial, valvular, or ventricular. Emergency physician sonographers used the transthoracic, subxiphoid, or the parasternal longaxis B-mode–view approaches to visualize the heart with 3.5-MHz curvilinear or sector probes. The end points of the resuscitations were death, return of spontaneous circulation (ROSC), admission to the hospital, survival to discharge from the hospital, and functional status after discharge. Return of spontaneous circulation was defined as resumption of a palpable pulse and blood pressure, even if transient. Recorded outcome variables were demographic data, the presence of PEA (evidence of electrical rhythm) or asystole, results of cardiac US examinations (evidence or absence of US-detected cardiac kinetic activity), whether the cardiac arrest occurred in the prehospital environment or the ED, estimated duration of prehospital cardiac arrest resuscitation, duration of ED cardiac arrest resuscitations, cardiac US evidence of pericardial effusion, and resuscitation outcomes. Whenever applicable for our inhospital study design, data were collected according to the recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest, the Utstein Style [9]. All data were contemporaneously noted on a standardized data collection forms along with a still image of the cardiac US. Statistical analysis included descriptive statistics, 95% confidence intervals based on proportions, univariate logistic regression, and receiver operating characteristic curve analysis; a was set at .05, 2-tailed.

3. Results 2. Methods This institutional review board–approved multicenter trial, carried out at 4 academic EDs with volumes of at least 40 000 visits per year and served by extensive emergency medical services systems, all with both basic life support and advanced life support units but with advanced life support units transporting cardiac arrest patients. These study sites prospectively enrolled a convenience sample of nontrauma pulseless adult subjects in cardiac arrest as part of the Sonography Outcomes Assessment Program. In all 4 centers, the policy and practice was to continue resuscitation once

During a 12-month period, 70 subjects were enrolled, each center with at least 10 subjects. Demographically, the cohort consisted of 61% males and 39% females with an age range of 16 to 94 years. Resuscitations were begun in the prehospital environment for 67 of 70 cardiac arrest subjects; the prehospital resuscitation intervals ranged from 5 to 77 minutes. Resuscitation duration intervals for 60 of the 70 subjects lasted less than 12 minutes in the ED. Within the asystole cohort, 31 of 36 resuscitation intervals lasted 12 minutes or less in the ED; for the PEA cohort, 29 of 34 lasted 12 minutes or less. As a point of comparison to the

Sonographically identified cardiac activity prediction of cardiac arrest resuscitation

461

Table 1

Rhythm ROSC + ROSC

+ Sonographic cardiac activity

+ Sonographic cardiac activity

Sonographic cardiac activity

Sonographic cardiac activity

PEA 8a 3

Asystole 0 0

PEA 0 23

Asystole 0 36

a

Only 1 of 8 survived to hospital discharge with scores of 1 on the Glasgow-Pittsburgh Cerebral Performance scale (good cerebral performance) and 1 in the Overall Performance category (capable of normal life) [9].

predictive abilities of cardiac sonography, a logistic regression model was conducted to control for cardiac arrest resuscitation intervals both in the prehospital and ED environments to correlate resuscitation times with ROSC. The univariate logistic regression model for prehospital and ED resuscitation intervals showed correlation for 4 (52%) of the 7 patients that had ROSC and subsequent admission to the hospital with an area under the curve of 0.26 (95% CI, 0-0.52; P = NS). Therefore, duration of resuscitation time intervals of our cardiac arrest cohort was not significantly associated with ROSC. The presenting rhythms before US were asystole in 36 and PEA in 34 (Table 1). Serial cardiac USs, more than a single US examination, were carried out in 53 of the subjects. Single cardiac US examinations were carried out in 17 of the subject: 9 for asystole subjects and 8 for PEA subjects. Eight subjects presenting with PEA had ROSC and survived long enough for admission to the intensive care unit; all subjects with asystole died. Fifty-nine subjects had no evidence of sonographically identifiable cardiac kinetic activity and their resuscitations were uniformly unsuccessful. Subjects with asystole on presenting rhythm strip uniformly lacked sonographic evidence myocardial contractions and did not have ROSC. Twentythree subjects (68%) presenting with PEA had no sonographic evidence of cardiac activity (true EMD) and did not have ROSC. The presence of sonographically identified cardiac kinetic activity (pseudo-EMD) during the resuscitation was identified in 11 PEA subjects (32%). Eight of this pseudo-EMD subgroup had sonographic evidence of cardiac activity and also had ROSC with 6 surviving for less than 12 hours after intensive care admission. Two subjects, who were noted to have cardiac activity both in their presenting sonographic examination and all subsequent examinations, survived a more than 24-hour stay in the intensive care unit. One of these subjects survived to hospital discharge after hemodialysis for renal failure–induced hyperkalemia with scores of 1 on the Glasgow-Pittsburgh Cerebral Performance scale (good cerebral performance) and 1 in the Overall Performance category (capable of normal life). None of the subjects in the PEA cohort were noted to have a pericardial effusion or tamponade as a potential cause of their cardiac arrest. The accuracy (true positives [US identification of cardiac activity with ROSC] + true negatives [US confirmation of no cardiac activity and no ROSC] / true positives + false

positives [US identification of cardiac activity with no ROSC] + false negatives [US identification of no cardiac activity with ROSC] + true negatives) of cardiac US to predict ROSC vs death in our cardiac arrest cohort was 0.95 (95% CI, 0.92-1.0). The accuracy of cardiac US to predict survival to hospital discharge vs death was 0.87 (95% CI, 0.79-0.95).

4. Discussion The resuscitation of ED patients with PEA or asystole may entail considerable time and effort with only a very small chance of success. Unfortunately, there is no consensus as to when the resuscitation should be terminated or continued. Time to cardiopulmonary resuscitation, time to defibrillation, comorbid diseases, precardiac arrest condition, initial arrest rhythm, and duration of resuscitative effort have been reported to have prognostic importance for cardiac arrest patients [10]. A time-efficient, reliable means of determining those patients who are potentially salvageable from cardiopulmonary resuscitation from those who are not would be of great clinical use. Although there are few studies on the use of transthoracic B-mode cardiac US to predict the outcome of cardiac arrest, this study is consistent with prior data that showed that patients with sonographically identified cardiac standstill do not have ROSC regardless of the initial cardiac electrical rhythm [7,8]. Cardiac sonography confirmed lack of cardiac kinetic motion in our entire asystole cohort. It also distinguished between pseudo-EMD and true EMD in our PEA cohort. Our entire true-EMD cohort like our asystole cohort did not have ROSC, which is consistent with previously published data regarding true EMD [5,7]. Our data further support previously published data that sonographically identifiable cardiac kinetic activity is associated with ROSC [8,10,11]. Specifically, our study suggests that sonographic evidence of cardiac kinetic motion during cardiopulmonary resuscitation is a potential predictor of ROSC (accuracy = 0.95, 95% CI, 0.92-1.0). This is evidenced by the fact that 8 of 11 subjects in our PEA cohort with cardiac kinetic motion on US examination had ROSC. The sonographic determination of cardiac kinetic motion was also an accurate predictor of survival to hospital discharge vs death (accuracy = 0.87, 95% CI, 0.79-0.95); 1 subject survived to hospital discharge with scores of 1 on the Glasgow-Pittsburgh

462 Cerebral Performance and Overall Performance scales [9]. Other physician investigators studying pseudo-EMD both invasively and with US have noted pseudo-EMD as proportion of the total PEA population to range from 42% to 86% [5,11-13]. It is notable that our pseudo-EMD population, which was primarily a population that had their resuscitations started in the prehospital environment, only constituted 23% of our PEA cohort. Previously suggested parameters for the prediction of ROSC, including prehospital resuscitation time intervals and ED resuscitation time intervals, were not reliable predictors for this cohort [14]. For this data set, there was no correlation between ROSC and time duration of resuscitation for patients with PEA. These data suggest that duration of resuscitation may be a more useful predictor in the setting of cardiac arrests secondary to ventricular tachyarrhythmias than for subjects with PEA and asystole. Important limitations of our study are notable. The total sample size of 70 was relatively modest and limits the confidence of our speculations that the presence or absence of sonographically detected cardiac activity can predict positive or negative outcomes. The predictive accuracy of cardiac standstill on US appears to be reliable; however, a large sample size of subjects with pseudo-EMD would be required to confidently associate ROSC with the sonographic detection of cardiac kinetic motion. Although the multicenter nature of our trial may enhance the universality of our results, the fact that our subjects were a convenience rather than a consecutive sample of cardiac arrest patients should be noted. This may be partly explained by (1) the availability of the US machine and (2) the availability of the study participant physician sonographers within their respective departments. Further 17 of 70 subjects did not receive sequential sonographic examinations after initial cardiac US examination, or the sequential examinations were not recorded on the data sheets. This is partially mitigated by the fact that of the 70 subjects, 60 had resuscitations lasting less than 12 minutes. The sonographic cardiac examinations were not videotaped and were not reviewed for accuracy. Finally, although there were instructions in the study protocol that cardiac US results do not influence the standard American Heart Association decision-making guidelines, the study was not blinded in that the sonographers were aware of the progress and results of the resuscitative efforts.

P. Salen et al.

5. Conclusion These data support the idea that transthoracic B-mode cardiac sonography can provide physicians involved in the resuscitation of pulseless patients with asystole and PEA with information that may help in making pertinent management decisions regarding cardiopulmonary resuscitation.

References [1] Stapleton ER, Aufderheide TP, Hazinski MF, Cummins RO. Adult CPR. BLS Healthcare Providers 2001:75. [2] Ochoa FJ, Ramalle-Gomara E, Carpintero JM, Garcia A, Saralegui I. Competence of health professionals to check the carotid pulse. Resuscitation 1998;37:173 - 5. [3] Flesche CW, Breuer S, Mandel LP, Beivik H, Tarnow J. The ability of health professionals to check the carotid pulse. Circulation 1994; 90:I - 288. [4] Eberle B, Dick WF, Schneider T, Wisser G, Doetsch S, Tzanova I. Checking the carotid pulse: diagnostic accuracy of first responders in patients with and without a pulse. Resuscitation 1996;33:107 - 16. [5] Paradis NA, Martin GB, Goetting MG, Rivers EP, Feingold M, Nowak RM. Aortic pressure during human cardiac arrest: identification of pseudo-electromechanical dissociation. Chest 1992;101(1):123 - 8. [6] Sanders AB, Kern KB, Berg RA. Searching for a predictive rule for terminating cardiopulmonary resuscitation. Acad Emerg Med 2001;8: 654 - 7. [7] Blaivas M, Fox JC. Outcome in cardiac arrest patients found to have cardiac standstill on the bedside emergency department echocardiogram. Acad Emerg Med 2001;8:616 - 21. [8] Salen P, O’Connor R, Sierzenski P, et al. Can cardiac sonography and oceanography be used independently and in combination to predict resuscitation outcomes? Acad Emerg Med 2001;8:610 - 5. [9] Cummins RO, Chamberlain DA, Abramson NS, et al. Recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein Style. Circulation 1991;84(2):960 - 75. [10] Part 6: Advanced Cardiovascular Life Support: Section 7 Algorithm Approach to ACLS Emergencies. August 2000. Circulation 2000:I-150 - 2 [Suppl I]. [11] Bocka JJ, Overton DT, Hauser A. Electromechanical dissociation in human beings; an echocardiographic evaluation. Ann Emerg Med 1988;17(5):450 - 2. [12] Mayron R, Gaudio FE, Plummer D, Asinger R, Elsperger J. Echocardiography performed by emergency physicians: impact on diagnosis and therapy. Ann Emerg Med 1988;17(2):73 - 7. [13] Tayal VS, Kline JA. Emergency echocardiography to detect pericardial effusions in patients in PEA and near PEA states. Resuscitation 2003;59:315 - 8. [14] Part 2: Ethical aspects of CPR and ECC. August 2000. Circulation 2000:I 14 - 9 [Suppl I].