The Mechanism of Blood Flow During Closed Chest Cardiac Massage in Humans: Transesophageal Echocardiographic Observations
STUART T. HIGANO, M.D., JAE K. OH, M.D., Division ofCardiovascular Diseases and Internal Medicine; GORDON A. EWY, M.D., Section of Cardiology, University ofArizona Health Sciences Center, Tucson, Arizona; JAMES B. SEWARD, M.D., Division ofCardiovascular Diseases and Internal Medicine
Despite years of research, the mechanism of forward blood flow during closed chest cardiac massage remains controversial. Two theories have been suggested: the cardiac pump theory and the thoracic pump theory. Transesophageal echocardiography offers a new approach for study of the flows and cardiac morphologic features during chest compressions in humans. Case reports are presented to illustrate the use of transesophageal echocardiography during cardiopulmonary resuscitation. The findings included right and left ventricular compression, closure of the mitral valve during compression, opening of the mitral valve during the release phase, and atrioventricular valvular regurgitation during compression, indicating a positive ventricular-to-atrial pressure gradient. These findings suggest that direct cardiac compression was the predominant mechanism of forward blood flow during cardiopulmonary resuscitation in these patients. An understanding of the actual mechanisms involved is necessary ifimproved cardiopulmonary resuscitative techniques or adjuncts are to be rationally developed for enhancing the outcome of resuscitation.
the cardiac pump theory and the thoracic pump theory. During the first 2 decades after closed chest "cardiac massage" had initially been described, the cardiac pump theory prevailed. In this model, forward blood flow results from compression of the heart between the sternum and the paraspinal structures (Fig. 1). Research and clinical observations during the past decade led to the thoracic pump theory, in which external chest compressions increase the intrathoracic pressure and thus intracardiac and intrathoracic vascular pressures.v" Forward blood flow results from the arteriovenous pressure gradiAddress reprint requests to Dr. J. K. Oh, Division of Car- ents thereby generated (Fig. 2). In the mid1980s, studies with use of extensive instrumendiovascular Diseases, Mayo Clinic, Rochester, MN 55905.
Emergency cardiac care was revolutionized with the introduction of closed chest cardiopulmonary resuscitation (CPR) in 1960.1 For the first time, a technique produced some forward blood flow in patients with cardiac arrest awaiting more definitive pharmacologic or electrical therapeutic intervention. Although closed chest CPR has been a clinically accepted procedure for 30 years, the mechanism of forward blood flow during external cardiac compression remains controversial. Two theories have been proposed:
Mayo Clin Proc 65:1432-1440, 1990
1432
Mayo Clin Proc, November 1990, Vol 65
ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE 1433
tation in dogs strongly supported the cardiac pump model. Other studies have suggested that both mechanisms may be operative in different clinical settings and with various CPR techniques. A clear understanding ofthe underlying mechanism or mechanisms in CPR is essential for the development of optimal CPR techniques. Unfortunately, studies in humans are difficult to perform. Almost all CPR research protocols reported to date have been done in patients in whom standard CPR protocols have failed, and the subjects were without adequate perfusion for extremely long periods before the experimental techniques were attempted. Transthoracic echocardiography has been reported in patients undergoing closed chest CPR,3,4 but the Thorax ability to perform adequate closed chest CPR and to obtain adequate echocardiographic images concomitantly might be questioned. Transesophageal echocardiography is now a practical method of studying the cardiac morphologic changes and flows in humans during external Fig. 1. Diagram of "cardiac pump theory," illustrating cardiac compressions.v? In transesophageal compression of the heart between sternum and paraspinal echocardiography, a side-viewing phased-array structures and ejection of blood into circulation during transducer is placed into the esophagus immedi- closed chest cardiac massage. (Heavy arrow = external compression.) Note that heart is compressed and ately adjacent and posterior to the heart. Excel- chest atrioventricular valve closes. Numbers represent theoretilent images of the myocardium can be obtained cal pressures in millimeters of mercury. (Modified from during external cardiac compressions. In addi- Weisfeldt ML, Chandra N: Physiology of cardiopulmonary tion to the two-dimensional images, Doppler and resuscitation. Annu Rev Med 32:435-442, 1981.) color flow imaging can be obtained for assessing to his local emergency room. His blood pressure flow. The following case reports demonstrate the was 80/50 mm Hg, and he was transferred to our use of transesophageal echoeardiography dur- medical center. On arrival, his vital signs showed ing CPR. This technique may help elucidate a blood pressure of 90/55 mm Hg, heart rate of the mechanism of blood flow during CPR in 106 beats/min, and respiratory rate of 36/min. humans. Thejugular venous pressure was increased at 20 em of water. The cardiac examination revealed a regular rhythm and normal first and second REPORT OF CASES heart sounds. A prominent third heart sound Case I.-A critically ill 72-year-old man was transferred to our institution because of near was noted. He had a grade 2 (on the basis of 1 to cardiovascular collapse, lactic acidosis, hyperka- 6) holosystolic murmur at the apex extending to lemia, and azotemia. One week before this the midclavicular line. Examination of the lungs admission, he had been admitted to his local disclosed bilateral adventitial sounds, consishospital with mild left-sided chest pain and tent with pulmonary edema. No peripheral dyspnea. After 5 days of treatment for conges- edema was present. The following laboratory findings were retive heart failure, he had been dismissed. Two days after dismissal, severe dyspnea and ported: serum potassium, 6.2 meq/liter; bicarleft-sided chest pain prompted the patient to go bonate, 12 meq/liter; creatinine, 3.6 mg/dl; and
1434 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
Thorax
0--+85
Fig. 2. Diagram of "thoracic pump theory," illustrating proposed events during closed chest cardiac massage. Increase in intrathoracic pressure caused by external compression is transmitted to all intrathoracic vessels. Pressure is transmitted unequally to arteries and veins as a result of venous valve located at thoracic outlet. The consequence is an arteriovenous pressure gradient and forward blood flow. During compression, left side of heart acts as a passive conduit for blood being ejected into systemic circulation. Mitral valve remains open during compression, and transmitral flow occurs. Numbers represent theoretical pressures in millimeters of mercury. (Modified from Weisfeldt ML, Chandra N: Physiology of cardiopulmonary resuscitation. Annu Rev Med 32:435-442, 1981.)
lactate, 15.6 ug/ml. The creatine kinase concentration was 124 U/liter (normal range, 23 to 99 U/liter) with 14% MB fraction. An electrocardiogram showed sinus rhythm at a rate of 88 beats/ min. A first-degree atrioventricular conduction delay and a left ventricular conduction defect with secondary ST- and T-wave abnormalities were noted. A chest roentgenogram disclosed bilateral pleural effusions with bilateral basilar interstitial infiltrates. Bedside echocardiography revealed the anterior wall to be thinned and akinetic with septal dyskinesia; mitral regurgitation was also noted. No shunts or pericardial
Mayo Clin Proc, November 1990, Vol 65
effusions were detected, and the papillary muscles appeared intact. Intravenous administration of fluids and pressor agents did not improve the hemodynamics of the patient. He was intubated, and mechanical ventilation was initiated. Insertion of an intra-aortic balloon pump improved his hemodynamics. The blood pressure increased to 115/60 mm Hg, the cardiac index increased to 2.2 liters/min per m", and the pulmonary wedge pressure declined to 20 mm Hg. Despite this initial improvement, on the second hospital day, sudden cardiovascular collapse and sinus rhythm developed, and CPR was initiated. Approximately 5 minutes after the resuscitative efforts were begun, transesophageal echocardiography was performed to search for potentially reversible mechanical complications of acute myocardial infarction. Chest compressions were withheld only momentarily during placement of the transesophageal probe and also after appropriate windows were obtained to record native cardiac motion. The transesophageal findings were consistent with severe myocardial dysfunction, the only functioning myocardium being the basal septum. The ventricular septum was intact, and no pericardial effusion was found. Transesophageal echocardiographic monitoring of chest compressions showed virtual obliteration of the right ventricular cavity, severe compression ofthe left ventricular cavity, closure of the mitral valve during compression, and subsequent opening of the mitral valve during the release phase (Fig. 3). Resuscitative attempts were discontinued after 30 minutes, and the patient died. Case 2.-A 70-year-old woman was brought to our institution by paramedics after she had a cardiac arrest. She had a history of hypertension, long-standing diabetes mellitus, a silent myocardial infarction in 1982, and severe left ventricular dysfunction; an echocardiogram 1 year earlier had shown an ejection fraction of 15 to 20%, apical dyskinesia, and inferior akinesia. On the day of admission, the patient had suffered sudden loss of consciousness. Bystanders at the scene immediately initiated CPR, and she was brought to our institution. Her initial
Mayo Clio Proc, November 1990, Vol 65
ECHOCARDlOGRAPHY DURING CLOSED CARDIAC MASSAGE 1435
Fig. 3 (case 1). Transesophageal echocardiographic images, showing sequential compression and release frames during closed chest cardiac massage. During compression, right ventricle (RY) is obliterated, left ventricle (L Y) is compressed, and mitral valve closes, an indication of cessation of transmitral flow. During release, mitral valve opens, and left ventricular filling occurs. ami = anterior mitral leaflet; LA = left atrium.
rhythm was fine ventricular fibrillation, which was converted to a stable rhythm by administration of epinephrine, sodium bicarbonate, lidocaine, and four countershocks. At this time, an electrocardiogram revealed an accelerated idioventricular rhythm, and a chest roentgenogram disclosed bilateral pulmonary infiltrates. Shortly after transfer to the coronary-care unit, the patient suffered another cardiac arrest with ventricular fibrillation that was resistant to repeated defibrillation. CPR was reinitiated. Transesophageal echocardiography was performed along with the resuscitative attempt to search for reversible causes of the collapse. Important findings, with chest compressions momentarily withheld, included no appreciable cardiac contraction and no evidence of tamponade, ventricular septal rupture, or papillary muscle rupture. Spontaneous echocardiographic contrast in the left ventricular outflow tract suggested a low-flow state. The chest compressions were also monitored during the resuscitation. The notable findings were compression ofthe right and left ventricular cavities and closure of the mitral and tricus-
pid valves during compression systole (Fig. 4). Color flow imaging revealed mild mitral regurgitation and severe tricuspid regurgitation during compressions. During the release phase, or compression diastole, the atrioventricular valves opened, and color flow imaging demonstrated rapid transvalvular flow (Fig. 5).
DISCUSSION CPR was first described in 1960 by Kouwenhoven and associates 1 as closed chest cardiac massage, an alternative to open cardiac massage. It was quickly disseminated as a technique to restore forward blood flow in victims of acute cardiac arrest, such as that occurring in drowning, electrocution, or the operating room. The heart was thought to be compressed, or "massaged," between the sternum and the spine, the result being ejection of blood from the left ventricle into the systemic circulation: The heart is limited anteriorly by the sternum and posteriorly by the vertebral bodies. Its lateral movement is restricted by the pericardium. Pressure on the sternum compresses the heart between it and the spine, forcing out blood. Relaxation of the pressure allows the heart to fill.'
1436 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
Mayo Cltn Proc, November 1990, Vol 65
Fig. 4 (case 2). Transesophageal echocardiographic images, showing sequential compression and release frames during closed chest cardiac massage. During compression, left ventricle (LV) is compressed, mitral valve closes (an indication of cessation oftransmitral flow), and aortic valve opens. With release, mitral valve opens, and left ventricular filling occurs. ami = anterior mitral leaflet; Ao = aorta; LA = left atrium; RV = right ventricle.
This forward stroke volume, which occurs tate. Conversely, victims with barrel chests, in during compression systole (compression phase), whom cardiac compression seemed unlikely, necessitates closure of the mitral valve and could be resuscitated. Furthermore, hemodyopening ofthe aortic valve. This mechanism has namic measurements obtained during CPR been termed the "cardiac pump modeL" In the revealed a substantial increase in right atrial cardiac pump model, the left ventricle then fills pressure during compression systole." Invesduring compression diastole (the release phase) tigators concluded that external cardiac comwith rapid transmitral flow (Fig. 1). Right heart pressions were inefficient because ofthese large hemodynamics were presumed to be analogous. venous pulse waves. Furthermore, palpable With use of this model, optimized CPR tech- femoral pulses during CPR only reflected presnique would include the maximal compression sure waves but did not predict forward blood rate while still allowing for adequate left ven- flow. Further evidence that challenged the cartricular filling-that is, stroke volume (thus diac pump model was provided by Criley and maximizing cardiac output, which is equiva- associates," who described "cough CPR" in 1976. lent to the compression rate times the stroke In their description of cough-induced cardiac volume). compression, asystolic patients in the catheSeveral subsequent clinical observations terization laboratory were kept conscious for 24 challenged this cardiac pump modeL Victims of to 39 seconds by repeated coughing. This techcardiac arrest with altered anatomy of the tho- nique led to the consideration of intrathoracic racic cage did not respond to CPR as predicted by pressure as the primary mechanism in the forthis modeL For example, victims with flail ward ejection of blood during CPR. This mechchests, in whom direct cardiac compression was anism has been termed the "thoracic pump easily performed, were often difficult to resusci- modeL"
Mayo Clin Proc, November 1990, Vol 65
ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE 1437
Fig. 5 (case 2). Transesophageal echocardiographic images with color flow mapping, showing sequential compression and release frames during closed chest cardiac massage. During compression, left ventricle (LV) is compressed, and mitral valve closes. Mitral regurgitation is demonstrated by mosaic color (double arrows) in left atrium (LA). A positive left ventricular-toleft atrial pressure gradient must exist for mitral regurgitation to occur. With release, mitral valve opens, and color flow imaging demonstrates left ventricular filling (brilliant blue area). RV = right ventricle.
In this thoracic pump model, external cardiac compression causes a general increase in intrathoracic pressure, which is transmitted to all cardiac chambers and vessels in the thorax (Fig. 2). This increased pressure generates an arteriovenous pressure gradient that results in forward flow. Critical to the production of forward flow is the development of the arteriovenous pressure gradients by the venous valves located at the thoracic outlet, which prevent transmission of increased thoracic pressure to the venous circulation. The presence of these valves has been substantiated echocardiographically, physiologically, and anatomically.v-" In this model, during compression systole, the left side ofthe heart acts as a passive conduit for thoracic blood being injected into the systemic circulation (that is, transmitral flow must occur during compression systole). The right side of the heart would presumably have no forward flow during compression systole, because of equal pressures in the right atrium, right ventricle,
and pulmonary artery. With use of this model, optimized CPR technique would include a fairly long duration of compression-that is, a duty cycle equal to or greater than 50%-to maintain the arteriovenous pressure gradient and forward flow. Direct evidence in support of the thoracic pump model was reported in 1980 by Rudikoff and associates," who described the hemodynamics of CPR in 15 dogs with cardiac arrest. They showed equal compression pressures in the left ventricle, aorta, right atrium, pulmonary artery, and esophagus, in conjunction with unequal transmission of pressure to the extrathoracic vessels. They postulated that the venous system had collapsed at the thoracic outlet. Subsequently, venous valves at the thoracic outlet were discovered. The arteriovenous pressure gradient generated across these valves resulted in forward blood flow. Furthermore, abdominal binding to prevent paradoxic diaphragmatic motion during compression led to an
1438 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
increase in aortic pressure and in carotid blood flow. In 1981, Niemann and colleagues'" performed cineangiography during CPR in 17 dogs with cardiac arrest and demonstrated forward flow without appreciable change in left ventricular dimensions. These findings signified- that the left ventricle was acting as a passive conduit and not a pump. These investigators also showed that simultaneous chest compressions and lung inflation increased blood flow in the left side of the heart and in the carotid artery. In 1981, Werner and co-workers! performed the first human studies with two-dimensional transthoracic echocardiographic visualization of the cardiac valves during CPR in five patients. They noted that the mitral valve remained open during the entire compression and release phase and that the aortic valve opened only during compression systole. Similarly, Rich and associates," who studied four patients during CPR, found no change in the left ventricular internal dimensions during chest compression and also aortic valve and mitral valve motion similar to that described by Werner and colleagues." These findings further supported the thoracic pump model. The thoracic pump model was first challenged by a group from Duke University. Maier and coworkers!' studied intact dogs with long-term instrumentation during CPR with various compression forces and rates. Increasing the compression force, measured by peak intrathoracic pressure, did not result in an increased stroke volume. With increasing compression rate, however, the stroke volume was constant, and cardiac output and coronary blood flow increased significantly. This finding was strong experimental evidence that direct cardiac compression was a major determinant of stroke volume during CPR. They emphasized the importance of high-impulse compressions, or high-impulse CPR, and hypothesized that the force transferred to the heart is a direct function of the initial momentum (defined as the compression mass times the velocity). They noted that previous investigators had used pneumatic compression devices, which result in low-momentum compressions. They also used high-fidelity mi-
Mayo Clin Proc, November 1990, Vol 65
cromanometer catheters to measure intrathoracic pressure during compressions; thus, they avoided possible pressure transients that may occur with saline-filled catheter manometers with imperfect frequency response characteristics. The intrathoracic pressures measured in this experimental model were lower than those previously reported and lower than the intracavitary pressures, evidence in support of the cardiac pump model. More recent studies from the Duke University group-s and others.P who used transthoracic and transesophageal echocardiography in dogs undergoing CPR, demonstrated mitral valve closure during compression in all but lowimpulse CPR. Subsequently, the Duke researchers demonstrated a pronounced influence of the rate of compression on the initial success of resuscitation after 30 minutes of CPR in dogs.P In these studies, the duty cycle was maintained at 50%, independent of the rate of compression. These data again supported the importance of the cardiac pump model of blood flow during CPR. Unfortunately, studies in humans are difficult to perform, and because of considerable differences in configuration of the chest wall, extrapolation of data from dogs to humans may not be justified. 15 The predominant mechanism of blood flow during CPR in humans remains controversial. The models are not mutually exclusive, and one may predominate over the other in any given resuscitation. Several factors may influence the mechanism of forward blood flow in any specific resuscitative effort, including technique of compression, thoracic anatomy, cardiac pathologic changes, underlying cardiac rhythm, and duration of resuscitation. A technique with a high rate and force may favor a direct compressive mechanismY Abnormalities of the thoracic cage, such as barrel chest, pectus excavatum, or flail chest, may play important roles. Underlying myocardial or atrioventricular valvular disease may impair passive ventricular filling during compression diastole or forward ventricular emptying during compression systole. These factors may favor an intrathoracic pressure mechanism.
Mayo Clin Proc, November 1990, Vol 65
ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE 1439
Sinus rhythm may improve mitral valve competence and thus favor direct compression." A lengthy resuscitative attempt may favor an intrathoracic pressure mechanism, perhaps because of changes in ventricular compliance during prolonged myocardial ischemia. Additionally, in some patients, direct right heart compression may predominate, whereas in other patients, left heart compressive mechanisms may predominate. In the two patients described in detail in our current report, transesophageal echocardiography demonstrated direct compression ofthe right and left ventricles during CPR compressions. Although some motion of the heart relative to the esophageal transducer may have occurred, such movement is likely to have been small and would not account for the -amount of compression seen. Additionally, the mitral valve was closed during compression systole. In the second patient,the presence of atrioventricular valvular regurgitation during compression suggested that a positive ventricular-to-atrial pressure gradient existed. These findings are consistent with a direct cardiac compressive mechanism of blood flow during CPR in these patients and are inconsistent with the thoracic pump mechanism. For the thoracic pump theory of CPR, the mitral valve must be open during compression. In both patients, the mitral valve generally remained closed throughout compression systole, occasionally opening toward the end of compression at a time when ventricular pressures were likely to be decreasing because of ventricular emptying. This mitral valve opening may have been due to intermittent increases in left atrial pressure from intrinsic atrial contraction. With release of the external compression, the mitral valve opened rapidly. During mid to late compression diastole, the mitral valve often closed, opening only intermittently presumably because of intrinsic atrial contraction. In the first patient, a less forceful compressive technique resulted in less direct cardiac compression; however, inadequate views of the mitral valve motion were obtained. Furthermore, transesophageal echocardiographic observations will be necessary to elucidate the predominant mechanism of forward
blood flow during CPR in humans in several clinical situations. These case reports demonstrate the utility of transesophageal echocardiography for studying CPR in humans.
CONCLUSION CPR has become a well-established technique to support circulation after cardiac arrest. Approximately 1,000 prehospital sudden deaths occur daily in the United States. An estimated 40% of patients with out-of-hospital cardiac arrests and substantiated ventricular fibrillation could be resuscitated; thus, approximately 100,000 lives could be saved annually. Although the primary goal during the past 3 decades has been to teach CPR techniques to as many people as possible, improved rates of resuscitation will also necessitate improved techniques or adjuncts (or both) to augment the low cardiac output, cerebral flow, and coronary blood flow generated by current CPR techniques. New techniques involve modifications in rate of compression, impulse of compression, duty cycle length, and, recently, simultaneous lung inflations and interposed abdominal compressions. ll ,14,16-18 Potential adjuncts to CPR include abdominal binding, military antishock trouser garments, and inflatable vests. 19,20 Each of these modifications was devised to improve the hemodynamics in either the cardiac pump model or the thoracic pump model of CPR. For further rational development of these new CPR techniques, or perhaps some as yet undevised techniques, an understanding of the predominant mechanism of blood flow during CPR in humans is necessary. With acquisition of such knowledge, the anticipated outcome will be evolution of new CPR techniques that will improve cardiac output and oxygen delivery to the brain and myocardium and thereby increase the number of patients successfully resuscitated. ACKNOWLEDGMENT We greatly appreciate the expert secretarial assistance of Kathleen A. Budensiek, who prepared the submitted manuscript.
1440 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
REFERENCES 1.
2. 3.
4.
5.
6.
7.
8.
9.
10.
11.
Kouwenhoven WE, Jude JR, Knickerbocker GG: Closed-chest cardiac massage. JAMA 173:10641067, 1960 MacKenzie GJ, Taylor SH, McDonald AH, Donald KW: Haemodynamic effects of external cardiac compression. Lancet 1:1342-1345, 1964 Rich S, Wix HL, Shapiro EP: Clinical assessment of heart chamber size and valve motion during cardiopulmonary resuscitation by two-dimensional echocardiography. Am Heart J 102:368-373, 1981 WernerJA, Greene HL, Janko CL, Cobb LA: Visualization of cardiac valve motion in man during external chest compression usingtwo-dimensional echocardiography: implications regarding the mechanism of blood flow. Circulation 63:1417-1421,1981 SewardJB,KhandheriaBK, OhJK,AbeIMD, Hughes RW Jr, Edwards WD, Nichols BA, Freeman WK, Tajik AJ: Transesophageal echocardiography: technique, anatomic correlations, implementation, and clinical applications. Mayo Clin Proc 63:649-680, 1988 Uenishi M, Sugimoto H, Sawada Y, Terai C, Yoshioka T, Sugimoto T: Transesophageal echocardiography during external chest compression in humans (letter to the editor). Anesthesiology 60:618, 1984 Clements FM, de Bruijn NP, Kisslo JA: Transesophageal echocardiographic observations in a patient undergoing closed-chest massage. Anesthesiology 64:826-828, 1986 Criley JM, Blaufuss AH, Kissel GL: Cough-induced cardiac compression: self-administeredform ofcardiopulmonary resuscitation. JAMA 236:1246-1250, 1976 Rudikoff MT, Maughan WL, Effron M, Freund P, Weisfeldt ML: Mechanisms of blood flow during cardiopulmonary resuscitation. Circulation 61:345352, 1980 Niemann JT, Rosborough JP, Hausknecht M, Garner D, Criley JM: Pressure-synchronized cineangiography during experimental cardiopulmonary resuscitation. Circulation 64:985-991, 1981 Maier GW, Tyson GS Jr, Olsen CO, Kernstein KH, Davis JW, Conn EH, Sabiston DC Jr, Rankin JS: The
12.
13.
14.
15. 16.
17.
18. 19.
20.
Mayo Clin Proc, November 1990, Vol 65
physiology of external cardiac massage: high-impulse cardiopulmonary resuscitation. Circulation 70:86-101, 1984 Feneley MP, Maier GW, Gaynor JW, Gall SA, Kisslo JA, Davis JW, Rankin JS: Sequence of mitral valve motion and transmitral blood flow during manual cardiopulmonary resuscitation in dogs. Circulation 76:363-375, 1987 Deshmukh HG, Weil MH, Rackow EC, Trevino R, Bisera J: Echocardiographic observations during cardiopulmonary resuscitation: a preliminary report. Crit Care Med 13:904-906, 1985 Feneley MP, Maier GW, Kern KB, Gaynor JW, Gall SA Jr, Sanders AB, Raessler K, Muhlbaier LH, Rankin JS, Ewy GA: Influence of compression rate on initial success of resuscitation and 24 hour survival after prolonged manual cardiopulmonary resuscitation in dogs. Circulation 77:240-250, 1988 Barsan WG, Levy RC: Experimental design for study of cardiopulmonary resuscitation in dogs. Ann Emerg Med 10:135-137,1981 Taylor GJ, Tucker WM, Greene HL, Rudikoff MT, Weisfeldt ML: Importance of prolonged compression during cardiopulmonary resuscitation in man. N Engl J Med 296: 1515-1517, 1977 Chandra N, RudikoffM, Weisfeldt ML: Simultaneous chest compression and ventilation at high airway pressure during cardiopulmonary resuscitation. Lancet 1:175-178, 1980 BabbsCF, TackerWAJr: Cardiopulmonary resuscitation with interposed abdominal compression. Circulation 74 (SuppI4):IV37-IV41, 1986 NiemannJT, RosboroughJP, CrileyJM: Continuous external counterpressure during closed-chest resuscitation: a critical appraisal of the military antishock trouser garment and abdominal binder. Circulation 74 (SuppI4):IV102-IV107, 1986 Halperin HR, Guerci AD, Chandra N, Herskowitz A, Tsitlik JE, Niskanen RA, Wurmb E, Weisfeldt ML: Vest inflation without simultaneous ventilation during cardiac arrest in dogs: improved survival from prolonged cardiopulmonary resuscitation. Circulation 74:1407-1415,1986
STUART T. HIGANO, M.D., JAE K. OH, M.D., Division ofCardiovascular Diseases and Internal Medicine; GORDON A. EWY, M.D., Section of Cardiology, University ofArizona Health Sciences Center, Tucson, Arizona; JAMES B. SEWARD, M.D., Division ofCardiovascular Diseases and Internal Medicine
Despite years of research, the mechanism of forward blood flow during closed chest cardiac massage remains controversial. Two theories have been suggested: the cardiac pump theory and the thoracic pump theory. Transesophageal echocardiography offers a new approach for study of the flows and cardiac morphologic features during chest compressions in humans. Case reports are presented to illustrate the use of transesophageal echocardiography during cardiopulmonary resuscitation. The findings included right and left ventricular compression, closure of the mitral valve during compression, opening of the mitral valve during the release phase, and atrioventricular valvular regurgitation during compression, indicating a positive ventricular-to-atrial pressure gradient. These findings suggest that direct cardiac compression was the predominant mechanism of forward blood flow during cardiopulmonary resuscitation in these patients. An understanding of the actual mechanisms involved is necessary ifimproved cardiopulmonary resuscitative techniques or adjuncts are to be rationally developed for enhancing the outcome of resuscitation.
the cardiac pump theory and the thoracic pump theory. During the first 2 decades after closed chest "cardiac massage" had initially been described, the cardiac pump theory prevailed. In this model, forward blood flow results from compression of the heart between the sternum and the paraspinal structures (Fig. 1). Research and clinical observations during the past decade led to the thoracic pump theory, in which external chest compressions increase the intrathoracic pressure and thus intracardiac and intrathoracic vascular pressures.v" Forward blood flow results from the arteriovenous pressure gradiAddress reprint requests to Dr. J. K. Oh, Division of Car- ents thereby generated (Fig. 2). In the mid1980s, studies with use of extensive instrumendiovascular Diseases, Mayo Clinic, Rochester, MN 55905.
Emergency cardiac care was revolutionized with the introduction of closed chest cardiopulmonary resuscitation (CPR) in 1960.1 For the first time, a technique produced some forward blood flow in patients with cardiac arrest awaiting more definitive pharmacologic or electrical therapeutic intervention. Although closed chest CPR has been a clinically accepted procedure for 30 years, the mechanism of forward blood flow during external cardiac compression remains controversial. Two theories have been proposed:
Mayo Clin Proc 65:1432-1440, 1990
1432
Mayo Clin Proc, November 1990, Vol 65
ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE 1433
tation in dogs strongly supported the cardiac pump model. Other studies have suggested that both mechanisms may be operative in different clinical settings and with various CPR techniques. A clear understanding ofthe underlying mechanism or mechanisms in CPR is essential for the development of optimal CPR techniques. Unfortunately, studies in humans are difficult to perform. Almost all CPR research protocols reported to date have been done in patients in whom standard CPR protocols have failed, and the subjects were without adequate perfusion for extremely long periods before the experimental techniques were attempted. Transthoracic echocardiography has been reported in patients undergoing closed chest CPR,3,4 but the Thorax ability to perform adequate closed chest CPR and to obtain adequate echocardiographic images concomitantly might be questioned. Transesophageal echocardiography is now a practical method of studying the cardiac morphologic changes and flows in humans during external Fig. 1. Diagram of "cardiac pump theory," illustrating cardiac compressions.v? In transesophageal compression of the heart between sternum and paraspinal echocardiography, a side-viewing phased-array structures and ejection of blood into circulation during transducer is placed into the esophagus immedi- closed chest cardiac massage. (Heavy arrow = external compression.) Note that heart is compressed and ately adjacent and posterior to the heart. Excel- chest atrioventricular valve closes. Numbers represent theoretilent images of the myocardium can be obtained cal pressures in millimeters of mercury. (Modified from during external cardiac compressions. In addi- Weisfeldt ML, Chandra N: Physiology of cardiopulmonary tion to the two-dimensional images, Doppler and resuscitation. Annu Rev Med 32:435-442, 1981.) color flow imaging can be obtained for assessing to his local emergency room. His blood pressure flow. The following case reports demonstrate the was 80/50 mm Hg, and he was transferred to our use of transesophageal echoeardiography dur- medical center. On arrival, his vital signs showed ing CPR. This technique may help elucidate a blood pressure of 90/55 mm Hg, heart rate of the mechanism of blood flow during CPR in 106 beats/min, and respiratory rate of 36/min. humans. Thejugular venous pressure was increased at 20 em of water. The cardiac examination revealed a regular rhythm and normal first and second REPORT OF CASES heart sounds. A prominent third heart sound Case I.-A critically ill 72-year-old man was transferred to our institution because of near was noted. He had a grade 2 (on the basis of 1 to cardiovascular collapse, lactic acidosis, hyperka- 6) holosystolic murmur at the apex extending to lemia, and azotemia. One week before this the midclavicular line. Examination of the lungs admission, he had been admitted to his local disclosed bilateral adventitial sounds, consishospital with mild left-sided chest pain and tent with pulmonary edema. No peripheral dyspnea. After 5 days of treatment for conges- edema was present. The following laboratory findings were retive heart failure, he had been dismissed. Two days after dismissal, severe dyspnea and ported: serum potassium, 6.2 meq/liter; bicarleft-sided chest pain prompted the patient to go bonate, 12 meq/liter; creatinine, 3.6 mg/dl; and
1434 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
Thorax
0--+85
Fig. 2. Diagram of "thoracic pump theory," illustrating proposed events during closed chest cardiac massage. Increase in intrathoracic pressure caused by external compression is transmitted to all intrathoracic vessels. Pressure is transmitted unequally to arteries and veins as a result of venous valve located at thoracic outlet. The consequence is an arteriovenous pressure gradient and forward blood flow. During compression, left side of heart acts as a passive conduit for blood being ejected into systemic circulation. Mitral valve remains open during compression, and transmitral flow occurs. Numbers represent theoretical pressures in millimeters of mercury. (Modified from Weisfeldt ML, Chandra N: Physiology of cardiopulmonary resuscitation. Annu Rev Med 32:435-442, 1981.)
lactate, 15.6 ug/ml. The creatine kinase concentration was 124 U/liter (normal range, 23 to 99 U/liter) with 14% MB fraction. An electrocardiogram showed sinus rhythm at a rate of 88 beats/ min. A first-degree atrioventricular conduction delay and a left ventricular conduction defect with secondary ST- and T-wave abnormalities were noted. A chest roentgenogram disclosed bilateral pleural effusions with bilateral basilar interstitial infiltrates. Bedside echocardiography revealed the anterior wall to be thinned and akinetic with septal dyskinesia; mitral regurgitation was also noted. No shunts or pericardial
Mayo Clin Proc, November 1990, Vol 65
effusions were detected, and the papillary muscles appeared intact. Intravenous administration of fluids and pressor agents did not improve the hemodynamics of the patient. He was intubated, and mechanical ventilation was initiated. Insertion of an intra-aortic balloon pump improved his hemodynamics. The blood pressure increased to 115/60 mm Hg, the cardiac index increased to 2.2 liters/min per m", and the pulmonary wedge pressure declined to 20 mm Hg. Despite this initial improvement, on the second hospital day, sudden cardiovascular collapse and sinus rhythm developed, and CPR was initiated. Approximately 5 minutes after the resuscitative efforts were begun, transesophageal echocardiography was performed to search for potentially reversible mechanical complications of acute myocardial infarction. Chest compressions were withheld only momentarily during placement of the transesophageal probe and also after appropriate windows were obtained to record native cardiac motion. The transesophageal findings were consistent with severe myocardial dysfunction, the only functioning myocardium being the basal septum. The ventricular septum was intact, and no pericardial effusion was found. Transesophageal echocardiographic monitoring of chest compressions showed virtual obliteration of the right ventricular cavity, severe compression ofthe left ventricular cavity, closure of the mitral valve during compression, and subsequent opening of the mitral valve during the release phase (Fig. 3). Resuscitative attempts were discontinued after 30 minutes, and the patient died. Case 2.-A 70-year-old woman was brought to our institution by paramedics after she had a cardiac arrest. She had a history of hypertension, long-standing diabetes mellitus, a silent myocardial infarction in 1982, and severe left ventricular dysfunction; an echocardiogram 1 year earlier had shown an ejection fraction of 15 to 20%, apical dyskinesia, and inferior akinesia. On the day of admission, the patient had suffered sudden loss of consciousness. Bystanders at the scene immediately initiated CPR, and she was brought to our institution. Her initial
Mayo Clio Proc, November 1990, Vol 65
ECHOCARDlOGRAPHY DURING CLOSED CARDIAC MASSAGE 1435
Fig. 3 (case 1). Transesophageal echocardiographic images, showing sequential compression and release frames during closed chest cardiac massage. During compression, right ventricle (RY) is obliterated, left ventricle (L Y) is compressed, and mitral valve closes, an indication of cessation of transmitral flow. During release, mitral valve opens, and left ventricular filling occurs. ami = anterior mitral leaflet; LA = left atrium.
rhythm was fine ventricular fibrillation, which was converted to a stable rhythm by administration of epinephrine, sodium bicarbonate, lidocaine, and four countershocks. At this time, an electrocardiogram revealed an accelerated idioventricular rhythm, and a chest roentgenogram disclosed bilateral pulmonary infiltrates. Shortly after transfer to the coronary-care unit, the patient suffered another cardiac arrest with ventricular fibrillation that was resistant to repeated defibrillation. CPR was reinitiated. Transesophageal echocardiography was performed along with the resuscitative attempt to search for reversible causes of the collapse. Important findings, with chest compressions momentarily withheld, included no appreciable cardiac contraction and no evidence of tamponade, ventricular septal rupture, or papillary muscle rupture. Spontaneous echocardiographic contrast in the left ventricular outflow tract suggested a low-flow state. The chest compressions were also monitored during the resuscitation. The notable findings were compression ofthe right and left ventricular cavities and closure of the mitral and tricus-
pid valves during compression systole (Fig. 4). Color flow imaging revealed mild mitral regurgitation and severe tricuspid regurgitation during compressions. During the release phase, or compression diastole, the atrioventricular valves opened, and color flow imaging demonstrated rapid transvalvular flow (Fig. 5).
DISCUSSION CPR was first described in 1960 by Kouwenhoven and associates 1 as closed chest cardiac massage, an alternative to open cardiac massage. It was quickly disseminated as a technique to restore forward blood flow in victims of acute cardiac arrest, such as that occurring in drowning, electrocution, or the operating room. The heart was thought to be compressed, or "massaged," between the sternum and the spine, the result being ejection of blood from the left ventricle into the systemic circulation: The heart is limited anteriorly by the sternum and posteriorly by the vertebral bodies. Its lateral movement is restricted by the pericardium. Pressure on the sternum compresses the heart between it and the spine, forcing out blood. Relaxation of the pressure allows the heart to fill.'
1436 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
Mayo Cltn Proc, November 1990, Vol 65
Fig. 4 (case 2). Transesophageal echocardiographic images, showing sequential compression and release frames during closed chest cardiac massage. During compression, left ventricle (LV) is compressed, mitral valve closes (an indication of cessation oftransmitral flow), and aortic valve opens. With release, mitral valve opens, and left ventricular filling occurs. ami = anterior mitral leaflet; Ao = aorta; LA = left atrium; RV = right ventricle.
This forward stroke volume, which occurs tate. Conversely, victims with barrel chests, in during compression systole (compression phase), whom cardiac compression seemed unlikely, necessitates closure of the mitral valve and could be resuscitated. Furthermore, hemodyopening ofthe aortic valve. This mechanism has namic measurements obtained during CPR been termed the "cardiac pump modeL" In the revealed a substantial increase in right atrial cardiac pump model, the left ventricle then fills pressure during compression systole." Invesduring compression diastole (the release phase) tigators concluded that external cardiac comwith rapid transmitral flow (Fig. 1). Right heart pressions were inefficient because ofthese large hemodynamics were presumed to be analogous. venous pulse waves. Furthermore, palpable With use of this model, optimized CPR tech- femoral pulses during CPR only reflected presnique would include the maximal compression sure waves but did not predict forward blood rate while still allowing for adequate left ven- flow. Further evidence that challenged the cartricular filling-that is, stroke volume (thus diac pump model was provided by Criley and maximizing cardiac output, which is equiva- associates," who described "cough CPR" in 1976. lent to the compression rate times the stroke In their description of cough-induced cardiac volume). compression, asystolic patients in the catheSeveral subsequent clinical observations terization laboratory were kept conscious for 24 challenged this cardiac pump modeL Victims of to 39 seconds by repeated coughing. This techcardiac arrest with altered anatomy of the tho- nique led to the consideration of intrathoracic racic cage did not respond to CPR as predicted by pressure as the primary mechanism in the forthis modeL For example, victims with flail ward ejection of blood during CPR. This mechchests, in whom direct cardiac compression was anism has been termed the "thoracic pump easily performed, were often difficult to resusci- modeL"
Mayo Clin Proc, November 1990, Vol 65
ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE 1437
Fig. 5 (case 2). Transesophageal echocardiographic images with color flow mapping, showing sequential compression and release frames during closed chest cardiac massage. During compression, left ventricle (LV) is compressed, and mitral valve closes. Mitral regurgitation is demonstrated by mosaic color (double arrows) in left atrium (LA). A positive left ventricular-toleft atrial pressure gradient must exist for mitral regurgitation to occur. With release, mitral valve opens, and color flow imaging demonstrates left ventricular filling (brilliant blue area). RV = right ventricle.
In this thoracic pump model, external cardiac compression causes a general increase in intrathoracic pressure, which is transmitted to all cardiac chambers and vessels in the thorax (Fig. 2). This increased pressure generates an arteriovenous pressure gradient that results in forward flow. Critical to the production of forward flow is the development of the arteriovenous pressure gradients by the venous valves located at the thoracic outlet, which prevent transmission of increased thoracic pressure to the venous circulation. The presence of these valves has been substantiated echocardiographically, physiologically, and anatomically.v-" In this model, during compression systole, the left side ofthe heart acts as a passive conduit for thoracic blood being injected into the systemic circulation (that is, transmitral flow must occur during compression systole). The right side of the heart would presumably have no forward flow during compression systole, because of equal pressures in the right atrium, right ventricle,
and pulmonary artery. With use of this model, optimized CPR technique would include a fairly long duration of compression-that is, a duty cycle equal to or greater than 50%-to maintain the arteriovenous pressure gradient and forward flow. Direct evidence in support of the thoracic pump model was reported in 1980 by Rudikoff and associates," who described the hemodynamics of CPR in 15 dogs with cardiac arrest. They showed equal compression pressures in the left ventricle, aorta, right atrium, pulmonary artery, and esophagus, in conjunction with unequal transmission of pressure to the extrathoracic vessels. They postulated that the venous system had collapsed at the thoracic outlet. Subsequently, venous valves at the thoracic outlet were discovered. The arteriovenous pressure gradient generated across these valves resulted in forward blood flow. Furthermore, abdominal binding to prevent paradoxic diaphragmatic motion during compression led to an
1438 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
increase in aortic pressure and in carotid blood flow. In 1981, Niemann and colleagues'" performed cineangiography during CPR in 17 dogs with cardiac arrest and demonstrated forward flow without appreciable change in left ventricular dimensions. These findings signified- that the left ventricle was acting as a passive conduit and not a pump. These investigators also showed that simultaneous chest compressions and lung inflation increased blood flow in the left side of the heart and in the carotid artery. In 1981, Werner and co-workers! performed the first human studies with two-dimensional transthoracic echocardiographic visualization of the cardiac valves during CPR in five patients. They noted that the mitral valve remained open during the entire compression and release phase and that the aortic valve opened only during compression systole. Similarly, Rich and associates," who studied four patients during CPR, found no change in the left ventricular internal dimensions during chest compression and also aortic valve and mitral valve motion similar to that described by Werner and colleagues." These findings further supported the thoracic pump model. The thoracic pump model was first challenged by a group from Duke University. Maier and coworkers!' studied intact dogs with long-term instrumentation during CPR with various compression forces and rates. Increasing the compression force, measured by peak intrathoracic pressure, did not result in an increased stroke volume. With increasing compression rate, however, the stroke volume was constant, and cardiac output and coronary blood flow increased significantly. This finding was strong experimental evidence that direct cardiac compression was a major determinant of stroke volume during CPR. They emphasized the importance of high-impulse compressions, or high-impulse CPR, and hypothesized that the force transferred to the heart is a direct function of the initial momentum (defined as the compression mass times the velocity). They noted that previous investigators had used pneumatic compression devices, which result in low-momentum compressions. They also used high-fidelity mi-
Mayo Clin Proc, November 1990, Vol 65
cromanometer catheters to measure intrathoracic pressure during compressions; thus, they avoided possible pressure transients that may occur with saline-filled catheter manometers with imperfect frequency response characteristics. The intrathoracic pressures measured in this experimental model were lower than those previously reported and lower than the intracavitary pressures, evidence in support of the cardiac pump model. More recent studies from the Duke University group-s and others.P who used transthoracic and transesophageal echocardiography in dogs undergoing CPR, demonstrated mitral valve closure during compression in all but lowimpulse CPR. Subsequently, the Duke researchers demonstrated a pronounced influence of the rate of compression on the initial success of resuscitation after 30 minutes of CPR in dogs.P In these studies, the duty cycle was maintained at 50%, independent of the rate of compression. These data again supported the importance of the cardiac pump model of blood flow during CPR. Unfortunately, studies in humans are difficult to perform, and because of considerable differences in configuration of the chest wall, extrapolation of data from dogs to humans may not be justified. 15 The predominant mechanism of blood flow during CPR in humans remains controversial. The models are not mutually exclusive, and one may predominate over the other in any given resuscitation. Several factors may influence the mechanism of forward blood flow in any specific resuscitative effort, including technique of compression, thoracic anatomy, cardiac pathologic changes, underlying cardiac rhythm, and duration of resuscitation. A technique with a high rate and force may favor a direct compressive mechanismY Abnormalities of the thoracic cage, such as barrel chest, pectus excavatum, or flail chest, may play important roles. Underlying myocardial or atrioventricular valvular disease may impair passive ventricular filling during compression diastole or forward ventricular emptying during compression systole. These factors may favor an intrathoracic pressure mechanism.
Mayo Clin Proc, November 1990, Vol 65
ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE 1439
Sinus rhythm may improve mitral valve competence and thus favor direct compression." A lengthy resuscitative attempt may favor an intrathoracic pressure mechanism, perhaps because of changes in ventricular compliance during prolonged myocardial ischemia. Additionally, in some patients, direct right heart compression may predominate, whereas in other patients, left heart compressive mechanisms may predominate. In the two patients described in detail in our current report, transesophageal echocardiography demonstrated direct compression ofthe right and left ventricles during CPR compressions. Although some motion of the heart relative to the esophageal transducer may have occurred, such movement is likely to have been small and would not account for the -amount of compression seen. Additionally, the mitral valve was closed during compression systole. In the second patient,the presence of atrioventricular valvular regurgitation during compression suggested that a positive ventricular-to-atrial pressure gradient existed. These findings are consistent with a direct cardiac compressive mechanism of blood flow during CPR in these patients and are inconsistent with the thoracic pump mechanism. For the thoracic pump theory of CPR, the mitral valve must be open during compression. In both patients, the mitral valve generally remained closed throughout compression systole, occasionally opening toward the end of compression at a time when ventricular pressures were likely to be decreasing because of ventricular emptying. This mitral valve opening may have been due to intermittent increases in left atrial pressure from intrinsic atrial contraction. With release of the external compression, the mitral valve opened rapidly. During mid to late compression diastole, the mitral valve often closed, opening only intermittently presumably because of intrinsic atrial contraction. In the first patient, a less forceful compressive technique resulted in less direct cardiac compression; however, inadequate views of the mitral valve motion were obtained. Furthermore, transesophageal echocardiographic observations will be necessary to elucidate the predominant mechanism of forward
blood flow during CPR in humans in several clinical situations. These case reports demonstrate the utility of transesophageal echocardiography for studying CPR in humans.
CONCLUSION CPR has become a well-established technique to support circulation after cardiac arrest. Approximately 1,000 prehospital sudden deaths occur daily in the United States. An estimated 40% of patients with out-of-hospital cardiac arrests and substantiated ventricular fibrillation could be resuscitated; thus, approximately 100,000 lives could be saved annually. Although the primary goal during the past 3 decades has been to teach CPR techniques to as many people as possible, improved rates of resuscitation will also necessitate improved techniques or adjuncts (or both) to augment the low cardiac output, cerebral flow, and coronary blood flow generated by current CPR techniques. New techniques involve modifications in rate of compression, impulse of compression, duty cycle length, and, recently, simultaneous lung inflations and interposed abdominal compressions. ll ,14,16-18 Potential adjuncts to CPR include abdominal binding, military antishock trouser garments, and inflatable vests. 19,20 Each of these modifications was devised to improve the hemodynamics in either the cardiac pump model or the thoracic pump model of CPR. For further rational development of these new CPR techniques, or perhaps some as yet undevised techniques, an understanding of the predominant mechanism of blood flow during CPR in humans is necessary. With acquisition of such knowledge, the anticipated outcome will be evolution of new CPR techniques that will improve cardiac output and oxygen delivery to the brain and myocardium and thereby increase the number of patients successfully resuscitated. ACKNOWLEDGMENT We greatly appreciate the expert secretarial assistance of Kathleen A. Budensiek, who prepared the submitted manuscript.
1440 ECHOCARDIOGRAPHY DURING CLOSED CARDIAC MASSAGE
REFERENCES 1.
2. 3.
4.
5.
6.
7.
8.
9.
10.
11.
Kouwenhoven WE, Jude JR, Knickerbocker GG: Closed-chest cardiac massage. JAMA 173:10641067, 1960 MacKenzie GJ, Taylor SH, McDonald AH, Donald KW: Haemodynamic effects of external cardiac compression. Lancet 1:1342-1345, 1964 Rich S, Wix HL, Shapiro EP: Clinical assessment of heart chamber size and valve motion during cardiopulmonary resuscitation by two-dimensional echocardiography. Am Heart J 102:368-373, 1981 WernerJA, Greene HL, Janko CL, Cobb LA: Visualization of cardiac valve motion in man during external chest compression usingtwo-dimensional echocardiography: implications regarding the mechanism of blood flow. Circulation 63:1417-1421,1981 SewardJB,KhandheriaBK, OhJK,AbeIMD, Hughes RW Jr, Edwards WD, Nichols BA, Freeman WK, Tajik AJ: Transesophageal echocardiography: technique, anatomic correlations, implementation, and clinical applications. Mayo Clin Proc 63:649-680, 1988 Uenishi M, Sugimoto H, Sawada Y, Terai C, Yoshioka T, Sugimoto T: Transesophageal echocardiography during external chest compression in humans (letter to the editor). Anesthesiology 60:618, 1984 Clements FM, de Bruijn NP, Kisslo JA: Transesophageal echocardiographic observations in a patient undergoing closed-chest massage. Anesthesiology 64:826-828, 1986 Criley JM, Blaufuss AH, Kissel GL: Cough-induced cardiac compression: self-administeredform ofcardiopulmonary resuscitation. JAMA 236:1246-1250, 1976 Rudikoff MT, Maughan WL, Effron M, Freund P, Weisfeldt ML: Mechanisms of blood flow during cardiopulmonary resuscitation. Circulation 61:345352, 1980 Niemann JT, Rosborough JP, Hausknecht M, Garner D, Criley JM: Pressure-synchronized cineangiography during experimental cardiopulmonary resuscitation. Circulation 64:985-991, 1981 Maier GW, Tyson GS Jr, Olsen CO, Kernstein KH, Davis JW, Conn EH, Sabiston DC Jr, Rankin JS: The
12.
13.
14.
15. 16.
17.
18. 19.
20.
Mayo Clin Proc, November 1990, Vol 65
physiology of external cardiac massage: high-impulse cardiopulmonary resuscitation. Circulation 70:86-101, 1984 Feneley MP, Maier GW, Gaynor JW, Gall SA, Kisslo JA, Davis JW, Rankin JS: Sequence of mitral valve motion and transmitral blood flow during manual cardiopulmonary resuscitation in dogs. Circulation 76:363-375, 1987 Deshmukh HG, Weil MH, Rackow EC, Trevino R, Bisera J: Echocardiographic observations during cardiopulmonary resuscitation: a preliminary report. Crit Care Med 13:904-906, 1985 Feneley MP, Maier GW, Kern KB, Gaynor JW, Gall SA Jr, Sanders AB, Raessler K, Muhlbaier LH, Rankin JS, Ewy GA: Influence of compression rate on initial success of resuscitation and 24 hour survival after prolonged manual cardiopulmonary resuscitation in dogs. Circulation 77:240-250, 1988 Barsan WG, Levy RC: Experimental design for study of cardiopulmonary resuscitation in dogs. Ann Emerg Med 10:135-137,1981 Taylor GJ, Tucker WM, Greene HL, Rudikoff MT, Weisfeldt ML: Importance of prolonged compression during cardiopulmonary resuscitation in man. N Engl J Med 296: 1515-1517, 1977 Chandra N, RudikoffM, Weisfeldt ML: Simultaneous chest compression and ventilation at high airway pressure during cardiopulmonary resuscitation. Lancet 1:175-178, 1980 BabbsCF, TackerWAJr: Cardiopulmonary resuscitation with interposed abdominal compression. Circulation 74 (SuppI4):IV37-IV41, 1986 NiemannJT, RosboroughJP, CrileyJM: Continuous external counterpressure during closed-chest resuscitation: a critical appraisal of the military antishock trouser garment and abdominal binder. Circulation 74 (SuppI4):IV102-IV107, 1986 Halperin HR, Guerci AD, Chandra N, Herskowitz A, Tsitlik JE, Niskanen RA, Wurmb E, Weisfeldt ML: Vest inflation without simultaneous ventilation during cardiac arrest in dogs: improved survival from prolonged cardiopulmonary resuscitation. Circulation 74:1407-1415,1986
