Ventricular Fibrillation Its Effect on Myocardial Flow, Distribution, and Perfomance Gerald D. Buckberg, M.D., a n d Christof E. Hottenrott, M.D. ABSTRACT Subendocardial ischemia develops in hearts that are fibrillated during cardiopulmonary bypass when: (1) the normal ventricle is fibrillated with a sustained electrical stimulus, (2) the hypertrophied ventricle is allowed to fibrillate spontaneously, (3)the fibrillating heart becomes distended, or (4) the perfusion pressure is reduced to approximately 50 mm Hg. Myocardial hypothermia reduces cardiac oxygen requirements during fibrillation but does not prevent ischemia when perfusion pressure falls to levels frequently attained during clinical open-heartoperations. The ischemia occurs because flow cannot rise sufficiently to meet the metabolic demands of ventricular fibrillation. The forces interacting to impede adequate flow to the subendocardium during ventricular fibrillation are: (1)the compressive forces exerted on subendocardial muscle by the strength of fibrillation, (2) the compressive forces resulting from raised intracavitary pressure due to occlusion or malfunction of the ventricular vent, and (3) the evolution of myocardial edema as ischemia is prolonged. We have abandoned the use of ventricular fibrillation in clinical open-heart operations and now allow the heart to beat continually with adequate perfusion pressure. We have not needed to use inotropic drugs postoperatively after aortic or mitral valve replacement since adopting this technique.
L
eft ventricular subendocardial necrosis, a major cause of fatal postoperative myocardial failure, results from ischemic injury to the heart during cardiopulmonary bypass. Ventricular fibrillation is used by many surgeons in an effort to prevent such damage during extracorporeal circulation. This form of myocardial preservation has been assumed safe because the heart remains continually perfused with oxygenated blood, air embolism is avoided, and a quiet operative field is achieved. Reis and associates [lo], however, have shown that continuous application of an AC fibrillating stimulus is deleterious to the metabolism and function of a normal heart. They could not determine the site and mechanism of fibrillation-induced ischemia because methodological limitations precluded regional flow m'easurements. Our studies, using the radioactive microsphere method to measure regional flow, were designed to deal with the following problems: (1) differences between electrical and spontaneous fibrillation in normal hearts; (2) effects of ventricular fibrillation in hypertrophied hearts; (3) effects of ventricular distention during fibrillation; (4)effects of changing perfusion pressure during fibrillation; (5)
From the Division of Thoracic Surgery, UCLA School of Medicine, Los Angeles, Calif. Supported by grants-in-aid from the Los Angeles County Heart Association,the Gilmore Foundation, the Blalock Foundation, and the U.S. Public Health Service. Address reprint requests to Dr. Buckberg, Division of Thoracic Surgery, UCLA School of Medicine, Los Angeles, Calif. 90024.
76
T H E ANNALS OF THORACIC SURGERY
Ventricular Fibrillation effects of myocardial hypothermia and duration of fibrillation; and (6) mechanisms of ischemia during fibrillation [3-73.
Technique Although regional flow measurements d o not provide evidence that sufficient oxygen is being delivered to meet metabolic demands, the observed flows (mV100 gm tissuehnin) and distribution (endocardiaVepicardia1 flow ratio) can be assessed in relation to simultaneous measurements of metabolism, histochemistry, and ventricular function in order to determine if flow is adequate. We believe barometers of adequate perfusion must be sufficiently sensitive to ensure that there is no metabolic, histochemical, or functional impairment before a method of myocardial preservation can be declared safe, meaning it causes no myocardial injury. Cell death, demonstrated by release of enzymes or structural changes on routine staining, is a late sign of irreversible ischemic damage. In our studies we measured coronary flow distribution with 8- to 10-p radioactive microspheres and determined biochemical impairment by assessing aortocoronary sinus differences in pH; lactate, to detect anaerobic metabolism and abnormal glycolysis; and potassium, to determine cell membrane injury. We looked for histochemical ischemia with the basic fuchsin stain, which shows ischemia when no structural change can be seen on normal microscopy, and determined ventricular performance by inscribing function curves. Studies were done in normothermic hearts (37°C) without hemodilution (which lowers blood oxygen content so that a higher coronary flow is necessary to maintain oxygen delivery) and at perfusion pressures of 100 mm Hg. ELECTRICAL VERSUS SPONTANEOUS FIBRILLATION
Cardiopulmonary bypass relieves the beating heart of its external work and reduces oxygen requirements by approximately 60% when the heart is perfused adequately in the beating, empty state [5]. Oxygen requirements are assessed by measuring left ventricular oxygen consumption ; oxygen delivery is considered adequate if no signs of metabolic, histochemical, or functional damage are present. When the left ventricle is fibrillated with an electrical stimulus and the stimulus is removed (spontaneous fibrillation), the ventricle consumes approximately 80%as much oxygen as it does in the working state (Fig. 1).Coronary blood flow increases to meet these requirements, and no important changes in biochemistry, histochemistry, or function occur. Our findings are comparable to those reported by Reis, Cohn, and Morrow [ 101. The oxygen requirements of the spontaneously fibrillating heart are perhaps greatest in the subendocardial muscle, because this region receives the highest proportion of left ventricular flow the endocardiaVepicardia1 flow ratio increases from 1.0 to 1.4 [5]. T h e elevated oxygen requirements during fibrillation are likely related to the frequent irregular myocardial contractions, since rate and extent of tension development are primary determinants of myocardial oxygen demands. With fibrillation electrically maintained by continuous application of a 1.5-to 7-volt stimulus, its vigor increases and oxygen requirements presumably rise. Our VOL. 20, NO. 1 , JULY, 1975
77
BUCKBERG AND HOTTENROTT
c
E
8-
\ 0
o 0 \ 8
64-
20-
FIG. 1. Oxygen consumption of the total heart and lejit ventricle (LV) during beating working, beating empty, spontaneously fibrillating, and electricallyfibrillating states. Note that lejit ventricular oxygen uptake during electricalfibrillation is appreciably lower in comparison with the values f o r spontaneous fibrillation. (From Hottenrott, Maloney, and Buckberg [5].)
studies show that electrically sustained fibrillating muscle prevents flow from increasing sufficiently to meet metabolic demands and causes ischemia. This flow impediment occurs with both AC and DC fibrillating stimuli [6]. Despite the increased vigor of fibrillation, left ventricular oxygen consumption is less than in spontaneous fibrillation. In contrast to the findings with spontaneous fibrillation, evidence of inadequate oxygen delivery during electrical fibrillation is provided by signs of anaerobic metabolism (decreased coronary sinus pH, myocardial lactate production), loss of cell membrane integrity (potassium efflux into coronary sinus blood; Fig. 2), and significant depression of myocardial performance after defibrillation (Fig. 3). Our findings are compatible with those of Reis and colleagues [ 101and occurred despite an increase in subendocardial flow -above
-
-+
FIG. 2. Biochemical evidence of left ventricular myocardial ischemia duringfibrillation. Note that there is no great difference in coronary sinus -aortic (CS-A) hydrogen ion, lactate, and potassium in spontaneouslyfibrillating hearts. With electricalfibrillation, hydrogen ion and lactate are produced and potassium is lost from the myocardium; an even greater rise occurred during the reactive hyperemia phase following deJfibrillation. (From Hottenrott, Maloney, and Buckberg [5].)
T
BEATING
WORKING
78
BEATING EMPTY
SPONTANEOUS ELECTRIC
FIBRILLATION FIBRILLATION
T H E ANNALS OF THORACIC SURGERY
Ventricular Fibrillation
45
1
1
?
40-1
I 0 II
Y- 30
u
P W Y
Li 3
I I
25-
0 I
I
I
0 I
20-
i I
0
$51
LEFT ATRIAL PRESSURE, mmHg
FIG. 3. Representative l&t ventricular (LV)function curves following 60 minutes of spontaneous fibrillation (left) and electricalfibrillation (right). Note the mild depression of ventricular function following spontaneous fibrillation and the marked depressionfollowing electricalfibrillation. There was no change in left ventricular function cumes following 60 minutes in the beating, nonworking state. (From Hottenrott, Maloney, and Buckberg [5].)
that of the beating, empty heart - and homogeneous flow distribution (endocardiaVepicardia1flow ratio = 1.0). These findings show that “normal” flow distribution (endo/epi ratio = 1.0) and increased subendocardial flow d o not necessarily signify adequate perfusion and that electrical fibrillation should be avoided because it causes ischemia. Another sign of ischemia caused by electrical fibrillation is the occurrence of reactive hyperemia after defibrillation. T h e shape of the postoperative function curve correlated closely with the time after defibrillation that it was inscribed. Function was depressed markedly if bypass was discontinued immediately (within 10 minutes) while the heart was still undergoing reactive hyperemia. Less myocardial depression was seen when the function curve was inscribed 30 minutes later and coronary flow had returned to control levels. T h e clinical correlate is that prolongation of bypass in hearts injured by hypoxia (aortic cross-clamping or ventricular fibrillation) may be beneficial to subsequent myocardial performance. EFFECTS OF LEFT VENTRICULAR HYPERTROPHY
Most studies of the effects of ventricular fibrillation have been performed in normal ventricles, although subendocardial necrosis occurs most commonly in hypertrophied hearts. It is well recognized that the hypertrophied heart is more vulnerable to ischemic injury than is the nonhypertrophied heart [9]. In contrast to our findings on the effects of spontaneous fibrillation on normal left ventricles, we observed that the hypertrophied heart becomes ischemic when it fibrillates [71. Subendocardial flow fails to rise sufficiently in spontaneously fibrillating hypertrophied hearts to provide adequate oxygen delivery (Fig. 4), and dogs with left ventricular hypertrophy develop the same metabolic and functional impairment as seen after electrical fibrillation in normal hearts [ 5 ] . At postmortem examina-
VOL. 20, NO. 1, JULY, 1975
79
BUCKBERG AND HOTTENROTT
160 -
T
Bwtinq empty
soot
140 -
120100 E
0" \
80-
0
60-
40 -
200
Normal L V
LVH
FIG. 4. BloodJlow through l&t ventricular ( L V )subendocardial muscle in the normal and hypertrophied heart ( L V H )when they are beating and nonworking and after 60 minutes of spontaneous ventricular fibrillation. Subendocardial p o w increases approximately 300 % during fibrillation in normal hearts. Conversely, coronaryflow to subendocardial muscle does not increase during fibrillation in hypertrophied hearts. (From Hottenrott, Towers, Kurkji, Maloney, and Buckberg [7].)
tion the hearts showed hemorrhagic necrosis of the left ventricle (Fig. 5). Recent studies by Isom and associates [8] provide evidence that in patients with hypertrophied heart, myocardial oxygen consumption fails to increase over that of the beating, empty heart during ventricular fibrillation; however, myocardial enzyme levels in coronary sinus blood in fibrillating, hypertrophied hearts increased significantly beyond the values observed when the heart was beating without doing external work. In the beating heart the subendocardium is most vulnerable to ischemic damage because it receives its flow in diastole; systolic compressive forces are greatest in the subendocardium and impede flow. Several factors may contribute to the inability of subendocardial flow to increase adequately in hypertrophied, fibrillating hearts. T h e fibrillating heart attains a chamber size comparable to that FIG. 5. Transuerse section of heart from a dog with chronic aorticstenosis and left ventricular hypertrophy who underwent 60 minutes of spontaneous ventricular fibrillation. Focal hemorrhages and venous congestion are present transmurally but are most prominent in the inner hayof the wall of the hypertrophied left ventricle. (From Hottenrott, Towers, Kurkji, Maloney and Buckberg [7].)
80
THE ANNALS OF THORACIC SURGERY
Ventricular Fibrillation FIG. 6. Distribuhoti of It$ veistriculnr ji’oi(i iii the f i v e experimetrtnl corrditioiis. Note tiint the proportioii of total l$t veiitriculnr j l o w rlelivQred to the subeirdocnr(liuiii zv highrst with spoiitniieousfibrillntioii mid redistributed nionQfroiti the subeiirlocnrdiuiir with disteiitioti. Electrically fibrilkrted Amrts do iiot receive proportionately nugiiieiited coroiicriy flow without disteiitioir, n i d the greatest redistributioii of flow nzcinpfroiir the subeiidocnrdiuni occurs wheii they are disteirderl. ( E N D O / E P I = eirdocnrdinlle~cnrdinl~. (From Hottetirott ntrd Buckberg [ 4 ] . )
2 00-
T
160-
a
EZj Nandsfended I Dislended
120-
0
z 8040-
0-
BEATING EMPTY
SPONTANEOUS
ELECTRICAL
F l BR I LL AT ION
at end-systole, so that subendocardial vessels may be compressed by fibrillating muscle fibers. Theoretical analysis of intramyocardial stress gradients suggests that these compressive forces are highest in the subendocardial region, especially in hypertrophied hearts [ 11. Studies of intramyocardial pressure by Baird and his co-workers [2] confirm this suggestion, as do our experimental and clinical studies of ventricular fibrillation in hypertrophied hearts. We conclude that the normal heart provides an inadequate model for studying clinical entities and that any form of normothermic fibrillation should be avoided in hypertrophied hearts. EFFECTS OF VENTRICULAR DISTENTION
T h e importance of adequate decompression of the left ventricle in hearts fibrillated during bypass is well recognized by all cardiac surgeons. Malfunction of the ventricular vent is associated with impaired postoperative myocardial performance, which has been attributed to overstretching of the myocardium. Our studies show that ventricular distention during fibrillation raises intracavitary pressure, reduces left ventricular driving pressure (difference between aortic and left ventricular pressure), and impedes subendocardial flow (Fig. 6). T h e flow impediment occurs in both spontaneously beating and electrically fibrillating normal hearts and results in ischemia and depressed myocardial performance [4]. T h e supply/demand discrepancy (ischemia) caused by fibrillation is accentuated by distention since oxygen requirements are raised simultaneously by increased wall tension. We believe that the ischemia caused by distention of fibrillating hearts may impair postoperative myocardial performance more than overstretching of muscle fibers. EFFECTS OF PERFUSION PRESSURE DURING FIBRILLATION
Our studies, which were performed at a perfusion pressure of 100 mm Hg, suggest that fibrillating muscle prevents flow from increasing sufficiently to meet metabolic demands. T h e obvious counterforce to the deleterious effects of ventricular fibrillation is the coronary driving pressure, which can be varied by changing aortic perfusion pressure during bypass. We found that raising perfusion pressure during the early phases of electrical fibrillation, when myocardial oxygen consumption is lowered because oxygen delivery is inadequate, increases
VOL. 20, NO. 1 , JULY, 1975
81
BUCKBERG AND HOTTENROTT both subendocardial flow and left ventricular oxygen consumption [6] and reverses anaerobic glycolysis. In other experiments [3], we found that reducing perfusion pressure from 100 to 50 mm Hg in spontaneously fibrillating normothermic hearts results in a significant fall in subendocardial oxygen delivery, redistribution of flow away from the subendocardium, abnormal glycolysis, and histochemical signs of ischemia. Baird and associates [2] reported similar changes in subendocardial flow and distribution when perfusion pressure is reduced in normal and hypertrophied hearts. Since patients with coronary artery disease frequently have a fairly large pressure drop across the atherosclerotic obstruction, it is likely that myocardial perfusion pressure will be substantially reduced below that in the aortic root during clinical bypass. All these studies suggest that despite the absence of ventricular hypertrophy, these hearts may be inadequately perfused if they are allowed to fibrillate spontaneously during coronary revascularization. EFFECTS OF HYPOTHERMIA AND DURATION OF FIBRILLATION
T h e left ventricle fibrillates more slowly and with less vigor when myocardial temperature is reduced. It is therefore understandable that the oxygen uptake of the adequately perfused, spontaneously fibrillating heart is reduced by approximately 40% when myocardial temperature is lowered from 37" to 28°C [3]. T h e hypothermic normal left ventricle, however, is not protected against subendocardial ischemia if perfusion pressures are reduced to 50 mm Hg; diminished subendocardial oxygen delivery, abnormal glycolysis, and histochemical damage result when perfusion pressure falls to levels frequently used during clinical open-heart operations. Our studies indicate that the extent of histochemical and biochemical ischemic damage is accentuated when the duration of electrical fibrillation and inadequate perfusion pressure is prolonged [3]. MECHANISMS OF ISCHEMIA
Our studies suggest at least three possible causes of ischemic subendocardial damage during ventricular fibrillation. First, it seems that the fibrillating muscle itself prevents flow from increasing sufficiently to meet metabolic demands. We found that increasing the strength of the fibrillating stimulus with either alternating or direct current caused a progressive fall in the peak achievable flow following maximum coronary vasodilation [6]. Second, distention of the fibrillating heart adds a counterforce that opposes flow by reducing coronary driving pressure and thus impairs subendocardial perfusion while increasing oxygen requirements due to raised wall tension. Third, in hypertrophied hearts fibrillating spontaneously and normal hearts fibrillating electrically, we observed that subendocardial flow fell progressively as fibrillation was continued for one hour. The reduced flow reflected a progressive rise in vascular resistance that occurred despite perfusion pressure, adequacy of venting, and fibrillating stimulus remaining constant. This observation led us to conclude that the myocardium may have become edematous because ischemic damage was prolonged during fibrillation. Our conclusion was supported by histological studies that showed increased
82
T H E ANNALS OF THORACIC SURGERY
Ventricular Fibrillation 350-
FIG. 7. Effects of mannitol on spontaneously fibrillating normal (LV) and hypertrophied ( L V H ) hearts. Left ventricular flow is augmented in the normal ventricle and depressed in the hypertrophied ventricle after 60 minutes of spontaneousfibrillation. Flow to the normal ventricle is reduced and that to hypertrophied ventricle increased immerliately following defibrillation. There is marked augmentation i n f l o w to hypertrophied heart and a smalljlow increment to normal heart following administration of mannitol after defibrillation. (From Hotten-
300
0
NORMAL L V
mLVH
250
kJ',;Cl gBdg,mIn
200-
150-
100-
50-
'-
BEATING
60 min SPONT
PREMANNlTOL POSTMANNlTOL
distance between myocardial fibers. We tested this hypothesis by infusing mannitol into the aortic root following defibrillation at a time when the coronary vessels were maximally dilated (no reactive hyperemia); a striking increase in coronary blood flow (Fig. 7) and improved ventricular performance resulted [6].
Clinical Correlation As a result of these experimental studies, we have completely abandoned the use of ventricular fibrillation during clinical open-heart operations. Since 1972 we have attempted, in all patients undergoing valve replacement, to allow the heart to beat continuously with an adequate coronary perfusion pressure. Because our studies strongly suggest that depressed myocardial performance after cardiopulmonary bypass is a sign of ischemic injury during extracorporeal circulation, we have compared our clinical results before and after 1972. The results are shown in the Table. Inotropic drugs have not been needed in any of 47 patients undergoing aortic valve replacement and in only 1 of 22 patients undergoing mitral valve replacement. This patient had prolonged intermittent ischemic arrest due to technical problems and since 1972 is the only person to have died of postoperative low-output syndrome following isolated valve replacement at our institution. NEED FOR POSTOPERATIVE INOTROPIC SUPPORT
Inotropic Dtugs Used MVR AVR
Method of Myocardial Preservation Ischemic arrest
47% (9/19) 54% (15/28) 5% (1/22)
Ventricular fibrillation Beating, empty MVR
=
mitral valve replacement; AVR
=
33 % (4/12) 41% (12/29) 0% (0/47)
aortic valve replacement.
VOL. 20, NO. 1, JULY, 1975
83
BUCKBERG AND HOTTENROTT
Dkcussion (Drs. Guyton, Buckberg, Michaelis, Brody, and Gay) We try to maintain a perfusion pressure of 100 mm Hg at the tip of each perfusion cannula. Of course the pressure drop along the tubing relative to flow rate must be taken into account as well as differences between the height of the pump and the height of the operating table. Flows average between 150 and 200 mumin in the left perfusion cannula and, depending on the size and distribution of the right coronary artery, anywhere between 75 and 125 mumin in the right. Perfusion is conducted at 32°C.If the heart fibrillates, we make sure the perfusion cannula has not been placed too far into the left coronary ostia. We try to defibrillate immediately, and if this is not possible, we suspect that flow is going preferentially into the circumflex system. We then reduce the perfusion temperature to 28"C, place ice on the surface of the heart, continue perfusing, and work more quickly. We use a balloon cannula except when there is a short left main coronary artery, in which case we use a basket cannula. We start perfusing immediatelyafter opening the aorta. The valve is removed while the coronary arteries are perfused so that calcific emboli are prevented and myocardial protection is achieved immediately. Since maintaining a beating, empty heart with adequate perfusion pressure, we have not had a single patient require postoperative inotropic support after aortic valve replacement.
References 1 . Archie, J. P., Jr. Determinants of regional intramyocardial pressure. J Surg Res 14:338. 1973. 2. Baird, R. J., Okumori, M., Dutka, F., d e la Rocha, A., and Goldbach, M. Surgical aspects of regional myocardial blood flow and myocardial pressure.J Thorac Cardiovasc Surg (In press.) 3. Brazier, J., McConnell, D., Cooper, N., DeLand, E., and Buckberg, G. T h e effects of temperature, time, and perfusion pressure on the adequacy and distribution of coronary flow during ventricular fibrillation. Surg Forum 25: 172, 1974. 4. Hottenrott, C., and Buckberg, G. Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow: 11. Effects of ventricular distention.] Thorac Cardiovasc Surg 68:626, 1974. 5. Hottenrott, C., Maloney, J. V., Jr., and Buckberg, G. Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow: I. Electrical vs. spontaneous fibrillation. J Thorac Cardiovasc Surg 68:6 15, 1974. 6. Hottenrott, C., Maloney, J. V., Jr., and Buckberg, G. Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow: 111. Mechanisms of ischemia. J Thorhc Cardiovasc Surg 68:634, 1974. 7. Hottenrott, C. E., Towers, B., Kurkji, H. J., Maloney, J. V., and Buckberg, G. T h e hazard of ventricular fibrillation in hypertrophied ventricles during cardiopulmonary bypass. J Thorac Cardiovasc Surg 66:742, 1973. 8. Isom, 0. W., Kutin, N. D., Falk, E. A., and Spencer, F. C. Patterns of myocardial metabolism during cardiopulmonary bypass and coronary perfusion. ] Thorac Cardiovasc Surg 66:705, 1973. 9. Iyengar, S. R. K., Ramchand, S., Charrette, E. J. P., Iyengar, C. K. S., and Lynn, R. B. Anoxic cardiac arrest: An experimental and clinical study of its effects.] Thorac Cardiovasc Surg 66:722, 1973.
84
THE ANNALS OF THORACIC SURGERY
Ventricular Fibrillation 1 0 . Keis, K.L., Cohn, L. H., and Morrow, A. G. Effects ofinduced ventricular fibrillation on ventricular perforiiiance and cardiac Inetabolism. Circulation 36:234, 1967.
Editors’ Note: It should be mentioned that many experienced surgeons have obtained excellent results in large numbers of patients by utilizing hypothermic anoxic arrest without coronmy perfusion. Subendocardial necrosis may occur in a number o f settings, not all o f which iiivolve ventricularfibrillation or cardiopulmonary bypass. Even duringperfirsion, its occurrence h a been shown recently to depend not only uponfibrillation or the lack there$, but also upon the adequacy of perfusion pressure.
VOL. 20, NO. 1, JULY, 1975
85
L
eft ventricular subendocardial necrosis, a major cause of fatal postoperative myocardial failure, results from ischemic injury to the heart during cardiopulmonary bypass. Ventricular fibrillation is used by many surgeons in an effort to prevent such damage during extracorporeal circulation. This form of myocardial preservation has been assumed safe because the heart remains continually perfused with oxygenated blood, air embolism is avoided, and a quiet operative field is achieved. Reis and associates [lo], however, have shown that continuous application of an AC fibrillating stimulus is deleterious to the metabolism and function of a normal heart. They could not determine the site and mechanism of fibrillation-induced ischemia because methodological limitations precluded regional flow m'easurements. Our studies, using the radioactive microsphere method to measure regional flow, were designed to deal with the following problems: (1) differences between electrical and spontaneous fibrillation in normal hearts; (2) effects of ventricular fibrillation in hypertrophied hearts; (3) effects of ventricular distention during fibrillation; (4)effects of changing perfusion pressure during fibrillation; (5)
From the Division of Thoracic Surgery, UCLA School of Medicine, Los Angeles, Calif. Supported by grants-in-aid from the Los Angeles County Heart Association,the Gilmore Foundation, the Blalock Foundation, and the U.S. Public Health Service. Address reprint requests to Dr. Buckberg, Division of Thoracic Surgery, UCLA School of Medicine, Los Angeles, Calif. 90024.
76
T H E ANNALS OF THORACIC SURGERY
Ventricular Fibrillation effects of myocardial hypothermia and duration of fibrillation; and (6) mechanisms of ischemia during fibrillation [3-73.
Technique Although regional flow measurements d o not provide evidence that sufficient oxygen is being delivered to meet metabolic demands, the observed flows (mV100 gm tissuehnin) and distribution (endocardiaVepicardia1 flow ratio) can be assessed in relation to simultaneous measurements of metabolism, histochemistry, and ventricular function in order to determine if flow is adequate. We believe barometers of adequate perfusion must be sufficiently sensitive to ensure that there is no metabolic, histochemical, or functional impairment before a method of myocardial preservation can be declared safe, meaning it causes no myocardial injury. Cell death, demonstrated by release of enzymes or structural changes on routine staining, is a late sign of irreversible ischemic damage. In our studies we measured coronary flow distribution with 8- to 10-p radioactive microspheres and determined biochemical impairment by assessing aortocoronary sinus differences in pH; lactate, to detect anaerobic metabolism and abnormal glycolysis; and potassium, to determine cell membrane injury. We looked for histochemical ischemia with the basic fuchsin stain, which shows ischemia when no structural change can be seen on normal microscopy, and determined ventricular performance by inscribing function curves. Studies were done in normothermic hearts (37°C) without hemodilution (which lowers blood oxygen content so that a higher coronary flow is necessary to maintain oxygen delivery) and at perfusion pressures of 100 mm Hg. ELECTRICAL VERSUS SPONTANEOUS FIBRILLATION
Cardiopulmonary bypass relieves the beating heart of its external work and reduces oxygen requirements by approximately 60% when the heart is perfused adequately in the beating, empty state [5]. Oxygen requirements are assessed by measuring left ventricular oxygen consumption ; oxygen delivery is considered adequate if no signs of metabolic, histochemical, or functional damage are present. When the left ventricle is fibrillated with an electrical stimulus and the stimulus is removed (spontaneous fibrillation), the ventricle consumes approximately 80%as much oxygen as it does in the working state (Fig. 1).Coronary blood flow increases to meet these requirements, and no important changes in biochemistry, histochemistry, or function occur. Our findings are comparable to those reported by Reis, Cohn, and Morrow [ 101. The oxygen requirements of the spontaneously fibrillating heart are perhaps greatest in the subendocardial muscle, because this region receives the highest proportion of left ventricular flow the endocardiaVepicardia1 flow ratio increases from 1.0 to 1.4 [5]. T h e elevated oxygen requirements during fibrillation are likely related to the frequent irregular myocardial contractions, since rate and extent of tension development are primary determinants of myocardial oxygen demands. With fibrillation electrically maintained by continuous application of a 1.5-to 7-volt stimulus, its vigor increases and oxygen requirements presumably rise. Our VOL. 20, NO. 1 , JULY, 1975
77
BUCKBERG AND HOTTENROTT
c
E
8-
\ 0
o 0 \ 8
64-
20-
FIG. 1. Oxygen consumption of the total heart and lejit ventricle (LV) during beating working, beating empty, spontaneously fibrillating, and electricallyfibrillating states. Note that lejit ventricular oxygen uptake during electricalfibrillation is appreciably lower in comparison with the values f o r spontaneous fibrillation. (From Hottenrott, Maloney, and Buckberg [5].)
studies show that electrically sustained fibrillating muscle prevents flow from increasing sufficiently to meet metabolic demands and causes ischemia. This flow impediment occurs with both AC and DC fibrillating stimuli [6]. Despite the increased vigor of fibrillation, left ventricular oxygen consumption is less than in spontaneous fibrillation. In contrast to the findings with spontaneous fibrillation, evidence of inadequate oxygen delivery during electrical fibrillation is provided by signs of anaerobic metabolism (decreased coronary sinus pH, myocardial lactate production), loss of cell membrane integrity (potassium efflux into coronary sinus blood; Fig. 2), and significant depression of myocardial performance after defibrillation (Fig. 3). Our findings are compatible with those of Reis and colleagues [ 101and occurred despite an increase in subendocardial flow -above
-
-+
FIG. 2. Biochemical evidence of left ventricular myocardial ischemia duringfibrillation. Note that there is no great difference in coronary sinus -aortic (CS-A) hydrogen ion, lactate, and potassium in spontaneouslyfibrillating hearts. With electricalfibrillation, hydrogen ion and lactate are produced and potassium is lost from the myocardium; an even greater rise occurred during the reactive hyperemia phase following deJfibrillation. (From Hottenrott, Maloney, and Buckberg [5].)
T
BEATING
WORKING
78
BEATING EMPTY
SPONTANEOUS ELECTRIC
FIBRILLATION FIBRILLATION
T H E ANNALS OF THORACIC SURGERY
Ventricular Fibrillation
45
1
1
?
40-1
I 0 II
Y- 30
u
P W Y
Li 3
I I
25-
0 I
I
I
0 I
20-
i I
0
$51
LEFT ATRIAL PRESSURE, mmHg
FIG. 3. Representative l&t ventricular (LV)function curves following 60 minutes of spontaneous fibrillation (left) and electricalfibrillation (right). Note the mild depression of ventricular function following spontaneous fibrillation and the marked depressionfollowing electricalfibrillation. There was no change in left ventricular function cumes following 60 minutes in the beating, nonworking state. (From Hottenrott, Maloney, and Buckberg [5].)
that of the beating, empty heart - and homogeneous flow distribution (endocardiaVepicardia1flow ratio = 1.0). These findings show that “normal” flow distribution (endo/epi ratio = 1.0) and increased subendocardial flow d o not necessarily signify adequate perfusion and that electrical fibrillation should be avoided because it causes ischemia. Another sign of ischemia caused by electrical fibrillation is the occurrence of reactive hyperemia after defibrillation. T h e shape of the postoperative function curve correlated closely with the time after defibrillation that it was inscribed. Function was depressed markedly if bypass was discontinued immediately (within 10 minutes) while the heart was still undergoing reactive hyperemia. Less myocardial depression was seen when the function curve was inscribed 30 minutes later and coronary flow had returned to control levels. T h e clinical correlate is that prolongation of bypass in hearts injured by hypoxia (aortic cross-clamping or ventricular fibrillation) may be beneficial to subsequent myocardial performance. EFFECTS OF LEFT VENTRICULAR HYPERTROPHY
Most studies of the effects of ventricular fibrillation have been performed in normal ventricles, although subendocardial necrosis occurs most commonly in hypertrophied hearts. It is well recognized that the hypertrophied heart is more vulnerable to ischemic injury than is the nonhypertrophied heart [9]. In contrast to our findings on the effects of spontaneous fibrillation on normal left ventricles, we observed that the hypertrophied heart becomes ischemic when it fibrillates [71. Subendocardial flow fails to rise sufficiently in spontaneously fibrillating hypertrophied hearts to provide adequate oxygen delivery (Fig. 4), and dogs with left ventricular hypertrophy develop the same metabolic and functional impairment as seen after electrical fibrillation in normal hearts [ 5 ] . At postmortem examina-
VOL. 20, NO. 1, JULY, 1975
79
BUCKBERG AND HOTTENROTT
160 -
T
Bwtinq empty
soot
140 -
120100 E
0" \
80-
0
60-
40 -
200
Normal L V
LVH
FIG. 4. BloodJlow through l&t ventricular ( L V )subendocardial muscle in the normal and hypertrophied heart ( L V H )when they are beating and nonworking and after 60 minutes of spontaneous ventricular fibrillation. Subendocardial p o w increases approximately 300 % during fibrillation in normal hearts. Conversely, coronaryflow to subendocardial muscle does not increase during fibrillation in hypertrophied hearts. (From Hottenrott, Towers, Kurkji, Maloney, and Buckberg [7].)
tion the hearts showed hemorrhagic necrosis of the left ventricle (Fig. 5). Recent studies by Isom and associates [8] provide evidence that in patients with hypertrophied heart, myocardial oxygen consumption fails to increase over that of the beating, empty heart during ventricular fibrillation; however, myocardial enzyme levels in coronary sinus blood in fibrillating, hypertrophied hearts increased significantly beyond the values observed when the heart was beating without doing external work. In the beating heart the subendocardium is most vulnerable to ischemic damage because it receives its flow in diastole; systolic compressive forces are greatest in the subendocardium and impede flow. Several factors may contribute to the inability of subendocardial flow to increase adequately in hypertrophied, fibrillating hearts. T h e fibrillating heart attains a chamber size comparable to that FIG. 5. Transuerse section of heart from a dog with chronic aorticstenosis and left ventricular hypertrophy who underwent 60 minutes of spontaneous ventricular fibrillation. Focal hemorrhages and venous congestion are present transmurally but are most prominent in the inner hayof the wall of the hypertrophied left ventricle. (From Hottenrott, Towers, Kurkji, Maloney and Buckberg [7].)
80
THE ANNALS OF THORACIC SURGERY
Ventricular Fibrillation FIG. 6. Distribuhoti of It$ veistriculnr ji’oi(i iii the f i v e experimetrtnl corrditioiis. Note tiint the proportioii of total l$t veiitriculnr j l o w rlelivQred to the subeirdocnr(liuiii zv highrst with spoiitniieousfibrillntioii mid redistributed nionQfroiti the subeiirlocnrdiuiir with disteiitioti. Electrically fibrilkrted Amrts do iiot receive proportionately nugiiieiited coroiicriy flow without disteiitioir, n i d the greatest redistributioii of flow nzcinpfroiir the subeiidocnrdiuni occurs wheii they are disteirderl. ( E N D O / E P I = eirdocnrdinlle~cnrdinl~. (From Hottetirott ntrd Buckberg [ 4 ] . )
2 00-
T
160-
a
EZj Nandsfended I Dislended
120-
0
z 8040-
0-
BEATING EMPTY
SPONTANEOUS
ELECTRICAL
F l BR I LL AT ION
at end-systole, so that subendocardial vessels may be compressed by fibrillating muscle fibers. Theoretical analysis of intramyocardial stress gradients suggests that these compressive forces are highest in the subendocardial region, especially in hypertrophied hearts [ 11. Studies of intramyocardial pressure by Baird and his co-workers [2] confirm this suggestion, as do our experimental and clinical studies of ventricular fibrillation in hypertrophied hearts. We conclude that the normal heart provides an inadequate model for studying clinical entities and that any form of normothermic fibrillation should be avoided in hypertrophied hearts. EFFECTS OF VENTRICULAR DISTENTION
T h e importance of adequate decompression of the left ventricle in hearts fibrillated during bypass is well recognized by all cardiac surgeons. Malfunction of the ventricular vent is associated with impaired postoperative myocardial performance, which has been attributed to overstretching of the myocardium. Our studies show that ventricular distention during fibrillation raises intracavitary pressure, reduces left ventricular driving pressure (difference between aortic and left ventricular pressure), and impedes subendocardial flow (Fig. 6). T h e flow impediment occurs in both spontaneously beating and electrically fibrillating normal hearts and results in ischemia and depressed myocardial performance [4]. T h e supply/demand discrepancy (ischemia) caused by fibrillation is accentuated by distention since oxygen requirements are raised simultaneously by increased wall tension. We believe that the ischemia caused by distention of fibrillating hearts may impair postoperative myocardial performance more than overstretching of muscle fibers. EFFECTS OF PERFUSION PRESSURE DURING FIBRILLATION
Our studies, which were performed at a perfusion pressure of 100 mm Hg, suggest that fibrillating muscle prevents flow from increasing sufficiently to meet metabolic demands. T h e obvious counterforce to the deleterious effects of ventricular fibrillation is the coronary driving pressure, which can be varied by changing aortic perfusion pressure during bypass. We found that raising perfusion pressure during the early phases of electrical fibrillation, when myocardial oxygen consumption is lowered because oxygen delivery is inadequate, increases
VOL. 20, NO. 1 , JULY, 1975
81
BUCKBERG AND HOTTENROTT both subendocardial flow and left ventricular oxygen consumption [6] and reverses anaerobic glycolysis. In other experiments [3], we found that reducing perfusion pressure from 100 to 50 mm Hg in spontaneously fibrillating normothermic hearts results in a significant fall in subendocardial oxygen delivery, redistribution of flow away from the subendocardium, abnormal glycolysis, and histochemical signs of ischemia. Baird and associates [2] reported similar changes in subendocardial flow and distribution when perfusion pressure is reduced in normal and hypertrophied hearts. Since patients with coronary artery disease frequently have a fairly large pressure drop across the atherosclerotic obstruction, it is likely that myocardial perfusion pressure will be substantially reduced below that in the aortic root during clinical bypass. All these studies suggest that despite the absence of ventricular hypertrophy, these hearts may be inadequately perfused if they are allowed to fibrillate spontaneously during coronary revascularization. EFFECTS OF HYPOTHERMIA AND DURATION OF FIBRILLATION
T h e left ventricle fibrillates more slowly and with less vigor when myocardial temperature is reduced. It is therefore understandable that the oxygen uptake of the adequately perfused, spontaneously fibrillating heart is reduced by approximately 40% when myocardial temperature is lowered from 37" to 28°C [3]. T h e hypothermic normal left ventricle, however, is not protected against subendocardial ischemia if perfusion pressures are reduced to 50 mm Hg; diminished subendocardial oxygen delivery, abnormal glycolysis, and histochemical damage result when perfusion pressure falls to levels frequently used during clinical open-heart operations. Our studies indicate that the extent of histochemical and biochemical ischemic damage is accentuated when the duration of electrical fibrillation and inadequate perfusion pressure is prolonged [3]. MECHANISMS OF ISCHEMIA
Our studies suggest at least three possible causes of ischemic subendocardial damage during ventricular fibrillation. First, it seems that the fibrillating muscle itself prevents flow from increasing sufficiently to meet metabolic demands. We found that increasing the strength of the fibrillating stimulus with either alternating or direct current caused a progressive fall in the peak achievable flow following maximum coronary vasodilation [6]. Second, distention of the fibrillating heart adds a counterforce that opposes flow by reducing coronary driving pressure and thus impairs subendocardial perfusion while increasing oxygen requirements due to raised wall tension. Third, in hypertrophied hearts fibrillating spontaneously and normal hearts fibrillating electrically, we observed that subendocardial flow fell progressively as fibrillation was continued for one hour. The reduced flow reflected a progressive rise in vascular resistance that occurred despite perfusion pressure, adequacy of venting, and fibrillating stimulus remaining constant. This observation led us to conclude that the myocardium may have become edematous because ischemic damage was prolonged during fibrillation. Our conclusion was supported by histological studies that showed increased
82
T H E ANNALS OF THORACIC SURGERY
Ventricular Fibrillation 350-
FIG. 7. Effects of mannitol on spontaneously fibrillating normal (LV) and hypertrophied ( L V H ) hearts. Left ventricular flow is augmented in the normal ventricle and depressed in the hypertrophied ventricle after 60 minutes of spontaneousfibrillation. Flow to the normal ventricle is reduced and that to hypertrophied ventricle increased immerliately following defibrillation. There is marked augmentation i n f l o w to hypertrophied heart and a smalljlow increment to normal heart following administration of mannitol after defibrillation. (From Hotten-
300
0
NORMAL L V
mLVH
250
kJ',;Cl gBdg,mIn
200-
150-
100-
50-
'-
BEATING
60 min SPONT
PREMANNlTOL POSTMANNlTOL
distance between myocardial fibers. We tested this hypothesis by infusing mannitol into the aortic root following defibrillation at a time when the coronary vessels were maximally dilated (no reactive hyperemia); a striking increase in coronary blood flow (Fig. 7) and improved ventricular performance resulted [6].
Clinical Correlation As a result of these experimental studies, we have completely abandoned the use of ventricular fibrillation during clinical open-heart operations. Since 1972 we have attempted, in all patients undergoing valve replacement, to allow the heart to beat continuously with an adequate coronary perfusion pressure. Because our studies strongly suggest that depressed myocardial performance after cardiopulmonary bypass is a sign of ischemic injury during extracorporeal circulation, we have compared our clinical results before and after 1972. The results are shown in the Table. Inotropic drugs have not been needed in any of 47 patients undergoing aortic valve replacement and in only 1 of 22 patients undergoing mitral valve replacement. This patient had prolonged intermittent ischemic arrest due to technical problems and since 1972 is the only person to have died of postoperative low-output syndrome following isolated valve replacement at our institution. NEED FOR POSTOPERATIVE INOTROPIC SUPPORT
Inotropic Dtugs Used MVR AVR
Method of Myocardial Preservation Ischemic arrest
47% (9/19) 54% (15/28) 5% (1/22)
Ventricular fibrillation Beating, empty MVR
=
mitral valve replacement; AVR
=
33 % (4/12) 41% (12/29) 0% (0/47)
aortic valve replacement.
VOL. 20, NO. 1, JULY, 1975
83
BUCKBERG AND HOTTENROTT
Dkcussion (Drs. Guyton, Buckberg, Michaelis, Brody, and Gay) We try to maintain a perfusion pressure of 100 mm Hg at the tip of each perfusion cannula. Of course the pressure drop along the tubing relative to flow rate must be taken into account as well as differences between the height of the pump and the height of the operating table. Flows average between 150 and 200 mumin in the left perfusion cannula and, depending on the size and distribution of the right coronary artery, anywhere between 75 and 125 mumin in the right. Perfusion is conducted at 32°C.If the heart fibrillates, we make sure the perfusion cannula has not been placed too far into the left coronary ostia. We try to defibrillate immediately, and if this is not possible, we suspect that flow is going preferentially into the circumflex system. We then reduce the perfusion temperature to 28"C, place ice on the surface of the heart, continue perfusing, and work more quickly. We use a balloon cannula except when there is a short left main coronary artery, in which case we use a basket cannula. We start perfusing immediatelyafter opening the aorta. The valve is removed while the coronary arteries are perfused so that calcific emboli are prevented and myocardial protection is achieved immediately. Since maintaining a beating, empty heart with adequate perfusion pressure, we have not had a single patient require postoperative inotropic support after aortic valve replacement.
References 1 . Archie, J. P., Jr. Determinants of regional intramyocardial pressure. J Surg Res 14:338. 1973. 2. Baird, R. J., Okumori, M., Dutka, F., d e la Rocha, A., and Goldbach, M. Surgical aspects of regional myocardial blood flow and myocardial pressure.J Thorac Cardiovasc Surg (In press.) 3. Brazier, J., McConnell, D., Cooper, N., DeLand, E., and Buckberg, G. T h e effects of temperature, time, and perfusion pressure on the adequacy and distribution of coronary flow during ventricular fibrillation. Surg Forum 25: 172, 1974. 4. Hottenrott, C., and Buckberg, G. Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow: 11. Effects of ventricular distention.] Thorac Cardiovasc Surg 68:626, 1974. 5. Hottenrott, C., Maloney, J. V., Jr., and Buckberg, G. Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow: I. Electrical vs. spontaneous fibrillation. J Thorac Cardiovasc Surg 68:6 15, 1974. 6. Hottenrott, C., Maloney, J. V., Jr., and Buckberg, G. Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow: 111. Mechanisms of ischemia. J Thorhc Cardiovasc Surg 68:634, 1974. 7. Hottenrott, C. E., Towers, B., Kurkji, H. J., Maloney, J. V., and Buckberg, G. T h e hazard of ventricular fibrillation in hypertrophied ventricles during cardiopulmonary bypass. J Thorac Cardiovasc Surg 66:742, 1973. 8. Isom, 0. W., Kutin, N. D., Falk, E. A., and Spencer, F. C. Patterns of myocardial metabolism during cardiopulmonary bypass and coronary perfusion. ] Thorac Cardiovasc Surg 66:705, 1973. 9. Iyengar, S. R. K., Ramchand, S., Charrette, E. J. P., Iyengar, C. K. S., and Lynn, R. B. Anoxic cardiac arrest: An experimental and clinical study of its effects.] Thorac Cardiovasc Surg 66:722, 1973.
84
THE ANNALS OF THORACIC SURGERY
Ventricular Fibrillation 1 0 . Keis, K.L., Cohn, L. H., and Morrow, A. G. Effects ofinduced ventricular fibrillation on ventricular perforiiiance and cardiac Inetabolism. Circulation 36:234, 1967.
Editors’ Note: It should be mentioned that many experienced surgeons have obtained excellent results in large numbers of patients by utilizing hypothermic anoxic arrest without coronmy perfusion. Subendocardial necrosis may occur in a number o f settings, not all o f which iiivolve ventricularfibrillation or cardiopulmonary bypass. Even duringperfirsion, its occurrence h a been shown recently to depend not only uponfibrillation or the lack there$, but also upon the adequacy of perfusion pressure.
VOL. 20, NO. 1, JULY, 1975
85
