demand balance with α-adrenergic drugs during cardiopulmonary resuscitation

Optimizing myocardial supply/demand balance with a-adrenergic drugs during cardiopulmonary resuscitation In 14 dogs the determinants of myocardial blood supply and metabolic demands were assessed during cardiopulmonary resuscitation (CPR) and under steady-state conditions in fibrillating hearts on cardiopulmonary bypass. During open chest cardiac massage (nine dogs), coronary diastolic blood pressure and blood flow were low. Vasopressor infusion (methoxamine or epinephrine) raised diastolic pressure from 33 ± 3 to 55 ± 3 mm. Hg* and increased coronary blood flow (CBF) 124 percent (from 38 ± 3 to 85 ± II c.c. per 100 Gm. per minute.* Comparison of these drugs in fibrillating hearts on cardiopulmonary bypass showed that epinephrine increased the "vigor of fibrillation" (intraventricular balloon pressure rose 24 percent and oxygen uptake increased 42 percent)* but impeded subendocardial flow 53 percent* (endocardiallepicardial flow ratio fell from 0.79 to 0.48*). In contrast, methoxamine did not significantly change intraventricular balloon pressure, oxygen uptake, coronary flow, or its distribution. We conclude that augmentation of diastolic pressure with alpha adrenergic drugs during CPR improves coronary perfusion and that inotropic drugs may worsen myocardial ischemia during CPR by raising oxygen demands while simultaneously impeding subendocardial blood supply.

James J. Livesay, M.D., David M. Follette, M.D., Klaus H. Fey, M.D., Roy L. Nelson, M.D., Edward C. DeLand, Ph.D., R. James Barnard, Ph.D., and Gerald D. Buckberg, M.D., Los Angeles, Calif.

.LVAyocardial ischemia, resulting in ventricular fibrillation, is the most common cause of cardiac arrest. This ischemia usually is due to an imbalance between myocardial oxygen demands and available oxygen supply and is most severe in left ventricular subendocardial muscle.1 The immediate success of any resuscitative attempt depends upon prompt restoration of a more favorable myocardial supply/demand balance. During most clinical cardiopulmonary resuscitations (CPR), attention usually is directed toward restoring a satisfactory systemic blood flow (adequate peripheral pulse); cardiotonic drugs (e.g., epinephrine) are frequently given to increase the "vigor of fibrillation" before countershock.2' 3 Resuscitation fails when the From the Department of Surgery, UCLA Medical Center, Los Angeles, Calif. 90024. This work was supported in part by funds from the Wilbur May Fund and the Blalock Foundation. Received for publication Feb. 7, 1978. Accepted for publication May 22, 1978. Address for reprints: G. D. Buckberg, M.D., Thoracic Surgery, UCLA Medical Center, Los Angeles, Calif. 90024. *p < 0.05.

244

myocardial ischemia causing the ventricular fibrillation is irreversible or when the supply/demand imbalance is not improved during open or closed chest cardiac massage. This study tests the hypotheses that (1) diastolic blood pressure is the key determinant of myocardial blood flow during CPR, (2) attention should be directed toward raising diastolic blood pressure with vasopressors, and (3) inotropic drugs which increase the "vigor of fibrillation" may worsen myocardial ischemia by impeding subendocardial blood supply while simultaneously increasing myocardial oxygen demands. Methods Experimental preparation. Fourteen mongrel dogs (weighing 20 to 35 kilograms) were premedicated with intramuscular morphine sulfate (2 mg. per kilogram), anesthetized with intravenous thiamylal (15 mg. per kilogram), and maintained on alpha chloralose (75 mg. per kilogram). Ventilation was controlled with a positive-pressure respirator. A median sternotomy was performed, heparin (2 mg. per kilogram) was given, and polyethylene catheters were placed into the aortic arch, femoral artery, left atrium, right atrium, coronary

0022-5223/78/0276-0244$00.80/0 © 1978 The C. V. Mosby Co.

Volume 76 Number 2 August, 1978

Myocardial supply/demand balance during CPR

sinus, and left ventricular cavity. The dogs were cannulated for cardiopulmonary bypass as previously described.4 Extracorporeal circulation was not instituted until after the resuscitation study was completed. Measurements. Aortic and left ventricular pressures were recorded continually through a Statham P23 db transducer on a Honeywell optical recorder. The diastolic pressure-time index (DPTI), an index of potential subendocardial blood supply, was calculated by planimetry of the area between aortic and left ventricular pressure curves as described previously.4 Regional flow. Total and regional coronary blood flow (CBF) were measured by injecting 8 to 10 /i radioactive microspheres (I41Ce, 85Sr, 125I, 46Sc) into the left atrium. During resuscitation, reference samples were collected for 6 minutes. This longer time interval was chosen after pilot studies during the low-output state of open chest cardiac massage showed that 99 percent of injected microspheres appeared in the reference sample within 5 minutes after injection. Our previous studies have validated the use of the microsphere method to measure flow accurately under low flow states (i.e., during CPR) so long as sufficient microspheres are present in the region of interest and in the reference sample.5 At the conclusion of the experiment, the heart was removed and the left ventricular wall was separated into three layers of equal thickness (subendocardium, midmyocardium, and subepicardium). Each layer was analyzed by gamma spectrometry and regional flows were calculated by a modification of the method of Rudolph and Heymann.6 Mean and phasic flow. Instantaneous mean and phasic coronary flows were measured by placing an electromagnetic flowmeter transducer (Statham 2200) around the left anterior descending coronary artery when alpha and beta adrenergic drugs were tested during the steady-state conditions in the second part of the study (i.e., cardiopulmonary bypass). Electrical zero flow calibration was checked frequently by mechanical occlusion of the distal vessel. Phasic flow recordings were not possible during cardiac massage because of the possible distortion and kinking of the flow transducer during manual compression. Calculations. Coronary vascular resistance (CVR). In the beating heart, CVR was calculated from the peak diastolic flow recorded by flowmeter and instantaneous aortic pressure. In fibrillating hearts, CVR was calculated from the mean CBF recorded by flowmeter and mean aortic pressure. CVR (dynes-sec-cm--') = Aortic pressure CBF (flowmeter)

x g()

245

Left ventricular oxygen consumption was calculated from measurements of left ventricular coronary flow (microspheres) and aortic and coronary sinus oxygen content (Severinghaus method7) by the equation: LV 0 2 consumption = CBF x (arterial-coronary sinus 0 2 content) Resuscitation study. Circulatory arrest was induced in nine dogs by fibrillating the heart with a brief 60 cycle alternating current electrical stimulus. Cardiopulmonary resuscitation (CPR) was begun after 1 minute of circulatory arrest by providing positive-pressure ventilation with 100 percent oxygen and instituting open chest cardiac massage. Open cardiac massage was chosen because (1) it is a widely accepted clinical technique for CPR, (2) a thoracotomy was necessary to measure coronary blood flow, and (3) closed chest massage in dogs does not reflect CPR in human beings because of the dissimilar thoracic anatomy.8 Interventions during resuscitation. CPR alone (control). All dogs underwent at least one episode of CPR without pharmacologic intervention. Regional flow measurements (microspheres) were made approximately 3 minutes after resuscitation was started. CPR plus vasopressors. Each dog underwent at least one additional episode of CPR while either methoxamine (Vasoxyl, 0.25 mg. per kilogram per minute) or epinephrine (Adrenalin, 10 meg. per kilogram per minute) was given by constant intravenous infusion. Defibrillation was carried out after the arterial reference samples were collected (6 minutes after injection of microspheres) and after the dog was placed on cardiopulmonary bypass. We used extracorporeal circulation between resuscitations in order to rest the heart (beating empty state), provide satisfactory perfusion pressure (100 mm. Hg), and correct acid-base abnormalities (restore pH to 7.40) so that subsequent resuscitative efforts in each dog could be carried out under comparable control conditions. Extracorporeal circulation was continued for 15 to 20 minutes, and ventricular fibrillation was reinduced only after the dog showed a minimum of 5 minutes of circulatory stability after discontinuation of cardiopulmonary bypass. Effects of vasopressors on the fibrillating heart. Since the resuscitation study was designed to simulate clinical CPR, it could not be performed under steadystate conditions. Therefore, we performed a separate study to assess the specific pharmacologic effects of vasopressors on the fibrillating heart under more highly controlled conditions than could be achieved in the CPR study. Five dogs did not undergo circulatory ar-

The Journal of Thoracic and Cardiovascular Surgery

2 4 6 Livesay et al.

Table I. Blood pressure and coronary flow during cardiopulmonary resuscitation LV coronary flow (c.c.l 100 Gm.lmin.)

Blood pressure (mm. Hg) Group Control (N = 9) CPR (N = 9) CPR + methoxamine (N = 6 ) CPR + epinephrine (N = 5)

DPTI (mm. Hg ■ sec./min.)

Total

Endocardial

Endolepi

4 ± 1

2,228 ± 366

82 ± 6

92 ± 7

1.13 ± 0.02

33 + 3

10 ± 2

855 ± 132

38 ± 3

34 ± 3

0.94 ± 0.03

83 ± 6*

54 + 5*

11 ± 3

1,911 + 295*

69 ± 7*

67 ± 7*

1.17 ± 0.11*

87 ± 4*

57 ± 6*

11+4

1,835 ± 137*

110 + 22*

105 + 24*

1.00 ± 0.02

Heart rate

Aortic systolic

Aortic diastolic

157+11

126 ± 9

79 ± 6

76 ± 3

68 ± 5

70 + 3

76 ± 5

Mean LV diastolic

Legend: Values are mean ± S.E. LV, Left ventricular. DPTI, Diastolic pressure-time index. CPR, Cardiopulmonary resuscitation. Endo/epi, Endocardial/ epicardial. *p < 0.05 compared to CPR.

rest; instead, they were placed on cardiopulmonary bypass in order (1) to control perfusion pressure at 65 mm. Hg (the approximate diastolic pressure during CPR in the resuscitation study), (2) to provide satisfactory tissue (whole body) perfusion to avoid the deleterious effects of peripheral metabolites on the myocardium, and (3) to maintain the fibrillating state which was present during resuscitation. We produced maximum coronary vasodilatation with dipyridamole (Persantine,* 1 mg. per kilogram) intravenously in order (1) to simulate clinical conditions of severe ischemia which exist after cardiac arrest and (2) to separate changes in resistance related to coronary autoregulation from the effects of increasing extravascular compression of coronary vessels by more vigorous ventricular fibrillation. The completeness of vasodilatation was repeatedly checked by ensuring that no coronary reactive hyperemia was present after a 30 second occlusion of the distal left anterior descending coronary artery. The systemic effects of dipyridamole were counteracted by adjusting the systemic flow rate of the extracorporeal circuit to maintain a constant aortic pressure. The vigor of fibrillation was assessed by recording left ventricular pressure within an intracavitary balloon inserted through an apical stab incision and inflated to a constant 10 ml. volume. The left ventricle was adequately vented so that intraventricular pressure was zero with the balloon deflated. Fibrillation without vasopressors. Control arterial pressure was kept at 65 mm. Hg by adjusting the flow rate of the extracorporeal circuit. Measurements of regional flow (microspheres were injected into the arterial *Dipyridamole (Persantine) was supplied by Boehringer Ingelheim Ltd., Elmsford, N. Y.

perfusion cannula of the extracorporeal circuit), left ventricular oxygen uptake, and intraventricular balloon pressure were made under steady-state conditions after 30 minutes of cardiopulmonary bypass. Fibrillation with vasopressors. Arterial pressure was lowered to 40 mm. Hg by reducing the flow rate of the extracorporeal circuit. With systemic flow rate kept constant, arterial pressure was then increased to 65 mm. Hg by raising peripheral vascular resistance with either methoxamine or epinephrine infusion. Regional flow, intraventricular pressure, and oxygen uptake measurements were repeated. Defibrillation was accomplished after each measurement, arterial pressure was raised to 100 mm. Hg, and pH correction (7.4) was made. Results Cardiac resuscitation. The results of CPR (cardiac massage) on blood pressure, DPTI, and regional coronary blood flow (CBF) are summarized in Table I and Figs. 1 to 4. Aortic systolic and diastolic blood pressures during open chest cardiac massage were significantly lower (52 percent, p < 0.001) than those recorded prior to arrest, whereas left ventricular diastolic pressure increased substantially during resuscitation (Fig. 1). Consequently, potential left ventricular subendocardial blood supply, estimated from the DPTI, was 62 percent lower during resuscitation. Subendocardial flow during CPR was only 34 c.c. per 100 Gm. per minute, a 63 percent reduction below the normal level of subendocardial perfusion seen in the working heart (Table I). Vasopressor infusion during CPR caused a significantly greater increase in diastolic than systolic blood pressure (67 versus 25 percent, p < 0.05); thus there

Volume 76 Number 2 August, 1978

BP

Myocardial supply/demand

CONTROL

CPR

balance during CPR

247

100-i

mmHg 120

80 j

6CH y

4OH 2CH

Fig. 1. Phasic aortic and left ventricular blood pressure during cardiopulmonary resuscitation (CPR). DPTI, Diastolic pressure-time index. Note the low aortic diastolic pressure, high left ventricular diastolic pressure, and low DPTI during conventional CPR. The variability in pressures reflects the non-steady-state conditions of CPR.

CPR

CPR + PRESSORS

Fig. 3. Left ventricular subendocardial blood flow during CPR. Note the significant increase in subendocardial flow with vasopressors. *p < 0.001. 130-1

2500 1.20 \

T

2000

1.10-1

1500

1.00 \

1000

.90 \

.80-1

500

ol CONTROL

CPR

CPR + PRESSORS

Fig. 2. Effect of vasopressors on diastolic pressure-time index during CPR. Note the significant increase in DPTI when vasopressors are given during CPR (p < 0.001). was a striking (120 percent) increment of DPTI (Fig. 2) and subendocardial flow (Fig. 3) (p < 0.001). Although left ventricular flow rose higher during resuscitation with epinephrine than methoxamine, the difference between the effects of these drugs did not reach statistical significance (p < 0.07). With methoxamine, however, the increased left ventricular flow became redistributed toward the subendocardium (endocardial/epicardial flow ratio 1.17, p < 0.05), whereas epinephrine did not change flow distribution from that occurring with cardiac massage alone (endocardial/ epicardial ratio 0.94 versus 1.0) (Fig. 4). Effect of vasopressors on fibrillating hearts. Conversion from the beating heart to ventricular fibrillation caused a 98 percent increase (p < 0.05) in coronary vascular resistance under the steady-state conditions of 65 mm. Hg perfusion pressure and fixed systemic flow rate (by extracorporeal circuit) (Fig. 5). The effects of methoxamine and epinephrine on regional CBF, oxygen uptake, and intraventricular balloon pressure measured in fibrillating hearts under steady-state conditions (no CPR) are summaried in Table II and Figs. 6 and 7.

CPR

Fig. 4. Left ventricular flow distribution to subendocardial (endo) and subepicardial (epi) muscle during CPR. Note the preferential flow redistribution toward the subendocardium with Vasoxyl. *p < 0.05. Epinephrine infusion caused ventricular fibrillation to become visibly more vigorous and rapid; the electrocardiographic pattern showed coarser fibrillation and intraventricular balloon pressure rose 24 percent (p < 0.01) (Fig. 6). Inotropic stimulation with epinephrine raised myocardial oxygen demands (oxygen uptake rose 43 percent, p < 0.05). Simultaneously, epinephrine caused a 36 percent reduction in total left ventricular flow and an even greater decrease (53 percent), in the subendocardial perfusion (p < 0.01). Consequently, total left ventricular flow became redistributed away from the subendocardial muscle (endocardial/epicardial flow ratio fell to 0.48, p < 0.02) (Fig. 7). In contrast, methoxamine did not change oxygen demands; left ventricular oxygen uptake was unaffected and the vigor of fibrillation (intraventricular balloon pressure) was unchanged. Nor did it significantly alter total or regional blood flow. Discussion Successful cardiopulmonary resuscitation (CPR) depends upon correction of myocardial ischemia by

The Journal of Thoracic and Cardiovascular Surgery

2 4 8 Livesay et al.

Table II. Vasopressor effects in fibrMating hearts LV coronary flow (c.c.1100 Gm./min.)

Oxygen

Group

Total

Endocardial

Epicardial

Endo/epi

LV balloon pressure (% of control)

Control (N = 5) Methoxamine (N = 5) Epinephrine (N = 5)

241 ± 38

202 ± 43

252 ± 33

0.79 ± 0.09

100

184 ± 29

143 ± 21

206 ± 34

0.72 ± 0.07

104 ± 5

7.8 ± 0.8

193 ± 14

0.48 ± 0.04*

124 ± 7*

13.0 ± 1.4*

155 ± 14*

94 ± 12*

uptake (c.c. 02l 100 Gm./min.) 9.5 ± 1.1

Legend: Values are mean ± S.E. For abbreviations see Table I. *p < 0.05 compared to control.

prompt restoration of sufficient coronary blood (oxygen) supply to meet myocardial metabolic requirements. Our study shows that myocardial perfusion in fibrillating ventricles during CPR, especially to subendocardial muscle, is maintained best when diastolic blood pressure is augmented by vasopressor drugs. However, subendocardial blood supply can be reduced severely when drugs with inotropic properties (e.g., epinephrine) are given to fibrillating hearts that have expended coronary vasodilator reserve. By testing the effects of alpha and beta adrenergic drugs under steady-state conditions of cardiopulmonary bypass and controlled perfusion pressure, we showed that pharmacologic augmentation of the force of fibrillation accentuates myocardial oxygen demands while simultaneously impeding blood flow to the inner shell of the left ventricle. In normally beating hearts, the subendocardium receives all of its flow during diastole4; systolic perfusion of this region is impeded by the contracting myocardium.9 Ventricular fibrillation simulates sustained systole and imposes an additional and continual counterforce to myocardial perfusion.10, n We have shown previously that the DPTI, the area between the aortic and left ventricular pressure curves in diastole, provides a reasonable estimate of potential subendocardial perfusion in conditions where coronary arteries are dilated maximally.4' 12, 13 DPTI undoubtedly overestimates potential subendocardial perfusion during cardiac massage because the fibrillating myocardium impedes flow when the resuscitator relaxes between compressions. As a result, coronary vascular resistance at the same perfusion pressure is greater in fibrillating than beating hearts, as shown in Fig. 5. It is apparent, therefore, that diastolic perfusion pressure and duration are the major determinants of subendocardial nourishment during CPR. With conventional techniques of CPR (cardiac mas-

sage alone), diastolic pressure could be maintained only at 33 mm. Hg; this resulted in a severe 62 percent reduction in the index of potential subendocardial flow (DPTI). This low driving pressure provided a left ventricular subendocardial flow of only 34 c.c. per 100 Gm. per minute, strikingly less than the level occurring in normally beating hearts (80 to 100 c.c. per 100 Gm. per minute).12 The failure of conventional CPR to provide a satisfactory diastolic pressure has been documented in clinical and experimental studies, but the critical importance of low diastolic perfusion pressure during CPR is rarely emphasized.8' 14~16 Augmentation of diastolic pressure during resuscitation has been accomplished by abdominal compression, aortic counterpulsation, occlusion of the descending aorta, and vasopressor infusion.17-21 Of these, vasopressors are available most readily. In our study, vasopressor infusion during CPR increased diastolic pressure more than systolic pressure, raising DPTI 120 percent and restoring it toward the normal range. Consequently, left ventricular flow rose 124 percent and reached levels necessary for adequate perfusion of the fibrillating heart, which requires almost as much oxygen as the beating working heart.22 These salutary effects of diastolic augmentation during CPR provide a reasonable explanation for the reported ease of defibrillation and improved survival in experimental studies where vasopressors are given during resuscitation.21 Coronary perfusion pressure rose comparably with epinephrine and methoxamine, but left ventricular flow was higher with epinephrine. This difference was not significant statistically because of the marked variability of flow during CPR in hearts capable of coronary autoregulation. The higher flow seen with epinephrine may reflect compensatory vasodilation to meet the increased metabolic demands caused by this drug. Methoxamine infusion redistributed left ventricular flow preferentially toward the subendocardial muscle; it re-

Volume 76

Myocardial supply/demand balance during CPR

Number 2

24 9

August, 1978

130

120

120 90

*o £ o o •B

60-

100

ss

30 0

110

90 J

Beating

0

Fibnllating

Fig. 5. Coronary vascular resistance in maximally vasodilated beating andfibnllatinghearts. Note: Vascular resistance is almost twice as great infibrillatingas in beating hearts at the same perfusion pressure. *p < 0.05. stored endocardial/epicardial flow ratio toward values (1.4:1) which we22, 23 have found necessary to avoid ischemia in fibrillating hearts. In contrast, epinephrine did not change flow distribution from that observed with CPR alone. These variations in regional coronary flow distribution suggested a difference between the alpha and beta adrenergic effects of these drugs and led us to design the second portion of this study, in which we tested their effects under more steady-state conditions than were possible during CPR. We studied the ways each vasopressor affected the determinants of supply and demand in fibrillating hearts by controlling perfusion pressure near levels achieved during CPR and abolishing vascular tone (with dipyridamole) to simulate the maximum vasodilatation occurring during severe ischemia. Fibrillation raised coronary vascular resistance 98 percent, an elevation reflecting increased extravascular compressive forces. The increased extravascular resistance during fibrillation may explain why coronary flows were not as high as previously reported during maximal vasodilatation in beating hearts, although the reduced perfusion pressure we selected for this study was an additional factor.24 Extravascular resistance in fibrillating hearts is even higher during CPR, because intermittent compression by the resuscitator impedes left ventricular flow further. Coronary flows were, however, significantly greater in fibrillating hearts with maximal vasodilatation than in the CPR study at comparable perfusion pressures. The lower flows during CPR may have been related to (1) the systolic opposition of flow during cardiac massage, (2) some element of autoregulation in hearts being resuscitated, and (3) the occurrence of myocardial edema which develops in fibrillating hearts when the perfusion pressure is inadequate.25 Epinephrine increased the vigor of fibrillation markedly and resulted in a significant rise in intraventricular balloon pressure. We used left ventricular balloon

CONTROL

VASOXYL

EPI

Fig. 6. Intraventricular balloon pressure during fibrillation with Vasoxyl or epinephrine (Epi) infusion. Note: Intraventricular balloon pressure is raised significantly by epinephrine, whereas Vasoxyl causes no change. *p < 0.05.

o Q

4020 0J

CONTROL VASOXYL

EPI

Fig. 7. Left ventricularflowdistribution in maximally vasodilated fibrillating hearts. (Endo/epi, Ratio of flow to subendocardial/subepicardial muscle layers). Note: Epinephrine (Epi) redistributed flow away from subendocardial muscle. *p < 0.05. pressure as a barometer of metabolic demands of fibrillation,26 and we found inotropic stimulation of the fibrillating hearts to result in a 43 percent rise in oxygen uptake. This pharmacologic augmentation of left ventricular oxygen demands raised oxygen uptake to 13 c.c. of oxygen per 100 Gm. per minute, a level which is 48 percent higher than necessary in hearts beating spontaneously and doing external work.22 The epinephrine-induced augmentation of myocardial oxygen demands was associated with a significant impedance of left ventricular subendocardial flow (53 percent reduction); the endocardial/epicardial flow ratio fell from 0.79 to 0.48. This finding is consistent with previous studies showing that increasing the vigor of fibrillation compromises coronary flow reserve25 and that flow impedance during fibrillation is greatest in the subendocardium.10' "• 2 2 , 2 7 These observations suggest that the primary salutary effects of epinephrine during CPR are due to its alpha adrenergic effect on peripheral resistance; its beta adrenergic action augments the vigor of fibrillation and imposes a needless increase in cardiac oxygen demands while further impeding subendocardial blood supply.

2 5 0 Livesay et al.

In contrast to epinephrine, methoxamine did not significantly alter myocardial blood supply or its distribution; it did not increase the vigor of fibrillation, intraventricular balloon pressure, or oxygen uptake. CBF was somewhat lower than control with methoxamine, and we suspect that this may reflect ischemic myocardial edema occurring as a consequence of our experimental design. We kept perfusion pressure low at 65 mm. Hg to simulate the conditions present during CPR, and we lowered it even further (to 40 mm. Hg) when we tested the drug effects. We28 have shown previously that fibrillating hearts become ischemic when perfusion pressure is low. It is unlikely that the lower subendocardial flow resulted from methoxamine-induced coronary constriction, since coronary flow recorded continuously by flowmeter did not change at the same perfusion pressure during methoxamine infusion. The absence of a coronary vasoconstrictor effect on subendocardial vessels by methoxamine was reported recently by Wechsler and associates.29 Our study suggests that methoxamine is better than epinephrine during CPR because its potent peripheral vasoconstrictor effect is not combined with simultaneously augmented myocardial oxygen demands or impeded blood supply. This observation is supported by reports showing methoxamine to be twice as effective as epinephrine for the initial resuscitation of fibrillating hearts.30 The clinical implications of our study are threefold. First, our findings suggest that reoxygenation of the heart during CPR can be accomplished most readily when the heart is flaccid rather than fibrillating between manual compressions. Intramyocardial compressive forces opposing CBF are lowest when the heart is flaccid rather than fibrillating between manual compressions. Intramyocardial compressive forces opposing CBF are lowest when the heart can undergo a more normal diastole. The systolic compression caused by the resuscitator provides for systemic oxygenation, but likely opposes coronary perfusion as much as a normal systole. This suggests that conversion of the fibrillating heart to a slow idioventricular rhythm provides the advantage of allowing better myocardial oxygen delivery during cardiac massage. The fibrillating heart, in addition to impeding its own oxygen supply, requires five times more oxygen than the arrested heart.31 However, if defibrillation cannot be accomplished immediately after encountering it, a supply/demand imbalance should be suspected and the primary emphasis directed toward its correction. Second, augmentation of diastolic pressure should be a primary goal of therapy, since CBF is low during

The Journal of Thoracic and Cardiovascular Surgery

conventional CPR due to an inadequate diastolic perfusion pressure and DPTI. Our findings suggest that vasopressors should be used early in the treatment of cardiopulmonary arrest, when peripheral vessels are still capable of vasoconstriction. In a hospital setting, direct arterial monitoring would be extremely useful in making beat-to-beat judgments about the effectiveness of diastolic augmentation. Third, epinephrine, the drug used most commonly during CPR, appears to have an adverse effect on myocardial oxygen supply/demand balance and may worsen ischemia during CPR. The inotropic properties of epinephrine increase metabolic demands at a time when supply is limited and reduce potential supply further by impeding subendocardial flow. Our study suggests that inotropic drugs which "increase the vigor of fibrillation" should be withheld during CPR. Instead, alpha adrenergic drugs (without inotropic effects) should be administered during CPR to improve the supply /demand balance. Our findings do not impugn the use of vasodilator or inotropic drugs, but rather suggest that their greatest usefulness is to maintain the supply/demand balance in the failing heart after cardiac massage is no longer necessary. It is likely that inotropic agents may be necessary after defibrillation, since depressed ventricular function is common following cardiac arrest.32 Similarly, peripheral vasodilators may be useful after a successful resuscitation to reverse the pharmacologically elevated afterload and to improve a low cardiac output.33 Although current mortality following resuscitation is high, it is hoped that the use of improved techniques of CPR and aggressive treatment of myocardial ischemia will improve survival in the event of a sudden cardiac death. We gratefully acknowledge the technical assistance provided by Marilee Lawrence, Jerry Leaf, Edward Dolendo, and Nancy Birman and the secretarial assistance of Barbara Voigt, Judi Wald, and Jill Kamon. REFERENCES 1 Lie JT, Titus JL: Pathology of the myocardium and the conduction system in sudden coronary death. Circulation 51:Suppl 3:41, 1975 2 Gordon AS, et al: Standards for cardiopulmonary resuscitation (CPR) and emergency cardiac care (ECC). JAMA 227:837, 1974 3 Goldberg AH: Cardiopulmonary arrest. N Engl J Med 290:381, 1974 4 Buckberg GD, Fixler DE, Archie JP, Hoffman JIE: Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ Res 30:67, 1972 5 Buckberg GD, Luck JC, Payne DB, Hoffman JIE, Archie

Volume 76 Number 2 August, 1978

6

7

8

9

10

JP, Fixler DE: Some sources of error in measuring regional blood flow with radioactive microspheres. J Appl Physiol 31:598, 1971 Rudolph AM, Heymann MA: Circulation of the fetus in utero. Methods for studying distribution of blood flow, cardiac output and oxygen blood flow. Circ Res 21:163, 1967 Behar MC, Severinghaus JW: Calibration and correction of blood 0 2 content measured by p 0 2 after CO saturation. J Appl Physiol 29:413, 1970 Weiser M, Adler LN, Kuhn LA: Hemodynamic effects of closed and open chest cardiac resuscitation in normal dogs and those with acute myocardial infarction. Am J Cardiol 10:555, 1962 Archie JP Jr: Mechanical determinants of myocardial blood flow and its distribution. Ann Thorac Surg 20:39, 1975 Baird RJ, Dutka R, Okumor M, de la Rocha A, Goldbach M, Hill TJ, MacGregor DC: Surgical aspects of regional myocardial blood flow and myocardial pressure. J THORAC CARDIOVASC SURG 69:17, 1975

11 Downey J: Compression of the coronary arteries by the fibrillating canine heart. Circ Res 39:53, 1976 12 Brazier J, Cooper N, Buckberg GD: The adequacy of subendocardial oxygen delivery. The interaction of determinants of flow, arterial oxygen content and myocardial oxygen need. Circulation 49:968, 1974 13 Buckberg GD, Towers B, Paglia DE, Mulder DG, Maloney JV Jr: Subendocardial ischemia after cardiopulmonary bypass. J THORAC CARDIOVASC SURG

14 15

16

17 18 19

20

64:669, 1972 Ikram H: Circulatory dynamics during manual and mechanical external massage in man. Resuscitation 1:229, 1972 Lillehei CW, Larvadia PG, DeWall RH, Seller RD: Four years' experience with external cardiac resuscitation. JAMA 193:651, 1965 Del Guercio LRM, Feins NR, Cohn JD, Coomaraswamy RP, Wellman SB, State D: Comparison of blood flow during external and internal cardiac massage in man. Circulation 31:Suppl 1:171, 1965 Redding JS: Abdominal compression in cardiopulmonary resuscitation. Anesth Analg 50:668, 1971 Molokhia FA, Ponn RB, Robinson J, Asimacoponlos PJ, Norman JC: A method of augmenting coronary perfusion during internal cardiac massage. Chest 62:610, 1972 Ohomoto T, Miura I, Konno S: A new method of external cardiac massage to improve diastolic augmentation and prolong survival time. Ann Thorac Surg 21:284, 1976 Menno AD, Schenk WG: Dynamics of coronary arterial flow. Flow alterations resulting from certain surgical pro-

Myocardial supply Idemand balance during CPR

25 1

cedures and drugs of surgical importance. Surgery 50:82, 1961 21 Pearson JW, Redding JS: Influence of peripheral vascular tone on cardiac resuscitation. Anesth Analg 44:746, 1965 22 Hottenrott C, Maloney JV Jr, Buckberg GD: Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow. I. Electrical vs. spontaneous fibrillation. J THORAC CARDIOVASC SURG 68:615, 1974

23 Hottenrott C, Buckberg GD: Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow. II. Effects of ventricular distension. J THORAC CARDIOVASC SURG 68:626, 1974

24 Holtz J, Restorff WV, Bard P, Bassenge E: Transmural distribution of myocardial blood flow and of coronary reserve in canine left ventricular hypertrophy. Basic Res Cardiol 72:286, 1977 25 Hottenrott C, Maloney JV Jr, Buckberg GD: Studies of the effects of ventricular fibrillation on the adequacy of regional myocardial flow. III. Mechanisms of ischemia. J THORAC CARDIOVASC SURG 68:634, 1974

26 Monroe RG, French G: Ventricular pressure-volume relationships and oxygen consumption in fibrillation and arrest. Circ Res 8:260, I960 27 Utley JR, Bryant LR, Mobin-Uddin K: Effects of hypercalcemia on subendocardial perfusion and ventricular wall during cardiopulmonary bypass. Surg Forum, p. 142, 1973. 28 Brazier JR, Cooper N, McConnell DH, Buckberg GD: Studies of the effects of hypothermia on regional myocardial blood flow and metabolism during cardiopulmonary bypass. III. Effects of temperature, time, and perfusion pressure in fibrillating hearts. J THORAC CARDIOVASC SURG 73:102, 1977

29 Wechsler AS, Symmonds JB, Kleinmen L: Effects of methoxamine on the coronary circulation during cardiopulmonary bypass. Circulation 54:Suppl 2:213, 1976 30 Redding JS, Pearson JW: Resuscitation from ventricular fibrillation. JAMA 203:93, 1968 31 Buckberg GD, Brazier JR, Nelson RL, Goldstein SM, McConnell DH, Cooper N: Studies of the effects of hypothermia on regional myocardial blood flow and metabolism during cardiopulmonary bypass. I. The adequately perfused beating, fibrillating, and arrested heart. J THORAC CARDIOVASC SURG 73:87, 1977

32 Liberthson RR, Nagel EL, Hirschman JC, Nussenfeld SR: Prehospital ventricular defibrillation. Prognosis and follow-up course. N Engl J Med 291:317, 1974 33 Forrester JS, da Luz PL, Chatterjee MB: Peripheral vasodilators in low cardiac output states. Surg Clin North Am 55:531, 1975