- Email: [email protected]
Mechanical assistance of the left ventricle: Acute effect on cardiac performance and coronary flow of different perfusion patterns
Mechanical assistance of the left ventricle: Acute effect on cardiac performance and coronary flow of different perfusion patterns Mechanical circulatory assistance by ventricular assist devices provides an opportunity to influence the aortic pressure pattern, which may affect ventricular loading and coronary perfusion. The effect of synchronous, pulsatile coronary perfusion of an assist device-supported left ventricle has not been studied. To analyze the effect of different perfusion patterns on left ventricular performance and on coronary flow, independent of pressure and volume loading, we used three different modes of aortic perfusion in an isometric, contracting, isolated canine heart model. The effect of nonpulsatile, counterpulsatile, and copulsatile coronary perfusion was analyzed in four subgroups to simulate different, clinically relevant situations (using two different ventricular end-diastolic volumes [normal and high] and two mean perfusion pressures [normal and critically low]). Our experiments demonstrated that total coronary flow is optimized by making the perfusion pressure pulsatile and by synchronously timing the pump systole with ventricular diastole (counterpulsation). Under identical conditions of preload and mean perfusion pressure, coronary flow and left ventricular contractility were decreased during nonpulsatile and copulsatile aortic perfusion when compared with counterpulsatile flow. There were no significant differences between the nonpulsatile and copulsatile modes. We conclude from these data that a nonejecting, but contracting, left ventricle will have improved systolic function and coronary blood flow if the coronary perfusion pressure is synchronized in a counterpulsatile manner. This is a significant implication for mechanical left ventricular assist devices when used to promote myocardial recovery. (J THORAC CARDIOVASC SURG 1992;104:561-8)
Fabio Bellotto, MD, Robert G. Johnson, MD, Jun Watanabe, MD, PhD, Marc J. Levine, MD, Alvin Franklin, MS, and Ronald M. Weintraub, MD, Boston, Mass.
Left ventricular (LV) mechanical assist devices (LVADs) are most commonly used clinically as a means of systemic circulatory support while awaiting cardiac transplantation. Under these circumstances myocardial recovery is not an issue, and factors associated with myocardial energetics (cardiac work and coronary perfusion) are irrelevant. As assist device technology improves, however, they are likely to be used to promote recovery of myocardium injured by ischemia or inflammation. In such situations the balance of cardiac work and coronary From the Division of Cardiothoracic Surgery and the Division of Cardiovascular Medicine, Beth Israel Hospital, Harvard Medical School, and the Charles A. Dana Research Center, Boston, Mass. Received for publication Oct. 3, 1990. Accepted for publication July 25, 1991. Address for reprints: Robert G. Johnson, MD, Division of Cardiothoracic Surgery, 330 Brookline Ave., Boston, MA 02215.
12/1/32837
blood flow (CBF) will be critically important. This issue may be even more important in the long-term use of these devices. Under normal conditions cardiac systole provides kinetic energy to blood and, for most organs, peak blood flow occurs during that phase of the cardiac cycle. The heart itself, however, receives the bulk of its total blood flow during diastole, as the blood "runs off" into the coronary circulation during myocardial relaxation. The use of a circulatory assist device may alter the relationship of cardiac mechanical cycles and aortic flow in a number of ways. If we assume that the device provides complete volume unloading of the left ventricle, no ejection will occur and aortic flow will be entirely determined by the device. The flow may be nonpulsatile or pulsatile. If pulsatile, the flow may be synchronized to the cardiac electricalmechanical cycles, or it may be asynchronous. Synchronous aortic pulsatile flow may be copulsatile, so that 561
56 2
The Journal of Thoracic and Cardiovascular Surgery
Bel/otto et al.
FLEXIBLE DIAPHRAGM
AIR
__
TO
DRIVE
CONSOLE
-::===::::::<~-p-RiA-AORTlC PRESSURE
::..
SALINE
Fig. 1. Schematic presentation of the model.
device systole occurs at the time of cardiac systole (even though an unloaded left ventricle may not be opening the aortic valve), or it may be counterpulsatile, so that device systole is coupled to cardiac diastole when aortic pressure would normally be decreasing. Although data from intraaortic balloon pump use demonstrate the benefits of diastolic pulsation, this may not be applicable to a situation in which the left ventricle is completely unloaded by an assist device and does not need to eject to sustain adequate systemic blood flow. It might be suggested that with an LV AD unloading the ventricle by copulsation is more physiologic than by counterpulsation since the pressure pulse comes during systole (nonejecting contraction) and coronary runoff will occur concomitantly with systemic pressure decay during diastole. If the LVAD is to be used in situations in which myocardial survival or recovery is desired, it would be beneficial to optimize the heart's blood supply and work demand ratio. Because significant engineering and clinical efforts are required to synchronize pulsatile flow, it is important that we determine whether any benefit is to be derived by the use of either copulsation or counterpulsation. These experiments were performed in an isolated, iso-
volumic-contracting canine heart preparation. This nonejecting left heart model eliminated the effect of ventricular unloading through ejection, while permitting independent control of mean coronary perfusion pressure. The three patterns of aortic flow were studied under four conditions of LV preload and coronary perfusion pressures. One of the four conditions mimicked the normal physiologic state of normal preload and coronary perfusion pressure, and the other three represented deranged physiology. We hypothesized that there would be no significant difference in CBF or cardiac performance among three clinically applicable patterns of perfusion: nonpulsatile, pulsatile-synchronous, and pulsatile-asynchronous under these four physiologic conditions.
Materials and methods Isolated perfused hearts from six mongrel dogs weighing between 19.9and 27.1 kg were prepared, with some modifications, according to a previously described model. I In each experiment a support dog was anesthetized with a-chloralose and urethane and the lungs were mechanically ventilated.Catheters wereplacedin the femoralartery and vein, and the animal was given intravenousheparin, 500 U/kg. The heart of a seconddog was prepared by cannulationof the right
Volume 104 Number 3 September 1992
Mechanical assistance of left ventricle
1.0 0.5
o
563
t-----=-±-f±f±fT-,.....-'--;-,.....-'--+-'--+-'--'-..:_ _
0'
:r
E
E
Q.
o
:i o ~
o
UJ
• . "' . • . . . , . • . ~ . • . . ' lr"! " •••.••.•.,.'. '..•.•.....,.• '. '.. '. · ' '.•. '. '. ' ' ' .t>'. . ~ G"J'I_/L Y· •..•.iUtJjJ .......................' i
:
..Hi··.····.····.····. .
. . . ..
.
.: . :' ---:-.<... ~ .-.; --: -;"':----::";'''':' ':"": :-. :-.: . -" ->:
'cOUNiER-iPULSATH1E·
..
.
)1
II
-
CO-PULSATILE
Fig. 2. Isolated examples of perfusion pressure modes. ECG, Electrocardiogram; CBF, coronary blood flow; PP, perfusion pressure; M, mean perfusion pressure; LVDP, left ventricular developed pressure.
ventricle and the carotid and subclavian arteries after heparinization. The heart was isolated by ligation of the superior and inferior venae cavae and descending thoracic aorta and was removed from the thorax. Coronary perfusion of the isolated heart was maintained via the carotid cannula, with blood obtained from the femoral artery of the support dog and circulated by a roller pump (Cardiovascular Instrument Corp., Wakefield, Mass.) through a bubble trap and a heat exchanger to maintain blood temperature at 37° C. A chamber, provided with a silicone rubber diaphragm dividing blood and air, was interposed between the perfusion line (carotid cannula) and a pneumatic blood pump drive console (Thermedics Inc" Woburn, Mass.), as shown in Fig. I. With the pump off, the chamber worked like a "damper," minimizing oscillations generated by the roller pump (usually smaller than 10 mm Hg) and so providing a virtually nonpulsatile perfusion pressure. Pulsations could be generated synchronously with different phases of the electrocardiogram. The heart was atrially paced at a rate of 130 beats/min. Pulsatile perfusion was achieved with a pulse pressure of 40 mm Hg, recorded in the aortic root through a different (the subclavian) cannula. Counterpulsation was achieved by starting the perfusion pressure upstroke simultaneously with the peak of the LV pressure. Copulsation was achieved by anticipating the perfusion pressure upstroke by 50 msec before the LVend-diastolic pressure. In this way the peak of pulse pressure coincided with the LV end-diastolic pressure in the counterpulsation pattern and with the peak
of LV pressure in the copulsation pattern (Fig. 2). The pulse pressure duration was set at 240 msec. Small adjustments of an in-circuit Starling resistor were necessary to maintain the mean coronary perfusion pressure (which is equal to "intraaortic pressure," as shown in Fig. I) constant in all three of the perfusion patterns. An in-line Doppler flowprobe was used to measure mean total CBF as the coronary return was drained from the isolated right side of the heart. CBF was expressed in milliliters of flow per minute per gram ventricular weight, and the flow probe was calibrated over the range of flows after each experiment. Coronary venous blood was drained into the femoral vein of the support dog. A drainage tube was inserted through a stab wound in the apex of the left ventricle to collect thebesian venous flow. A silicone rubber mount balloon was placed in the left ventricle through the mitral anulus and was secured with sutures and connected to a Statham P23Dp pressure transducer. The isolated heart was then submerged in a warm bath and allowed to equilibrate for 30 minutes at a coronary nonpulsatile mean perfusion pressure of 100 mm Hg at a heart rate of 130 beats/ min. Baseline mechanical and biochemical data were then obtained; LV end-diastolic pressure was measured and LV developed pressure (DP) was determined (DP = LV systolic pressure minus LV end-diastolic pressure) as indices of diastolic compliance and systolic performance, respectively. During the course of a study, changes in LV end-diastolic pressure were
564
The Journal of Thoracic and Cardiovascular Surgery
Bel/otto et at.
Table I
NoP/NN CnP/NN CoP/NN NoP/NL CnP/NL CoP/NL NoP/HN CnP/HN CoP/HN NoP/HL CnP/HL CoP/HL
BV
CPP
EDP
SP
DP
CBF
(ml)
(mmHg)
(mmHg)
(mmHg)
(mm Hg)
(mltminfgm)
92±1O 91 ± 10 93 ± 10 49 ± 4 49 ± 4 49 ± 3 94 ± 10 94 ± 10 94 ± 13 46 ± 7 47 ± 7 47 ± 7
4±2 4±2 4±1 4±2 3± 2 4 ± 2 II ± 3 12 ± 4 II ± 3 12 ± 6 10 ± 5 II ± 4
103 ± 114 ± 101 ± 88 ± 97 ± 87 ± 144 ± 150 ± 142 ± 109 ± 116 ± 97 ±
99 ± 110 ± 97 ± 83 ± 94 ± 83 ± 133 ± 139 ± 132 ± 97 ± 106 ± 86 ±
14 ± 14 ± 14 ± 13 ± 13 ± 13 ± 23 ± 23 ± 23 ± 24 ± 24 ± 24 ±
7 7 7 7 7 7 II II II 10 10 10
28 31 24 23 27 21 40 44 40 27 34 24
28 32 25 23 27 23 42 44 44 32 37 26
1.02 ± 1.28 ± 1.23 ± 0.65 ± 0.80 ± 0.64 ± 1.34 ± 1.49 ± 1.33 ± 0.56 ± 0.71 ± 0.55 ±
0.7 0.7 0.7 0.3 0.5 0.3 0.6 0.7 0.6 0.3 0.3 0.4
CBFr* 100 ± 107 ± 101 ± 60 ± 71 ± 60 ± 135 ± 148 ± 118 ± 57 ± 69 ± 55 ±
0 7 4 26 26 26 33 30 24 28 27 22
BV, Balloon volume; CPP, coronary perfusion pressure; EDP, left ventricular end-diastolic pressure; SP, left ventricular systolic pressure; DP, left ventricular developed pressure; CBF, coronary blood flow ml/rnin/gm of myocardium; CBF r: CBF ratio; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, copulsatile perfusion pressure; NN, normal left ventricular end-diastolic pressure, normal pulmonary pressure; NL, normal left ventricular end-diastolic pressure, low pulmonary pressure; HN, high left ventricular end-diastolic pressure, normal pulmonary pressure; HL, high left ventricular end-diastolic pressure, low pulmonary pressure. *Ratio = CBF at the particular experimental condition/CBF at baseline (NoP and NN).
taken to represent changes in diastolic stiffness, and changes in DP were interpreted as variations in contractility, Each of six isolated hearts was studied in a protocol consisting of a sequence of four hemodynamic conditions simulating four different pathophysiologic situations. In the first, during nonpulsatile perfusion, we inflated the LV balloon to reach an LV systolic pressure of 100 mm Hg while perfusing the aorta with a mean coronary perfusion pressure of 100 mm Hg. Using a nonpulsatile pattern as a control, we alternated the sequence of counterpulsation and copulsation patterns; DP, LV end-diastolic pressure, and CBF were measured after at least 3 minutes of stabilization during each perfusion pattern. In the second series of tests, we abruptly decreased the mean coronary perfusion pressure to 50 mm Hg and then repeated the preceding patterns and measurements. In the third and fourth series of hemodynamic conditions, we increased the ventricular balloon volume to reach an LV systolic pressure of 150 mm Hg and then repeated the whole procedure as in the first and second series, with both normal and low perfusion pressures. In this way, we were able to analyze the impact of nonpulsatile, counterpulsatile, and copulsatile perfusion under four different conditions produced by a combination of two preloads (normal and high) at two perfusion pressures (normal and critically low). Although the peak systolic and diastolic pressures varied with the mode of perfusion, the mean aortic pressure (in this model, coronary perfusion pressure) was held constant at either 100 mm Hg or 50 mm Hg according to the preceding protocol. Myocardial metabolic parameters were tested and compared under what we considered the most physiologically stressful conditions during the fourth part of the experiment, that is, with the highest LV end-diastolic pressure and the lowest coronary perfusion pressure, Blood was sampled from the arterial and coronary venous lines for oxygen and carbon dioxide tension measurements, as well as for determination of lactate production or extraction or both. The hemoglobin, pH, oxygen pressure, carbon dioxide pressure, oxygen content, and saturation were determined with a blood gas analyzer (Instrumentational Laboratory. Lexington, Mass.). Arterial-venous oxygen content difference was calculated; myocardial oxygen consumption was expressed as the
product of CBF and arterial-venous oxygen content difference indexed per 100 gm of ventricular weight. Transmyocardial lactate production/extraction was measured by the technique described by Apstein, Puchaer, and Brachfeld.? All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the N atinal Institutes of Health (NIH Publication No. 80-23, revised 1978), Statistics. Values are expressed as mean ± standard deviation and were compared for statistically significant differences by repeated-measures analysis of variance and appropriate multiple-sample comparison tests (blocked Neuman-Keuls test and Friedman rank sum test). Group NoP (nonpulsatile mode) served as the basis for all intergroup comparisons. If no significant differences were found, the mean values of group CnP (counterpulsation) were compared with those of group CoP (copulsation). Results As shown in Table I, the balloon volume and mean perfusion pressure were maintained at constant values within each test series, In all the conditions tested, we observed higher CBF and DP during counterpulsation. These differences reached statistical significance only when the LV end-diastolic pressure was high and the coronary perfusion pressure was low. Higher CBF and DP were observed to occur simultaneously and almost immediately with alterations in the mode of perfusion (Fig. 3). Note that the rapid paper speed tracing shows the pulsatile coronary perfusion pressure tracings and the slower speed demonstrates the mean (constant for all modes) coronary perfusion pressure, Total CBF. With the exception of the first condition
Volume 104 Number 3 September 1992
Mechanical assistance of left ventricle
START Co-P
No-P WVUmlY'lIIIWlI)
565
• ~~ ~-.,.J---'T 'IIUWU'•••UIII"W.. '.IMU'. UMlMNBVllvr."IU'MMU"""'I"IlII1U"'llJ/'mM¥l"! I
CooP IIU'Y' l'IA'U'WlUUP'W"IlIWUM'll!''¥'WJMMM9111MN'MMU''V!fll'Mn'UI
START Cn-P
_illlitllilllINihIiil_
Cn-P Fig. 3. Transitions of perfusion pressure modes. ECG, Electrocardiogram; LVDP, left ventricular developed pressure; CBF, mean coronary blood flow; PP, perfusion pressure; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, copulsatiie PP. normal LVEDP. normal PP
normal LVEDP. low PP
2.0
2.0
~ 1.5
~ 1.5
e
'E E
c
:g
1.0
E
0.5 0.0
No-P
Cn-P
1.0 0.5 0.0
COoP
I
~ 1.5
C
'E E
~
-'E C
E
0.5
No-P
Cn-P
COoP
2.0
1.0
0.0
Cn-P
high LVEDP. low PP
high LVEDP. normal PP 2.0
No-P
COop
1.5 1.0 0.5 0.0
No-P
Cn-P
COoP
Fig. 4. Effect of different perfusion modes on mean coronary blood flow. LVEDP, Left ventricular end-diastolic pressure; PP, perfusion pressure; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, co-pulsatile perfusion pressure. Comparisons were made: NaP versus CnP and NaP versus CoP.
(normal LV end-diastolic pressure, normal coronary perfusion pressure), the CBF in all the remaining subgroups wassignificantly higher in the counterpulsatile than in the nonpulsatile mode (Fig. 4). The greatest relative CBF
difference between the nonpulsatile and counterpulsatile patterns and (21%) was recorded during conditions we considered most physiologically stressful (high LV enddiastolic pressure, low coronary perfusion pressure). The
The Journal of Thoracic and Cardiovascular Surgery
5 6 6 Bellotto et al.
normal LVEDP, normal PP
normal LVEDP, low PP
160
160
:120 E E 80
120
Cl
:I:
E E 80
40
o
40 No-P
Cn-P
0
COOP
No-P
bigb LVEDP, normal PP
160
:120 E E 80
:120
40
40
E
E
No-P
Cn-P
COop
bigb LVEDP, low PP
160
o
Cn-P
80
o
COOp
No-P
Cn-P
COOp
Fig. 5. Effectof different perfusion modeson left ventriculardeveloped pressure. LVEDP, Left ventricularend-diastolicpressure; PP, perfusion pressure; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, copulsatileperfusion pressure. Comparisons were made: NaP versus CnP and NaP versus CoP. LACTATE EXTRACTION 5.0
4.0 !?!
.s ~
3.0
E 1.0
0.0 No-P
0.0 Cn-P
Co-P
No-P
Cn-P
Co·p
Fig. 6. Effect of different perfusion patterns on myocardial metabolism during conditions of high left ventricular end-diastolic pressureand lowmean coronary perfusion pressure. NaP, Nonpulsatileperfusion pressure; CnP, counterpulsatileperfusionpressure;CoP, copulsatile perfusion pressure.Statistical significance: NaP versusCnP and CnP versus CoP. CBF during copulsation was always lower than in the non pulsatile pattern, but the difference did not reach statistical significance. LV developed pressure. During counterpulsation, the L V developed pressure was always higher than during the nonpulsatile perfusion (Fig. 5). The greatest statistical difference in developed pressure between perfusion modes was noted during the most physiologic condition of a normal LV end-diastolic pressure and normal coronary perfusion pressure (nonpulsatile versus copulsatile: p < 0.005). In contrast, the intermode differences in the DP seen during the stressful situation of a high LV enddiastolic pressure and low coronary perfusion pressure did not reach statistical significance. We could not demonstrate any significant difference in the DP achieved with
copulsatile versus nonpulsatile modes of perfusion under any of the hemodynamic conditions studied. LV end-diastolic pressure. The LV end-diastolic pressure did not show any significant variation among the three perfusion patterns in the four conditions examined, whether with high or low balloon volume (see Table I). Myocardial metabolism. In the only condition in which we evaluated myocardial metabolism (high LV end-diastolic pressure and low coronary perfusion pressure), counterpulsation was associated with a rise in both the myocardial oxygen consumption and lactate extraction when compared with nonpulsation and copulsation. Only the difference in lactate extraction found between the counterpulsation and copulsation modes reached statistical significance (Fig. 6).
Volume 104 Number 3 September 1992
Discussion This study was not intended to add to the extensive data on the effect of pulsatile versus nonpulsatile perfusion of the myocardium during cardiopulmonary bypass or in the presence of ventricular fibrillation.l'!" Instead, we were interested in looking at three commonly used modes of perfusion in a working heart model as they might pertain to intermediate or long-term assistance of the failing left ventricle. When dealing with a working heart, the benefit of synchronized, pulsatile CBF is determined by the desirability of providing an augmented aortic pressure during the period of myocardial relaxation (coun terpulsation) versus the more physiologic condition of an increased aortic pressure immediately before myocardial relaxation (copulsation) versus the maintenance of constant aortic pressure (permitting CBF to be regulated independent of pressure pulses [nonpulsatile]). The diastolic augmentation of counterpulsation is widely assumed to increase CBF in the situation in which the native ventricle is ejecting, but this is not well documented. Counterpulsation provided by intraaortic balloon pumping has been shown to improve the subendocardial flow only in the presence of either critical coronary stenosisl ' or severe hypotension,16-19 when autoregulation is impaired. The intraaortic balloon, however, unlike the LV AD, can be effective only in supporting an ejecting heart (under these conditions there are two positive-pressure waves). Copulsatile perfusion has not been studied previously, because its use in tandem with LV ejection would seem to result in an unfavorable increase in the work of competitive ejection. Additionally, Downey and Kirk-" have suggested that any process that causes the heart to rely on delivery of a greater percentage of its CBF during systole has the potential for resulting in decreased CBF to the subendocardium. Mechanical assistance of the failing left ventricle, however, would theoretically support the entire systemic circulation while unloading the ventricle and precluding ejection. Under these conditions copulsation is an option that rna y closely mimic the ph ysiologic decrease in aortic pressure that occurs in association with the decrease in the wall tension of diastole. In our model the intracoronary blood pressure during diastole in all three modes of perfusion was always higher than the LV enddiastolic pressure. In evaluating the impact on beating hearts of pulsatile pressure augmentation induced by an LV AD, two different factors have to be taken into consideration: (I) pressure and volume unloading of the left ventricle and (2) possible "intrinsic" effects of pressure waves on coronary resistances and flow. Our isolated dog heart model controls the former, and it is therefore ideal for separating the
Mechanical assistance of left ventricle
567
"intrinsic" effect of different perfusion patterns on myocardial performance and CBF from the effects that these perfusion patterns have on ventricular loading. Despite a common bias that diastolic pressure augmentation increases CBF in the presence of ejection, we expected to find no difference under these controlled conditions. Our assumption was that local coronary autoregulation would be the primary determinant of CBF and that the modes of perfusion, without their work-altering effects, would not affect CBF or LV performance differently. End-diastolic pressures were not different between the modes of perfusion in any of the hemodynamic states. Since systolic function and CBF did vary significantly, we suspect that the lack of apparent diastolic dysfunction is explained by the acute and brief changes in perfusion patterns and pressures. An evaluation of the long-term effects of different modes of perfusion should be undertaken in the future. We assume that the demonstrated differences in CBF and L V performance were the consequence of altering the temporal interaction of isometric rhythmic contractions and the mode of perfusion, since mean perfusion pressure, baseline preload, and heart rate were controlled at constant values. In our experiments counterpulsation resulted in greater CBF and LVDP than did the other modes of perfusion. With counterpulsation, CBF was significantly greater during all conditions except the physiologic conditions of normal LV end-diastolic pressure and normal coronary perfusion pressure. LVDP was significantly higher in all conditions except those most physiologically perturbed by a high LV end-diastolic pressure and a critically low mean coronary perfusion pressure. The obvious explanation for better CBF during counterpulsation than during non pulsatile or copulsatile flow is the increased coronary perfusion pressure provided during diastole when coronary resistance is lower. In copulsation, the perfusion pressure falls during diastole when the coronary resistance is low (although a positive gradient from aorta to coronary sinus is always maintained), and nonpulsatile flow does not have as high a diastolic-peak pressure gradient as does counterpulsation. Our CBF values, with those of nonpulsatile flow being intermediate between copulsation and counterpulsation, would seem to support the explanation that diastolic pressure gradient is the primary determinant of CBF under these conditions. In all conditions the DP appropriately parallels the CBF. Similarly, although tested only during conditions of high L Vend-diastolic pressure and critically low coronary perfusion pressure, the myocardial oxygen consumption paralleled the DP for each mode of perfusion. Under these critical hemodynamic conditions, the ability to increase
5 6 8 Bellotto et al.
myocardial oxygen consumption with counterpulsatile perfusion, although not significant statistically, suggests that oxygen consumption is limited by the significant difference in CBF that occurs with copulsatile perfusion. These data are supported by the mean lactate extraction determination, which was significantly greater during counterpulsation than during copulsation, again suggesting less-than-optimal CBF under these conditions with copulsatile perfusion. On the basis of these data, if the goal is LV myocardial recovery, counterpulsatile synchronization of CBF is beneficial in terms of CBF and DP. In the intact circulation, it is possible that the ventricular unloading provided by an LV AD will affect the determinants of LV work such that satisfactory CBF can be obtained independent of the mode of perfusion. In the presence of less than perfect unloading (e.g., when the LV cannula is partially occluded) or in the presence of low perfusion pressures (e.g., in states oflow peripheral resistance), it is clear that CBF can be limited by the mode of aortic perfusion under a variety of conditions in which determinants of LV work are held constant. Data organization and analysis were performed using the data analysis facilities of the Beth Israel Hospital Clinical Research Center. We thank Bernard J. Ransil, MD, from the Charles A. Dana Research Institute and the HarvardThorndike Laboratory of Beth Israel Hospital, Department of Medicine, Boston, for his expert statistical advice. REFERENCES I. Gaasch WH, Bing OHL, Pine MB, et al. Myocardial contracture during prolonged ischemic arrest and reperfusion. Am J PhysioI1978;235:H619. 2. Apstein CS, Puchaer E, Brachfeld N. Improved automated lactate determination. Ann Biochem 1970;38:20. 3. Ogata T, Nonoyama lA, Takeda J, Sasaki H. A comparative study on the effectiveness of pulsatile and non-pulsatile blood flow in extracorporeal circulation. Nippon Geka Hokan 1960;29:59. 4. Wilkens H, Regelson W, Hoffmeister FS. The physiologic importance of pulsatile blood flow. N Engl J Med 1962; 267:443. 5. Rainer WG. Pulsatile versus nonpulsatile pumping. In: Mechanical devices to assist the failing heart. Washington, D.C.: National Academy of Sciences, National Research Council, 1966:59-64. 6. Trinkle JK, Helton NE, Wood RE, Bryant LR. Metabolic comparison of a new pulsatile pump and a roller pump for
The Journal of Thoracic and Cardiovascular Surgery
cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1969;58:562. 7. Salerno TA, Shizgal HM, Dobell ARC. Blood flow with roller versus pulsatile pump during cardiopulmonary bypass in hypertrophied hearts. Surg Forum 1974;25:141. 8. Pappas G, Winter SD, Kopriva CJ, Steel PP. Improvements of myocardial and other vital organ function and metabolism with simple method of pulsatile flow (IABP) during clinical cardiopulmonary bypass. Surgery 1975; 77:34. 9. Maddux G, Pappas G, Jenkins M, et al. Comparison of pulsatile and nonpulsatile flow during cardiopulmonary bypass on left ventricular ejection fraction early after aortocoronary surgery. Am J Cardiol 1976;37:1000. 10. Habal SM, Weiss MB, Spotnitz HM, et al. Effects ofpulsatile and nonpulsatile coronary perfusion on performance of the canine left ventricle. J THoRAc CARDIOVASC SURG 1976;72:742. II. Dunn J, Peterson A, Kirsh MM. Effects of pulsatile perfusion upon left ventricular function. J Surg Res 1978;25:211. 12. Levine FH, Grotte GJ, Fallon JT, et al. Effects of pulsatile and nonpulsatile reperfusion on the post ischemic myocardium [Abstract]. Am J Cardiol 1980;45:394. 13. Wakabayashi A, Kibo T, Gilman P, et al. Pulsatile pressure-regulated coronary perfusion during ventricular fibrillation. Arch Surg 1981;105:36. 14. Levine FH, Phillips HR, Carter JE, et al. The effect of pulsatile perfusion on preservation of left ventricular function after aortocoronary bypass grafting. Circulation 1981; 64:40. 15. Gill CC, Wechsler AS, Newman GE, Oldham HN Jr. Augmentation and redistribution of myocardial blood flow during acute ischemia by intraaortic balloon pumping. Ann Thorac Surg 1973;16:445. 16. Powell WJ Jr, Daggett WM, Magro AE, et al. Effects of intra-aortic balloon counterpulsation on cardiac performance, oxygen consumption, and coronary blood flow in dogs. Circ Res 1970;26:753. 17. Mueller H, Ayres SM, Conklin EF, et al. The effects of intra-aortic counterpulsation on cardiac performance and metabolism in shock associated with acute myocardial infarction. J Clin Invest 1971;50:1885. 18. Scheidt S, Wilner G, Mueller H, et al. Intra-aortic balloon counterpulsation in cardiogenic shock. N Engl J Moo 1973;288:979. 19. Whittle JL, Feldman RL, Pepine CJ, et al. Effects of intraaortic balloon pumping on regional and total coronary flow in patients with coronary disease. Am J Cardiol 1980; 45:395. 20. Downey JM, Kirk ES. Distribution of the coronary blood flow across the canine heart wall during systole. Circ Res 1974;34:251.
Fabio Bellotto, MD, Robert G. Johnson, MD, Jun Watanabe, MD, PhD, Marc J. Levine, MD, Alvin Franklin, MS, and Ronald M. Weintraub, MD, Boston, Mass.
Left ventricular (LV) mechanical assist devices (LVADs) are most commonly used clinically as a means of systemic circulatory support while awaiting cardiac transplantation. Under these circumstances myocardial recovery is not an issue, and factors associated with myocardial energetics (cardiac work and coronary perfusion) are irrelevant. As assist device technology improves, however, they are likely to be used to promote recovery of myocardium injured by ischemia or inflammation. In such situations the balance of cardiac work and coronary From the Division of Cardiothoracic Surgery and the Division of Cardiovascular Medicine, Beth Israel Hospital, Harvard Medical School, and the Charles A. Dana Research Center, Boston, Mass. Received for publication Oct. 3, 1990. Accepted for publication July 25, 1991. Address for reprints: Robert G. Johnson, MD, Division of Cardiothoracic Surgery, 330 Brookline Ave., Boston, MA 02215.
12/1/32837
blood flow (CBF) will be critically important. This issue may be even more important in the long-term use of these devices. Under normal conditions cardiac systole provides kinetic energy to blood and, for most organs, peak blood flow occurs during that phase of the cardiac cycle. The heart itself, however, receives the bulk of its total blood flow during diastole, as the blood "runs off" into the coronary circulation during myocardial relaxation. The use of a circulatory assist device may alter the relationship of cardiac mechanical cycles and aortic flow in a number of ways. If we assume that the device provides complete volume unloading of the left ventricle, no ejection will occur and aortic flow will be entirely determined by the device. The flow may be nonpulsatile or pulsatile. If pulsatile, the flow may be synchronized to the cardiac electricalmechanical cycles, or it may be asynchronous. Synchronous aortic pulsatile flow may be copulsatile, so that 561
56 2
The Journal of Thoracic and Cardiovascular Surgery
Bel/otto et al.
FLEXIBLE DIAPHRAGM
AIR
__
TO
DRIVE
CONSOLE
-::===::::::<~-p-RiA-AORTlC PRESSURE
::..
SALINE
Fig. 1. Schematic presentation of the model.
device systole occurs at the time of cardiac systole (even though an unloaded left ventricle may not be opening the aortic valve), or it may be counterpulsatile, so that device systole is coupled to cardiac diastole when aortic pressure would normally be decreasing. Although data from intraaortic balloon pump use demonstrate the benefits of diastolic pulsation, this may not be applicable to a situation in which the left ventricle is completely unloaded by an assist device and does not need to eject to sustain adequate systemic blood flow. It might be suggested that with an LV AD unloading the ventricle by copulsation is more physiologic than by counterpulsation since the pressure pulse comes during systole (nonejecting contraction) and coronary runoff will occur concomitantly with systemic pressure decay during diastole. If the LVAD is to be used in situations in which myocardial survival or recovery is desired, it would be beneficial to optimize the heart's blood supply and work demand ratio. Because significant engineering and clinical efforts are required to synchronize pulsatile flow, it is important that we determine whether any benefit is to be derived by the use of either copulsation or counterpulsation. These experiments were performed in an isolated, iso-
volumic-contracting canine heart preparation. This nonejecting left heart model eliminated the effect of ventricular unloading through ejection, while permitting independent control of mean coronary perfusion pressure. The three patterns of aortic flow were studied under four conditions of LV preload and coronary perfusion pressures. One of the four conditions mimicked the normal physiologic state of normal preload and coronary perfusion pressure, and the other three represented deranged physiology. We hypothesized that there would be no significant difference in CBF or cardiac performance among three clinically applicable patterns of perfusion: nonpulsatile, pulsatile-synchronous, and pulsatile-asynchronous under these four physiologic conditions.
Materials and methods Isolated perfused hearts from six mongrel dogs weighing between 19.9and 27.1 kg were prepared, with some modifications, according to a previously described model. I In each experiment a support dog was anesthetized with a-chloralose and urethane and the lungs were mechanically ventilated.Catheters wereplacedin the femoralartery and vein, and the animal was given intravenousheparin, 500 U/kg. The heart of a seconddog was prepared by cannulationof the right
Volume 104 Number 3 September 1992
Mechanical assistance of left ventricle
1.0 0.5
o
563
t-----=-±-f±f±fT-,.....-'--;-,.....-'--+-'--+-'--'-..:_ _
0'
:r
E
E
Q.
o
:i o ~
o
UJ
• . "' . • . . . , . • . ~ . • . . ' lr"! " •••.••.•.,.'. '..•.•.....,.• '. '.. '. · ' '.•. '. '. ' ' ' .t>'. . ~ G"J'I_/L Y· •..•.iUtJjJ .......................' i
:
..Hi··.····.····.····. .
. . . ..
.
.: . :' ---:-.<... ~ .-.; --: -;"':----::";'''':' ':"": :-. :-.: . -" ->:
'cOUNiER-iPULSATH1E·
..
.
)1
II
-
CO-PULSATILE
Fig. 2. Isolated examples of perfusion pressure modes. ECG, Electrocardiogram; CBF, coronary blood flow; PP, perfusion pressure; M, mean perfusion pressure; LVDP, left ventricular developed pressure.
ventricle and the carotid and subclavian arteries after heparinization. The heart was isolated by ligation of the superior and inferior venae cavae and descending thoracic aorta and was removed from the thorax. Coronary perfusion of the isolated heart was maintained via the carotid cannula, with blood obtained from the femoral artery of the support dog and circulated by a roller pump (Cardiovascular Instrument Corp., Wakefield, Mass.) through a bubble trap and a heat exchanger to maintain blood temperature at 37° C. A chamber, provided with a silicone rubber diaphragm dividing blood and air, was interposed between the perfusion line (carotid cannula) and a pneumatic blood pump drive console (Thermedics Inc" Woburn, Mass.), as shown in Fig. I. With the pump off, the chamber worked like a "damper," minimizing oscillations generated by the roller pump (usually smaller than 10 mm Hg) and so providing a virtually nonpulsatile perfusion pressure. Pulsations could be generated synchronously with different phases of the electrocardiogram. The heart was atrially paced at a rate of 130 beats/min. Pulsatile perfusion was achieved with a pulse pressure of 40 mm Hg, recorded in the aortic root through a different (the subclavian) cannula. Counterpulsation was achieved by starting the perfusion pressure upstroke simultaneously with the peak of the LV pressure. Copulsation was achieved by anticipating the perfusion pressure upstroke by 50 msec before the LVend-diastolic pressure. In this way the peak of pulse pressure coincided with the LV end-diastolic pressure in the counterpulsation pattern and with the peak
of LV pressure in the copulsation pattern (Fig. 2). The pulse pressure duration was set at 240 msec. Small adjustments of an in-circuit Starling resistor were necessary to maintain the mean coronary perfusion pressure (which is equal to "intraaortic pressure," as shown in Fig. I) constant in all three of the perfusion patterns. An in-line Doppler flowprobe was used to measure mean total CBF as the coronary return was drained from the isolated right side of the heart. CBF was expressed in milliliters of flow per minute per gram ventricular weight, and the flow probe was calibrated over the range of flows after each experiment. Coronary venous blood was drained into the femoral vein of the support dog. A drainage tube was inserted through a stab wound in the apex of the left ventricle to collect thebesian venous flow. A silicone rubber mount balloon was placed in the left ventricle through the mitral anulus and was secured with sutures and connected to a Statham P23Dp pressure transducer. The isolated heart was then submerged in a warm bath and allowed to equilibrate for 30 minutes at a coronary nonpulsatile mean perfusion pressure of 100 mm Hg at a heart rate of 130 beats/ min. Baseline mechanical and biochemical data were then obtained; LV end-diastolic pressure was measured and LV developed pressure (DP) was determined (DP = LV systolic pressure minus LV end-diastolic pressure) as indices of diastolic compliance and systolic performance, respectively. During the course of a study, changes in LV end-diastolic pressure were
564
The Journal of Thoracic and Cardiovascular Surgery
Bel/otto et at.
Table I
NoP/NN CnP/NN CoP/NN NoP/NL CnP/NL CoP/NL NoP/HN CnP/HN CoP/HN NoP/HL CnP/HL CoP/HL
BV
CPP
EDP
SP
DP
CBF
(ml)
(mmHg)
(mmHg)
(mmHg)
(mm Hg)
(mltminfgm)
92±1O 91 ± 10 93 ± 10 49 ± 4 49 ± 4 49 ± 3 94 ± 10 94 ± 10 94 ± 13 46 ± 7 47 ± 7 47 ± 7
4±2 4±2 4±1 4±2 3± 2 4 ± 2 II ± 3 12 ± 4 II ± 3 12 ± 6 10 ± 5 II ± 4
103 ± 114 ± 101 ± 88 ± 97 ± 87 ± 144 ± 150 ± 142 ± 109 ± 116 ± 97 ±
99 ± 110 ± 97 ± 83 ± 94 ± 83 ± 133 ± 139 ± 132 ± 97 ± 106 ± 86 ±
14 ± 14 ± 14 ± 13 ± 13 ± 13 ± 23 ± 23 ± 23 ± 24 ± 24 ± 24 ±
7 7 7 7 7 7 II II II 10 10 10
28 31 24 23 27 21 40 44 40 27 34 24
28 32 25 23 27 23 42 44 44 32 37 26
1.02 ± 1.28 ± 1.23 ± 0.65 ± 0.80 ± 0.64 ± 1.34 ± 1.49 ± 1.33 ± 0.56 ± 0.71 ± 0.55 ±
0.7 0.7 0.7 0.3 0.5 0.3 0.6 0.7 0.6 0.3 0.3 0.4
CBFr* 100 ± 107 ± 101 ± 60 ± 71 ± 60 ± 135 ± 148 ± 118 ± 57 ± 69 ± 55 ±
0 7 4 26 26 26 33 30 24 28 27 22
BV, Balloon volume; CPP, coronary perfusion pressure; EDP, left ventricular end-diastolic pressure; SP, left ventricular systolic pressure; DP, left ventricular developed pressure; CBF, coronary blood flow ml/rnin/gm of myocardium; CBF r: CBF ratio; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, copulsatile perfusion pressure; NN, normal left ventricular end-diastolic pressure, normal pulmonary pressure; NL, normal left ventricular end-diastolic pressure, low pulmonary pressure; HN, high left ventricular end-diastolic pressure, normal pulmonary pressure; HL, high left ventricular end-diastolic pressure, low pulmonary pressure. *Ratio = CBF at the particular experimental condition/CBF at baseline (NoP and NN).
taken to represent changes in diastolic stiffness, and changes in DP were interpreted as variations in contractility, Each of six isolated hearts was studied in a protocol consisting of a sequence of four hemodynamic conditions simulating four different pathophysiologic situations. In the first, during nonpulsatile perfusion, we inflated the LV balloon to reach an LV systolic pressure of 100 mm Hg while perfusing the aorta with a mean coronary perfusion pressure of 100 mm Hg. Using a nonpulsatile pattern as a control, we alternated the sequence of counterpulsation and copulsation patterns; DP, LV end-diastolic pressure, and CBF were measured after at least 3 minutes of stabilization during each perfusion pattern. In the second series of tests, we abruptly decreased the mean coronary perfusion pressure to 50 mm Hg and then repeated the preceding patterns and measurements. In the third and fourth series of hemodynamic conditions, we increased the ventricular balloon volume to reach an LV systolic pressure of 150 mm Hg and then repeated the whole procedure as in the first and second series, with both normal and low perfusion pressures. In this way, we were able to analyze the impact of nonpulsatile, counterpulsatile, and copulsatile perfusion under four different conditions produced by a combination of two preloads (normal and high) at two perfusion pressures (normal and critically low). Although the peak systolic and diastolic pressures varied with the mode of perfusion, the mean aortic pressure (in this model, coronary perfusion pressure) was held constant at either 100 mm Hg or 50 mm Hg according to the preceding protocol. Myocardial metabolic parameters were tested and compared under what we considered the most physiologically stressful conditions during the fourth part of the experiment, that is, with the highest LV end-diastolic pressure and the lowest coronary perfusion pressure, Blood was sampled from the arterial and coronary venous lines for oxygen and carbon dioxide tension measurements, as well as for determination of lactate production or extraction or both. The hemoglobin, pH, oxygen pressure, carbon dioxide pressure, oxygen content, and saturation were determined with a blood gas analyzer (Instrumentational Laboratory. Lexington, Mass.). Arterial-venous oxygen content difference was calculated; myocardial oxygen consumption was expressed as the
product of CBF and arterial-venous oxygen content difference indexed per 100 gm of ventricular weight. Transmyocardial lactate production/extraction was measured by the technique described by Apstein, Puchaer, and Brachfeld.? All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the N atinal Institutes of Health (NIH Publication No. 80-23, revised 1978), Statistics. Values are expressed as mean ± standard deviation and were compared for statistically significant differences by repeated-measures analysis of variance and appropriate multiple-sample comparison tests (blocked Neuman-Keuls test and Friedman rank sum test). Group NoP (nonpulsatile mode) served as the basis for all intergroup comparisons. If no significant differences were found, the mean values of group CnP (counterpulsation) were compared with those of group CoP (copulsation). Results As shown in Table I, the balloon volume and mean perfusion pressure were maintained at constant values within each test series, In all the conditions tested, we observed higher CBF and DP during counterpulsation. These differences reached statistical significance only when the LV end-diastolic pressure was high and the coronary perfusion pressure was low. Higher CBF and DP were observed to occur simultaneously and almost immediately with alterations in the mode of perfusion (Fig. 3). Note that the rapid paper speed tracing shows the pulsatile coronary perfusion pressure tracings and the slower speed demonstrates the mean (constant for all modes) coronary perfusion pressure, Total CBF. With the exception of the first condition
Volume 104 Number 3 September 1992
Mechanical assistance of left ventricle
START Co-P
No-P WVUmlY'lIIIWlI)
565
• ~~ ~-.,.J---'T 'IIUWU'•••UIII"W.. '.IMU'. UMlMNBVllvr."IU'MMU"""'I"IlII1U"'llJ/'mM¥l"! I
CooP IIU'Y' l'IA'U'WlUUP'W"IlIWUM'll!''¥'WJMMM9111MN'MMU''V!fll'Mn'UI
START Cn-P
_illlitllilllINihIiil_
Cn-P Fig. 3. Transitions of perfusion pressure modes. ECG, Electrocardiogram; LVDP, left ventricular developed pressure; CBF, mean coronary blood flow; PP, perfusion pressure; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, copulsatiie PP. normal LVEDP. normal PP
normal LVEDP. low PP
2.0
2.0
~ 1.5
~ 1.5
e
'E E
c
:g
1.0
E
0.5 0.0
No-P
Cn-P
1.0 0.5 0.0
COoP
I
~ 1.5
C
'E E
~
-'E C
E
0.5
No-P
Cn-P
COoP
2.0
1.0
0.0
Cn-P
high LVEDP. low PP
high LVEDP. normal PP 2.0
No-P
COop
1.5 1.0 0.5 0.0
No-P
Cn-P
COoP
Fig. 4. Effect of different perfusion modes on mean coronary blood flow. LVEDP, Left ventricular end-diastolic pressure; PP, perfusion pressure; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, co-pulsatile perfusion pressure. Comparisons were made: NaP versus CnP and NaP versus CoP.
(normal LV end-diastolic pressure, normal coronary perfusion pressure), the CBF in all the remaining subgroups wassignificantly higher in the counterpulsatile than in the nonpulsatile mode (Fig. 4). The greatest relative CBF
difference between the nonpulsatile and counterpulsatile patterns and (21%) was recorded during conditions we considered most physiologically stressful (high LV enddiastolic pressure, low coronary perfusion pressure). The
The Journal of Thoracic and Cardiovascular Surgery
5 6 6 Bellotto et al.
normal LVEDP, normal PP
normal LVEDP, low PP
160
160
:120 E E 80
120
Cl
:I:
E E 80
40
o
40 No-P
Cn-P
0
COOP
No-P
bigb LVEDP, normal PP
160
:120 E E 80
:120
40
40
E
E
No-P
Cn-P
COop
bigb LVEDP, low PP
160
o
Cn-P
80
o
COOp
No-P
Cn-P
COOp
Fig. 5. Effectof different perfusion modeson left ventriculardeveloped pressure. LVEDP, Left ventricularend-diastolicpressure; PP, perfusion pressure; NaP, nonpulsatile perfusion pressure; CnP, counterpulsatile perfusion pressure; CoP, copulsatileperfusion pressure. Comparisons were made: NaP versus CnP and NaP versus CoP. LACTATE EXTRACTION 5.0
4.0 !?!
.s ~
3.0
E 1.0
0.0 No-P
0.0 Cn-P
Co-P
No-P
Cn-P
Co·p
Fig. 6. Effect of different perfusion patterns on myocardial metabolism during conditions of high left ventricular end-diastolic pressureand lowmean coronary perfusion pressure. NaP, Nonpulsatileperfusion pressure; CnP, counterpulsatileperfusionpressure;CoP, copulsatile perfusion pressure.Statistical significance: NaP versusCnP and CnP versus CoP. CBF during copulsation was always lower than in the non pulsatile pattern, but the difference did not reach statistical significance. LV developed pressure. During counterpulsation, the L V developed pressure was always higher than during the nonpulsatile perfusion (Fig. 5). The greatest statistical difference in developed pressure between perfusion modes was noted during the most physiologic condition of a normal LV end-diastolic pressure and normal coronary perfusion pressure (nonpulsatile versus copulsatile: p < 0.005). In contrast, the intermode differences in the DP seen during the stressful situation of a high LV enddiastolic pressure and low coronary perfusion pressure did not reach statistical significance. We could not demonstrate any significant difference in the DP achieved with
copulsatile versus nonpulsatile modes of perfusion under any of the hemodynamic conditions studied. LV end-diastolic pressure. The LV end-diastolic pressure did not show any significant variation among the three perfusion patterns in the four conditions examined, whether with high or low balloon volume (see Table I). Myocardial metabolism. In the only condition in which we evaluated myocardial metabolism (high LV end-diastolic pressure and low coronary perfusion pressure), counterpulsation was associated with a rise in both the myocardial oxygen consumption and lactate extraction when compared with nonpulsation and copulsation. Only the difference in lactate extraction found between the counterpulsation and copulsation modes reached statistical significance (Fig. 6).
Volume 104 Number 3 September 1992
Discussion This study was not intended to add to the extensive data on the effect of pulsatile versus nonpulsatile perfusion of the myocardium during cardiopulmonary bypass or in the presence of ventricular fibrillation.l'!" Instead, we were interested in looking at three commonly used modes of perfusion in a working heart model as they might pertain to intermediate or long-term assistance of the failing left ventricle. When dealing with a working heart, the benefit of synchronized, pulsatile CBF is determined by the desirability of providing an augmented aortic pressure during the period of myocardial relaxation (coun terpulsation) versus the more physiologic condition of an increased aortic pressure immediately before myocardial relaxation (copulsation) versus the maintenance of constant aortic pressure (permitting CBF to be regulated independent of pressure pulses [nonpulsatile]). The diastolic augmentation of counterpulsation is widely assumed to increase CBF in the situation in which the native ventricle is ejecting, but this is not well documented. Counterpulsation provided by intraaortic balloon pumping has been shown to improve the subendocardial flow only in the presence of either critical coronary stenosisl ' or severe hypotension,16-19 when autoregulation is impaired. The intraaortic balloon, however, unlike the LV AD, can be effective only in supporting an ejecting heart (under these conditions there are two positive-pressure waves). Copulsatile perfusion has not been studied previously, because its use in tandem with LV ejection would seem to result in an unfavorable increase in the work of competitive ejection. Additionally, Downey and Kirk-" have suggested that any process that causes the heart to rely on delivery of a greater percentage of its CBF during systole has the potential for resulting in decreased CBF to the subendocardium. Mechanical assistance of the failing left ventricle, however, would theoretically support the entire systemic circulation while unloading the ventricle and precluding ejection. Under these conditions copulsation is an option that rna y closely mimic the ph ysiologic decrease in aortic pressure that occurs in association with the decrease in the wall tension of diastole. In our model the intracoronary blood pressure during diastole in all three modes of perfusion was always higher than the LV enddiastolic pressure. In evaluating the impact on beating hearts of pulsatile pressure augmentation induced by an LV AD, two different factors have to be taken into consideration: (I) pressure and volume unloading of the left ventricle and (2) possible "intrinsic" effects of pressure waves on coronary resistances and flow. Our isolated dog heart model controls the former, and it is therefore ideal for separating the
Mechanical assistance of left ventricle
567
"intrinsic" effect of different perfusion patterns on myocardial performance and CBF from the effects that these perfusion patterns have on ventricular loading. Despite a common bias that diastolic pressure augmentation increases CBF in the presence of ejection, we expected to find no difference under these controlled conditions. Our assumption was that local coronary autoregulation would be the primary determinant of CBF and that the modes of perfusion, without their work-altering effects, would not affect CBF or LV performance differently. End-diastolic pressures were not different between the modes of perfusion in any of the hemodynamic states. Since systolic function and CBF did vary significantly, we suspect that the lack of apparent diastolic dysfunction is explained by the acute and brief changes in perfusion patterns and pressures. An evaluation of the long-term effects of different modes of perfusion should be undertaken in the future. We assume that the demonstrated differences in CBF and L V performance were the consequence of altering the temporal interaction of isometric rhythmic contractions and the mode of perfusion, since mean perfusion pressure, baseline preload, and heart rate were controlled at constant values. In our experiments counterpulsation resulted in greater CBF and LVDP than did the other modes of perfusion. With counterpulsation, CBF was significantly greater during all conditions except the physiologic conditions of normal LV end-diastolic pressure and normal coronary perfusion pressure. LVDP was significantly higher in all conditions except those most physiologically perturbed by a high LV end-diastolic pressure and a critically low mean coronary perfusion pressure. The obvious explanation for better CBF during counterpulsation than during non pulsatile or copulsatile flow is the increased coronary perfusion pressure provided during diastole when coronary resistance is lower. In copulsation, the perfusion pressure falls during diastole when the coronary resistance is low (although a positive gradient from aorta to coronary sinus is always maintained), and nonpulsatile flow does not have as high a diastolic-peak pressure gradient as does counterpulsation. Our CBF values, with those of nonpulsatile flow being intermediate between copulsation and counterpulsation, would seem to support the explanation that diastolic pressure gradient is the primary determinant of CBF under these conditions. In all conditions the DP appropriately parallels the CBF. Similarly, although tested only during conditions of high L Vend-diastolic pressure and critically low coronary perfusion pressure, the myocardial oxygen consumption paralleled the DP for each mode of perfusion. Under these critical hemodynamic conditions, the ability to increase
5 6 8 Bellotto et al.
myocardial oxygen consumption with counterpulsatile perfusion, although not significant statistically, suggests that oxygen consumption is limited by the significant difference in CBF that occurs with copulsatile perfusion. These data are supported by the mean lactate extraction determination, which was significantly greater during counterpulsation than during copulsation, again suggesting less-than-optimal CBF under these conditions with copulsatile perfusion. On the basis of these data, if the goal is LV myocardial recovery, counterpulsatile synchronization of CBF is beneficial in terms of CBF and DP. In the intact circulation, it is possible that the ventricular unloading provided by an LV AD will affect the determinants of LV work such that satisfactory CBF can be obtained independent of the mode of perfusion. In the presence of less than perfect unloading (e.g., when the LV cannula is partially occluded) or in the presence of low perfusion pressures (e.g., in states oflow peripheral resistance), it is clear that CBF can be limited by the mode of aortic perfusion under a variety of conditions in which determinants of LV work are held constant. Data organization and analysis were performed using the data analysis facilities of the Beth Israel Hospital Clinical Research Center. We thank Bernard J. Ransil, MD, from the Charles A. Dana Research Institute and the HarvardThorndike Laboratory of Beth Israel Hospital, Department of Medicine, Boston, for his expert statistical advice. REFERENCES I. Gaasch WH, Bing OHL, Pine MB, et al. Myocardial contracture during prolonged ischemic arrest and reperfusion. Am J PhysioI1978;235:H619. 2. Apstein CS, Puchaer E, Brachfeld N. Improved automated lactate determination. Ann Biochem 1970;38:20. 3. Ogata T, Nonoyama lA, Takeda J, Sasaki H. A comparative study on the effectiveness of pulsatile and non-pulsatile blood flow in extracorporeal circulation. Nippon Geka Hokan 1960;29:59. 4. Wilkens H, Regelson W, Hoffmeister FS. The physiologic importance of pulsatile blood flow. N Engl J Med 1962; 267:443. 5. Rainer WG. Pulsatile versus nonpulsatile pumping. In: Mechanical devices to assist the failing heart. Washington, D.C.: National Academy of Sciences, National Research Council, 1966:59-64. 6. Trinkle JK, Helton NE, Wood RE, Bryant LR. Metabolic comparison of a new pulsatile pump and a roller pump for
The Journal of Thoracic and Cardiovascular Surgery
cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1969;58:562. 7. Salerno TA, Shizgal HM, Dobell ARC. Blood flow with roller versus pulsatile pump during cardiopulmonary bypass in hypertrophied hearts. Surg Forum 1974;25:141. 8. Pappas G, Winter SD, Kopriva CJ, Steel PP. Improvements of myocardial and other vital organ function and metabolism with simple method of pulsatile flow (IABP) during clinical cardiopulmonary bypass. Surgery 1975; 77:34. 9. Maddux G, Pappas G, Jenkins M, et al. Comparison of pulsatile and nonpulsatile flow during cardiopulmonary bypass on left ventricular ejection fraction early after aortocoronary surgery. Am J Cardiol 1976;37:1000. 10. Habal SM, Weiss MB, Spotnitz HM, et al. Effects ofpulsatile and nonpulsatile coronary perfusion on performance of the canine left ventricle. J THoRAc CARDIOVASC SURG 1976;72:742. II. Dunn J, Peterson A, Kirsh MM. Effects of pulsatile perfusion upon left ventricular function. J Surg Res 1978;25:211. 12. Levine FH, Grotte GJ, Fallon JT, et al. Effects of pulsatile and nonpulsatile reperfusion on the post ischemic myocardium [Abstract]. Am J Cardiol 1980;45:394. 13. Wakabayashi A, Kibo T, Gilman P, et al. Pulsatile pressure-regulated coronary perfusion during ventricular fibrillation. Arch Surg 1981;105:36. 14. Levine FH, Phillips HR, Carter JE, et al. The effect of pulsatile perfusion on preservation of left ventricular function after aortocoronary bypass grafting. Circulation 1981; 64:40. 15. Gill CC, Wechsler AS, Newman GE, Oldham HN Jr. Augmentation and redistribution of myocardial blood flow during acute ischemia by intraaortic balloon pumping. Ann Thorac Surg 1973;16:445. 16. Powell WJ Jr, Daggett WM, Magro AE, et al. Effects of intra-aortic balloon counterpulsation on cardiac performance, oxygen consumption, and coronary blood flow in dogs. Circ Res 1970;26:753. 17. Mueller H, Ayres SM, Conklin EF, et al. The effects of intra-aortic counterpulsation on cardiac performance and metabolism in shock associated with acute myocardial infarction. J Clin Invest 1971;50:1885. 18. Scheidt S, Wilner G, Mueller H, et al. Intra-aortic balloon counterpulsation in cardiogenic shock. N Engl J Moo 1973;288:979. 19. Whittle JL, Feldman RL, Pepine CJ, et al. Effects of intraaortic balloon pumping on regional and total coronary flow in patients with coronary disease. Am J Cardiol 1980; 45:395. 20. Downey JM, Kirk ES. Distribution of the coronary blood flow across the canine heart wall during systole. Circ Res 1974;34:251.