| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2000;70:614-620
© 2000 The Society of Thoracic Surgeons
a Departments of Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA
Address reprint requests to Dr Torchiana, Surgical Cardiovascular Unit, Massachusetts General Hospital, BUL-119, 55 Fruit St, Boston, MA 02114
e-mail: torchiana.david{at}mgh.harvard.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentReferences |
|---|
Methods. Canine hearts were arrested by antegrade cold and warm blood cardioplegia in random order. Cardioplegic distribution was measured using radiolabeled microspheres before and just after induction of each period of arrest.
Results. With cold cardioplegia, perfusion of left ventricular subepicardial and midwall regions decreased. Subendocardial to subepicardial perfusion ratios increased significantly in the left ventricle as a whole, the anterior and posterior regions of the left ventricular free wall, and the interventricular septum. With warm arrest, transmural flow distribution was not significantly altered from preceding prearrest values. At constant coronary flow, coronary perfusion pressure was initially similar after induction of arrest at both temperatures, but it rose subsequently during warm cardioplegia.
Conclusions. The data suggest that during normothermic arrest, vasomotor tone regulates cardioplegic distribution, and hyperkalemic vasoconstriction is of slow onset. In the absence of beating and with vasomotion inhibited by hypothermia, cardioplegic distribution during cold arrest appears to be primarily dependent on vascular anatomy. There was no evidence of subendocardial underperfusion during cooling contracture.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Cooling contracture generates substantial force and can be sustained for 2 or 3 minutes (see Figure 4 in Cannon and associates [4]), a period comparable to the initial, and sometimes only, infusion of cold cardioplegia in clinical practice. Subendocardial blood flow is impeded during systolic contraction [7–10]. Optimal cardioplegic protection depends on adequate distribution of perfusion to all myocardial regions. We therefore asked whether cooling contracture might result in maldistribution of cardioplegic solution and particularly in underperfusion of the subendocardium.
Studies were conducted in canine hearts with radiolabeled microspheres to evaluate regional perfusion. We measured subendocardial to subepicardial perfusion (I/O) ratios just after induction of arrest by antegrade cold blood cardioplegia (CBC) and compared them with I/O ratios during right heart bypass (RHBP) preceding arrest and with those just after induction of arrest by antegrade warm blood cardioplegia (WBC). We used a modified high-potassium Fremes’s blood cardioplegic solution, similar to that used clinically [11], which is known to elicit cooling contracture in the canine heart [4].
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Seven mongrel dogs of either sex weighing 15 to 30 kg were anesthetized intravenously with a warm solution of 
-chloralose (75 to 150 mg/kg) and urethane (0.75 to 1.5 g/kg). After endotracheal intubation, ventilation was established with oxygen-enriched room air. A median sternotomy was performed, the heart was supported in a pericardial cradle, and the azygos vein was ligated. Aortic pressure was continuously monitored with a catheter placed in the aortic arch through the right internal mammary artery. Right heart bypass was established as previously described [12] using a bubble oxygenator and perfusion tubing primed with the blood of 1 or 2 donor dogs anesthetized with methohexital sodium (12.5 mg/kg). Systemic venous blood from the cannulated venae cavae drained by gravity into the oxygenator. The right heart drained separately, allowing timed volumetric collection of coronary venous blood, which was returned to the oxygenator, or of effluent cardioplegic solution, which was discarded. A calibrated roller pump controlled cardiac output by infusing warm oxygenated blood from the oxygenator into the cannulated pulmonary artery. Aortic pressure was regulated by a second calibrated roller pump that withdrew or infused warm oxygenated blood through cannulas in the left subclavian and femoral arteries. A double-lumen catheter implanted in the aortic root enabled simultaneous monitoring of aortic root pressure and infusion of cardioplegic solution. A 15-cm polyvinyl cannula (inner diameter, 0.125 cm [0.05 inch]) was placed in the left atrium for microsphere injection during RHBP. A similar cannula for withdrawing reference blood samples was introduced through the brachiocephalic trunk so that its tip lay in the aortic arch. A short wide-bore metal cannula was placed through the left ventricular (LV) apical dimple for measurement of LV pressure and to vent the left ventricle (LV) during cardioplegia. The electrocardiogram was monitored using a right ventricular epicardial lead. After the sinoatrial node was crushed, heart rate was held constant by pacing through wires sewn to the right atrium.
All pressures were measured with strain-gauge transducers (DTX; Viggo-Spectramed, Oxnard, CA) and recorded with the electrocardiogram on a multichannel polygraph (Hewlett-Packard model 8877A). Arterial carbon dioxide tension and pH were held in the normal range, and arterial oxygen tension was kept higher than 100 mm Hg [12]. Body temperature was between 36°C and 38°C except as required for cold cardioplegia. Myocardial septal temperature was measured with a needle probe (Sensortek, Clifton, NJ). Anticoagulation was maintained with heparin sodium, 6,000 units initially and 1,000 U/h added to the circulating blood. To maintain stable anesthesia for the duration of the experiments, -chloralose and urethane to a total dose of 300 mg/kg and 3 g/kg, respectively, were slowly added to the oxygenator reservoir when cardiopulmonary bypass had been established.
Protocol
Hearts underwent two periods of cardioplegic arrest induced by antegrade blood cardioplegia delivered at 4°C or 37°C in random order during which cardioplegic distribution was measured. Baseline regional myocardial blood flow (RMBF) measurements were obtained on RHBP before each period of arrest.
Baseline
Right heart bypass was established at a heart rate of 165 beats per minute and a mean aortic pressure of 85 mm Hg, and the cardiac output gradually increased until end-diastolic pressure reached 6 to 8 cm H2O. After stabilization, coronary blood flow was measured by a 1-minute timed volumetric collection, and RMBF was evaluated as will be described.
Cardioplegia
Total cardiopulmonary bypass was instituted, pacing was discontinued, and the LV was vented against a pressure of 5 cm H2O through the LV apical cannula. For CBC, systemic temperature was lowered by decreasing the temperature of the circulating blood to 28°C. Aortic pressure was decreased to 50 mm Hg, the aorta was cross-clamped, and blood from the oxygenator continuously mixed with a modified Fremes’s cardioplegic solution (4 parts blood to 1 part solution), described later, was immediately delivered at 4°C or 37°C at a constant flow rate for 6 minutes through the aortic root cannula by way of a heat exchanger and an in-line static mixer with a diameter of 0.625 cm (0.25 inch) (Cole-Parmer, Niles, IL). After arrest was confirmed by direct inspection of the heart, cardioplegic distribution was measured by microspheres as will be described, and coronary flow was measured. We noted the time from arrest (or onset of ventricular fibrillation) to initiation of the microsphere injection. Coronary perfusion pressure was continuously recorded during cardioplegia.
Reperfusion
Lidocaine hydrochloride, 50 mg, was added to the oxygenator reservoir. The aortic cross-clamp was removed to start reperfusion with normokalemic blood at 37°C, the heart was defibrillated if necessary by electrical countershock, and atrial pacing was resumed. The heart was allowed to beat while vented for 30 minutes at a mean arterial pressure of 85 mm Hg, after which the LV vent was closed, RHBP was resumed, and cardiac output gradually rose to its baseline value. When hemodynamics were stable, a second control RMBF evaluation was obtained as before at baseline cardiac output and a mean aortic pressure of 85 mm Hg. Then arrest was induced at the second temperature, and cardioplegic distribution was again evaluated. At the end of the experiment, the heart was excised, the LV was sectioned for radionuclide counting, zero pressure corrections were obtained by exposing the tips of cannulas in situ to air, and cannula positions were confirmed.
Regional myocardial blood flow
During RHBP, approximately 2 million microspheres, 10 to 12 µm in diameter, labeled with the radioactive isotopes cerium 141 or ruthenium 103 (New England Nuclear, Billerica, MA) in 2 mL of 0.9% saline solution plus 0.01% Tween 80 were placed in an injection vial [13], suspended by a vortex mixer immediately before injection, and flushed into the left atrium by the steady injection of 20 mL of warm saline solution over 20 seconds. During this infusion, the left subclavian artery catheter was clamped, with aortic pressure supported through the femoral arteries by an infusion of less than 2,000 mL/min or by withdrawal. These conditions prevent admixture of blood that does not contain microspheres from the pump oxygenator with blood supplying the coronary arteries or with the reference blood sample [12]. The reference blood sample comprising three sequential 40-second collections was withdrawn from the aortic arch into weighed vials by a Holter pump starting 5 seconds before the start of the microsphere injection.
Cardioplegic flow distribution
Approximately 200,000 microspheres labeled with tin 113 or scandium 46 in 2 mL of 0.9% saline solution plus 0.01% Tween 80 were placed in an injection vial, suspended by a vortex mixer, and flushed, while the cardioplegic solution was flowing, into the aortic root cardioplegia line through a sidearm before the in-line mixer by the steady injection of 20 mL of saline solution at 4°C or 37°C over 20 seconds. This injection did not appreciably affect coronary perfusion pressure.
Before the LV was sectioned, valvular tissue, epicardial fat, the atria and right ventricular free wall, and the right ventricular papillary muscles on the interventricular septum were discarded. The LV papillary muscles were removed for radionuclide counting. The interventricular septum was divided into anterior and posterior segments, which were sectioned into right ventricular (regarded as subepicardial), midmyocardial, and LV subendocardial layers. The LV free wall was divided into anterior, posterior, and lateral segments, which were separated into basal, equatorial, and apical portions, each sectioned into subendocardial, midmyocardial, and subepicardial layers. Radionuclides in reference blood samples and in the LV tissue samples were counted in a multichannel pulse-height analyzer–gamma spectrometer (Packard Instruments, Chicago, IL). A computer program on a Vax 11/780 computer (Digital Equipment, Maynard, MA) separated overlapping spectra of radioisotopes using the stripping method [13]. Subendocardial to subepicardial perfusion ratios were calculated for the entire LV and for anterior, lateral, posterior, and septal regions. The ratio of flow per gram in the posterior papillary muscle to flow per gram in the whole LV was also calculated. During cardioplegia when no reference samples were obtained, RMBF (100 gm LV-1) was calculated as the product of counts per minutes for the region of interest and coronary blood flow (milliliters per minute) divided by the product of counts per minute for the total LV and the weight of the region of interest. For consistency, this method was used for all experimental periods.
Blood cardioplegic solution
Before admixture with blood, the modified high-potassium Fremes’s solution [11] contained the following in millimoles per liter: K+, 100; Mg2+, 9; Na+, 5.4; Cl-, 75; HCO3-, 25; SO42-, 9; citrate, 1.8; citric acid, 0.3; and glucose, 278. After one part of this solution was mixed with four parts of canine pump blood for delivery to the heart, the solution contained approximately 18 mEq/L K+ and 0.75 mmol/L ionized Ca2+ [4].
Calculations
Coronary vascular resistance (millimeters of mercury per milliliter per minute per 100 g of LV) was calculated as the product of 100 and the mean perfusion pressure divided by the product of coronary flow and LV wet weight. As coronary venous return was drained through a wide-bore tube that has multiple side-holes and was placed across the tricuspid valve opened to atmosphere, downstream pressure was assumed to be zero.
Statistical analysis
Data are expressed as the mean ± the standard error of the mean. Data underwent repeated-measures analysis of variance with subsequent contrasts, if indicated, by paired t test. Continuous variables measured only once were compared by Student’s t test. Values of p less than 0.05 were taken to indicate significance.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Coronary flow and coronary vascular resistance on RHBP did not differ significantly between the control periods preceding CBC and WBC (Table 1). Coronary vascular resistance fell slightly during induction of cold arrest and significantly during induction of warm arrest (see Table 1); coronary flow and perfusion pressure at the time of the microsphere injection did not differ significantly between CBC and WBC. Coronary perfusion pressure remained stable during CBC. During WBC, it rose slowly from 47 ± 8 mm Hg to a peak of 119 ± 11 mm Hg during the last minute of WBC (p < 0.002), and coronary vascular resistance rose from 0.39 ± 0.03 mm Hg · mL-1 · min-1 · 100 g LV-1 to 0.98 ± 0.10 mm Hg · mL-1 · min-1 · 100 g LV-1 (p = 0.005).
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Left ventricular myocardial regional blood flow distribution depends on the interaction between coronary perfusion pressure gradients and coronary vascular geometry, which is primarily the result of coronary anatomy, the mechanical results of extravascular compressive or tethering forces, and vascular smooth muscle tone. It is unclear whether coronary anatomy itself, evaluated by regional capillary and arteriolar density, favors perfusion of subendocardium or subepicardium [14].
In the beating heart, the pressure gradients and extravascular forces controlling myocardial perfusion are time dependent. During diastole, flow is regulated primarily by vasomotor tone. Flow is decreased by compressive forces during systole [7–10]. The contributions of LV chamber pressure and myocardial stiffness to this decrement remain unclear [7, 9]. Although there is little or no forward flow in coronary epicardial arteries during systole, subepicardial flow is increased by beating compared with arrest [8]; recent data [8] suggest that subepicardial perfusion is augmented by blood expressed from the subendocardium by compression during systole.
A functional evaluation of the influence on RMBF of coronary anatomy per se can be obtained when beating is arrested and vasomotor tone removed by maximal vasodilation. Under these circumstances, regional flow is determined only by vascular anatomy, and I/O ratios rise to approximately 2 [8, 14], although variability of the ratios among small myocardial regions increases [10].
The effects of contracture on flow distribution appear to vary with its cause and the species studied. In the rabbit with normothermic barium contracture during maximal vasodilation, subendocardial flow was lower than subepicardial flow, but this flow gradient was reversed during diastolic arrest [14]. In isolated rat hearts, whether isovolumic or not, we [2, 15] found that ischemic contracture during arrest at 8°C greatly increased global coronary vascular resistance, probably mainly by extravascular compression. In isovolumic canine hearts at 37°C, subendocardial flow and I/O ratios measured using intermittent perfusion during ischemic contracture were greater than in perfused beating hearts [16]. This supports the present study in that contracture in the canine heart can be associated with increased I/O ratios.
In addition to its extrinsic mechanical effects on the coronary circulation, cardioplegic arrest is accompanied by mixed vasomotor influences. Coronary vasomotor tone, when present, is the dominant determinant of RMBF [10]. Cardioplegia stops electromechanical activity and, with the exception of cooling contracture, which is energy consuming [2, 3], thereby reduces metabolic requirements [4, 17, 18]. The consequent decrease in vasodilative metabolites, which adjust coronary flow to oxygen demand, would tend to cause vasoconstriction [18]. In addition, certain constituents of cardioplegic solutions are vasoactive. Potassium at concentrations inducing rapid arrest causes vasoconstriction [19] by increasing calcium influx in vascular smooth muscle [20]. In Fremes’s solution [11], the reduction of ionized Ca2+ by citrate [21] and the increased Mg2+ concentration [22] are vasodilative. Hypothermia paralyzes vascular smooth muscle, progressively overriding the effects of vasoconstrictive agonists as the temperature is lowered; cold reduces resting tension and K+-induced tension in human coronary artery strips [23] and reverses vasoconstriction in K+-arrested canine hearts [18, 19]. Thus, the balance of the mechanical and vasomotor influences, both chemical and thermal, varies with specific cardioplegic conditions.
Early studies of the distribution of cardioplegic flow used a higher K+ concentration [18, 24]—then customary—than in high-K+ Fremes’s solution [11]. Leicher and coworkers [24] arrested canine hearts with a blood solution at 16°C containing 30 mEq/L K+. Coronary vascular resistance increased from 0.9 mm Hg · mL-1 · min-1 · 100 g-1 during the arresting dose to 2.6 mm Hg · mL-1 · min-1 · 100 g-1 during the second dose. These values are higher than those we obtained with CBC or WBC using Fremes’s solution with a lower [K+] and added vasodilative effects (see Table 1), but the delayed rise in resistance agrees with our observations during WBC. Myocardial cooling in the study of Leicher and colleagues may have been insufficient to override the vasoconstrictive effect of the higher [K+]. Chitwood and colleagues [18], using continuous cardioplegia with a [K+] of 32 ± 9 mEq/L, found similar I/O ratios during arrest at 37°C as in the empty beating heart, in general accord with our results for WBC. Stepwise reductions in temperature to 15°C, (involving prolonged exposure to 32 mEq/L K+, did not significantly increase I/O ratios [18]), a finding in contrast to our results during CBC after only brief exposure to milder hyperkalemia.
Other reports indicate an increasing proportion of flow to the subendocardium during cold cardioplegia. Buckberg and associates [17] found an I/O ratio of 1.1 in the empty beating heart at 37°C and of 2.0 during arrest at 22°C with blood cardioplegia. Duarte and coauthors [25] also reported that the subendocardium received twice the flow of the subepicardium with cold crystalloid cardioplegia; however, the I/O ratio during tepid blood cardioplegia (28°C) was greater than for cold crystalloid. Aldea and colleagues [26] found a mean I/O ratio just above unity during normothermic cardioplegia, similar to our results (see Table 3). However, they used carbochromen to induce maximal vasodilation, and together with arrest, this might have been expected to double the I/O ratio [8, 14].
Our results were obtained during experiments designed to mimic clinical cardioplegia but are limited in certain respects. There are potential species differences. The occurrence of cooling contracture has not been documented in human hearts. Cooling contracture observed in our previous canine studies [4] using cardioplegia of the same composition as in this study was reproducible and consistent. We endeavored to time the microsphere injection so that cooling contracture would be established before the injection started and would be sustained during the period required for all injected microspheres to lodge in capillaries; this may not have been entirely achieved. Cardioplegic flow ratios might have differed from the results obtained here if pressure rather than flow had been held constant during the cardioplegia infusion. As we have discussed, there are multiple influences at play on the coronary circulation at the time of cardioplegic arrest. There may still be a compressive effect of cooling contracture on subendocardial blood flow that is too brief or insignificant for us to detect with radiolabeled microspheres. A more instantaneous measure of coronary flow distribution such as myocardial contrast echocardiography might have been superior.
In conclusion, during warm arrest, I/O ratios remained at control levels, a finding suggesting that autoregulation was conserved; the observed increase in resistance, which was delayed, is most likely attributable to hyperkalemia. Cold arrest relaxes vascular smooth muscle, overriding vasoconstrictive or compressive effects; the combination of vasomotor paralysis and cessation of beating would favor perfusion determined primarily by vascular anatomy with increased I/O ratios, as we observed. During cold contracture, any redistribution of blood from the subendocardium to the subepicardium comparable to that occurring with beating during systole would be a single event taking place only as contracture develops and not while it is sustained. Although it may be unexpected that contracture did not impede subendocardial flow, our results concur with observations in the canine heart during ischemic contracture [16], when strong vasodilative influences would also be expected. It is clear that cooling contracture does not decrease subendocardial perfusion and so impair myocardial protection. Further, because cardioplegic flow to the subepicardium during hypothermic arrest is only half of the flow to the subendocardium, adequate cardioplegic flows are essential to ensure its optimal preservation.
|
Acknowledgments |
|---|
Doctor Vine was supported by Research Fellowship 13-401-912 from the Massachusetts Affiliate of the American Heart Association.
We thank Alvin Denenberg, MS, and Cheng-Zai Lu, MS, for radionuclide and blood gas analysis and Diane Barbarisi for technical assistance.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. L. Sloots, J. Vinten-Johansen, and G. P. Dobson Warm nondepolarizing adenosine and lidocaine cardioplegia: Continuous versus intermittent delivery J. Thorac. Cardiovasc. Surg., May 1, 2007; 133(5): 1171 - 1178. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. R. Mallidi and S. E. Fremes Reply to the Editor J. Thorac. Cardiovasc. Surg., December 1, 2003; 126(6): 2111 - 2112. [Full Text] [PDF]
|
|
|
|
 |
|
 A.C.L. Graffigna, G. Nollo, C. Pederzolli, P. Ferrari, L. Widesott, and R. Antolini Continuous monitoring of myocardial acid-base status during intermittent warm blood cardioplegia Eur. J. Cardiothorac. Surg., June 1, 2002; 21(6): 995 - 1001. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
