- Email: [email protected]
The microvascular distribution of cardioplegic solution in the piglet heart
The microvascular distribution of cardioplegic solution in the piglet heart Retrograde versus antegrade delivery The uniform distribution of cardioplegic solution to all areas of the microvasculature is considered critical for myocardial protection. Despite this, little information exists regarding the ability of retrogradely infused cardioplegic solution to perfuse the microvasculature of the heart. In this report, the microvascular distribution of retrogradely delivered cardioplegic solution was studied by means of a technique to quantitatively demonstrate capillary perfusion. Duroc piglet hearts were subjected to either antegrade (n = 4) or retrograde (n = 8) perfusion fixation with 2.5 % glutaraldehyde and subsequently perfused with NTB-2 (an intracapillary marker). The results indicate that retrogradely delivered NTB-2 consistently perfused the anterior half of the intraventricular septum and the anterior and lateral free walls of the left ventricle but inconsistently perfused the posterior half of the intraventricular septum, the posterior wall of the left ventricle, and a small paraseptal region of the right ventricle. The remainder of the right ventricle was not perfused. In contradistinction, aU regions of the heart were consistently perfused by the antegrade technique. In regions of the heart in which retrograde microvascular perfusion occurred, no statistical difference was found in the quantitative degree of capillary perfusion achieved by either the antegrade or retrograde technique. These results have important implications for planning strategies of myocardial protection and suggest that further investigation concerning the microvascular distribution of retrogradely delivered cardioplegic solution in human beings is merited. (J THORAC CARDIOVASC SURG 1993;105:845-53)
Richard N. Gates, MD, Hillel Laks, MD, Davis C. Drinkwater, MD, Jeffrey Pearl, MD, Ana Maria Zaragoza, MD, Elias Kaczer, BS, and Paul Chang, BS, Los Angeles, Calif.
h e clinical application of retrograde cardioplegia is increasing and several reports attest to its efficacy.I>' Retrograde cardioplegia is particularly suited for adult coronary artery bypass operations when acute or critical obstruction of coronary arteries exists. Furthermore, many pediatric procedures necessitating aortotomy are facilitated by retrograde techniques. Despite this, little information exists as to the ability of retrogradely infused
From The Divisionof Cardiothoracic Surgery, Department of Surgery and Department of Pathology, University of California at Los Angeles, Medical Center, Los Angeles, Calif. Read at the Eighteenth Annual Meeting of The Western Thoracic Surgical Association, Kauai, Hawaii, June 24-27, 1992. Address for reprints: Hillel Laks, MD, Division of Cardiothoracic Surgery,UCLA Medical Center, CHS 62-182,10833 LeConte Ave., LosAngeles, CA 90024. Copyright
1993 by Mosby-Year Book, Inc.
0022-5223/93 $\.00 + .10
12/6/44661
cardioplegic solution to perfuse the microvasculature of the heart. Two experimental approaches have been applied in an attempt to determine the distribution of retrogradely infused cardioplegic solution. The first consists of the retrograde injection of radiolabeled microspheres that are subsequently trapped in the microvasculature.v" Although this method provides qualitative data concerning regional distribution, it provides no quantitative information regarding the perfusion of microvascular beds. The second approach involves the retrograde infusion of resins or dyes and subsequent maceration or radiography of the heart. 7,8 These methods reveal regional anatomic distribution in a qualitative manner but fail to provide quantitative information regarding microvascular beds that have not been fixed in resin or dye. The efficacy of oxygenated cardioplegia is dependent on delivery of the cardioplegic agent to all areas of the microvasculature." In view of the increasing reliance on
845
The Journal of Thoracic and Cardiovascular Surgery May 1993
8 4 6 Gates et al.
IVS
RV
RV
LV LV Epi
Epi
Mid-ventricle
Apex
Fig. 1. Diagrammatic representation of the sections of ventricles taken for histologic analysis. RV, Right ventricle; LV, left ventricle; IVS, interventricular septum; Epi, epicardium; Endo, endocardium.
retrograde cardioplegia for myocardial protection, we have developed a technique to examine the microvascular distribution of retrogradely delivered cardioplegic solution in a quantitative manner. This was achieved by capillary fixation by retrograde glutaraldehyde perfusion and retrograde perfusion of the intracapillary marker NTB-2 to demonstrate capillary patency. These experiments were performed in a piglet model under conditions simulating those used in clinical practice. They provide insight into the anatomic distribution and degree of capillary perfusion achieved by retrograde cardioplegia in clinically relevant conditions.
Materials and methods Twelve Duroe piglets were divided into three experimental groups. Group 1 (n = 4) was composed of neonatal piglets (I to 3 days old, 2 to 2.5 kg). Group 2 (n = 4) and group 3 (n = 4) consisted of infant piglets (25 to 35 days old, 5 to 7 kg). Piglets in group 1 were subjected to antegrade crystalloid cardioplegia, retrograde perfusion of glutaraldehyde fixative via the coronary sinus, and NTB-2 perfusion. Group 2 piglets were subjected to antegrade blood cardioplegia, antegrade perfusion of glutaraldehyde fixative, and NTB-2 perfusion. Group 3 piglets received antegrade blood cardioplegia, retrograde perfusion of glutaraldehyde fixative, and NTB-2 perfusion. Group I, 2, and 3 piglets were anesthetized with intramuscular ketamine (l00 rug/kg) and acepromazine (0.1 rug/kg). Tracheostomy was then performed and the animals were mechanically ventilated (Bird Mark 7, Bird Corp., Palm Springs, Calif.) with an inspired oxygen fraction of 1.0, a rate
of 15 to 20 cycles/min, and an inspiratory pressure of 25 mm Hg. Through a median sternotomy, the left hemiazygos veinand left subclavian artery were ligated. Heparin (1000 U) was administered. The innominate artery was cannulated with a modified 12-gauge catheter (Angiocath; Deseret Medical, Sandy, Utah) for group 1 and a modified 16F arterial cannula (DLP Inc., Grand Rapids, Mich) for groups 2 and 3. The central arterial catheter was attached to the cardioplegia delivery tubing, where pressure was monitored. The superior and inferior venae cavae were then ligated and an aortic crossclamp was placed on the descending aorta. The heart was vented via incisions in the pulmonary artery and left atrial appendage. Cardioplegic solution at 4° C was delivered via the catheter in the innominate artery for 2 minutes at 80 mm Hg. Once arrest was obtained, topical hypothermia was applied with iced saline. The inferior pulmonary ligaments were divided and the heart-lung block was excised. In groups I and 3, the inferior vena cava was opened to the level of the atria. The previously placed tie on the left hemiazygos vein was cut and a 12-gauge Angioeath catheter was introduced into the left hemiazygos vein and advanced into the coronary sinus. This was possible because the left hemiazygos vein drains into the coronary sinus in approximately 95% of piglets. The catheter was tied in place and flushed with heparinized saline solution to remove any air from within the cardiac venous drainage system. The coronary sinus ostium was carefully oversewn with 7-0 or 8-0 Prolene suture (Ethicon, Inc., Somerville, N.J.). In all groups glutaraldehyde fixation began 15 minutes after initial cardioplegic arrest. For both the antegrade and retrograde groups, all solutions were delivered at a pressure of 50 to 60 mm Hg. In groups I and 3, the aortic clamp was released and the heart was perfused retrogradely for 2 minutes with phosphate buffer, 0.1 mol/L at 37° C (Sigma Chemical Company, St. Louis, Mo.). This was followed by 5 minutes of retrograde perfusion with 2.5% glutaraldehyde at 37° C (Sigma), (diluted to concentration with phosphate buffer, 0.1 mol/L), Retrograde perfusion was continued for 10 minutes with phosphate buffer, 0.1 mol/L at 37° C. Finally, the heart was perfused retrogradely for 2 minutes at 37° C with autoradiographic emulsion type NTB-2 (Eastman Kodak Co., Rochester, N.Y.) that had been diluted I: 1 with phosphate buffer, 0.1 mol/L. For group 2 hearts, the protocol was the same but the perfusion times were different. Phosphate buffer was perfused antegradely for 2 minutes, 2.5% glutaraldehyde was perfused antegradely for 2 minutes, and phosphate buffer was perfused antegradely for 6 minutes. The antegrade perfusion times were shorter because flow rates were higher. Adjusting perfusion times allowed similar volumes of fixative and perfusate to be administered both antegradely and retrogradely. Finally, NTB-2 was perfused antegradely for 2 minutes. After perfusion of NTB-2, hearts were placed in a 1:10 formalin solution (Medical Chemical Corp., Santa Monica, Calif.) and stored for 24 hours at 4° to 10° C. For group 2 and 3 piglets, blood for the cardioplegic solution was obtained from adolescent (3 to 6 months, 33 to 65 kg) Duroc pigs. Blood was mixed with crystalloid cardioplegic solution in a 4:1 ratio. The crystalloid cardioplegic solution was composed of 500 ml of 5% dextrose in 0.2NS, 200 ml of 0.3 molar tromethamine (Abbott Labs, Chicago, Ill.), 50 ml of citratephosphate-dextrose solution (Abbott), and 60 mEq of potassium chloride. The blood cardioplegic solution was directed by a roller pump (model 572400 I-A, Sarns/ 3M, Ann Arbor, Mich.)
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gates et al.
847
Fig. 2. Photomicrograph of the anterior interventricular septum and anterior right ventricle. Note the presence of NTB-2 in the microvasculature of the interventricular septum and its absence from the right ventricle. (Hematoxylin and eosin stain; original magnification X16.1.) to a heated reservoir/bubble trap (BCD Plus, Shiley Inc., Irvine, Calif.). The blood cardioplegic solution was cooled to 40 C by counterflow with a heater/cooler pump (Sarns/3M) beforedelivery. For the group I piglets, 4 0 C Viaspan (DuPont Pharmaceuticals, Wilmington, Del.) cardioplegic solution was used. After 24 hours of formalin storage, group I hearts were horizontally transected at the mid ventricular level. A thin horizontal slice was taken through the left ventricle, interventricular septum, and right ventricle. Because these hearts were small, the entire horizontal specimen was able to be placed into one Tissue-Tek III unicassette (Miles Laboratories, Elkhart, Ind.) for histologic processing. For group 2 and 3 hearts, a thin horizontalslice was taken from the apex and placed into a cassette. The heart was then horizontally divided approximately I em abovethe mid ventricular level. Thin longitudinal sections were then taken from the left ventricular free wall, the anterior-mid interventricular septum, and the right ventricular free wall (Fig. I). Specimens were routinely stained with hematoxylin and eosin and the tissue was cut at a depth of 6 mm and placed onto slides. Group I slides (which contained the entire horizontal section through the left ventricle, intraventricular septum, and right ventricle) were examined under high-power (I61.2X) and lowpower (16.1X) light microscopes to determine gross regional perfusion. Group 2 and 3 hearts were examined with a high-power (312.5X) light microscope in the following regions of the heart: (I) left ventricular free wall epicardium, (2) left ventricular free wall endocardium, (3) interventricular septum at anterior-mid septum, (4) right ventricular free wall epicardium,(5) right ventricular free wall endocardium, (6) apex at mid left ventricular free wall, (7) apex at mid interventricular septum,and (8) apex at mid right ventricular free wall. Four separatemicroscopicfields of each region were examined by a light microscope attached to a videocamera (model XPC501, HitachiMedical Corp., Tokyo, Japan). The video image was trans-
fered to computer matrix form and stored on a hard drive by means of a video digitalizer (SuperVIA digitizing board, Jovian Logic Corporation, Fremont, Calif.). Each image was reviewed on a color computer screen and the number of capillaries filled or not filled with NTB-2 was hand counted with computer assistance. Hand counting was not done in a blinded fashion, but computer co-counting was of course blinded. Statistical analysis was performed between groups and regions by means of the unpaired two-tailed Student's t test. All animals used for these studies received humane care in compliance with the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH publication No. 88-23, revised 1985). In addition, the experimental protocol was reviewed and approved by the University of California at Los Angeles Chancellor's Animal Research Committee (ARC No. 91-254).
Results Gross microvascular distribution of NTB-2. Examination of group 1 and 3 hearts revealed the gross microvascular distribution of retrogradely delivered NTB-2 to be as follows: (1) consistently perfused-the anterior half of the interventricular septum and the anterior and lateral free walls of the left ventricle (8/8); (2) inconsistently perfused-the posterior half of the interventricular septum (2/4), the posterior free wall of the left ventricle (2/4), and a small area of the anterior free wall of the right ventricle next to the interventricular septum (1/4); (3) not perfusedorrarelyperfused-all oftheright ventricular free wall except a small area of the anterior free wall next to the interventricular septum (0/8). Although the microvasculature of the right ventricle was not perfused, NTB-2 could frequently be identified in
848
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gates et al.
Posterior
RV LV
Anterior
Not perfused Inconsistently perfused _
Consistently perfused
Fig. 4. Diagrammatic representation of a horizontal section throughthe mid ventricle of group 1 hearts. Thisdemonstrates regions of the heart where retrograde perfusion occurred. R V, Right ventricle; LV, left ventricle.
Fig. 3. Photomicrograph of the right ventricle. Note the presenceof NTB-2 in the largeepicardial vessel and in a few large intramyocardial veins, but not in the microvasculature. (Hematoxylin and eosin stain; original magnification X16.1.)
ventricular septum (p = 0.363). However, retrogradely delivered NTB-2 failed to perfuse the right ventricular free wall epicardium or endocardium and the apex at the mid right ventricular free wall (p < 0.05). Antegradely delivered NTB-2 successfully perfused these regions. Discussion
large epicardial and intramyocardial veins, as well as in the cavity of the right ventricle (Figs. 2, 3, and 4). In group 2 hearts, all eight regions of the ventricle sampled were consistently perfused by antegradely delivered NTB-2 (4/4). Quantitative microvascular perfusion of NTB-2. Examination of groups 2 and 3 allowed for the determination of the percentage of capillaries perfused (NTB-2 filled/NTB-2 filled plus unfilled) (Figs. 5 and 6). An average of 218 capillaries per region of the heart were counted. Sections cut transversely or near transversely were used for the purpose of counting. Results are summarized in Table I. Capillary perfusion of antegradely delivered versus retrogradely delivered NTB-2 was compared at the eight ventricular sites sampled. No significant difference in NTB-2 content between antegrade and retrograde delivery routes was noted in the left ventricular free wall epicardium (p = 0.649), the left ventricular free wall endocardium (p = 0.168), the anterior-mid interventricular septum (p = 0.622), the apex at the mid left ventricular free wall (p = 0.455), and the apex at the mid inter-
Van Reempts, Haseldonckx, and Borgers'" were the first group to report the use of a photographic emulsion agent to study the microvasculature in its functional state. They used Ilford L4 nuclear research emulsion to study the microvascular perfusion of rat brains subjected to experimentally induced ischemia. Subsequently, Sage and Gavin!' developed a technique of glutaraldehyde perfusion fixation of the myocardium that allowed a heart to be anatomically fixed in its physiologic functioning state. Using an acrylic resin as an intracapillary marker, they demonstrated in the procaine-arrested heart that 95.2% of capillaries were perfusable (i.e., functional) and therefore contained the acrylic resin. Sheppard and Gavinl- further refined the technique of glutaraldehyde perfusion fixation and began using NTB-2 as an intracapillary marker. NTB-2 diluted with phosphate buffer has the advantage of a viscosity of about 4 centistrokes, which is similar to that of blood. Furthermore, NTB-2 is easily visualized within capillaries by either plain hematoxylin and eosin staining or electron microscopy. Using these methods, Sheppard and Gavin'? were able to demonstrate the transmural progression of the no-reflow phe-
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gates et al.
849
Fig. 5. Photomicrographof a sectionof the left ventricular endocardium after antegrade fixation and NT B-2 perfusion. The sectioncuts transverselyacrossmusclebundlesand capillaries. NTB-2 particlescan be seen eccentrically placed in most capillaries. (Hematoxylin and eosin stain; original magnification X 161.2.)
Fig. 6. Photomicrograph of a sectionof the left ventricular endocardium after retrograde fixation and NTB-2 perfusion. The sectioncuts longitudinallyacrossmusclebundlesand capillaries.NTB-2 particlescan be seen within the capillaries. (Hematoxylin and eosin stain; original magnification X 161.2.) nomenon (nonperfused capillaries) in globally ischemic hearts. We have modified this technique to study the microvascular distribution of retrogradely delivered cardioplegic solution. The results of this study demonstrate that, in regions of the heart to which solutions are delivered retrogradely (i.e., the anterior and lateral free wall of the left ventricle and the anterior half of the intraventricular septum), they are capable of perfusing capillary beds to the same degree
as antegradely delivered solutions. This is the first quantitative proof of the ability of retrogradely delivered solutions to perfuse capillary beds within the heart. Furthermore, these studies show the gross anatomic pattern of distribution of retrogradely infused cardioplegic solution in the piglet heart. The left ventricular anterior and lateral free walls, as well as the anterior half of the interventricular septum, were consistently perfused. The left ventricular posterior free wall, the posterior half of the
850
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gates et al.
Table I. Degree of capillary perfusion Perfusion
1%
mean ± SD)
Location
Antegrade In = 4)
Retrograde
p Value
LV free wall epicardium LV free wall endocardium Intraventricular septum at anterior-mid septum RV free wall epicardium R V free wall endocardium Apex at mid-LV free wall Apex at mid-intraventricular septum Apex at mid-RV free wall
90.5 ± 4.2 89.75 ± 2.22 89 ± 2.16
91.75 ± 3.1 91.75 ± 1.26 91 ± 7.39
p = 0.649, NS p=0.168, NS p= 0.622, NS
90.5 92.75 88.75 89
± 3.51 ± 1.5 ± 2.5 ± 1.15
Not perfused Not perfused 85 ± 9.06 77 ± 24.34
88.75 ± 3.3
Three hearts not perfused; one heart 35% perfused
p<0.05 p<0.05 p= 0.455, NS p= 0.363, NS
p > 0.5
SD, Standard deviation; LV, left ventricular; RV, right ventricular; NS, not significant.
interventricular septum, and a small anterior portion of the right ventricular free wall were inconsistently perfused. Finally, the remainder of the right ventricular free wall was rarely perfused (and when perfused, only slightIy). When a region of the heart was not grossly perfused, essentially no capillaries in the region contained NTB-2. Nonetheless, NTB-2 could frequently be seen in large epicardial and intramyocardial veins of the right ventricle, but it did not extend into the capillary network. The regional distribution of retrogradely delivered solutions in the piglet heart appears to parallel the distribution in other previously reported mammalian models. Qualitative anatomic studies performed on dogs by Lolley and Hewittj' as well as Shiki and associates," demonstrated that the left ventricle was well supplied with resins and contrast materials that were infused via the retrograde route. In comparison, however, the septum, and particularly the right ventricle, were poorly perfused. Furthermore, Lolley and Hewitt'' suggested that as much as 75% of retrogradely delivered solutions were shunted through thebesian veins into the cavity of the right ventricle. Using radiolabeled microspheres and a dog model, Crooke and associates" demonstrated that cardioplegic solution delivered by way of the coronary sinus was distributed better to the left ventricle than to the septum." The right ventricle was found to be poorly perfused, even when the right atrial approach for retrograde delivery of the cardioplegic solution was used. Stirling and colleagues" also performed coronary sinus cardioplegia studies using radiolabeled microspheres in a dog model. They found poor delivery to the right ventricle and significantly reduced delivery to the posterior septum and posterior free wall of the left ventricle. Only 16.5% of delivered microspheres were trapped within the right ventricle. Our study provides an anatomic explanation for the
findings reported by previous investigators. It suggests that, in animal models, the right ventricle fails to trap a significant percentage of microspheres because retrogradely delivered solutions do not perfuse the capillary beds of the right ventricle. The identification of NTB-2 in the body and epicardial veins of the right ventricle, but not its capillaries, has led to the following hypothesis: A percentage of retrogradely delivered cardioplegic solution is shunted from right ventricular epicardial veins, through thebesian veins, into the body of the right ventricle. Hence the underlying microvasculature is not perfused. This is speculation, however, because the design of this study does not allow this hypothesis to be confirmed or denied. The ability of coronary sinus-delivered NTB-2 to perfuse microvascular beds as well as antegradely delivered NTB-2 in the distribution of the left anterior descending and circumflex coronary arteries implies that retrograde cardioplegia should be ideally suited for myocardial protection in cases of acute or critical obstruction of these vessels. Laboratory work in dogs by Gundry and Kirsh.P as well as work in pigs by Haan and colleagues, 14 suggests that this is the case. Both of these groups demonstrated that superior myocardial protection was obtained when retrograde cardioplegia was used in the presence of acute arterial obstructions of the left coronary arteries. Recently, retrograde continuous warm blood cardioplegia has been reported as a technique of myocardial protection during coronary artery bypass operations. Salerno and associates! have stated that the theoretic advantages of the approach include "the continuous supply of oxygen and substrates to the arrested heart while avoiding the side effects of hypothermia." However, for the entire myocardium to benefit from warm metabolism, the blood cardioplegic solution must be delivered throughout the microvasculature. This study (in the pig-
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
let model) suggests that retrogradely delivered blood cardioplegic solution effectively perfuses the microvasculature of the heart only in the general distribution of the left anterior descending and the circumflex coronary arteries. The microvasculature of the right ventricle and frequently the posterior septum and posterior left ventricle would not be perfused. Hence these regions of the piglet heart would be susceptible to a warm anaerobic arrest during continuous warm retrograde blood cardioplegia. The work of Goldstein and coworkers 15 support this concept. This group delivered continuous cold retrograde blood cardioplegia to adult dogs for 2 hours and assessed ventricular function and adenosine triphosphate levels I hour after reperfusion. They found that both left and right ventricular function had returned to normal; however, adenosine triphosphate levels were significantly lower in the right ventricle than in the left ventricle. Nonetheless, the reported experience and general clinical impression of many surgeons using retrograde continuous warm blood cardioplegia are that myocardial protection is improved and that right ventricular dysfunction is uncommon.!'' With respect to our experiments reported herein, three explanations could account for these observations. First, the model of NTB-2 perfusion is inaccurate. We consider this to be unlikely for several reasons: (I) Histologicexamination ofthe hearts indicates that they are well preserved in their functional state without significant artifact; (2) the general technique has been previously used to demonstrate microvascular perfusion, and the specifictechnique has been used to demonstrate the nonreflow phenomenon in ischemic hearts; (3) the anatomic resultsobtained have not conflicted with (and in fact have supported) nonanatomic studies of retrograde cardioplegia. Second, the vasculature of the piglet's right ventricle may not be analogous to that of the human. This explanation is certainly possible. Although the thebesian venous system in man parallels that of other higher mammals, the factors that determine whether retrogradely infused blood cardioplegic solution traverses the microvasculature or is shunted through thebesian vessels are unknown. Species- or age-related differences could account for preferential cardioplegic flow through one system or the other. Third, the microvasculature of the human right ventricleis not perfused by blood cardioplegic solutions administered via the retrograde route. However, the salutary benefits of a warm aerobic arrest on left ventricular function outweigh the negative effects of a warm anaerobic arrest on right ventricular function. This is certainly possible, inasmuch as there have been no experimental or
Gates et al.
851
clinical studies examining right ventricular function after continuous warm retrograde blood cardioplegia. Indeed, nonrandomized, retrospective clinical reports on warm continuous blood cardioplegia have focused on measurements ofleft ventricular function (i.e., intraaortic balloon pump requirement, pressor requirement, cardiac output) and not right ventricular function. I. 17. 18 As such, subtle right ventricular dysfunction may not be appreciated. We conclude that further research is necessary to determine what degree of microvascular perfusion occurs in the human right ventricle during warm continuous retrograde cardioplegia. Furthermore, caution should be exercised before this technique is used as a sole means of myocardial protection in patients with right ventricular hypertrophy. We thank Saeedeh Terhoni and Isabell Rose for their assistance in the histologic preparation of specimens. REFERENCES 1. Salerno TA, Houck JP, Barrozo CAM, et al. Retrograde continuous warm blood cardioplegia: a new concept in myocardial protection. Ann Thorac Surg 1991 ;51 :245-7. 2. Menasche P, Subayi JB, Veyssie L, Le Dref 0, Chevret S, Piwnica A. Efficacy of coronary sinus cardioplegia in patients with complete coronary artery occlusions. Ann Thorac Surg 1991;51:418-23. 3. Bhayana IN, Kalmbach T, Booth FV, Mentzer RM, Schimert G. Combinedantegradejretrograde cardioplegia for myocardialprotection: a clinicaltrial. J THORAC CARDIOVASC SURG 1989;98:956-60. 4. Crooke GA, Harris LJ, Grossi EA, Baumann FG, Galloway AC, Colvin S. Biventricular distributionof cold blood cardioplegic solution administered by different retrograde techniques. J THORAC CARDIOVASC SURG 1991;102: 631-8. 5. Partington MT, Acar C, Buckberg GO, Julia P, Kofsky ER, Bugyi HI. Studies of retrograde cardioplegia. I. Capillary blood flow distribution to myocardium supplied by open and occluded arteries. J THORAC CARDIOVASC SURG 1989;97:605-12. 6. Stirling MC, McClanahan TB, Schott RJ, et al. Distribution of cardioplegic solution infusedantegradely and retrogradely in normal canine hearts. J THORAC CARDIOVASC SURG 1989;98: 1066-76. 7. Shiki K, Masuda M, Yonenaga K, Asou T, Tokunaga K. Myocardial distribution of retrograde flow through the coronary sinus of the excised normal canine heart. Ann Thorac Surg 1986;41 :265-71. 8. LolleyOM, Hewitt RL. Myocardial distribution of asanguineous solutions retroperfused under low pressure through the coronary sinus. J Cardiovasc Surg 1980; 21:287-94. 9. BuckbergGO. Recent advances in myocardial protection
8 5 2 Gates et ai.
10.
II.
12.
13.
14.
15.
16. 17.
18.
using antegradejretrograde blood cardioplegia. Eur Heart J 1989;lO(suppl H):43-8. Van Reempts J, Haseldonckx M, Borgers M. A simple technique for the microscopic study of microvascular geometry and tissue perfusion, allowing simultaneous histopathologic evaluation. Microvasc Res 1983;25:300-6. Sage MD, Gavin JB. Morphological identification offunctional capillaries in the myocardium. Anat Rec 1984; 208:283-9. Sheppard AJ, Gavin JB. The transmural progression of the no-reflow phenomenon in globally ischemic hearts. Basic Res CardioI1988;83:611-7. Gundry SR, Kirsh MM. A comparison of retrograde cardioplegia versus antegrade cardioplegia in the presence of coronary artery obstruction. Ann Thorac Surg 1984; 38:2:125-7. Haan C, Lazar HL, Bernard S, Rivers S, Zallnick J, Shemin RJ. Superiority of retrograde cardioplegia after acute coronary artery occlusion. Ann Thorac Surg 1991; 51:408-12. Goldstein JP, Salter DR, Murphy CE, Abd-Elfattah AS, Morris JJ, Wechsler AS. The efficacy of blood versus crystalloid coronary sinus cardioplegia during global myocardial ischemia. Circulation I986;74(Pt 2):II199-104. Engelman RM. Retrograde continuous warm blood cardioplegia. Ann Thorac Surg 1991;51:180-1. Lichtenstein SV, Ashe KA, EI Dalati H, Cusimano RJ, Panos A, Slutsky AS. Warm heart surgery. J THORAC CARDIOVASC SURG 1991;101:269-74. Lichtenstein SV, Abel JG, Salerno T. Warm heart surgery and results of operations for recent myocardial infarction. Ann Thorac Surg 1991;52:455-60.
Discussion Dr. Ances Razzouk (Lorna Linda, Calif). My associate Dr. Steven Gundry, regrets very much not being able to be here to discuss this paper in person. Dr. Gates and his colleagues have chosen yet another method of investigating the capillary level distribution of fluids administered in either antegrade or retrograde fashion. Their findings showing inadequate distribution of retrogradely infused solutions into the right ventricle and posterior interventricular septum are consistent with what every other investigator has found in the animal model. They offer in the manuscript three explanations for these findings, and it is to these that we will direct the discussion. First, they state one explanation is that the NTB-2 method is inaccurate, and then they dismiss this assumption out-ofhand. Closer inspection of the experimental model, however, shows that all three groups actually received only antegrade cardioplegia, group I receiving cold antegrade crystalloid cardioplegia and groups 2 and 3 receiving cold antegrade blood cardioplegia. The hearts were then excised. Unfortunately, that was the end of the cardioplegia. Thereafter, and in vitro, groups I and 3 received, via the coronary sinus, 37° C phosphate buffer, then 37° C glutaraldehyde, and then 37° C phosphate buffer, followed by 37° C NTB-2. Group 2 received the same compounds via an antegrade route but for a slightly different time
The Journal of Thoracic and Cardiovascular Surgery May 1993
period. Note that at no time during the experiment did these hearts receive retrograde cardioplegia. What the authors have shown is the retrograde distribution of NTB-2 in previouslycold hearts suddenly perfused and fixed with warm glutaraldehyde, and this unfortunately bears no relation to clinical circumstances. The second explanation the authors give is that the animals are not like human beings. I'm sure we all agree, as did Dr. Gott in 1957, when he reported on the first 10 patients have successful operations with continuous retrograde perfusion. He reported a dramatic difference in the distribution of blood perfusion in human hearts versus animal models. Indeed, at the 1992 meeting of the Society of Thoracic Surgeons, we presented the first quantitative study of capillary delivery of continuous warm retrograde cardioplegia in humans and found no difference in myocardial oxygen consumption between the right and left ventricles. The third explanation given in the manuscript is that the right ventricular portion of the human heart is not protected by retrograde blood cardioplegia. Again, unfortunately, the authors did not test this theory, for no retrograde blood cardioplegia was given in this experiment. In over 200 consecutive human hearts studied by transesophageal echocardiography at our institution in Loma Linda, we have seen well preserved right ventricular function in hearts protected exclusively with continuous warm blood cardioplegia. My first question to Dr. Gates is this: Why wasn't retrograde cardioplegia given to any group, as your title implies? Dr. Gates. First, I would like to thank the co-reviewers, Drs. Razzouk and Gundry, for their comments and questions and I will attempt to respond to as many of these as possible. As Dr. Razzouk states, this is indeed another animal model that demonstrates inadequate distribution of solutions delivered retrogradely to the right ventricle and posterior intraventricular septum. However, this study differs from others in an important way. This is an anatomic study that allows for the gross and microscopic examination of the entire intact heart. Microsphere studies, resin studies, and dye studies do not allow one to histologically study the intact heart. With regard to your concerns that the technique is unreliable and invalid, we too share reservations when a "new" scientific method of study is presented. However, the technique of glutaraldehyde perfusion fixation and intracapillary marking to assess capillary perfusion is a well-accepted approach to studying the microvasculature. The general technique has been used in multiple peer-reviewed publications, and the specific technique using NTB-2 has been used by Sheppard and Gavin! in their report demonstrating the transmural progression of the no-reflow phenomenon in globally ischemic hearts. Indeed, the technique has been extremely effective in demonstrating the widely known distribution of antegradely delivered solutions. Hence we believe this technique (which uses a solution of phosphate buffer and NTB-2 to mimic blood cardioplegia) is valid and that the results represent the distribution of retrograde cardioplegia in piglet hearts. Are piglets like human beings with regard to the distribution of solutions delivered by way of the coronary sinus? This report cannot answer that question. However, we have begun to apply this technique to explanted human hearts from transplant recipients. Our preliminary results have demonstrated that capillaries within the right ventricle of the human heart are perfused by retrograde cardioplegia. However, of retrogradely
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
delivered warm blood cardioplegic solution, approximately 65% was found exiting thebesian veins, 32% the left coronary ostia, and 2% to 3% the right coronary ostia. This makes one wonder what the true capillary or "nutritive" flow to the human right ventricle is during retrograde cardioplegia. We are aware of the Loma Linda group's abstract presented at the twenty-eighth Society of Thoracic Surgeons meeting where they reported their measurements of pH, oxygen tension, carbon dioxide tension, the bicarbonate radical, base excess, and oxygencontent in the aortic root (combined right and left coronary artery) effluent and coronary sinus effluent during and after warm continuous retrograde blood cardioplegia.? On the basis of this experiment, they have concluded that capillary deliveryof cardioplegic solution to the right ventricle is sufficient to maintain myocardial homeostasis and eliminate myocardial ischemia. However, if one is aware that more than 90% of aortic root effluent during warm retrograde cardioplegia reflects blood that has traversed the myocardium supplied by the left coronary artery, it becomes difficult to draw conclusions about the right ventricle. Similarly, if one is aware that the vast majority of right ventricular postcapillary venous return is through thebesian veins (not the coronary sinus), it is again difficult to draw any meaningful conclusions about right ventricular"nutritive" flowwhen using such a technique. Therefore, we are not convinced that one can make any statement regarding the metabolic status of the right ventricle based on the data presented within the aforementioned abstract. We have read Dr. Gundry's 1990 report on the clinical use ofcoldcontinuous retrograde blood cardioplegia for myocardial protection using the Gundry RCSP catheter (DLP, Inc., Grand Rapids, Mich.), which he has designed.' The excellent right ventricular cooling and protection this technique affords is an experience we too have shared. Our current study supports the concept that excellent right ventricular cooling can be achieved byretrograde delivery of cold solutions. Anatomically, multiple
Gates et at.
853
thebesian veins (through which the majority of retrogradely delivered solutions flow) penetrate the right ventricular myocardium and are able to act as conduction radiators to effectively cool the right ventricle. Dr. Razzouk's statement that the recent Loma Linda experience with warm continuous retrograde blood cardioplegia has demonstrated good right ventricular myocardial protection as examined by transesophageal echocardiography is of great interest to us all and we certainly await the published results. Dr. Razzouk. You mentioned that you used a perfusion pressure of 50 to 60 mm Hg but no monitoring catheter was placed in the coronary sinus. We all know that the actual pressure measured in the tubing is not the real pressure in the coronary sinus. Can you comment on that please? Dr. Gates. In early pilot studies we did puncture the coronary sinus and directly monitored coronary sinus pressures. Because direct pressure correlated well with line pressure, we then routinely measured line pressure. REFERENCES I. Sheppard AJ, Gavin JB. The transmural progression of the no-reflow phenomenon in globally ischemic hearts. Basic Res CardioI1988;83:611-7. 2. Gundry SR, Wang N, Bannon D, et al. Continuous warm retrograde blood cardioplegia: maintenance of myocardial homeostasis and elimination of myocardial ischemia [Abstract]. Twenty-eighth Annual Meeting of the Society of Thoracic Surgeons, Orlando, Fla. 3. Gundry SR, Sequiera A, Razzouk AM, McLaughlin JS, Bailey LL. Facile retrograde cardioplegia: transatrial cannulation of the coronary sinus. Ann Thorac Surg 1990; 50:882-7.
Richard N. Gates, MD, Hillel Laks, MD, Davis C. Drinkwater, MD, Jeffrey Pearl, MD, Ana Maria Zaragoza, MD, Elias Kaczer, BS, and Paul Chang, BS, Los Angeles, Calif.
h e clinical application of retrograde cardioplegia is increasing and several reports attest to its efficacy.I>' Retrograde cardioplegia is particularly suited for adult coronary artery bypass operations when acute or critical obstruction of coronary arteries exists. Furthermore, many pediatric procedures necessitating aortotomy are facilitated by retrograde techniques. Despite this, little information exists as to the ability of retrogradely infused
From The Divisionof Cardiothoracic Surgery, Department of Surgery and Department of Pathology, University of California at Los Angeles, Medical Center, Los Angeles, Calif. Read at the Eighteenth Annual Meeting of The Western Thoracic Surgical Association, Kauai, Hawaii, June 24-27, 1992. Address for reprints: Hillel Laks, MD, Division of Cardiothoracic Surgery,UCLA Medical Center, CHS 62-182,10833 LeConte Ave., LosAngeles, CA 90024. Copyright
1993 by Mosby-Year Book, Inc.
0022-5223/93 $\.00 + .10
12/6/44661
cardioplegic solution to perfuse the microvasculature of the heart. Two experimental approaches have been applied in an attempt to determine the distribution of retrogradely infused cardioplegic solution. The first consists of the retrograde injection of radiolabeled microspheres that are subsequently trapped in the microvasculature.v" Although this method provides qualitative data concerning regional distribution, it provides no quantitative information regarding the perfusion of microvascular beds. The second approach involves the retrograde infusion of resins or dyes and subsequent maceration or radiography of the heart. 7,8 These methods reveal regional anatomic distribution in a qualitative manner but fail to provide quantitative information regarding microvascular beds that have not been fixed in resin or dye. The efficacy of oxygenated cardioplegia is dependent on delivery of the cardioplegic agent to all areas of the microvasculature." In view of the increasing reliance on
845
The Journal of Thoracic and Cardiovascular Surgery May 1993
8 4 6 Gates et al.
IVS
RV
RV
LV LV Epi
Epi
Mid-ventricle
Apex
Fig. 1. Diagrammatic representation of the sections of ventricles taken for histologic analysis. RV, Right ventricle; LV, left ventricle; IVS, interventricular septum; Epi, epicardium; Endo, endocardium.
retrograde cardioplegia for myocardial protection, we have developed a technique to examine the microvascular distribution of retrogradely delivered cardioplegic solution in a quantitative manner. This was achieved by capillary fixation by retrograde glutaraldehyde perfusion and retrograde perfusion of the intracapillary marker NTB-2 to demonstrate capillary patency. These experiments were performed in a piglet model under conditions simulating those used in clinical practice. They provide insight into the anatomic distribution and degree of capillary perfusion achieved by retrograde cardioplegia in clinically relevant conditions.
Materials and methods Twelve Duroe piglets were divided into three experimental groups. Group 1 (n = 4) was composed of neonatal piglets (I to 3 days old, 2 to 2.5 kg). Group 2 (n = 4) and group 3 (n = 4) consisted of infant piglets (25 to 35 days old, 5 to 7 kg). Piglets in group 1 were subjected to antegrade crystalloid cardioplegia, retrograde perfusion of glutaraldehyde fixative via the coronary sinus, and NTB-2 perfusion. Group 2 piglets were subjected to antegrade blood cardioplegia, antegrade perfusion of glutaraldehyde fixative, and NTB-2 perfusion. Group 3 piglets received antegrade blood cardioplegia, retrograde perfusion of glutaraldehyde fixative, and NTB-2 perfusion. Group I, 2, and 3 piglets were anesthetized with intramuscular ketamine (l00 rug/kg) and acepromazine (0.1 rug/kg). Tracheostomy was then performed and the animals were mechanically ventilated (Bird Mark 7, Bird Corp., Palm Springs, Calif.) with an inspired oxygen fraction of 1.0, a rate
of 15 to 20 cycles/min, and an inspiratory pressure of 25 mm Hg. Through a median sternotomy, the left hemiazygos veinand left subclavian artery were ligated. Heparin (1000 U) was administered. The innominate artery was cannulated with a modified 12-gauge catheter (Angiocath; Deseret Medical, Sandy, Utah) for group 1 and a modified 16F arterial cannula (DLP Inc., Grand Rapids, Mich) for groups 2 and 3. The central arterial catheter was attached to the cardioplegia delivery tubing, where pressure was monitored. The superior and inferior venae cavae were then ligated and an aortic crossclamp was placed on the descending aorta. The heart was vented via incisions in the pulmonary artery and left atrial appendage. Cardioplegic solution at 4° C was delivered via the catheter in the innominate artery for 2 minutes at 80 mm Hg. Once arrest was obtained, topical hypothermia was applied with iced saline. The inferior pulmonary ligaments were divided and the heart-lung block was excised. In groups I and 3, the inferior vena cava was opened to the level of the atria. The previously placed tie on the left hemiazygos vein was cut and a 12-gauge Angioeath catheter was introduced into the left hemiazygos vein and advanced into the coronary sinus. This was possible because the left hemiazygos vein drains into the coronary sinus in approximately 95% of piglets. The catheter was tied in place and flushed with heparinized saline solution to remove any air from within the cardiac venous drainage system. The coronary sinus ostium was carefully oversewn with 7-0 or 8-0 Prolene suture (Ethicon, Inc., Somerville, N.J.). In all groups glutaraldehyde fixation began 15 minutes after initial cardioplegic arrest. For both the antegrade and retrograde groups, all solutions were delivered at a pressure of 50 to 60 mm Hg. In groups I and 3, the aortic clamp was released and the heart was perfused retrogradely for 2 minutes with phosphate buffer, 0.1 mol/L at 37° C (Sigma Chemical Company, St. Louis, Mo.). This was followed by 5 minutes of retrograde perfusion with 2.5% glutaraldehyde at 37° C (Sigma), (diluted to concentration with phosphate buffer, 0.1 mol/L), Retrograde perfusion was continued for 10 minutes with phosphate buffer, 0.1 mol/L at 37° C. Finally, the heart was perfused retrogradely for 2 minutes at 37° C with autoradiographic emulsion type NTB-2 (Eastman Kodak Co., Rochester, N.Y.) that had been diluted I: 1 with phosphate buffer, 0.1 mol/L. For group 2 hearts, the protocol was the same but the perfusion times were different. Phosphate buffer was perfused antegradely for 2 minutes, 2.5% glutaraldehyde was perfused antegradely for 2 minutes, and phosphate buffer was perfused antegradely for 6 minutes. The antegrade perfusion times were shorter because flow rates were higher. Adjusting perfusion times allowed similar volumes of fixative and perfusate to be administered both antegradely and retrogradely. Finally, NTB-2 was perfused antegradely for 2 minutes. After perfusion of NTB-2, hearts were placed in a 1:10 formalin solution (Medical Chemical Corp., Santa Monica, Calif.) and stored for 24 hours at 4° to 10° C. For group 2 and 3 piglets, blood for the cardioplegic solution was obtained from adolescent (3 to 6 months, 33 to 65 kg) Duroc pigs. Blood was mixed with crystalloid cardioplegic solution in a 4:1 ratio. The crystalloid cardioplegic solution was composed of 500 ml of 5% dextrose in 0.2NS, 200 ml of 0.3 molar tromethamine (Abbott Labs, Chicago, Ill.), 50 ml of citratephosphate-dextrose solution (Abbott), and 60 mEq of potassium chloride. The blood cardioplegic solution was directed by a roller pump (model 572400 I-A, Sarns/ 3M, Ann Arbor, Mich.)
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gates et al.
847
Fig. 2. Photomicrograph of the anterior interventricular septum and anterior right ventricle. Note the presence of NTB-2 in the microvasculature of the interventricular septum and its absence from the right ventricle. (Hematoxylin and eosin stain; original magnification X16.1.) to a heated reservoir/bubble trap (BCD Plus, Shiley Inc., Irvine, Calif.). The blood cardioplegic solution was cooled to 40 C by counterflow with a heater/cooler pump (Sarns/3M) beforedelivery. For the group I piglets, 4 0 C Viaspan (DuPont Pharmaceuticals, Wilmington, Del.) cardioplegic solution was used. After 24 hours of formalin storage, group I hearts were horizontally transected at the mid ventricular level. A thin horizontal slice was taken through the left ventricle, interventricular septum, and right ventricle. Because these hearts were small, the entire horizontal specimen was able to be placed into one Tissue-Tek III unicassette (Miles Laboratories, Elkhart, Ind.) for histologic processing. For group 2 and 3 hearts, a thin horizontalslice was taken from the apex and placed into a cassette. The heart was then horizontally divided approximately I em abovethe mid ventricular level. Thin longitudinal sections were then taken from the left ventricular free wall, the anterior-mid interventricular septum, and the right ventricular free wall (Fig. I). Specimens were routinely stained with hematoxylin and eosin and the tissue was cut at a depth of 6 mm and placed onto slides. Group I slides (which contained the entire horizontal section through the left ventricle, intraventricular septum, and right ventricle) were examined under high-power (I61.2X) and lowpower (16.1X) light microscopes to determine gross regional perfusion. Group 2 and 3 hearts were examined with a high-power (312.5X) light microscope in the following regions of the heart: (I) left ventricular free wall epicardium, (2) left ventricular free wall endocardium, (3) interventricular septum at anterior-mid septum, (4) right ventricular free wall epicardium,(5) right ventricular free wall endocardium, (6) apex at mid left ventricular free wall, (7) apex at mid interventricular septum,and (8) apex at mid right ventricular free wall. Four separatemicroscopicfields of each region were examined by a light microscope attached to a videocamera (model XPC501, HitachiMedical Corp., Tokyo, Japan). The video image was trans-
fered to computer matrix form and stored on a hard drive by means of a video digitalizer (SuperVIA digitizing board, Jovian Logic Corporation, Fremont, Calif.). Each image was reviewed on a color computer screen and the number of capillaries filled or not filled with NTB-2 was hand counted with computer assistance. Hand counting was not done in a blinded fashion, but computer co-counting was of course blinded. Statistical analysis was performed between groups and regions by means of the unpaired two-tailed Student's t test. All animals used for these studies received humane care in compliance with the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH publication No. 88-23, revised 1985). In addition, the experimental protocol was reviewed and approved by the University of California at Los Angeles Chancellor's Animal Research Committee (ARC No. 91-254).
Results Gross microvascular distribution of NTB-2. Examination of group 1 and 3 hearts revealed the gross microvascular distribution of retrogradely delivered NTB-2 to be as follows: (1) consistently perfused-the anterior half of the interventricular septum and the anterior and lateral free walls of the left ventricle (8/8); (2) inconsistently perfused-the posterior half of the interventricular septum (2/4), the posterior free wall of the left ventricle (2/4), and a small area of the anterior free wall of the right ventricle next to the interventricular septum (1/4); (3) not perfusedorrarelyperfused-all oftheright ventricular free wall except a small area of the anterior free wall next to the interventricular septum (0/8). Although the microvasculature of the right ventricle was not perfused, NTB-2 could frequently be identified in
848
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gates et al.
Posterior
RV LV
Anterior
Not perfused Inconsistently perfused _
Consistently perfused
Fig. 4. Diagrammatic representation of a horizontal section throughthe mid ventricle of group 1 hearts. Thisdemonstrates regions of the heart where retrograde perfusion occurred. R V, Right ventricle; LV, left ventricle.
Fig. 3. Photomicrograph of the right ventricle. Note the presenceof NTB-2 in the largeepicardial vessel and in a few large intramyocardial veins, but not in the microvasculature. (Hematoxylin and eosin stain; original magnification X16.1.)
ventricular septum (p = 0.363). However, retrogradely delivered NTB-2 failed to perfuse the right ventricular free wall epicardium or endocardium and the apex at the mid right ventricular free wall (p < 0.05). Antegradely delivered NTB-2 successfully perfused these regions. Discussion
large epicardial and intramyocardial veins, as well as in the cavity of the right ventricle (Figs. 2, 3, and 4). In group 2 hearts, all eight regions of the ventricle sampled were consistently perfused by antegradely delivered NTB-2 (4/4). Quantitative microvascular perfusion of NTB-2. Examination of groups 2 and 3 allowed for the determination of the percentage of capillaries perfused (NTB-2 filled/NTB-2 filled plus unfilled) (Figs. 5 and 6). An average of 218 capillaries per region of the heart were counted. Sections cut transversely or near transversely were used for the purpose of counting. Results are summarized in Table I. Capillary perfusion of antegradely delivered versus retrogradely delivered NTB-2 was compared at the eight ventricular sites sampled. No significant difference in NTB-2 content between antegrade and retrograde delivery routes was noted in the left ventricular free wall epicardium (p = 0.649), the left ventricular free wall endocardium (p = 0.168), the anterior-mid interventricular septum (p = 0.622), the apex at the mid left ventricular free wall (p = 0.455), and the apex at the mid inter-
Van Reempts, Haseldonckx, and Borgers'" were the first group to report the use of a photographic emulsion agent to study the microvasculature in its functional state. They used Ilford L4 nuclear research emulsion to study the microvascular perfusion of rat brains subjected to experimentally induced ischemia. Subsequently, Sage and Gavin!' developed a technique of glutaraldehyde perfusion fixation of the myocardium that allowed a heart to be anatomically fixed in its physiologic functioning state. Using an acrylic resin as an intracapillary marker, they demonstrated in the procaine-arrested heart that 95.2% of capillaries were perfusable (i.e., functional) and therefore contained the acrylic resin. Sheppard and Gavinl- further refined the technique of glutaraldehyde perfusion fixation and began using NTB-2 as an intracapillary marker. NTB-2 diluted with phosphate buffer has the advantage of a viscosity of about 4 centistrokes, which is similar to that of blood. Furthermore, NTB-2 is easily visualized within capillaries by either plain hematoxylin and eosin staining or electron microscopy. Using these methods, Sheppard and Gavin'? were able to demonstrate the transmural progression of the no-reflow phe-
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gates et al.
849
Fig. 5. Photomicrographof a sectionof the left ventricular endocardium after antegrade fixation and NT B-2 perfusion. The sectioncuts transverselyacrossmusclebundlesand capillaries. NTB-2 particlescan be seen eccentrically placed in most capillaries. (Hematoxylin and eosin stain; original magnification X 161.2.)
Fig. 6. Photomicrograph of a sectionof the left ventricular endocardium after retrograde fixation and NTB-2 perfusion. The sectioncuts longitudinallyacrossmusclebundlesand capillaries.NTB-2 particlescan be seen within the capillaries. (Hematoxylin and eosin stain; original magnification X 161.2.) nomenon (nonperfused capillaries) in globally ischemic hearts. We have modified this technique to study the microvascular distribution of retrogradely delivered cardioplegic solution. The results of this study demonstrate that, in regions of the heart to which solutions are delivered retrogradely (i.e., the anterior and lateral free wall of the left ventricle and the anterior half of the intraventricular septum), they are capable of perfusing capillary beds to the same degree
as antegradely delivered solutions. This is the first quantitative proof of the ability of retrogradely delivered solutions to perfuse capillary beds within the heart. Furthermore, these studies show the gross anatomic pattern of distribution of retrogradely infused cardioplegic solution in the piglet heart. The left ventricular anterior and lateral free walls, as well as the anterior half of the interventricular septum, were consistently perfused. The left ventricular posterior free wall, the posterior half of the
850
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gates et al.
Table I. Degree of capillary perfusion Perfusion
1%
mean ± SD)
Location
Antegrade In = 4)
Retrograde
p Value
LV free wall epicardium LV free wall endocardium Intraventricular septum at anterior-mid septum RV free wall epicardium R V free wall endocardium Apex at mid-LV free wall Apex at mid-intraventricular septum Apex at mid-RV free wall
90.5 ± 4.2 89.75 ± 2.22 89 ± 2.16
91.75 ± 3.1 91.75 ± 1.26 91 ± 7.39
p = 0.649, NS p=0.168, NS p= 0.622, NS
90.5 92.75 88.75 89
± 3.51 ± 1.5 ± 2.5 ± 1.15
Not perfused Not perfused 85 ± 9.06 77 ± 24.34
88.75 ± 3.3
Three hearts not perfused; one heart 35% perfused
p<0.05 p<0.05 p= 0.455, NS p= 0.363, NS
p > 0.5
SD, Standard deviation; LV, left ventricular; RV, right ventricular; NS, not significant.
interventricular septum, and a small anterior portion of the right ventricular free wall were inconsistently perfused. Finally, the remainder of the right ventricular free wall was rarely perfused (and when perfused, only slightIy). When a region of the heart was not grossly perfused, essentially no capillaries in the region contained NTB-2. Nonetheless, NTB-2 could frequently be seen in large epicardial and intramyocardial veins of the right ventricle, but it did not extend into the capillary network. The regional distribution of retrogradely delivered solutions in the piglet heart appears to parallel the distribution in other previously reported mammalian models. Qualitative anatomic studies performed on dogs by Lolley and Hewittj' as well as Shiki and associates," demonstrated that the left ventricle was well supplied with resins and contrast materials that were infused via the retrograde route. In comparison, however, the septum, and particularly the right ventricle, were poorly perfused. Furthermore, Lolley and Hewitt'' suggested that as much as 75% of retrogradely delivered solutions were shunted through thebesian veins into the cavity of the right ventricle. Using radiolabeled microspheres and a dog model, Crooke and associates" demonstrated that cardioplegic solution delivered by way of the coronary sinus was distributed better to the left ventricle than to the septum." The right ventricle was found to be poorly perfused, even when the right atrial approach for retrograde delivery of the cardioplegic solution was used. Stirling and colleagues" also performed coronary sinus cardioplegia studies using radiolabeled microspheres in a dog model. They found poor delivery to the right ventricle and significantly reduced delivery to the posterior septum and posterior free wall of the left ventricle. Only 16.5% of delivered microspheres were trapped within the right ventricle. Our study provides an anatomic explanation for the
findings reported by previous investigators. It suggests that, in animal models, the right ventricle fails to trap a significant percentage of microspheres because retrogradely delivered solutions do not perfuse the capillary beds of the right ventricle. The identification of NTB-2 in the body and epicardial veins of the right ventricle, but not its capillaries, has led to the following hypothesis: A percentage of retrogradely delivered cardioplegic solution is shunted from right ventricular epicardial veins, through thebesian veins, into the body of the right ventricle. Hence the underlying microvasculature is not perfused. This is speculation, however, because the design of this study does not allow this hypothesis to be confirmed or denied. The ability of coronary sinus-delivered NTB-2 to perfuse microvascular beds as well as antegradely delivered NTB-2 in the distribution of the left anterior descending and circumflex coronary arteries implies that retrograde cardioplegia should be ideally suited for myocardial protection in cases of acute or critical obstruction of these vessels. Laboratory work in dogs by Gundry and Kirsh.P as well as work in pigs by Haan and colleagues, 14 suggests that this is the case. Both of these groups demonstrated that superior myocardial protection was obtained when retrograde cardioplegia was used in the presence of acute arterial obstructions of the left coronary arteries. Recently, retrograde continuous warm blood cardioplegia has been reported as a technique of myocardial protection during coronary artery bypass operations. Salerno and associates! have stated that the theoretic advantages of the approach include "the continuous supply of oxygen and substrates to the arrested heart while avoiding the side effects of hypothermia." However, for the entire myocardium to benefit from warm metabolism, the blood cardioplegic solution must be delivered throughout the microvasculature. This study (in the pig-
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
let model) suggests that retrogradely delivered blood cardioplegic solution effectively perfuses the microvasculature of the heart only in the general distribution of the left anterior descending and the circumflex coronary arteries. The microvasculature of the right ventricle and frequently the posterior septum and posterior left ventricle would not be perfused. Hence these regions of the piglet heart would be susceptible to a warm anaerobic arrest during continuous warm retrograde blood cardioplegia. The work of Goldstein and coworkers 15 support this concept. This group delivered continuous cold retrograde blood cardioplegia to adult dogs for 2 hours and assessed ventricular function and adenosine triphosphate levels I hour after reperfusion. They found that both left and right ventricular function had returned to normal; however, adenosine triphosphate levels were significantly lower in the right ventricle than in the left ventricle. Nonetheless, the reported experience and general clinical impression of many surgeons using retrograde continuous warm blood cardioplegia are that myocardial protection is improved and that right ventricular dysfunction is uncommon.!'' With respect to our experiments reported herein, three explanations could account for these observations. First, the model of NTB-2 perfusion is inaccurate. We consider this to be unlikely for several reasons: (I) Histologicexamination ofthe hearts indicates that they are well preserved in their functional state without significant artifact; (2) the general technique has been previously used to demonstrate microvascular perfusion, and the specifictechnique has been used to demonstrate the nonreflow phenomenon in ischemic hearts; (3) the anatomic resultsobtained have not conflicted with (and in fact have supported) nonanatomic studies of retrograde cardioplegia. Second, the vasculature of the piglet's right ventricle may not be analogous to that of the human. This explanation is certainly possible. Although the thebesian venous system in man parallels that of other higher mammals, the factors that determine whether retrogradely infused blood cardioplegic solution traverses the microvasculature or is shunted through thebesian vessels are unknown. Species- or age-related differences could account for preferential cardioplegic flow through one system or the other. Third, the microvasculature of the human right ventricleis not perfused by blood cardioplegic solutions administered via the retrograde route. However, the salutary benefits of a warm aerobic arrest on left ventricular function outweigh the negative effects of a warm anaerobic arrest on right ventricular function. This is certainly possible, inasmuch as there have been no experimental or
Gates et al.
851
clinical studies examining right ventricular function after continuous warm retrograde blood cardioplegia. Indeed, nonrandomized, retrospective clinical reports on warm continuous blood cardioplegia have focused on measurements ofleft ventricular function (i.e., intraaortic balloon pump requirement, pressor requirement, cardiac output) and not right ventricular function. I. 17. 18 As such, subtle right ventricular dysfunction may not be appreciated. We conclude that further research is necessary to determine what degree of microvascular perfusion occurs in the human right ventricle during warm continuous retrograde cardioplegia. Furthermore, caution should be exercised before this technique is used as a sole means of myocardial protection in patients with right ventricular hypertrophy. We thank Saeedeh Terhoni and Isabell Rose for their assistance in the histologic preparation of specimens. REFERENCES 1. Salerno TA, Houck JP, Barrozo CAM, et al. Retrograde continuous warm blood cardioplegia: a new concept in myocardial protection. Ann Thorac Surg 1991 ;51 :245-7. 2. Menasche P, Subayi JB, Veyssie L, Le Dref 0, Chevret S, Piwnica A. Efficacy of coronary sinus cardioplegia in patients with complete coronary artery occlusions. Ann Thorac Surg 1991;51:418-23. 3. Bhayana IN, Kalmbach T, Booth FV, Mentzer RM, Schimert G. Combinedantegradejretrograde cardioplegia for myocardialprotection: a clinicaltrial. J THORAC CARDIOVASC SURG 1989;98:956-60. 4. Crooke GA, Harris LJ, Grossi EA, Baumann FG, Galloway AC, Colvin S. Biventricular distributionof cold blood cardioplegic solution administered by different retrograde techniques. J THORAC CARDIOVASC SURG 1991;102: 631-8. 5. Partington MT, Acar C, Buckberg GO, Julia P, Kofsky ER, Bugyi HI. Studies of retrograde cardioplegia. I. Capillary blood flow distribution to myocardium supplied by open and occluded arteries. J THORAC CARDIOVASC SURG 1989;97:605-12. 6. Stirling MC, McClanahan TB, Schott RJ, et al. Distribution of cardioplegic solution infusedantegradely and retrogradely in normal canine hearts. J THORAC CARDIOVASC SURG 1989;98: 1066-76. 7. Shiki K, Masuda M, Yonenaga K, Asou T, Tokunaga K. Myocardial distribution of retrograde flow through the coronary sinus of the excised normal canine heart. Ann Thorac Surg 1986;41 :265-71. 8. LolleyOM, Hewitt RL. Myocardial distribution of asanguineous solutions retroperfused under low pressure through the coronary sinus. J Cardiovasc Surg 1980; 21:287-94. 9. BuckbergGO. Recent advances in myocardial protection
8 5 2 Gates et ai.
10.
II.
12.
13.
14.
15.
16. 17.
18.
using antegradejretrograde blood cardioplegia. Eur Heart J 1989;lO(suppl H):43-8. Van Reempts J, Haseldonckx M, Borgers M. A simple technique for the microscopic study of microvascular geometry and tissue perfusion, allowing simultaneous histopathologic evaluation. Microvasc Res 1983;25:300-6. Sage MD, Gavin JB. Morphological identification offunctional capillaries in the myocardium. Anat Rec 1984; 208:283-9. Sheppard AJ, Gavin JB. The transmural progression of the no-reflow phenomenon in globally ischemic hearts. Basic Res CardioI1988;83:611-7. Gundry SR, Kirsh MM. A comparison of retrograde cardioplegia versus antegrade cardioplegia in the presence of coronary artery obstruction. Ann Thorac Surg 1984; 38:2:125-7. Haan C, Lazar HL, Bernard S, Rivers S, Zallnick J, Shemin RJ. Superiority of retrograde cardioplegia after acute coronary artery occlusion. Ann Thorac Surg 1991; 51:408-12. Goldstein JP, Salter DR, Murphy CE, Abd-Elfattah AS, Morris JJ, Wechsler AS. The efficacy of blood versus crystalloid coronary sinus cardioplegia during global myocardial ischemia. Circulation I986;74(Pt 2):II199-104. Engelman RM. Retrograde continuous warm blood cardioplegia. Ann Thorac Surg 1991;51:180-1. Lichtenstein SV, Ashe KA, EI Dalati H, Cusimano RJ, Panos A, Slutsky AS. Warm heart surgery. J THORAC CARDIOVASC SURG 1991;101:269-74. Lichtenstein SV, Abel JG, Salerno T. Warm heart surgery and results of operations for recent myocardial infarction. Ann Thorac Surg 1991;52:455-60.
Discussion Dr. Ances Razzouk (Lorna Linda, Calif). My associate Dr. Steven Gundry, regrets very much not being able to be here to discuss this paper in person. Dr. Gates and his colleagues have chosen yet another method of investigating the capillary level distribution of fluids administered in either antegrade or retrograde fashion. Their findings showing inadequate distribution of retrogradely infused solutions into the right ventricle and posterior interventricular septum are consistent with what every other investigator has found in the animal model. They offer in the manuscript three explanations for these findings, and it is to these that we will direct the discussion. First, they state one explanation is that the NTB-2 method is inaccurate, and then they dismiss this assumption out-ofhand. Closer inspection of the experimental model, however, shows that all three groups actually received only antegrade cardioplegia, group I receiving cold antegrade crystalloid cardioplegia and groups 2 and 3 receiving cold antegrade blood cardioplegia. The hearts were then excised. Unfortunately, that was the end of the cardioplegia. Thereafter, and in vitro, groups I and 3 received, via the coronary sinus, 37° C phosphate buffer, then 37° C glutaraldehyde, and then 37° C phosphate buffer, followed by 37° C NTB-2. Group 2 received the same compounds via an antegrade route but for a slightly different time
The Journal of Thoracic and Cardiovascular Surgery May 1993
period. Note that at no time during the experiment did these hearts receive retrograde cardioplegia. What the authors have shown is the retrograde distribution of NTB-2 in previouslycold hearts suddenly perfused and fixed with warm glutaraldehyde, and this unfortunately bears no relation to clinical circumstances. The second explanation the authors give is that the animals are not like human beings. I'm sure we all agree, as did Dr. Gott in 1957, when he reported on the first 10 patients have successful operations with continuous retrograde perfusion. He reported a dramatic difference in the distribution of blood perfusion in human hearts versus animal models. Indeed, at the 1992 meeting of the Society of Thoracic Surgeons, we presented the first quantitative study of capillary delivery of continuous warm retrograde cardioplegia in humans and found no difference in myocardial oxygen consumption between the right and left ventricles. The third explanation given in the manuscript is that the right ventricular portion of the human heart is not protected by retrograde blood cardioplegia. Again, unfortunately, the authors did not test this theory, for no retrograde blood cardioplegia was given in this experiment. In over 200 consecutive human hearts studied by transesophageal echocardiography at our institution in Loma Linda, we have seen well preserved right ventricular function in hearts protected exclusively with continuous warm blood cardioplegia. My first question to Dr. Gates is this: Why wasn't retrograde cardioplegia given to any group, as your title implies? Dr. Gates. First, I would like to thank the co-reviewers, Drs. Razzouk and Gundry, for their comments and questions and I will attempt to respond to as many of these as possible. As Dr. Razzouk states, this is indeed another animal model that demonstrates inadequate distribution of solutions delivered retrogradely to the right ventricle and posterior intraventricular septum. However, this study differs from others in an important way. This is an anatomic study that allows for the gross and microscopic examination of the entire intact heart. Microsphere studies, resin studies, and dye studies do not allow one to histologically study the intact heart. With regard to your concerns that the technique is unreliable and invalid, we too share reservations when a "new" scientific method of study is presented. However, the technique of glutaraldehyde perfusion fixation and intracapillary marking to assess capillary perfusion is a well-accepted approach to studying the microvasculature. The general technique has been used in multiple peer-reviewed publications, and the specific technique using NTB-2 has been used by Sheppard and Gavin! in their report demonstrating the transmural progression of the no-reflow phenomenon in globally ischemic hearts. Indeed, the technique has been extremely effective in demonstrating the widely known distribution of antegradely delivered solutions. Hence we believe this technique (which uses a solution of phosphate buffer and NTB-2 to mimic blood cardioplegia) is valid and that the results represent the distribution of retrograde cardioplegia in piglet hearts. Are piglets like human beings with regard to the distribution of solutions delivered by way of the coronary sinus? This report cannot answer that question. However, we have begun to apply this technique to explanted human hearts from transplant recipients. Our preliminary results have demonstrated that capillaries within the right ventricle of the human heart are perfused by retrograde cardioplegia. However, of retrogradely
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
delivered warm blood cardioplegic solution, approximately 65% was found exiting thebesian veins, 32% the left coronary ostia, and 2% to 3% the right coronary ostia. This makes one wonder what the true capillary or "nutritive" flow to the human right ventricle is during retrograde cardioplegia. We are aware of the Loma Linda group's abstract presented at the twenty-eighth Society of Thoracic Surgeons meeting where they reported their measurements of pH, oxygen tension, carbon dioxide tension, the bicarbonate radical, base excess, and oxygencontent in the aortic root (combined right and left coronary artery) effluent and coronary sinus effluent during and after warm continuous retrograde blood cardioplegia.? On the basis of this experiment, they have concluded that capillary deliveryof cardioplegic solution to the right ventricle is sufficient to maintain myocardial homeostasis and eliminate myocardial ischemia. However, if one is aware that more than 90% of aortic root effluent during warm retrograde cardioplegia reflects blood that has traversed the myocardium supplied by the left coronary artery, it becomes difficult to draw conclusions about the right ventricle. Similarly, if one is aware that the vast majority of right ventricular postcapillary venous return is through thebesian veins (not the coronary sinus), it is again difficult to draw any meaningful conclusions about right ventricular"nutritive" flowwhen using such a technique. Therefore, we are not convinced that one can make any statement regarding the metabolic status of the right ventricle based on the data presented within the aforementioned abstract. We have read Dr. Gundry's 1990 report on the clinical use ofcoldcontinuous retrograde blood cardioplegia for myocardial protection using the Gundry RCSP catheter (DLP, Inc., Grand Rapids, Mich.), which he has designed.' The excellent right ventricular cooling and protection this technique affords is an experience we too have shared. Our current study supports the concept that excellent right ventricular cooling can be achieved byretrograde delivery of cold solutions. Anatomically, multiple
Gates et at.
853
thebesian veins (through which the majority of retrogradely delivered solutions flow) penetrate the right ventricular myocardium and are able to act as conduction radiators to effectively cool the right ventricle. Dr. Razzouk's statement that the recent Loma Linda experience with warm continuous retrograde blood cardioplegia has demonstrated good right ventricular myocardial protection as examined by transesophageal echocardiography is of great interest to us all and we certainly await the published results. Dr. Razzouk. You mentioned that you used a perfusion pressure of 50 to 60 mm Hg but no monitoring catheter was placed in the coronary sinus. We all know that the actual pressure measured in the tubing is not the real pressure in the coronary sinus. Can you comment on that please? Dr. Gates. In early pilot studies we did puncture the coronary sinus and directly monitored coronary sinus pressures. Because direct pressure correlated well with line pressure, we then routinely measured line pressure. REFERENCES I. Sheppard AJ, Gavin JB. The transmural progression of the no-reflow phenomenon in globally ischemic hearts. Basic Res CardioI1988;83:611-7. 2. Gundry SR, Wang N, Bannon D, et al. Continuous warm retrograde blood cardioplegia: maintenance of myocardial homeostasis and elimination of myocardial ischemia [Abstract]. Twenty-eighth Annual Meeting of the Society of Thoracic Surgeons, Orlando, Fla. 3. Gundry SR, Sequiera A, Razzouk AM, McLaughlin JS, Bailey LL. Facile retrograde cardioplegia: transatrial cannulation of the coronary sinus. Ann Thorac Surg 1990; 50:882-7.