CARDIOPULMONARYBYPASS, MYOCARDIAL MANAGEMENT, AND SUPPORT TECHNIQUES

RETROGRADE Surgeons often rely primarily on retrograde cardioplegia for myocardial protection, CARDIOPLEGIADOES NOT because it provides adequate left ventricular even in the presence of ADEQUATELY PERFUSE THE coronary disease. Clinically, however, adequate right ventricular perfusion by RIGHT retrograde delivery has not been demonstrated. Using intraoperative transesopha- geal , we examined retrograde delivery of cardioplegic solutions by contrast echocardiography, which directly assesses myocardial perfusion. In 15 patients (seven having coronary bypass and eight having valve operations), 4 ml of sonicated Isovue medium was injected retrograde via a coronary sinus catheter. Myocardial perfusion was assessed quantitatively by visual inspection and back- ground-subtracted videodensitometric analysis. In five patients undergoing aortic , right and left coronary ostial drainage was estimated during retrograde infusion. Before the aortic erossclamp was removed, myocardial oxygen extraction was calculated in all 15 patients by first delivering warm blood cardiople- gic solution for 2 minutes in a retrograde fashion and then taking samples from the cardioplegia line and aortic root. This determined the oxygen extraction ratio across the myocardium at the end of retrograde delivery. Warm blood cardioplegic solution was next given antegrade, and 15 seconds later samples were taken from the cardioplegia line and a right ventricular (acute marginal) to determine the oxygen extraction ratio across the right ventricle. As assessed by contrast echocar- diography, retrograde infusion resulted in almost four times more perfusion to the left ventricular free wall and septum than to the right ventricular free wall (74 - 2 versus 69 _ 2 versus 20 -+ 2, p < 0.05). In those five patients with an aortotomy the right ostial drainage was less than 5 ml/min whereas left ostial drainage was estimated at 80 ml/min during retrograde administration. Oxygen extraction across the myocardium supplied by retrograde infusion was low after 2 minutes. Con- versely, when antegrade cardioplegia was started, right ventricular oxygen extrac- tion rose fourfold (42% --- 5% versus 11% -+ 1%, p < 0.05), demonstrating that retrograde cardioplegia had not adequately perfused the right ventricular myocar- dium. Conclusions: 1. Retrograde cardioplegia provides poor right ventricular myocardial perfusion as assessed by contrast echocardiography and coronary ostial drainage. (2) This poor perfusion is inadequate to meet myocardial demands as

Bradley S. Allen, MD (by invitation), Jacqueline W. Winkelmann, BS (by invitation), Hanafy Hanafy, MD (by invitation), Renee S. Hartz, MD, Kirk S. Bolling, MD, MPH (by invitation), Jongwok Ham, BS (by invitation), and Steven Feinstein, MD (by invitation), Chicago, Ill.

From the Divisions of , and Cardiology, The odern cardioplegic solutions afford surgeons University of Illinois, Chicago, Ill. the ability to provide improved myocardial pro- Read at the Seventy-fourth Annual Meeting of The American M tectionJ' 2 However, to be effective, they must be Association for Thoracic Surgery, New York, N.Y., April 24-27, 1994. adequately distributed to all areas of the . 1' 3, 4 Address for reprints: Bradley S. Allen, MD, University of Illinois, The assurance of adequate cardioplegic distribution Cardiothoracic Surgery Department, 840 S. Wood Street, 518-H (M/C 958), Chicago, IL 60612. Copyright © 1995 by Mosby-Year Book, Inc. J THORA¢CARDIOVASC SURG 1995;109:1116-26 0022-5223/95 $3.00 + 0 12/6/63006

1116 The Journal of Thoracic and Cardiovascular Surgery Allen et al. 1117 Volume 109, Number 6

demonstrated by the high right ventricular oxygen extraction after a prolonged retrograde infusion. (3) Therefore surgeons must not rely solely on retrograde cardioplegia for right ventricular myocardial protection. This concept is especially important if continuous warm blood cardioplegia is used, because myocardial requirements are then higher. (J THORAC CARDIOVASC SURG 1995;109:1116-26) is especially important in patients with coronary We therefore studied right ventricular myocardial disease, because maldistribution of flow is the rea- perfusion in patients during retrograde cardioplegic son for the operation. Studies show it is safer to delivery using (1) contrast echocardiography, a safe clamp the for 4 hours with good cardioplegic reliable method of assessing transmural and seg- distribution than for as little as 30 minutes when the mental blood flow distribution patterns in myocar- same cardioplegic solution is given without attempts dium, (2) observations of coronary ostial effluent, to deliver it beyond coronary stenosis. 4 This prob- and (3) metabolic analysis of right and left ventric- lem of maldistribution of cardioplegic solutions is ular oxygen uptake. particularly evident with antegrade delivery, because nonhomogenenous distribution of cardioplegic solu- Methods tion is known to occur in the presence of coronary Institutional approval and informed written consent stenosis) -5 This has resulted in surgeons pursuing were obtained in 15 patients (eight men, seven women; alternative methods of cardioplegic administration aged 27 through 79 years, mean age 54 years). From this to ensure adequate perfusion to all areas of the group eight patients underwent valve operations (five , two , and one myocardium. mitral and aortic valve replacement), four underwent Several studies have demonstrated the distribu- coronary artery bypass grafting (CABG), and three tion advantages of retrograde administration in the CABG and valve operations. Two-dimensional trans- settings of left coronary stenosis and aortic valve esophageal echocardiography was used during the entire regurgitation and when internal mammary artery surgical procedure in all patients. After systemic heparinization, grafts are used. 5-7 Partington and associates 7' s stud- was established by means of an aortic cannula and either ied the distribution of retrograde perfusion in dogs single right atrial (coronary bypass or aortic valve proce- by mixing radioactive microspheres with a blood dures) or bicaval cannulation (mitral valves). In all pa- cardioplegic solution and administering it via the tients an antegrade cannula (DLP, Ann Arbor, Mich.) was placed in the aortic root and a retrograde cannula (Ret- coronary sinus. Despite coronary stenosis, they ob- roplegia; Research Medical, Salt Lake City, Utah) was served good left ventricular perfusion, although the positioned via the coronary sinus in the midportion of the right ventricle received inconsistent and sometimes great cardiac vein as previously described. 5 Patients were scant flow.7' s However, the clinical significance of cooled to 28° C, and the heart was arrested with either this poor right ventricular perfusion has been ques- warm (patients in unstable condition) or cold blood cardioplegic solution by means of methods reported pre- tioned because of differences in coronary venous viously.5, 13,14 anatomy between animals and human beings. Simi- Retrograde contrast injections were performed during larly, most clinical studies have also focused primar/- the initial cold blood cardioplegic infusion in all 15 ily on the ability of retrograde delivery to provide patients. In five patients without coronary heart disease (mitral valve repair only) an antegrade contrast infusion cardioplegic solution to the left ventricle, 5' 9, 10 leav- was also performed. Sonicated Isovue medium (E. R. ing the question of right ventricular perfusion un- Squibb & Sons, Inc., Princeton, N.J.) was used as the answered. However, inadequate right ventricular contrast agent. Sonication of the solution was performed protection may lead to an unexpectedly low post- according to the methods first introduced by Feinstein and associates. 15 A commercial sonicator (Heat Systems, operative cardiac output despite good preservation Ultrasonics, Plainfield, N.J.) was used to sonicate 8 ml of of left ventricular function, rl' 12 Although the pa- Isovue medium and produced a medium for contrast tient usually recovers, this complication may pro- enhancement, which was kept sterile in a 10 ml syringe long the postoperative period, increasing hospital and passed to the surgeon by pouring the sonicated costs. In addition, as more surgeons begin to rely contents into a sterile basin. The surgeon then aspirated the contents into another sterile syringe and manually primarily on retrograde delivery for cardioplegic injected 4 ml of contrast agent into the retrograde cannula distribution, the question of right ventricular perfu- via a stopcock. The contrast medium was flushed into the sion assumes increasing importance. muscle by the blood cardioplegi c solution, which was The Journal of Thoracic and 1118 Allen et al. Cardiovascular Surgery June 1995 given at a rate of 200 to 250 ml/min (pressure < 50 mm infusions. In the five patients having aortic valve opera- Hg). A similar method was used for the antegrade injec- tions the amount of blood draining from the right and left tions. coronary ostia during retrograde infusion was noted. Transesophageal images of the right ventricular free After the last coronary anastomosis (CABG), aorto- wall, interventricular septum, and left ventricular free wall tomy closure (aortic valve), or left atrial closure (mitral), were obtained with a 5.0 MHz transducer by means of a an infusion of warm (37 ° C) substrate-enriched blood commercial ultrasound scanner (Acuson, Mountainview, cardioplegic solution (200 to 250 ml/min) was given Calif.). Venous return was partially occluded to slightly retrogradely for 2 minutes (pressure < 50 mm Hg). At the distend the right ventricle for improved ultrasonic visual- end of 2 minutes, 5 ml of blood was taken simultaneously ization. Images were recorded on half-inch videotape from the aortic root and cardioplegia line to determine before and for approximately 2 minutes after each con- the oxygen uptake across the myocardium supplied by trast injection. Optimal power and gain settings were retrograde delivery. Antegrade cardioplegia was then obtained for each patient and were kept constant through- started at a rate of 200 ml/min and 15 seconds later a out the rest of the study. In some cases, contrast injections sample was obtained from the cardioplegia line and a were repeated. Contrast enhancement was immediately coronary vein (1 to 2 ml) on the surface of the right noted by the surgeon, but recorded images were analyzed ventricle (acute marginal vein). This determined the oxy- off-line by videodensitometric analysis. gen uptake across the right ventricle. The warm antegrade An ultrasonic video acquisition system was used to infusion was then continued for a total of 2 minutes. The measure pixel intensity in a given region of interest. Video crossclamp was removed, and the operation was finished. data from the half-inch videotape was analyzed by a Samples were analyzed for hemoglobin and oxygen by the framegrabber computer. The computer system used a use of an IL Co-Oximeter analyzer (Instrumentation series 151 image processor (ITEX-150; Imaging Technol- Laboratories, Lexington, Mass.) and Radiometer blood ogy, Inc., Bedford, Mass.), which converts the video image gas analyzer (Radiometer A/S, Copenhagen, Denmark). into a pixel matrix (512 by 512 pixels per frame), making Oxygen content was calculated as follows: it possible to quantify the pixel intensity in gray levels of a predetermined region of interest. These intensities were 0 2 content = 1.36 ml Ojgm × Hgb × 02 saturation generated within the region of interest in real time at a rate of 30 frames per second. Time versus intensity curves + (0.003 × Po2) were generated for each region. In this study, regions of interest were placed in the right ventricular wall, the where Hgb is hemoglobin and Po2 is oxygen tension. interventricular septum, and the left ventricular free wall. Oxygen extraction ratio across the myocardium was cal- Background-subtracted analysis was performed; peak culated as follows: pixel intensity in gray levels was obtained during contrast injections and was subtracted from baseline (before injec- 0 2 extraction ratio tion) intensity, and the difference in pixel intensity was reported. 02 content (cardioplegia) - 0 2 (distal sample) The operation was then conducted according to a cardioplegic strategy described previously. 5' 13,14 The de- 0 2 contents (cardioplegia) tails will only be summarized here to illustrate a slight where the distal sample was the aortic root when retro- variation in cardioplegic administration to that previously grade cardioplegia was in effect and the acute marginal reported. All patients received cold blood cardioplegic (right ventricle) coronary vein when antegrade cardio- solution for 4 minutes (2 minutes antegrade and 2 minutes plegia was in effect. retrograde) either as an induction dose or, in patients in Statistical analysis was done with the aid of unstable condition, after warm induction. 14" 16 In patients Statistics. the Statistical Analysis Systems (Cary, N.C.). Data are undergoing CABG each distal anastomosis was per- expressed as mean + the standard error of the mean. formed during cold ischemic arrest. Next, the proximal Group data were compared by Student's t test. Significant anastomosis was constructed with the use of a continuous differences were defined as a probability for each test of retrograde infusion of cold (4 ° C) blood cardioplegic p < 0.05. solution. After completion of the proximal aortic anasto- mosis, a 1-minute infusion of cold blood cardioplegic solution was given antegradely before the next distal Results anastomosis. This procedure was repeated for each graft, All patients survived the procedure and none had with the internal mammary artery graft being done last to allow the infusion of a warm blood cardioplegic reperfu- major postoperative complications. sate ("hot shot") after this anastomosis. In patients with Eehocardiographic observations. Perfusion was valve disease cold (4 °) blood cardioplegic solution was immediately assessed by observing the ultrasound continuously infused (150 to 200 ml/min) in a retrograde monitor. This assessment alone clearly demon- fashion for almost the entire operation. The only time strated better distribution of cardioplegic solution to retrograde administration was stopped was during place- ment of the posterior leaflet (mitral) or left commissure the septum and left ventricular free wall than to the sutures (aortic) or for intermittent (approximately every right ventricle. In one patient a retrograde contrast 20 to 30 minutes) cold antegrade blood cardioplegic injection did not produce any myocardial enhance- The Journal of Thoracic and Cardiovascular Surgery Allen et al. 1119 Volume 109, Number 6

80 _ ""T-- --V-

60

Videointensity 40 (Gray levels)

"/c 20

Freewall Septum Freewall LEFT VENTRICLE RIGHT VENTRICLE

Fig. 1. Myocardial perfusion (assessed by contrast echocardiography) of the left ventricle, septum, and right ventricle during retrograde delivery of cardioplegic solution in 12 patients. Perfusion of the left ventricle and septum is three to four times greater than that of the right ventricle (mean -+ standard error). *p < 0.05.

ment. After the retrograde cannula had been repo- values were not statistically different from those in sitioned, a second contrast injection did produce patients with coronary disease who had retrograde enhancement, confirming that the catheter origi- perfusion (see Figs. 1 and 3). nally had been malpositioned. In three patients, Ostial observations. In all five patients undergo- complete analysis was not possible because of sub- ing aortic valve replacement, flow emanating from optimal images of the left ventricular free wall. the right coronary ostium was less than 5 ml/min Background-subtracted pixel intensity was mea- during retrograde cardioplegic infusion. In two pa- sured in the right ventricular free wall, septum, and tients the ostial effluent was essentially zero. Con- left ventricular free wall in 12 patients. In the 12 versely, all five patients had abundant flow from the cases analyzed, perfusion to the left ventricular free left coronary ostium which was measured in two wall and septum was three to four times greater than patients and found to be 74 and 82 ml/min. perfusion to the right ventricular free wall (74 _+ 2 Metabolic observations. Fourteen patients were versus 69 _ 2 versus 21 _+ 2, p < 0.05) (Fig. 1). available for metabolic studies during the terminal In five patients with no coronary disease under- warm infusion ("hot shot"); in one patient an ade- going mitral valve repair, an antegrade contrast quate right coronary vein was not identified. At the injection was performed through the aortic root end of the 2-minute retrograde infusion of warm immediately after the retrograde injection. In all five cardioplegic solution, the oxygen extraction was patients, antegrade injection of cardioplegic solu- 11% _+ 1% across the myocardium supplied by tion perfused all regions (left ventricle, right ventri- retrograde administration. Conversely, when ante- cle, and septum) equally (Fig. 2). Conversely, in grade delivery of cardioplegic solution was begun these same five patients retrograde administration immediately after this infusion, oxygen extraction by demonstrated poor perfusion of the right ventricular the right ventricle increased fourfold (42% _ 5%, free wall when compared with the septum and left p < 0.05) (Fig. 4). ventricular free wall (p < 0.05) (Fig. 3). Pixel intensities for the left ventricular free wall, the Discussion septum, and the right ventricular free wall were This study confirms animal experimental evi- 76 _+ 2, 71 _ 3, and 21 + 3, respectively. These dence 7, s, 17 and demonstrates that retrograde car- The Journal of Thoracic and 1120 Allen et aL CardiovascularSurgery June 1995

80_

60 - --r--

Videointensity 40 (Gray levels)

20

i

Freewall Septum Freewall LEFT VENTRICLE RIGHT VENTRICLE Fig. 2. Myocardial perfusion (assessed by contrast echocardiography) of the left ventricle, septum, and right ventricle during antegrade cardioplegic delivery in five patients without coronary disease. All areas of the myocardium are perfused equally (mean + standard error).

80- --T--

60-

Videointensity 40- (Gray levels)

~e 20-

Freewall Septum Freewall LEFT VENTRICLE RIGHT VENTRICLE

Fig. 3. Myocardial perfusion (assessed by contrast echocardiography) of the left ventricle, septum, and right ventricle during retrograde cardioplegic delivery in five patients without . The right ventricle receives one third to one fourth the peffusion of the septum and left ventricular free wall despite the absence of coronary disease (mean _+ standard error). *p < 0.05. dioplegic administration provides inadequate and demonstrated by (1) contrast echocardiography (see poor right ventricular perfusion in the clinical set- Figs. 1 and 3), (2) the lack of right coronary ostial ting. The inability of retrograde delivery to provide effluent during retrograde infusion, and (3) the high substantial right ventricular myocardial perfusion is oxygen uptake by the right ventricle when antegrade The Journal of Thoracic and Cardiovascular Surgery Allen et aL 1121 Volume 109, Number 6

60m

I Oxygen 40 Extraction Ratio (%)

20

End Retrograde Start Antegrade (RV Vein) Fig. 4. Oxygen extraction across the myocardium supplied by retrograde administration at the end of 2 minutes of warm blood cardioplegic infusion (stippled bar) and by the right ventricular myocardium when antegrade cardioplegia was started immediately after this retrograde infusion (solid bar). Note the fourfold increase in myocardial oxygen extraction by the right ventricle when antegrade infusion is started, demonstrating the lack of adequate right ventricular perfusion with retrograde delivery. See text for details (mean +- standard error). *p < 0.05.

delivery was given after a prolonged retrograde ular myocardial perfusion with retrograde car- infusion had repaid the energy debt to its distribu- dioplegic administration. Because retrograde deliv- tion (see Fig. 4). ery will perfuse left ventricular myocardium in the Intraoperative contrast echocardiography has presence of this coronary stenosis, 5' 7-9 one would been shown to be a safe, reliable method of assess- expect a similar phenomenon in the right ventricle. ing myocardial perfusion. 9' lo Sonicated contrast However, to discount this as the reason for poor agents (i.e., Isovue medium) produce small, highly right ventricular perfusion, we examined five pa- reflective microbubbles that, unlike microspheres, tients without coronary artery disease undergoing are capable of crossing the capillary microcircula- only mitral valve repair. These same five patients tion, making it possible to directly assess myocardial also had an antegrade contrast injection performed perfusion. 15 Recently, Aronson and colleagues 9 to validate our method of imaging the right ventri- used contrast echocardiography to evaluate myocar- cle. As expected, in the absence of coronary artery dial distribution of cardioplegic solution delivered in disease, antegrade administration results in homo- an antegrade or retrograde manner in human geneous distribution of cardioplegic solution to all . This study confirmed previous laboratory areas of the heart (see Fig. 2), confirming antegrade experimentation 7, s, 17, 18 and concluded that retro- perfusion of the right ventricle and our method of grade cardioplegia provides perfusion to all left imaging. Conversely, despite the absence of coro- ventricular myocardial regions, regardless of coro- nary disease, right ventricular perfusion was poor nary stenosis. However, assessment of right ventric- with retrograde delivery in all five of these patients ular myocardium was not possible because of inad- (see Fig. 3). Indeed, right ventricular perfusion was equate echocardiographic images. identical to that seen in patients with coronary The present study differed in that we primarily disease (see Figs. 1 and 3), indicating that coronary used contrast echocardiography to examine right stenosis does not play a role in distribution of ventricular perfusion by slightly distending the right cardioplegic solution to the right ventricle via retro- ventricle to obtain good images. As shown in Fig. 1, grade infusion. contrast echocardiography shows poor right ventric- In the five patients undergoing aortic valve re- The Journal of Thoracic and 1 1 2 2 Allen et al. Cardiovascular Surgery June 1995 placement, coronary ostial flow was measured dur- this right ventricular energy debt inasmuch as some ing retrograde cardioplegic delivery. No effluent retrograde flow drains into the right ventricle via from the right coronary ostium was apparent in two thebesian , y' 17,19, 22 Indeed, we did see (by con- patients, and in the other three drainage was less trast echocardiography) some drainage into the right than 5 ml/min. This lack of flow from the right ventricle, but we were unsure whether it was coming coronary ostium during retrograde infusion has from thebesian veins or simply cardioplegic solution been documented experimentally 17' 19 and is consis- that escaped the coronary sinus and traversed the tent with the poor right ventricular perfusion dem- tricuspid valve. This thebesian drainage helps cool the onstrated by contrast echocardiography. Conversely, fight ventricle, thereby reducing its energy demands all patients had abundant flow emanating from the and limiting ischemic damage. 5' 7 However, it cannot left coronary ostium. Flow was measured in two and repay the energy debt, inasmuch as it is nonnutri- found to be approximately 80 ml/min, confirming tive,5, 7 reaching the thesbesian veins only through good retrograde catheter placement and left ventric- venovenous collaterals. Therefore, we would have ular perfusion. expected the energy debt incurred by the fight ventri- Although contrast echocardiography and the lack cle to be even greater had these patients received of right coronary ostial effluent both demonstrate continuous warm retrograde cardioplegia, because the poor right ventricular perfusion, they cannot predict metabolic demands are increased at 37° C, making this whether this small amount of right ventricular flow nonnutritive flow less important, a' s, 17, 22 is sufficient to meet the demands of the arrested Cardiac venous anatomy plays a key role in un- heart. To address this question, we performed met- derstanding our results. According to a postmortem abolic studies. At the end of each operation a study of 280 hearts by Ludinghausen, z3 only 21% to 2-minute infusion of warm (37 ° C) substrate-en- 25% of all hearts have a small cardiac vein. This vein riched blood cardioplegic solution was infused ret- courses in the sulcus between the right and rograde via the coronary sinus (200 ml/min, pressure the right ventricle and drains to the dorsal wall of < 50 mm Hg). Samples taken from the cardioplegia the right ventricle, entering the coronary sinus at its line and aortic root at the end of this 2-minute origin. Thus it is difficult to perfuse this vein with a infusion demonstrated low oxygen extraction (11%) balloon-tipped catheter. When the small cardiac across the myocardium supplied by retrograde ad- vein is missing (75% of all cases), drainage of the ministration. Indeed, the amount of oxygen extrac- right ventricle is limited to thebesian veins, which tion in this sample was equivalent to what we 16 drain directly into the ventricle, y' 17, 23 Therefore, in previously reported to represent basal metabolic the majority of patients, retrograde perfusion to the demands of the warm arrested heart. This indicates right ventricular free wall may be technically impos- that the left ventricle is sufficiently perfused by sible to accomplish via any type of coronary sinus retrograde cardioplegia to allow all the cellular cannula. processes that were damaged or depleted during Antegrade delivery avoids right ventricular isch- to be repaired. Conversely, when ante- emia and limits cardioplegic volume by produc- grade cardioplegia was started immediately after ing prompt myocardial arrest. 1' 5 However, the solu- this 2-minute infusion, samples taken from a right tion is distributed poorly in the presence of severe coronary vein 15 seconds later demonstrated almost coronary artery disease, especially in patients a fourfold increase in oxygen extraction by the right who receive arterial (internal mammary artery) ventricular myocardium (see Fig. 4). This high oxy- grafts.5, 7, 8 Other limitations of antegrade cardiople- gen uptake is consistent with what has been shown gia included inadequate distribution with aortic to occur when reversible ischemic myocardium is regurgitation, .the need to interrupt mitral valve reperfused and indicates that the right ventricular procedures to remove retractors, and ostial damage myocardium was insufficiently perfused by retro- during cardioplegic infusions in aortic valve opera- grade cardioplegia 1' 2o, 21 to allow the same cellular tions. 5 Retrograde (coronary sinus) cardioplegia off- process to be repaid. In addition, this right ventric- sets some of these problems and has been associated ular energy debt occurred despite a cardioplegic with a lower risk of myocardial damage and im- strategy that used prolonged periods of continuous proved postoperative performance. 1'5'24 Retro- cold (4 ° C) retrograde cardioplegie perfusion and grade cardioplegia also avoids trauma to the coro- intermittent antegrade infusions. nary and interferes minimally with the Our use of cold cardioplegia may have lessened surgical procedures. However, this study demon- The Journal of Thoracic and Cardiovascular Surgery Allen et aL 1123 Volume 109, Number 6 strates that only antegrade delivery provides ade- 8. Partington M, Acar C, Buckberg G, Julia P. Studies of quate nutritional right ventricular perfusion. There- retrograde cardioplegia. II. Advantages of antegrade/ fore both routes of administration are required to retrograde cardioplegia to optimize distribution in maximize the advantages while limiting the inherent jeopardized myocardium. J THOP,AC CARDIOVASC inadequacies each method possesses. We believe SURG 1989;97:613-22. 9. Aronson S, Lee B, Zaroff J, et al. Myocardial distri- adapting a rigid approach of using only one method bution of cardioplegic solution after retrograde deliv- of cardioplegia delivery is unwarranted and poten- ery in patients undergoing cardiac surgical proce- tially harmful, because it does not ensure distribu- dures. J THORAC CARDIOVASC SURG 1993;105:214-21. tion to all areas of the heart. In addition, because 10. Goldman M, Mindich B. Intraoperative cardioplegia only antegrade delivery adequately perfuses the contrast echocardiography for assessing myocardial right ventricle, we currently recommend that the perfusion during open heart surgery. J Am Coll right coronary artery be grafted first in patients with Cardiol 1984;4:1029-34. total occlusion so that cardioplegic solution can be 11. Mullen J, Fremes S, Weisel R, et al. Right ventricular distributed via the graft. function: a comparison between blood and crystalloid In summary, retrograde cardioplegia provides in- eardioplegia. Ann Thorac Surg 1987;43:17-24. adequate protection of the right ventricular myocar- 12. Morris JIII, Wechsler A. Right ventricular perfor- dium as shown by contrast echocardiography, obser- mance and protection. In: Cardiac surgery: state of vations of poor right coronary ostial effluent, and the art reviews. Philadelphia: Hanley & Belfus, Inc., metabolic analysis. Therefore surgeons must not 1988:303-29. rely solely on this method of myocardial protection 13. Allen B, Buckberg G, Fontan F. Superiority of con- trolled surgical reperfusion versus percutaneous if they are to assure optimal myocardial protection. transluminal coronary in acute coronary We thank Greg Mork, Jim Murray, and Nancy Bourtsos occlusion. J THORAC CARDIOVASCSURG 1993;105:864- for technical support and Felicia Mitchell for organiza- 84. tional assistance. 14. Allen B, Rosenkrantz E, Buckberg G, et al. Studies of prolonged regional ischemia. VI. Myocardial infarc- REFERENCES tion with left ventricular power failure: a medical/ 1. Buckberg G. Myocardial protection during adult car- surgical emergency requiring urgent diac operations. In: Baue A, Geha A, Hammond G, with maximal protection of remote muscle. J THORAC Laks H, Naunheim K, eds. Glenn's thoracic and CARDIOVASC aURa 1989;98:691-703. cardiovascular surgery. 5th ed. East Norwalk, Conn. 15. Feinstein S, Ten Cate F, Zwehl W, et al. Two- Appleton & Lange, 1991:1417-42. dimensional contrast echocardiography. I. In vitro 2. Buckberg G. A proposed "solution" to the cardiople- development and quantitative analysis of echo con- gic controversy. J THORACCARDIOVASC SURG 1979;77: trast agents. J Am Col Cardiol 1984;3:14-20. 803-15. 16. Hanafy H, Allen B, Winkelmann JW, Ham J, Hartz 3. Hilton C, Teubl W, Acker M, McEnany M. Inade- R. Warm blood cardioplegic induction: an underused quate cardioplegic protection with obstructed coro- modality. Ann Thorac Surg 1994;58:1589-94. nary arteries. Ann Thorac Surg 1979;28:323. 17. Gates R, Laks H, Drinkwater D, et al. The microvas- 4. Becker H, Vinten-Johansen J, Buckberg G, et al. cular distribution of cardioplegic solution in the piglet Critical importance of ensuring cardioplegic delivery heart: retrograde versus antegrade delivery. J THORAC with coronary stenoses. J THORAC CARDIOVASC SURG CARDIOVASC SURG 1993;105:845-53. 1981;81:507-15. 18. Aronson S, Lee B, Liddicoat J, et al. Assessment of 5. Buckberg G. Antegrade/retrograde blood cardiople- retrograde cardioplegia distribution using contrast gia to ensure cardioplegic distribution: operative tech- echocardiography. Ann Thorac Surg 1991;52:810-4. niques and objectives. J Cardiac Surg 1989;4:216-38. 19. Gates R, Laks H, Drinkwater D, et al. Gross and 6. Drinkwater D, Cushen C, Laks H, Buckberg G. The microvascular distribution of retrograde cardioplegia use of combined antegrade-retrograde infusion of in explanted human hearts. Ann Thorac Surg 1993; blood cardioplegic solution in pediatric patients un- 56:410-7. dergoing heart operations. J THORAC CARDIOVASC 20. Rosenkrantz E, Okamoto F, Buckberg G, et al. Safety SURG 1992;104:1349-55. of prolonged aortic clamping with blood cardioplegia. 7. Partington M, Acar C, Buckberg G, Julia P, Kofsky E, III. Aspartate enrichment of glutamate blood cardio- Bugyi H. Studies of retrograde cardioplegia. I. Capil- plegia in energy-depleted hearts after ischemic and lary blood flow distribution to myocardium supplied reperfusion injury. J THORAC CARDIOVASCSURG 1986; by open and occluded arteries. J THORACCARDIOVASC 91:428-35. SURO 1989;97:605-12. 21. Rosenkrantz E, Vinten-Johansen J, Buckberg G, et al. The Journal of Thoracic and 1124 Allen et aL Cardiovascular Surgery June 1995

Benefits of normothermic induction of cardioplegia in can be preserved despite the absence of substantial nutri- energy-depleted hearts, with maintenance of arrest by tive perfusion. However, we still believe all patients multidose cold blood cardioplegic infusions. J THO- should be given antegrade cardioplegia for two reasons. RAC CARDIOVASCSURG 1982;84:667-76. First, a terminal warm reperfusate ("hot shot") has been 22. Salerno T, Houck J, Barrozo C, et al. Retrograde shown to lower operative mortality and improve postop- continuous warm blood cardioplegia: a new concept erative function by counteracting any deficiencies in myo- in myocardial protection. Ann Thorac Surg 1991;51: cardial protection and preventing a reperfusion injury. As demonstrated by this study, only antegrade administration 245-7. provides the right ventricle with sufficient flow to benefit 23. Lundinghausen M. Nomenclature and distribution from this infusion. Second, antegrade cardioplegia results pattern of cardiac veins in man. In: Mohl W, Faxon D, in quicker myocardial arrest, thereby limiting cardioplegic Wolher E, eds. Proceedings of the Second Interna- volumes and right ventricular ischemia during cardiople- tional Symposium on Myocardial Protection via the gic induction. Coronary Sinus. New York: Springer-Verlag, 1986:13- Dr. Lawrence I. Bonchek (Lancaster, Pa.). Since the 32. early days of cold cardioplegia ill the 1970s, we have been 24. Buckberg GD, Beyersdorf F, Kato NS. Technical concerned about right ventricular protection, particularly considerations and logic of antegrade and retrograde ill patients with pulmonary hypertension. Our concern was blood cardioplegic delivery. Semin Thorac Cardiovasc heightened in those days by two patients who died after Surg 1993;5:125-33. double valve replacement with clear-cut right ventricular failure despite the use of topical saline hypothermia. We therefore devised a cooling jacket now known as the DLP Discussion Lancaster cooling jacket, which surrounds and cools the Dr. John M. Kratz (Charleston, S.C.). Your comments entire heart and eliminates many of the difficulties of regarding warm continuous cardioplegia are well taken, topical hypothermia and the need for frequent or ante- and I agree that retrograde infusion may not deliver an grade perfusion. adequate flow to the right ventricle. However, most of us I realize your preference may be for warm cardioplegia. are practicing intermittent cold retrograde perfusion to Have your findings in this study affected your clinical the right ventricle. Although we all know from our simple practice? What is your current clinical practice? That is of fingertip thermistors that are used every day in the interest, particularly for patients who require prolonged operating room that it takes a little longer to cool the right procedures such as replacement of the . side of the heart than the left, we know the right side does Dr. Allen. Yes, this study has changed our clinical cool. practice. We now believe that the right ventricle is rela- I would like to remind you of a study that my colleagues tively underperfused by retrograde cardioplegia and so and I presented before the Southern Thoracic Surgical always use antegrade administration to ensure adequate Association and published in The Annals of Thoracic perfusion. The only time that we allow more than 20 Surgery in 1992. This study evaluated right ventricular minutes to elapse without an antegrade infusion is if we ejection fractions during use of retrograde versus ante- are using continuous cold retrograde perfusion, because grade cold intermittent blood cardioplegia. this probably provides enough protection for the right We demonstrated that right ventricular ejection frac- ventricle by maintaining hypothermia. However, we al- tion falls mildly over the first hour or so after separation ways try to deliver some cardioplegic solution antegradely, from bypass with both antegrade and retrograde cardio- especially during the terminal warm reperfusion period plegia. With retrograde cardioplegia right ventricular ("hot shot"), because this study has shown that retrograde ejection fractions were just as good and at some points administration does not adequately repay the right ven- slightly better than when antegrade cardioplegia alone tricular energy debt that develops despite good hypother- was used. I think that although the flow to the right mic cardioplegic protection. As demonstrated in Fig. 4, ventricle is slower, that problem can be overcome with just even though all of our patients received prolonged con- a bit more volume of eardioplegic solution being given, tinuous cold retrograde and intermittent antegrade infu- and one can rely on intermittent cold retrograde cardio- sions of cardioplegic solution, we still saw a right ventric- plegia alone to protect the right ventricle. ular energy debt at the end of the operation, which was Dr. Allen. In general, Dr. Kratz, I would agree. We repaid only during antegrade administration. This energy recently studied continuous retrograde cardioplegia at debt probably occurred because (1) even though we used various flow rates and examined recovery of right ventric- prolonged periods of continuous cold retrograde cardio- ular function after 2 hours of aortic crossclamping. In plegia that cooled the myocardium, it still did not provide animals protected with continuous warm retrograde car- much nutritive flow, and (2) we still had some short dioplegia, right ventricular function was depressed despite intervals of cold ischemic arrest when optimal visualiza- flow rates up to 400 ml/min. Conversely, when we used tion was needed. Regarding topical hypothermia, we continuous cold retrograde blood cardioplegia, even at published several years ago the detrimental effects of much lower flow rates, right ventricular stroke work index topical ice on the phrenic nerve. In addition, this same was well maintained. We therefore concluded that al- study demonstrated no improvement in myocardial pro- though the right ventricle is poorly perfused by the tection when topical cooling was added to combined retrograde cardioplegia technique, if the metabolic de- antegrade/retrograde cardioplegic protection. We there- mands are kept low by cooling the myocardium, function fore believe topical cooling, is necessary only when the The Journal of Thoracic and Cardiovascular Surgery Allen et al. 1 1 2 5 Volume 109, Number 6 patient does not have an open right coronary artery plete preservation of right ventricular function. Second, because at other times it is very easy to cool the right most studies of continuous warm retrograde cardioplegia ventricle via the antegrade route. We therefore use topical have examined primarily determinants of left ventricular cooling (l) when we are unable to use antegrade cardio- function (i.e., use of intraaortic balloon pump, postoper- plegia (i.e., severe aortic insufficiency) or (2) in patients ative cardiac output, inotropic agents). However, many with total occlusion of the right coronary artery, and then surgeons have seen high central venous pressures in the only until we have constructed a vein graft, at which point range of 12 to 14 cm H20 after continuous cardioplegic we use this new graft to distribute our antegrade car- protection has been used. This indicates that although the dioplegic solution. right ventricle is functioning, it is not functioning normally Dr. William D. Spotnitz (Charlottesville, Va.). I congrat- and is probably shifted to a point further out on its ulate Dr. Allen and his coauthors for their clinical use of Starling function curve. We believe that this poor right the emerging technology of contrast echocardiography. At ventricular protection is totally avoidable by simply using the University of Virginia, we have demonstrated in a dog the antegrade route to infuse cardioplegic solution inter- model of retrograde coronary sinus cardioplegia that the mittently and as a terminal warm infusion. We do not interventricular septum is underperfused as measured by understand why surgeons continue to take the adversarial a gold standard of radiolabeled microspheres, as well as by approach of using only antegrade or retrograde delivery. contrast echocardiography. This work confirmed the abil- Each method of administration has advantages and dis- ity of contrast echocardiography to qualitatively evaluate advantages and only by using both methods can the ad- the retrograde delivery of cardioplegic solution. vantages of each be maximized. Therefore, we use retro- Use of radiolabeled microspheres, at our institution, grade delivery to distribute cardioplegic solution beyond to quantitatively evaluate microvascular flow has con- coronary stenoses, especially with arterial grafts, and firmed that all cardioplegic solution delivered antegradely antegrade cardioplegia to assure perfusion of the right reaches the microvasculature. However, significant de- ventricle, obtain rapid cardiac arrest, and avoid right creases in microvascular flow were noted with retrograde ventricular ischemia during induction. delivery of cardioplegic solution. In fact, using extraction Dr. Robert A. Guyton (Atlanta, Ga.). I have a question fraction of thallium and technetium as markers of tissue about the oxygen extraction levels. The normal oxygen perfusion revealed even larger reductions of cellular flow extraction in a normal beating heart is 60% to 65%. I with retrograde cardioplegia. This occurs as a result of contend that the 11% oxygen extraction that you obtained thebesian veins. between the cardioplegia line and the aorta is a reflection Although retrograde perfusion results in decreased of the presence of noncoronary collateral flow, mixing of microvascular and cellular flow, measurements of myocar- blood that has come through the pulmonary vessels with dial temperature revealed similarly excellent cooling by cardioplegic blood. The oxygen extraction of 42% in the retrograde and antegrade techniques. right ventricle gives us an indication that the protection of Thus my question concerns the excellent clinical results the right ventricle is quite adequate, because it is better obtained with retrograde cardioplegia. Is this improve- than the extraction of the normal beating heart. Can you ment the result of good cooling of the right ventricle or respond to that? I agree with you that the right ventricle some other phenomenon that allows adequate right ven- is relatively underperfused, but your data support ade- tricular protection despite the decreased perfusion of quate metabolic flow in the right ventricle, not inadequate contrast material you have demonstrated? metabolic flow. Dr. Mien. Dr. Spotnitz, I believe that one must examine Dr. Mien. Dr. Guyton, with regard to the oxygen the distribution of retrograde cardioplegic solution to the extraction across the myocardium supplied by retrograde right and left ventricles separately. Most experimental and cardioplegia, we believe this represents primarily the clinical studies demonstrate that in the presence of coro- metabolic demands of the left ventricle, because only the nary stenosis, retrograde administration provides good left coronary ostium had significant drainage during ret- perfusion of the left ventricle. Therefore, since surgeons rograde administration. As to the 11% oxygen extraction are performing an increasing number of arterial grafts in obtained during retrograde administration, these mea- which antegrade cardioplegic solution cannot be distrib- surements were taken in the arrested flaccid heart, not the uted down the new graft, we believe retrograde cardiople- beating working heart, so one would expect the oxygen gia is warranted in most cases. extraction to be approximately 11%. This is what we and The question is how well retrograde administration others have shown represents basal metabolic demands of protects the right ventricle and why some surgeons are the warm arrested heart. Therefore, we believe that reporting good clinical results with continuous warm retrograde cardioplegia provides sufficient flow to the left retrograde cardioplegia. First, although this was not a ventricle to allow the depleted cellular processes to be functional study, we did examine right ventricular function repaired, so that now it requires only enough oxygen to by having an independent cardiologist review the echocar- meet its basal metabolic demands. diograms postoperatively in all patients. Right ventricular Oxygen extraction across the right ventricle was mea- wall motion was then graded from zero (normal) to four sured during antegrade cardioplegic administration im- (dyskinesis) before CABG and immediately after bypass mediately after the retrograde infusion, also in the ar- was discontinued. The results demonstrated improvement rested heart. Therefore, we once again did not expect the in right ventricular wall motion after bypass. Therefore we oxygen extraction to be 65%, but approximately the same can conclude that a cardioplegic strategy that combines 11% we saw at the end of the retrograde cardioplegic both antegrade and retrograde delivery results in corn- infusion. However, the fourfold increase in myocardial The Journal of Thoracic and 1126 Allen et aL Cardiovascular Surgery June 1995 oxygen extraction indicates that, unlike the left ventricle, patients underwent only a valve procedure. We did this the right ventricle still had an energy debt that had not because we wanted to exclude coronary stenosis as a been repaid, despite the prolonged retrograde infusion. problem. Later in the study we did include seven patients I should also mention that although we did not measure undergoing CABG, but we did not specifically examine lactate levels, because of lack of sample, we did measure each heart to see which had a small cardiac vein. However, pH across the myocardium during the warm cardioplegia as you have pointed out, those that have a small cardiac infusions ("hot shot"). At the end of the retrograde vein do tend to have fewer right ventricular acute mar- infusion, the pH was unchanged across the left ventricular ginal veins, as do patients with a small right coronary myocardium and the oxygen extraction was at basal met- artery. abolic levels. Conversely, when antegrade cardioplegia Dr. Randas J. V. Batista (Curitiba, Brazil). Our experi- was next begun there was an acidotic washout from the ence has been different from yours. We use retrograde right ventricle and the oxygen extraction rose fourfold. continuous warm blood cardioplegia with a Foley cathe- This further supports our contention that the right ven- ter, and we routinely open the right atrium and place a tricular myocardium was underperfused by retrograde purse-string suture around the ostium of the coronary administration and could only have it's energy debt repaid sinus. We can see a posterior coronary vein and other by antegrade administration. right ventricular veins draining into the coronary sinus To make sure we were not having sampling errors from right at the mouth of the coronary sinus. the aortic root (i.e., sampling pulmonary bronchial collater- From your presentation, I believe you use balloon als or noncoronary collateral flow), we analyzed hematocrit catheters. These catheters must be placed distally into values for each sample. Because we used cardioplegic solu- the coronary sinus; otherwise these right ventricular tion at a 4 to 1 ratio, the hematocrit value of cardioplegic veins are not perfused and the right ventricle is not blood was slightly lower than that of systemic blood. There- protected. fore, if the hematocrit value was higher, indicating that we Dr. Allen. Dr. Batista, I will answer your question in two had some mixing, we discarded the sample. different ways. First, Dr. Laks, Dr. Drinkwater, and the Because of the possibility of this occurring, two samples group from UCLA have done two studies in which they were taken from the aorta root in each patient and were put a purse-string suture around the coronary sinus and compared to make sure the results were similar. In perfused the heart in a retrograde fashion. Despite not addition, many of these patients had left ventricular vents using a balloon-tipped catheter, they still saw relatively (i.e., aortic valves) or the left atrium was still opened poor perfusion of the right ventricle in human hearts. (mitral valves) during the retrograde infusion, which Second, if you look at postmortem studies, only 21% to further prevented aortic sampling problems. 25% of people have a small cardiac vein, in which case you Dr. John L. Oehsner (New Orleans, La.). In these 15 are going to have to do what you are saying and cannulate patients, did you look at the pattern of the coronary a lot of right atrial veins separately. To me it seems very anatomy? In the case of a diminutive right coronary artery complicated to double cannulate, put a purse-string suture and hyperdominant left, in which the left posterior de- in the coronary sinus, look for multiple draining right scending artery branches off and the veins are always atrial veins, and still worry about whether you have present, you would expect to have much better flow to the cannulated all of the important right ventricular drainage right side on retrograde perfusion than you would in vessels. In contrast, it seems much simpler to give inter- patients with a large right coronary artery. mittent antegrade infusions. This approach is important Dr. Allen. No, Dr. Ochsner, we did not look at varia- not only for right ventricular protection, but for limiting tions in coronary anatomy. We originally started this study cardioplegic volume and preventing right ventricular isch- in patients with valve disease. As you saw, eight of the 15 emia during cardioplegic induction.