Is Reperfusion Injury from Multiple Aortic Cross-Clamping a Current Myth of Cardiac Surgery?

Transcription

1 ORIGINAL ARTICLES Is Reperfusion Injury from Multiple Aortic Cross-Clamping a Current Myth of Cardiac Surgery? David M. Lolley, M.D., Jefferson F. Ray, 111, M.D., William 0. Myers, M.D., Richard D. Sautter, M.D., and Gregory Sheldon, P.A. ABSTRACT Four hundred eighty adult patients undergoing cardiac operations had systemic and topical hypothermic anoxic arrest supplemented with potassium chloride pharmacological cardioplegia in a prospective randomized study. Group 1 (217 patients) had continuous aortic cross-clamping and one single anoxic arrest period during the cardiac portion of the operation; which resulted in a transmural myocardial infarction rate of 8.3%, myocardial injury incidence of 12.4% 4.6% cardiacrelated deaths, 11.5% and 24.8% severe and malignant ventricular arrhythmias, 21.7% rate of severe vasopressor usage, a mean group serum glutamic oxaloacetic transaminase (SGOT) of 140 f 39 IU, and a mean group lactic dehydrogenase (LDH) of 636 * 78.2 IU. Group 2 (263 patients) had intermittent aortic cross-clamping with multiple reperfusion intervals, which resulted in a significantly lower incidence of transmural myocardial infarction at 1.9% (p < 0.01), rate of myocardial injury at 5.66% (p < 0.02), number of cardiac deaths at 0.76% ( p < 0.02), 8.7% and 16.0% severe and malignant ventricular arrhythmias (p < 0.01), severe vasopressor utilization rate of 14.3% (p < 0.05), mean group SGOT at 72.0 f 3.1 IU (p < 0.01), and mean group LDH at f 12.3 IU (p < 0.05) than Group 1. These results do not support the contention that intermittent aortic cross-clamping in conjunction with hypothermia and pharmacological cardioplegia leads to increased clinical cardiac damage compared with continuous aortic cross-clamping. The converse is implied, in that the anoxic heart may benefit from the physiological effects of briefly reperfused oxygenated blood. From the Department of Cardiovascular and Thoracic Surgery, Marshfield Clinic, St. Joseph Hospital, and the Marshfield Medical Foundation, Inc., Marshfield, WI. Presented at the Twenty-sixth Annual Meeting of the Southern Thoracic Surgical Association, Nov 1-3, 1979, San Antonio, TX. Address reprint requests to Dr. Lolley, Trover Clinic, Clinic Dr, Madisonville, KY With the advent of cardiac hypothermia and pharmacological cardioplegia, workers in many centers have advocated the use of prolonged uninterrupted anoxic cardiac arrest utilizing one period of aortic cross-clamping without interval reperfusion during cardiac surgery [l, 2, 4, 5, 11, 12, 141. Intermittent aortic crossclamping, in spite of its being the time-honored method of cardiac preservation [201, has been abandoned, presumably to avoid increased reperfusion edema and injury alleged to occur after multiple perfusion intervals. This new approach is based on the belief that the summation of cardiac injury sustained during multiple cross-clampings and brief reperfusion intervals is greater than that occurring for the same anoxic arrest time during continuous aortic cross-clamping 11, 4, 7, 81. Some of the clinical reports in support of this concept are based on sequential series that are not comparable because of continued advancement of the state of the art [l, 11, 141. These studies have usually compared hypothermic, continuously cross-clamped, anoxic arrested hearts to normothermic, fibrillating hearts or normothermic, intermittently arrested hearts with interval fibrillation [7,8,11, 12,141. Hearts treated with hypothermic pharmacological cardioplegia and subjected to continuous aortic cross-clamping have likewise been compared to normothermic, fibrillating, unclamped hearts, to normothermic, intermittently cross-clamped hearts with interval fibrillation, or to intermittently cross-clamped hearts with hypothermia and interval fibrillation [l, 2, 51. In none of these clinical studies were myocardial preservation factors kept constant while intermittent cross-clamping alone was varied with continuous cross-clamping. Most of the laboratory experiments have been patterned after clinical reports and are lacking in the same respects or else involve extreme by The Society of Thoracic Surgeons

2 111 Lolley et al: Reperfusion Injury from Aortic Cross-Clamping conditions in which animals having intermittent cross-clamping were subjected to repeated and prolonged anoxic or ischemic arrest periods at normothermia with high-pressure reperfusion intervals and normothermic, fibrillating hearts [4, 7,8, 161. No studies have compared intermittent to continuous aortic cross-clamping utilizing modern pharmacological cardioplegia, hypothermia, and avoidance of gross interval fibrillation. Only Follette and associates [lo] have conducted a relatively comparative experiment; this was in their study on blood cardioplegia, which may be considered a modified form of intermittent aortic crossclamping. There is evidence that hypothermia, when combined with intermittent aortic crossclamping, noticeably reduces reperfusion injury occurring during normothermic intermittent cross-clamping [3]. In addition, it is known that potassium chloride cardioplegia with substrate, hyperosmotic cardioplegia, and coronary arterial infusates also can negate reperfusion edema [9, 15, 211. Because these factors have not been studied in combination with one another, it became imperative that a clinical study be conducted to compare intermittent with continuous aortic cross-clamping utilizing a modern form of pharmacological cardioplegia, which includes base, potassium chloride, an osmotic agent, and metabolic substrate. Therefore this study was undertaken to see whether brief multiple reperfusion periods with oxygenated blood were clinically harmful to the anoxic arrested heart compared with hearts made anoxic by continuous aortic cross-clamping. Clinical harm was assessed by the incidence of myocardial necrosis sequelae such as transmural myocardial infarction, myocardial injury pattern, cardiac-related death, and elevated cardiac enzyme levels, and by rates of atrial arrhythmias, severity of ventricular arrhythmias, utilization of vasopressors to keep blood pressure within a preassigned range, and incidence of low cardiac output syndrome necessitating intraaortic balloon support. Diminution of the relative risk of cardiac surgery through anoxic arrest imposed by either continuous or intermittent aortic cross-clamping could be assessed by reduced incidences of these occurrences. In a practical sense, this improvement would be evident in better clinical cardiac function and lessened evidence of myocardial necrosis. Material and Methods Four hundred eighty adult patients undergoing cardiac operations were initially assigned in a random fashion to two groups that differed as to whether intermittent aortic cross-clamping was employed during procedures utilizing anoxic cardiac arrest. These patients included all those having elective or urgent operations during which aortic cross-clamping is commonly interrupted to allow interval reperfusion periods, such as multiple coronary artery bypass grafting, multiple valve replacement, coronary artery bypass plus cardiac valve replacement or ventricular aneurysm repair, or combinations of these procedures. Instances of emergency salvage, acute myocardial infarction, and procedures for which it is a common practice to employ one aortic cross-clamping, such as single coronary artery bypass grafting, ventricular aneurysm resection, or single valve replacement, were not included in the series. The study was prospective and included the period January, 1977, through January, In Group 1 (217 patients), the aorta was cross-clamped continuously without reperfusion intervals before final release of the clamp, and the patients had one uninterrupted anoxic period for completion of the cardiac procedure. The 263 patients in Group 2 had the aorta cross-clamped intermittently with multiple reperfusion intervals before final release of the clamp at completion of the cardiac portion of the operation and of the anoxic arrest periods. Usually distal coronary arterial grafts were done in sequence, allowing 3- to 5-minute perfusion intervals between each anastomosis. For more extensive procedures with anoxic arrest periods longer than 30 minutes, such as valve replacements in combination with coronary, aneurysm, or other valve procedures, 15- to 20- minute reperfusion intervals were employed between major portions of the cardiac procedure and anoxic arrest times. During the brief reperfusion periods, hearts usually remained in standstill with minimal stiffening and fasciculation.

3 112 The Annals of Thoracic Surgery Vol 30 No 2 August 1980 Coronary arterial perfusion with oxygenated pump blood was not utilized during aortic valve procedures in either group. Cardiopulmonary bypass was carried out according to standard techniques, utilizing systemic (25" to 28 C) and topical (iced saline slush at 4" to 7 C) hypothermia in the pericardial well, cold cardiac standstill, and routine venting of the left ventricle if left atrial pressure rose above 20 mm Hg. The pump perfusate was Normosol-R (ph 7.4) containing 4 meq of potassium chloride per liter. The hemodilution factor was 16 ml per kilogram of body weight. A bubble oxygenator and roller pump were utilized, with high-flow perfusion of 70 ml per kilogram and maintenance of mean preoperative arterial pressure for standard operations judged not to involve more than 60 to 90 minutes of cardiopulmonary bypass time. For extended procedures a membrane oxygenator was employed. Perfusion pressures ranged from 65 to 90 mm Hg and were not allowed to exceed 100 mm Hg. Distal anastomoses were performed before proximal anastomoses. Warming was effected during the last aortic cross-clamping, at the beginning of the last anastomosis, or during closure of the heart or aorta, and immediate defibrillation was accomplished, either spontaneously or electrically, as soon as the aortic cross-clamp was released. Central venous, left atrial, and arterial pressures were monitored and kept within preoperative or physiological range. All patients in both groups had the aortic root cannulated with a large-bore 12F polyethylene catheter, through which a solution of 1,000 ml of 5% glucose in water containing 20 meq of potassium chloride, 4.4 meq of sodium bicarbonate, 12.5 gm of mannitol, and 20 units of regular insulin was infused at the beginning of the aortic cross-clamping. The solution was injected in a 200 ml bolus over a 1- to 2-minute interval and subsequently allowed to drip in continuously, under gravity pressure (in line, 40 rnm Hg), at a rate of 300 ml every 15 minutes throughout the duration of aortic crossclamping. This was repeated for each crossclamp period or at 15- to 20-minute intervals during continuous aortic cross-clamping. In operations 'involving an opened aortic root, the solution was infused or dripped in through cannulas loosely inserted in the coronary artery orifices. The clear solution escaping around the loosely applied cannulas did not interfere with performance of the procedure; and because of the low viscosity of the asanguineous solution, little holding or infusion pressure was needed to infuse the coronary circulation effectively, thereby avoiding damage to the coronary arteries. The rationale for these approaches was that after the initial bolus of solution was administered to achieve a flaccid and electromechanically quiescent heart, a continuous drip could replenish anaerobic substrate (since a form of "multidose" cardioplegia would keep the heart electromechanically quiescent) and achieve coronary washout of lactic acid and anaerobic intermediate metabolites to maximize glycolytic flux. A mean volume of 1,250 ml per patient was utilized (range was 450 to 2,500 ml). The solution had a ph of 7.78 and osmolality of 348 mosm, and was at 4 C. As the study progressed, individual surgeon preference was established in extensive procedures, and critically ill poor-risk patients were assigned to intermittent aortic cross-clamping in violation of the protocol. This biased the study against Group 2 by enlarging its numbers to include patients in New York Heart Association Functional Class 4 and Canadian Heart Association Angina Class 4. Including a larger number of worse cases in Group 2 should have resulted in greater evidence of poorer postoperative clinical performance. The contrary occurred in this group, indicating that useful information was available in the study despite a break in the randomization process. The need for vasopressor (dopamine) to maintain postbypass pressures above preoperative levels or at 120 mm Hg systolic in the later recovery period was observed and recorded. Subsequent vasopressor usage was quantitated according to the following schedule: none if the patient had postbypass pressures above either 120 mm Hg systolic or the preoperative mean aortic pressure and required no dopamine; mild if the patient subsequently utilized less than 0.6 pg of dopamine per kilogram per minute at any time; and severe if the patient needed more than 0.6 pg of dopamine per kilogram per minute to maintain a pressure greater than 120 mm Hg systolic or the preoperative mean aortic pressure. Perfu-

4 113 Lolley et al: Reperfusion Injury from Aortic Cross-Clamping sion pressures were kept at this range in the belief that higher levels would aid in preventing thrombosis of fresh grafts. Arrhythmias were observed and recorded for at least five days postoperatively using a Hewlett-Packard Model HP B computerized arrhythmia monitoring system. Atrial arrhythmias were graded as follows: none if the patient had no atrial disturbances other than an occasional paroxysmal atrial contraction (PAC); mild if the patient experienced an episode of paroxysmal atrial tachycardia (PAT), atrial flutter, or atrial fibrillation that was hemodynamically minor, lasted less than one day, and resolved on minimal drug therapy; and severe if an episode of PAT, atrial fibrillation, or atrial flutter resulted in hypotension, lasted more than one day, or necessitated extensive drug intervention, electrical cardioversion, or both. Ventricular disturbances were graded as follows: none if no paroxysmal ventricular contractions (PVCs) occurred; mild if PVCs occurred but were unifocal and no greater than 10 per minute; severe if the patient had episodes of multifocal PVCs, PVCs greater than 10 per minute, or events such as bigeminy, trigeminy, or paired PVCs; and malignant if the patient had an episode of salvos of PVCs, ventricular tachycardia, or ventricular fibrillation. Serum glutamic oxaloacetic transaminase (SGOT), lactic dehydrogenase (LDH), and creatine phosphokinase (CPK) were monitored 12, 24, and 48 hours postoperatively, and peak levels were recorded. Heart muscle band fraction (CPK-MB) was monitored every 8 hours for 48 hours, and peak levels were compared between the two groups. The myocardial infarction index for each patient was computed from these values according to the method of Shell and Sobol [18, 191, and the mean group levels were compared. Because patients with acute infarction were excluded from the study, preoperative enzyme levels in the two groups were normal and similar. In Group 1, preoperative mean enzyme levels were 23.2 f 1.28 international units (IU) for SGOT, 166 f 4.48 IU for LDH, and 32.9 f 3.14 IU for CPK. Mean preoperative enzyme levels in Group 2 were 20.1 * 1.12 IU for SGOT, 168 f 4.29 IU for LDH, and f 8.1 for CPK. The incidence of unequivocal transmural myocardial infarctions was recorded. This was assessed postmortem or by the appearance of new or deeper Q waves (greater than 2 mm depression) on two subsequent postoperative electrocardiograms with SGOT or CPK-MB above 100 IU. Myocardial injury or damage was also assessed and was assumed to have occurred if electrocardiographic changes such as currents of injury were observed but were unassociated with enzyme changes, if enzyme changes were detected without electrocardiographic changes, or if clinical cardiac performance was poor. The need for intraaortic balloon counterpulsation to counteract low cardiac output syndrome was recorded, as was the incidence of cardiac-related deaths. Group mean values were compared statistically by means of Student s t test. The incidence of cardiac-related events was compared using chi-square contingency tables. Results The two groups established for this study were of similar mean age: 57.0? 0.7 years for Group 1 and 57.0 f 0.6 years for Group 2 (all group means are given with standard error of the mean). Group 1 had 23.0% female patients (50 of 217) and Group 2 had 23.6% (62 of 263 patients). Because of the protocol violations that resulted in more clinically poor risk patients in Group 2, that group had a higher incidence of preoperative myocardial infarction (63.9%; 168 of 263 patients) than Group 1 (40.1%; 87 of 217 patients). Mean percentage left ventricular ejection fractions in Group 2 were somewhat lower, at 59.0 f l.oo/~, than in Group 1, 62.0 f 1.2%. Left ventricular end-diastolic pressures were similar for the two groups: 14.0 * 0.6 mm Hg in Group 1 versus 15.0 f 0.6 mm Hg in Group 2. Of 183 Group 1 patients undergoing coronary artery bypass, 4, 35, 111, and 33 were in Canadian Heart Association* Angina Classes l, 2, 3, and 4, respectively; in Group 2, 0, 28, 104, and *Canadian Heart Association angina classification: Class &no angina pectoris; rest, 0; Class 1-angina only with strenuous or prolonged activity; rest, 0; Class 2-angina with rapid walking up stairs or uphill, with cold, or with emotion; rest, 0; Class 3-angina with walking one to two blocks or a flight of stairs; rest, 0; Class Pangina with any physical activity; rest, 0. (From the Canadian Cardiovascular Society Committee: Nomenclature and Grading for Coronary Artery Surgery, 1972.)

5 114 The Annals of Thoracic Surgery Vol 30 No 2 August 1980 Table 1. Total Cardiac Procedures Performed in the Two Groups of Patients Group 1: Continuous Group 2: Intermittent Procedure Crossclamping Crossclamping Double bypass Triple bypass Quadruple bypass Quintuple bypass 2 11 Sextuple bypass 0 3 Mitral and coronary 6 9 Aortic and coronary 7 12 Ventricular aneurysm and coronary Multiple valves 6 1 Valve, aneurysm, and 0 1 coronary Total patients of 208 patients were in Canadian Heart Association Classes 1 through 4. Of 34 Group 1 patients having combined procedures, 0, 7, 18, and 9 were in New York Heart Association* Functional Classes 1, 2, 3, and 4, respectively, before operation; of the 55 in Group 2, 0, 9, 20, and 26 were in Functional Classes 1,2,3, and 4. The range of cardiac procedures was similar in both groups (Table 1). There were more multiple valve procedures in Group 1, but Group 2 patients had more quintuple and sextuple coronary artery bypass graftings and combined coronary artery bypass with cardiac valve or aneurysm repair or both. Group 1 had a mean total cardiopulmonary bypass time of 93.0 k 2.5 minutes (range, 47 to 292 minutes), significantly (p < 0.001) shorter than the time of f 2.4 minutes (range, 48 to 274 minutes) in Group 2. This reflects the added reperfusion time during intermittent aortic cross-clamping in Group 2. However, the mean aortic cross-clamp time for Group 1, 52.0 k 1.3 minutes (range, 22 to 144 minutes), was nearly the same as that for Group 2, 50 f 1.0 *New York Heart Association Cardiac Functional Classification: Class 1-uncompromised; Class 2-slightly compromised; Class 3-moderately compromised; Class 4- severely compromised. (From New York Heart Association: Nomenclature and Criteria for Diseases of the Heart and Great Vessels. Eighth edition. Boston, Little, Brown, 1979.) minutes (range, 20 to 112 minutes), indicating that anoxic arrest time was not prolonged when intermittent aortic cross-clamping and pharmacological cardioplegia were utilized. The incidence of immediate spontaneous defibrillation was similar in both groups: 52% (113 of 217 patients) for Group 1 and 58% (152 of 263 patients) for Group 2. Group 2, despite having a longer cardiopulmonary bypass time, had a significantly lower (p < 0.01) mean peak SGOT than Group 1, 72.0 f 3.1 versus f 39.0 IU. The mean peak LDH of k 12.3 IU in Group 2 was significantly lower ( p < 0.05) than the 636 f 78.2 IU in Group 1. However, Group 1 s postoperative mean peak CPK (498.0 f 55.2 IU), CPK-MB (17.1 k 1.05 IU), and CPK-MB myocardial infarct index (8.2 k 0.6 gram-equivalents) were similar to the Group 2 mean peak CPK (482.0 f 33.2 IU), CPK-MB (17.8 f 0.71 IU), and CPK-MB myocardial infarct index (8.1 k 0.4 gram-equivalents). Group 1 patients demonstrated a significantly higher incidence of myocardial necrosis sequelae by having more transmural myocardial infarctions (p < 0.01), instances of myocardial injury (p < 0.02), and cardiac-related deaths ( p < 0.02) (Table 2). This trend was even more striking for the subgroup of patients having multiple, extensive procedures such as combined coronary arterial bypass grafting with valve replacement or aneurysm repair or both, or multiple valve replacements. Seven of the 34 Group 1 patients undergoing combined procedures had transmural myocardial infarctions compared with 2 of the 55 Group 2 patients, and 9 of the 34 Group 1 patients died cardiac-related deaths compared with 1 of the 55 in Group 2. In comparison with Group 2, Group 1 patients demonstrated significantly more serious ventricular arrhythmias (p < 0.01) in the postoperative period as detected by computerized arrhythmia monitoring (Table 3). However, neither intermittent nor continuous crossclamping affected the incidence of postoperative atrial disturbances, as the rates and severity of PAT, atrial fibrillation, and atrial flutter were similar for the two groups (Table 4). Postbypass contractility was decidedly better for Group 2, as significantly less dopamine was

6 ~~ 115 Lolley et al: Reperfusion Injury from Aortic Cross-Clamping Table 2. Myocardial Necrosis Sequelae, All Cardiac Procedures Incidence (%) Event Group 1: Group 2: Continuous Intermittent Cross-clamping Cross-clamping Significance Transmural myocardial infarction 8.30 (181217) 1.90 (51263) p < 0.01 Myocardial injury (271217) 5.70 (151263) p < 0.02 Cardiac death 4.60 (101217) 0.76 (21263) p < 0.02 Need for intraaortic balloon support 4.15 (91217) 2.66 (7/263) NS NS = not significant. Table 3. Ventricular Arrhythmias, All Cardiac Proceduresa Incidence (YO) Group 1: Group 2: Continuous Intermittent Arrhythmias'' Cross-clamping Cross-clamping None 39.2 (851217) 61.2 ( ) Mild 24.9 (541217) 14.1 (371263) Severe 11.5 (251217) 8.7 (231263) Malignant 24.4 (53/217) 16.0 (42/263) "Computer tabulation for five days. "p < 0.01 by chi-square contingency table. Table 4. Atrial Arrhythmias, All Cardiac Proceduresa Incidence (YO) Group 1: Group 2: Continous Intermittent Arrhythmias" Cross-clamping Cross-clamping None 70.5 ( ) 76.4 ( ) Mild 16.6 (361217) 16.0 (421263) Severe 12.9 (28/217) 7.6 (201263) "Computer tabulation for five days. bnot significant by chi-square contingency table. utilized in that group ( p < 0.05) than in Group 1 to keep mean postbypass blood pressures above preoperative levels or above 120 mm Hg systolic (Table 5). Comment Our results in a large concurrent series of patients do not support the contention that continuous aortic cross-clamping causes less Table 5. Peak Drug Use, All Cardiac Proceduresa Incidence (Yo) Group 1: Group 2: Vasopressorh Continuous Intermittent Usage Cross-clamping Cross-clamping None 60.8 ( ) 60.1 ( ) Mild 17.5 (381217) 25.5 (671263) Severe 21.7 (471217) 14.4 (38/263) adopamine was used to keep mean pressure above either the preoperative level or 120 mm Hg systolic. bp < 0.05 by chi-square contingency table. clinical cardiac damage than intermittent cross-clamping during hypothermia and pharmacological cardioplegia. The theoretical increases in reperfusion edema, injury, or both in patients having intermittent aortic crossclamping were not seen. A greater frequency of myocardial necrosis sequelae, such as transmural myocardial infarction, myocardial injury, cardiac-related deaths, and higher group mean cardiac enzymes such as SGOT and LDH, implies the converse: that the hearts subjected to intermittent aortic cross-clamping may have benefited from the brief periods of reperfusion with oxygenated blood. A statistically significant higher rate of severe ventricular arrhythmias and vasopressor usage in the group that had continous cross-clamping reinforces this concept. It does not detract from information derived from this study that violations of protocol caused the intermittent cross-clamp group to contain a larger number of poor-risk patients who required more extensive cardiac surgery,

7 116 The Annals of Thoracic Surgery Vol 30 No 2 August 1980 as the resulting bias is against this group and should have worsened the results. In fact, better clinical results were obtained despite a longer cardiopulmonary bypass time and less favorable operative cases. It could be speculated that during procedures in which continuous aortic cross-clamping was employed, there might have been a tendency to hurry the operation as the anoxic period became prolonged, which could have resulted in more clinically harmful errors in this group, accounting for the poorer results. The identical mean aortic crossclamping times for the two groups and the absence in this study of cases involving technical problems do not support this point. Mechanisms to explain why clinical cardiac damage is not increased with intermittent cross-clamping have been extensively investigated in the laboratory. Brown and associates [3] demonstrated significant reduction of reperfusion edema during intermittent aortic cross-clamping with hypothermia alone. Wright and associates [21] showed nearly complete preservation of left ventricular function when potassium chloride cardioplegia with substrate was utilized during each aortic crossclamping in animals subjected to intermittent anoxia. Their study also demonstrated conservation of adenosine nucleotides, especially if the hearts remained inactive during reperfusion periods. In the present study, electromechanical inactivity of the cold hearts during the brief reperfusion periods may have replicated this situation. Also, reperfusion pressures did not approach the 100 mm Hg level that is usually associated with perfusion injury [8]. Hyperosmotic or hyperoncotic cardioplegic solutions (mannitol or colloid) have been shown by Levitsky and co-workers [15] and Foglia and associates [9] to reduce myocardial reperfusion edema significantly in canine models. Our use of a hyperosmotic (348 mosm), asanguineous coronary infusate may explain why reperfusion injury was not increased in Group 2. Perhaps the reduction in reperfusion edema allowed the reperfused blood to exert its beneficial physiological effects (replenishing substrate, providing buffer, and eluting toxic metabolites) without the deleterious effects of reperfusion edema. Indeed, the beneficial ef- fects of so-called blood cardioplegia may be due to a modified form of intermittent aortic crossclamping with hypothermic or pharmacological cardioplegia that negates reperfusion edema. With this technique, electrolyte-, calcium-, and ph-altered blood is perfused into the coronary circulation at 15- to 30-minute intervals under reduced pressure to interrupt the anoxic arrest periods and provide oncotic reduction of reperfusion edema, cause electromechanical quiescence of the heart, supply buffer, and replenish substrate [4, 91. In summary, the frequency of clinical cardiac damage was not increased in a large group of patients subjected to intermittent aortic crossclamping during anoxic cardiac arrest compared with a group that had continuous aortic cross-clamping. The utilization of hypothermia and asanguineous coronary infusates containing buffer, substrate, and potassium chloride in addition to being hyperosmotic probably reduced reperfusion edema and injury to a level that did not differ between the two groups. The better clinical cardiac performance in the group receiving intermittent aortic cross-clamping implies that hearts freed from the reperfusion injury of multiple aortic cross-clampings benefit from the salutory effects of oxygenated blood. References 1. Adappa MG, Jacobsen LB, Hetzer R, et al: Cold hyperkalemic cardiac arrest versus intermittent aortic cross-clamping for coronary artery bypass surgery. J Thorac Cardiovasc Surg 75:171, Behrendt DM, Kirsh MM, Jochim KE, et al: Effects of cardioplegia solution on human contractile element velocity. Ann Thorac Surg 26:499, Brown AH, Braimbridge MV, Darracott S, et al: An experimental evaluation of continuous normothermic, intermittent hypothermic, and intermittent normothermic coronary perfusion. Thorax 29:38, Chitwood WR, Hill RC, Kleinman LH, et al: The effects of intermittent ischemic arrest on the perfusion of myocardium supplied by collateral coronary arteries. Ann Thorac Surg 26:535, Conti VR, Bertranou EG, Blackstone EH, et al: Cold cardioplegia versus hypothermia for myocardial protection. J Thorac Cardiovasc Surg 76:577, Cunningham JN, Adams PX, Knapp EA, et al:

8 117 Lolley et al: Reperfusion Injury from Aortic Cross-Clamping Preservation of ATP, ultrastructure, and ventricular function following aortic crossclamping and reperfusion: clinical use of potassium plus blood cardioplegia. J Thorac Cardiovasc Surg 78:708, Engelman RM, Adler S, Gouge TH, et al: The effect of normothermic anoxic arrest and ventricular fibrillation on the coronary flow distribution of the pig. J Thorac Cardiovasc Surg , Engelman RM, Chandra R, Baumann FG, et al: Myocardial reperfusion, a cause of ischemic injury during cardiopulmonary bypass. Surgery 80:266, Foglia RP, Steed DL, Follette DM, et al: Iatrogenic edema with potassium cardioplegia. J Thorac Cardiovasc Surg 78:217, Follette DM, Mulder DG, Maloney JV, et al: Advantages of blood cardioplegia over continuous coronary perfusion or intermittent ischemia. J Thorac Cardiovasc Surg 76:604, Griepp RB, Stinson EB, Oyer E, et al: The superiority of aortic cross-clamping with profound local hypothermia for myocardial protection during aortocoronary bypass grafting. J Thorac Cardiovasc Surg 70:995, Griepp RB, Stinson EB, Shumway NE: Profound local hypothermia for myocardial protection during open heart surgery. J Thorac Cardiovasc Surg 66:731, Hearse DJ, Stewart DA, Braimbridge MV: Myocardial protection during ischemic cardiac arrest. J Thorac Cardiovasc Surg 76:16, Koster JK, Cohn LH, Collins JJ, et al: Continuous hypothermic arrest versus intermittent ischemia for myocardial protection during coronary revascularization. Ann Thorac Surg 24:330, Levitsky S, Mullin ED, Sloane RE, et al: Effects of a hyperosmotic perfusate of extended preservation of the heart. Circulation 43,44:suppl 1:124, Levitsky S, Wright RN, Rao KS, et al: Does intermittent coronary perfusion offer greater myocardial protection than continuous aortic crossclamping? Surgery 82:51, Ross D: Cardiac surgery-the golden years. Honored Speaker address, 58th Annual Meeting of the American Association of Thoracic Surgery, New Orleans, LA, May 9, Shell WE, Sobel BE: Biochemical markers of ischemic injury. Circulation 53:Suppl 1:98, Sobel BE, Bresnahan GF, Shell WE, et al: Estimation of infarct size in man and its relationship to prognosis. Circulation 46:640, Spencer FC: Surgical procedures for coronary atherosclerosis. Prog Cardiovasc Dis 14:399, Wright RN, Levitsky S, Holland C, et al: Beneficial effects of potassium cardioplegia during intermittent aortic cross-clamping and reperfusion. J Surg Res 24:201, 1978 Notice of Correction In the announcement "Statement on Extracorporeal Technology" in the February issue of The Annals (Ann Thorac Surg 29:101, 1980), the last sentence should read as follows: However, it must be understood that the surgeon in charge is morally and ethically responsible for the care of the patient.