Hemolysis is a well known complication caused by

Transcription

1 Free Hemoglobin Impairs Cardiac Function in Neonatal Rabbit Hearts Shintaro Nemoto, MD, Mitsuru Aoki, MD, Chang Dehua, MD, and Yasuharu Imai, MD Department of Pediatric Cardiovascular Surgery, The Heart Institute of Japan, Tokyo Women s Medical College, Tokyo, Japan Background. Hemolysis caused by cardiopulmonary bypass causes renal dysfunction and other organ failure presumably by superoxide production catalyzed by iron derived from free hemoglobin (f-hb). It might also impair cardiac function by the same mechanism, especially in the ischemia-reperfusion period and in neonates where serum antioxidant activity is lower than adults. Methods. We evaluated effects of f-hb on cardiac function with or without ischemia and reperfusion using a newborn (7 days old) rabbit crystalloid-perfused Langendorff model. After baseline measurements, the hearts were divided into the following four groups (8 hearts per group): (1) those perfused with regular Krebs-Henseleit bicarbonate buffer, (2) those perfused 30 minutes with KH buffer containing 1 mg/ml of f-hb obtained from osmotic hemolysis, (3) those subjected to 180 minutes of cold global ischemia with infusion of crystalloid cardioplegia and reperfused with Krebs-Henseleit buffer, and (4) those subjected to the same ischemia and reperfused with Krebs-Henseleit buffer containing 1 mg/ml of f-hb. The left ventricular function (using conductance catheter and isovolumic balloon) and coronary flow were measured. Results. Free hemoglobin significantly impaired not only left ventricular function but also coronary flow even without ischemia (p < 0.05). When ischemia and reperfusion were involved, the group reperfused with f-hb showed the worst left ventricular function and coronary flow among the groups. Conclusions. This study shows that f-hb directly impaired cardiac function and coronary flow in neonatal hearts especially in ischemia and reperfusion. (Ann Thorac Surg 2000;69:1484 9) 2000 by The Society of Thoracic Surgeons Accepted for publication Sept 14, Address reprint requests to Dr Nemoto, Laboratories of Cardiac Molecular and Cellular Physiology, Department of Medicine-Cardiology, Veterans Affairs Medical Center, Baylor College of Medicine, Room 243, Building 110, 2002 Holcombe Blvd, Houston, TX 77030; snemoto@bcm.tmc.edu. Hemolysis is a well known complication caused by cardiopulmonary bypass. It has been associated with various causes, such as increased shear stress generated by blood pumps and suction systems [1], activated compliment C5b-9 depositing on the surface of erythrocytes [2], artificial materials such as mechanical valves, and oxygen-derived free radicals [3]. It is widely known that hemolysis not only impairs renal function [1] but also causes multiple organ failure [4]. The precise mechanism of these injuries, however, still remains to be determined. Presumably, superoxide anion production catalyzed by iron derived from free hemoglobin has an important role in these injuries, because intense peroxidation reactions within biological membranes initiated with Fenton s reaction begin instantly upon addition of Fe 2 and precede detactable OH formation [5]. Although fetal rat myocardial tissues show a steady and progressive increase of antioxidant enzyme with age [6], serum antioxidant activity interacting with iron is lower in neonates than adults [7, 8]. Therefore, free hemoglobin caused by hemolysis might also impair cardiac function in neonates by the same mechanism, especially when ischemia and reperfusion are involved. The purpose of the current study was to investigate the effect of free hemoglobin on cardiac function in neonatal rabbit hearts with or without ischemia-reperfusion injury. Material and Methods Experimental Preparation An isolated crystalloid-perfused Langendorff model was used to study 32 hearts from neonatal rabbits (Japanese white rabbit, 7 days old). Rabbits were anesthetized with an intraperitoneal injection of 60 mg/kg sodium pentobarbital and 100 IU heparin. When the rabbit totally lacked sensation, the thoracic cavity was opened and the heart was excised quickly and placed in ice-cold normal saline. The aorta was cannulated retrogradely for isolated heart perfusion. The heart was perfused at 40 cm H 2 O aortic pressure by gravity. Modified Krebs-Henseleit bicarbonate buffer (NaCl 118 mm/l, NaHCO 3 25 mm/l, KH 2 PO mm/l, KCl 4 mm/l, MgSO mm/l, CaCl mm/l, and glucose 11.1 mm/l) was used as perfusate and oxygenated with a mixture of 95% oxygen and 5% carbon dioxide. The perfusate and water bath were controlled at 37 C by heater-circulator except during the hypothermic phase, which was produced by circulating ice water. A latex balloon containing a microtip pressure transducer (SPC-350; Millar Instruments Inc, 2000 by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)

2 Ann Thorac Surg NEMOTO ET AL 2000;69: FREE HEMOGLOBIN AND CARDIAC FUNCTION 1485 Houston, TX) was placed into the left ventricle (LV) through the left atrium to measure LV function. Measurements Left ventricular function was measured during isovolumic contraction by inflating the intraventricular balloon by stepwise increments (0.01 ml). Left ventricular developed pressure (DP) and its first derivative (dp/dt) were recorded at each volume. The systolic function was evaluated by measuring the maximum DP, positive maximum dp/dt, peak DP at a constant balloon volume (V10), and peak dp/dt at V10. We defined V10 as the balloon volume to give an end-diastolic pressure of 10 mm Hg during baseline measurement. Negative maximum dp/dt was measured to evaluate diastolic function. Coronary efflux was allowed to drip through the open pulmonary artery onto a glass cylinder. The fluid was collected and measured. Experimental Protocol Baseline measurements were made after a 15-minute stabilization period. Then, in the nonischemia groups, the hearts were perfused with modified Krebs-Henseleit buffer with or without free hemoglobin for 30 minutes. The LV function was assessed at 15 and 30 minutes after baseline measurements. In ischemia-reperfusion groups, the perfusate and water bath were cooled to 20 C. At 10 minutes after the start of cooling, when both temperatures reached 20 C, the heart was subjected to cold cardioplegic arrest by infusion of St. Thomas cardioplegic solution every 30 minutes (3 ml initial dose and 1.5 ml subsequent dose) and to topical cooling. The composition of the cardioplegic solution was NaCl 110 mm/l, NaHCO 3 10 mm/l, KCl 16 mm/l, MgCl 2 16 mm/l, and CaCl mm/l. After 180 minutes of cold ischemia, reperfusion was begun with the perfusate with or without free hemoglobin at 20 C followed by rewarming to normothermia. The LV function was assesed at 15 and 30 minutes of reperfusion. Fig 1. Study protocol. Experimental Groups The hearts were divided into four groups. In the control nonischemia group K (n 8), the hearts were perfused with modified Krebs-Henseleit solution alone without ischemia. In group H (n 8), the hearts were perfused with the Krebs-Henseleit solution containing 1 mg/ml of free hemoglobin for 30 minutes without ischemia. In the control ischemic group K-I (n 8), the hearts were reperfused with the Krebs-Henseleit solution alone. In group H-I (n 8), the hearts were reperfused with the perfusate containing 1 mg/ml of free hemoglobin. The free hemoglobin was obtained from osmotic hemolysis of heparinized fresh homologus whole blood (Fig 1). All animals in this study received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the National Institutes of Health (NIH Publication No , revised 1985). Statistical Analyses All values were expressed as mean standard error and analyzed by a statistical analysis system (Stat View version 4.5; Abacus Concepts Inc., Berkeley, California). Repeated-measures of analysis of variance (ANOVA) was used to compare the differences in recovery between groups. Data were further compared by the Student t test if ANOVA was significant. A p value less than 0.05 was considered significant. Results There were no significant differences among the four groups at baseline (Table 1). Therefore, further results are given as percentage of the baseline values (Tables 2 and 3). Figure 2 shows percentage change of maximum LV DP and positive maximum dp/dt. Group H showed a lower recovery of maximum DP than group K, but it was not significant. Group H showed a significantly lower recovery of maximum LV dp/dt than group K. Group H-I showed significantly lower recovery of these indices than group K-I and the worst impairment among the four groups. Figure 3 shows percentage change of maximum LVDP at V10 and maximum positive dp/dt at V10. These variables could be considered systolic function indices under an equal preload condition over the course of this study. Group H could not produce the same DP and positive dp/dt values as at baseline under the equal preload status. Group H showed a significantly lower recovery than group K. Moreover, group H-I had significantly worse impairment than group K-I. With regard to maximum negative dp/dt, indicating diastolic function, group H showed significantly lower recovery than group K, as shown in Figure 4. Group H-I showed significantly lower recovery than group K-I and the worst impairment among the four groups. Figure 5 shows percentage change of coronary flow. Group H had significantly lower coronary flow than group K. Group H-I showed significantly lower coronary flow than group K-I at both 15 and 30 minutes of

3 1486 NEMOTO ET AL Ann Thorac Surg FREE HEMOGLOBIN AND CARDIAC FUNCTION 2000;69: Table 1. Baseline Measurements a K H K-I H-I Heart rate (bpm) DP (max) (mm Hg) DP (V10) (mm Hg) dp/dt (max) (mm Hg/sec) dp/dt (V10) (mm Hg/sec) dp/dt (max) (mm Hg/sec) CF (ml/min) a Values are mean standard error. There are no significant differences among the groups by analysis of variance. bpm beats per minute; CF coronary flow; DP(max) maximum peak LVDP; DP(V10) peak LVDP at a constant balloon volume (V10); dp/dt (max) maximum of peak positive LV dp/dt; dp/dt (V10) LV dp/dt at V10; V10 volume to produce an end-diastolic pressure of 10 mm Hg; dp/dt (max) maximum of peak negative LV dp/dt. reperfusion and the worst recovery among the four groups. Comment There were three major findings of this study. First, free hemoglobin significantly impaired not only left ventricular systolic function but also diastolic function even without ischemia and reperfusion. Second, when ischemia and reperfusion were involved, free hemoglobin impaired cardiac function more severely than without ischemia and reperfusion. Third, although the coronary perfusion pressure was maintained at the same level throughout the experiments, free hemoglobin significantly reduced the coronary flow, especially when ischemia and reperfusion were involved. This suggests that free hemoglobin directly causes coronary endothelial damage. In this study, we did not examine the concentrations of free ion and free Fe 3 in the perfusate containing free hemoglobin. However, we believe that large amounts of free iron would be present in the perfusate, because it is known that a certain degree of hemolysis always occurs during cardiopulmonary bypass [9] and that patients who have open heart operations have a transient elevation of plasma free hemoglobin that peaks several hours postoperatively [10]. Furthermore, it is known that reperfusion after ischemic arrest causes an increase in free hemoglobin and free hemoglobin concentrations, simultaneously releasing free iron and generating hydroxy radicals [3]. Although previous studies have found iron-mediated oxidative injury to cardiac function, our study shows the effects of the more clinically important free hemoglobin on cardiac function. Previous studies demonstrated that either exposure of intact hearts to an iron-catalyzed OH generating system or higher intracellular reactive iron levels increased myocardial lipid peroxide levels and decreased LV function [11, 12]. Lipid peroxidation is initiated by production of OH, which occurs through the iron-catalyzed Fenton reaction (H 2 O 2 Fe 2 3 OH OH Fe 3 ) and the formation of [Fe 2 O 2 3 Fe 3 O 2 ] chelate complex [5], especially during hemolysis [13]. Under the condition that iron is bound to or attached to cell membranes, these reactions from OH production to lipid peroxidation occur rapidly and continuously [14]. Table 2. Percentage Recovery of Left Ventricular Function a K H K-I H-I DP (max) 15 min min DP (V10) 15 min min dp/dt (max) 15 min min dp/dt (V10) 15 min min dp/dt (max) 15 min min a Values are mean standard error. Abbreviations as in Table 1.

4 Ann Thorac Surg NEMOTO ET AL 2000;69: FREE HEMOGLOBIN AND CARDIAC FUNCTION 1487 Table 3. Percentage Recovery of Coronary Flow K H K-I H-I 15 min min Values are mean standard error. This might be a major cause of impaired cardiac function, especially when ischemia and reperfusion are involved. Although many efforts to prevent oxidant-induced myocardial damage by using antioxidative agents, such as U74006F, a 21-aminosteroid, to prevent lipid peroxydation [15, 16]; selenium to reduce intercellular free iron in heart tissue [17]; and deferoxamine, an iron chelator [18], have been tested in various experimental models, their effects on contractile function were partial and additive. Fig 3. Percentage changes of maximum DP (A) and maximum positive dp/dt (B) at the volume to produce an end-diastolic pressure of 10 mm Hg at baseline. Group H could not produce the same DP and positive dp/dt values as at baseline under the equal preload status and had significantly lower recovery than group K. Group H-I had significantly worse impairment than group K-I. Fig 2. (A) Percentage changes of maximum left ventricular developed pressure. Group H showed a lower recovery than group K, but was not significant. Group H-I showed significantly lower recovery than group K-I. (B) Percentage changes of positive maximum dp/dt. Group H had significantly reduced dp/dt than group K. Group H-I had the worst impairment among the four groups. Therefore, to prevent iron-related oxidant-induced myocardial dysfunction, it should be most important to reduce hemolysis. In the current study, free hemoglobin administration reduced the coronary flow in nonischemic hearts. This same deleterious effect on coronary perfusion was recently reported by use of Fe 2 and adenosine diphosphate, which increased ion-catalyzed oxidant injury [11]; Fe 3 and 8-hydroxyquinolene, which increased intracellular reactive iron level [12]; and H 2 O 2 and Fe 2 adenosine diphosphate chelete, a OH-generating system [15]. Coronary endothelial injury resulting from membrane lipid peroxidation directly caused by the oxidants was suspected as the main mechanism of reduced coronary flow. However, the precise mechanism still remains unknown. The worst recovery of coronary flow was in the group that received free hemoglobin during reperfusion

5 1488 NEMOTO ET AL Ann Thorac Surg FREE HEMOGLOBIN AND CARDIAC FUNCTION 2000;69: without any blood component in the perfusate, the presence of which might attenuate and diminish the unfavorable effects of free hemoglobin. Additional studies using an isolated blood-perfused heart model will be necessary. This work was supported in part by the Japanese Ministry of Education, Science, and Culture Grants (Shintaro Nemoto and Mitsuru Aoki). Fig 4. Percentage changes of maximum negative dp/dt are shown. Free hemoglobin significantly impaired diastolic function, which is active dilation here, in groups H and H-I. after ischemia. Previous studies emphasize that the coronary endothelium has an important role on cardiac function in ischemia-reperfusion injury [19 21]. It was reported that ischemia and reperfusion caused coronary endothelial damage and reduced production of endogenous coronary vasodilators, such as nitric oxide [20], and that reduced coronary flow and LV function were restored by induction of nitric oxide [21]. In the current study, free hemoglobin might have amplified the postischemic endothelial injury. Alternatively, the low coronary flow might have been appropriate to the decreased mechanical work produced during ischemia, since it was reported that even when the coronary flow was increased by infusion of nitrotriglycerin to maximize the flow, LV function remained depressed [15]. We conclude that free hemoglobin directly impaired LV function and coronary blood flow in neonatal rabbit hearts, especially when ischemia and reperfusion were involved. These results emphasize the effects of free hemoglobin on neonates who have open heart operations. The conclusions are limited by the fact that experiments were performed in the isolated heart Fig 5. Percentage changes of coronary flow. Group H-I showed significantly lower coronary flow than group K-I at both 15 and 30 minutes of reperfusion and the worst recovery among the groups. References 1. Kirklin JW, Barratt-Boyes BG. Cardiac surgery. New York: Churchill Livingstone, 1993: Salama A, Hugo F, Heinrich D, et al. Deposition of terminal C5b-9 compliment complexes on erythrocytes and leukocytes during cardiopulmonary bypass. N Engl J Med 1988; 318: Das DK, Engelman RM, Liu X, et al. Oxygen-derived free radicals and hemolysis during open heart surgery. Mol Cell Biochem 1992;111: Seyr M, Hasibeder W, Furtwaegler W, Antretter H, Mutz NJ. Multiple organ failure after mitral valve repair with intravascular hemolysis and its recovery due to mitral valve replacement. Intens Care Med 1993;19: Braughler JM, Duncan LA, Chase RL. The involvement of iron in lipid peroxidation. J Biol Chem 1986;261: Hayashibe H, Asayama K, Dobashi K, Kato K. Prenatal development of antioxidant enzymes in rat lung, kidney, and heart: marked increase in immunoreactive superoxide dismutase, glutathione peroxidase, and catalase in the kidney. Pediatr Res 1990;27: Sullivan JL. Iron, plasma antioxidants, and the oxygen radical disease of prematurity. Am J Dis Child 1988;142: Sullivan JL, Newton RB. Serum antioxidant activity in neonates. Arch Dis Child 1988;63: Lohrer RM, Trammer AR, Deitrich W, Hagl S, Linderkamp O. The influence of extracorporeal circulation and hemoseparation on red cell deformability and membrane proteins in coronary artery disease. J Thorac Cardiovasc Surg 1990;99: Lamon EW, Powell TJ, Blackstone EH. Hemolysis after cardiopulmonary bypass. J Thorac Cardiovasc Surg 1990;99: Lesnefsky EJ, Allen KG, Carrea FP, Horwitz LD. Ironcatalyzed reactions cause lipid peroxydation in the intact heart. J Mol Cell Cardiol 1992;24: Oubidar M, Marie C, Mossiat C, Bralet J. Effects of increasing intracellular reaction iron level on cardiac function and oxidative injury in the isolated rat heart. J Mol Cell Cardiol 1996;28: Ribrov SR, Benov LC. Relationship between the hemolytic action of heavy metals and lipid peroxidation. Biochem Biophys Acta 1981;640: Girotti AW, Thomas JP. Superoxide and hydrogen peroxidedependent lipid peroxydation in intact and triton-dispersed erythrocyte membranes. Biochem Biophys Res Commun 1984;118: Kong Y, Lesnefsky EJ, Ye J, Horwitz LD. Prevention of lipid peroxidation does not prevent oxidant-induced myocardial contractile dysfunction. Am J Physiol 1994;267:H Holzgrefe HH, Buchanan LV, Gibson JK. Effects of U74006F, a novel inhibitor of lipid peroxidation, in stunned reperfused canine myocardium. J Cardiovasc Pharmacol 1990;15: Simoni J, Simoni G, Garcia EL, et al. Protective effect of selenium on hemoglobin mediated lipid peroxidation in vivo. Artif Cells Blood Substit Immobil Biotechnol 1995;23: Dulchavsky SA, Davidson SB, Cullen WJ, et al. Effects of deferoxamine on H 2 O 2 -induced oxidative stress in isolated rat heart. Basic Res Cardiol 1996;91:

6 Ann Thorac Surg NEMOTO ET AL 2000;69: FREE HEMOGLOBIN AND CARDIAC FUNCTION Kawata H, Sawatari K, Mayer JE Jr. Evidence of the role of neutrophils in reperfusion injury after cold cardioplegic ischemia in neonatal lambs. J Thorac Cardiovasc Surg 1992; 103: Sawatari K, Kadoba K, Berger KA, Daich JA, Mayer JE Jr. Influence on initial reperfusion pressure after hypothermic cardioplegia on endothelial modulation of coronary tone in neonatal lambs: impaired coronary vasodilation response to acetylcholine. J Thorac Cardiovasc Surg 1991;101: Hiramatsu T, Forbess JM, Miura T, Mayer JE Jr. Effects of L-arginine and L-NAME on recovery of neonatal lamb hearts after cold cardioplegic ischemia: evidence for an important role for endothelial production of nitric oxide. J Thorac Cardiovasc Surg 1995;109:81 7. INVITED COMMENTARY Nemoto and colleagues examined the effect of free hemoglobin on the cardiac function in both intact heart and in reperfused heart after cold cardioplegic ischemia using an isolated heart model. They showed that free hemoglobin had some detrimental effects on ventricular function and coronary flow. Although this finding is new and very important for the clinical practice in pediatric cardiac surgery, there may be some problems and limitations in this article. The authors speculate that superoxide production by the iron derived from free hemoglobin might cause ventricular dysfunction and coronary flow reduction. However, they do not examine iron content or superoxide production, and the mechanism of these phenomena is entirely unclear. This adverse effect on cardiac function might be caused by free hemoglobin itself, as the authors expected, but it might also be caused by another unknown effect of osmotic hemolysis used in this experiment. To eliminate this possibility, the authors should repeat a similar experiment using some substance that might reverse the effect of free hemoglobin; for example, haptoglobin or iron chelater. This reversal experiment could exclude the possibility of an unknown effect of osmotic hemolysis and could also elucidate important information revealing the precise mechanism of dysfunction of the heart perfused with free hemoglobin. The concentration of free hemoglobin in this experiment is 1 mg per ml ( 100 mg/dl). This concentration is slightly higher than the usual during hemolysis in clinical situation as reported by Siderys [1]. The authors should described how they decide the concentration of free hemoglobin in this experiment. I have not encountered cardiac dysfunction caused by hemolysis in a clinical case so far. In this experiment, there was no blood component in the perfusate, which might attenuate the effect of free hemoglobin. The authors should be aware that this phenomenon might not be a duplication of hemolysis in the normal circulation with blood components. The authors should perform some additional experiments to reveal the mechanism of this adverse effect of free hemoglobin and to examine actual effect in a more clinically relevant model. Takuya Miura, MD Department of Cardiovascular Surgery Osaka Medical Center and Research Institute for Maternal and Child Health 840 Murodo-cho, Izumi Osaka , Japan takuyamiura@hotmail.com. Reference 1. Siderys H. Pericardial fluid and hemolysis [Letter]. Ann Thorac Surg 1993;56: by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)