Recent experimental studies have documented that 1

Transcription

1 Detrimental Effects of Cardiopulmonary Bypass in Cyanotic Infants: Preventing the Reoxygenation Injury Bradley S. Allen, MD, Shaikh Rahman, MS, Michel N. Ilbawi, MD, Michael Kronon, MD, Kirk S. Bolling, MD, Ari O. Halldorsson, MD, and Harold Feinberg, PhD Division of Cardiothoracic Surgery, The University of Illinois, Chicago, Illinois Background. Recent experimental studies have shown that acute hypoxia followed by abrupt reoxygenation using cardiopulmonary bypass (CPB) results in an unintended injury mediated by oxygen free radicals, which can be modified by initiating CPB at a lower fraction of inspired oxygen (FiO 2 ) or by leukocyte filtration. However, the clinical relevance of these experimental studies has been questioned because chronic hypoxia may allow compensatory changes to occur. Methods. Seven acyanotic infants had CPB initiated at an FiO 2 of 1.0. Of 21 cyanotic infants, 7 (group 1) had CPB initiated at an FiO 2 of 1.0, 6 (group 2) at an FiO 2 of 0.21, and 8 (group 3) underwent CPB using leukocyte filtration. Biopsy of right atrial tissue was performed before and 10 to 20 minutes after the initiation of CPB. The tissue was incubated in 4-mmol/L t-butylhydroperoxide (a strong oxidant), and the malondialdehyde (MDA) level was measured to determine the antioxidant reserve capacity. The more MDA produced, the greater was the depletion of tissue antioxidants secondary to oxygen free radical formation during reoxygenation. Results. There was no difference in the prebypass antioxidant reserve capacity between cyanotic and acyanotic hearts ( versus nmol MDA/g protein). However, after the initiation of CPB without leukocyte filtration, MDA production rose markedly in the cyanotic (groups 1 and 2) as compared with the acyanotic hearts (322% versus 40%; p < 0.05), indicating a depletion of antioxidants. In cyanotic hearts, initiating CPB at an FiO 2 of 1.0 (group 1) resulted in increased MDA production (407% versus 227%) as compared with hearts in which CPB was initiated at an FiO 2 of 0.21 (group 2), indicating a greater generation of oxygen free radicals in group 1. Conversely, there was only a minimal increase in MDA production in 8 of the 21 infants (group 3) in whom white blood cells were effectively filtered (19% versus 322%; p < 0.05). Conclusions. First, increased amounts of oxygen free radicals are generated in cyanotic infants with the initiation of CPB. Second, this production is reduced by initiating CPB at an FiO 2 of 0.21 or by effectively filtering white blood cells. Third, these changes parallel those seen in the acute experimental model, validating its use for future study. (Ann Thorac Surg 1997;64:1381 8) 1997 by The Society of Thoracic Surgeons Recent experimental studies have documented that 1 to 2 hours of acute hypoxia followed by abrupt reoxygenation results in an injury characterized by a decrease in systolic contractility, an increase in diastolic stiffness, and elevated pulmonary vascular resistance [1 4]. This injury, which has been referred to as the reoxygenation injury, is mediated by oxygen free radicals and can be modified by leukocyte depletion or by reoxygenating at a lower oxygen concentration [4, 5]. However, the relevance of these experimental findings has been questioned, because chronic hypoxia secondary to congenital heart disease may allow compensatory changes to develop that can prevent or modify the reoxygenation injury. We therefore used the same biochemical tests as used in our experimental studies to examine the effects of abrupt reoxygenation using cardiopulmonary bypass in cyanotic and acyanotic infants undergoing operative repair (1) to determine whether a similar reoxygenation injury occurs clinically and (2) to determine whether it can be modified by leukocyte depletion or a lower fraction of inspired oxygen. Material and Methods Institutional and informed consent was obtained in 28 patients. Seven had a normal oxygen saturation and were considered to be acyanotic, whereas 21 had an oxygen saturation of less than 85% and were considered to be cyanotic (Table 1). After systemic heparinization, cannulas were placed in the ascending aorta and right atrium Presented at the Thirty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Feb 3 5, Address reprint requests to Dr Allen, Division of Cardiothoracic Surgery, University of Illinois at Chicago, 840 S. Wood St, 417 CSB (M/C 958), Chicago, IL This article has been selected for the open discussion forum on the STS Web site: by The Society of Thoracic Surgeons /97/$17.00 Published by Elsevier Science Inc PII S (97)

2 1382 ALLEN ET AL Ann Thorac Surg PREVENTING REOXYGENATION INJURY 1997;64: Table 1. Preoperative Diagnosis Diagnosis No. of Patients Acyanotic patients Atrial septal defect 3 Ventricular septal defect 4 Cyanotic patients Transposition of great arteries 5 Hypoplastic left heart syndrome 5 Tetralogy of Fallot 6 Total anomolous pulmonary venous return 3 Univentricular heart 2 for single or bicaval cannulation. The bypass circuit was heparinized, primed with packed red blood cells, which were first washed (Dideco Shiley auto transfusion system; Shiley Corp, Irvine, CA), one unit of fresh frozen plasma, and 200 mg of calcium gluconate; the ph was adjusted with NaHCO 3. After cardiopulmonary bypass was instituted, the aortic pressure was maintained at 30 to 50 mm Hg by adjusting flow from 2.0 to 2.5 L min 1 m 2. Experimental Groups acyanotic infants. In the 7 acyanotic infants the cardiopulmonary bypass prime was circulated using 100% oxygen (partial pressure of oxygen [PO 2 ], 400 to 550 mm Hg). After the initiation of bypass this was decreased slowly to maintain an arterial PO 2 of 200 to 300 mm Hg. cyanotic infants. In 7 cyanotic infants (group 1) the cardiopulmonary bypass prime was circulated using 100% oxygen (PO 2, 400 to 550 mm Hg), and after initiating bypass this was decreased slowly to keep the arterial PO 2 between 200 and 300 mm Hg. In 6 infants (group 2) the cardiopulmonary bypass prime was circulated using 21% oxygen, resulting in a PO 2 of 140 to 155 mm Hg. Bypass was then initiated, and 5 to 10 minutes later the oxygen level was increased slowly to maintain an arterial PO 2 of 200 to 300 mm Hg. Eight other cyanotic patients (group 3) had blood for the bypass prime passed through a Pall RC-400 leukocyte filter (Pall Biomedical, Glencoe, NY) before it was added to the bypass circuit. A continuous leukocyte-depleting BC-1 filter (Pall Biomedical) was placed in the arterial line circuit. All these patients were newborns (5 to 21 days old) weighing between 2.5 and 4.0 kg and therefore required bypass flows of less than 700 ml/min. The white blood cell (WBC) filters (BC-1) were used for the first 30 to 60 minutes of bypass and then removed by redirecting the arterial line (using a Y connector) across a standard Pall 40- m (LPE-1440) arterial filter for the remainder of the operation. Pressure across the BC-1 and standard arterial filters was measured at 5, 10, and 30 minutes. The protocol described for group 1 was followed in 3 of these infants, initially circulating the prime using 100% oxygen, and the protocol for group 2 was followed in the other 5 infants, circulating the prime using 21% oxygen. Tissue and Serum Measurements A small piece of right atrial tissue was removed in all patients during venous cannulation, and another piece was removed 10 to 20 minutes after initiating cardiopulmonary bypass and before aortic cross-clamping. The tissue was immediately frozen in liquid nitrogen, and the antioxidant reserve capacity in the tissue was analyzed. Blood was obtained from the patient before bypass, 5 and 30 minutes after initiating cardiopulmonary bypass, and on arrival in the intensive care unit. This was analyzed for the total WBC count (Cell-dyne 3000; Abbott Laboratories, Deerfield, IL), neutrophil count (manual differential), and platelet count (expressed per milliliter times 1,000). Biochemical Measurements The myocardial antioxidant reserve capacity was assessed according to the method of Godin and Garnett [6] by determining in vitro lipid peroxidation in cardiac tissue that was homogenized and incubated with t- butylhydroperoxide at a concentration of 4 mmol/l for 15 minutes at 37 C. Lipid peroxidation was determined by measuring thiobarbituric acid reactive substances spectrophotometrically at 532 nm. A standard curve is run simultaneously and lipid peroxidation is expressed as nanomoles malondialdehyde (MDA) per gram protein of heart tissue. Ten to 15 mg of dry tissue was homogenized at 4 C in 1.75 ml of 0.5-mol/L Tris buffer and 0.1-mmol/L EDTA (ph, 7.6). Aliquots of the particle-free homogenate were incubated with 4-mmol/L t-butylhydroperoxide for 30 minutes at 37 C. The mixture was then deproteinized with an equal volume of 28% trichloroacetic acid containing 0.1-mmol/L sodium arsenite. The concentration of MDA in the protein-free supernatant was determined by heating it in boiling water for 15 minutes with 0.5% thiobarbituric acid in mmol/L sodium hydroxide and reading the color at 532 nm. The antioxidant reserve capacity is expressed as the percentage increase in MDA production as compared with prebypass levels. This allows each child to act as its own control, because it is the change in tissue antioxidants that quantitates exposure to oxygen free radicals. Statistics Data were analyzed using JMP V2.0 (SAS Institute, Carey, NC) on a Macintosh IIVX computer (Apple Inc, Cupertino, CA). Paired Student s t test was used for comparing variables among experimental groups at a probability level of less than Group data are expressed as the mean standard error of the mean. Results The average infant age was 8 2 months (range, 4 days to 37 months), and the preoperative diagnoses are listed in Table 1. The pressure across the BC-1 filter ranged between 14 and 45 mm Hg (28 4 mm Hg) at flows of 200 to 600 ml/min ( ml/min) and did not increase during the 30 to 60 minutes the filter was in place. The blood was then redirected across a standard arterial filter,

3 Ann Thorac Surg ALLEN ET AL 1997;64: PREVENTING REOXYGENATION INJURY 1383 Fig 2. Antioxidant reserve capacity in acyanotic and cyanotic infants after reoxygenation using cardiopulmonary bypass without leukofiltration. Note that there is a marked loss of the antioxidant reserve capacity in cyanotic infants, indicating exposure to a large amount of oxygen free radicals during reoxygenation (see text for description). (p 0.05; MDA malondialdehyde.) Fig 1. (A) Total white blood cell (WBC) count and (B) neutrophil count, with and without leukofiltration, before bypass, 5 and 30 minutes after initiating bypass, and on arrival in the intensive care unit. Note that, in infants undergoing leukofiltration, the total white blood cell and neutrophil counts decrease much more abruptly after 5 minutes of bypass and remain low 30 minutes later. However, they have returned to the baseline value by the time the infants arrive in the intensive care unit (*p 0.05.) and the pressure ranged from 5 to 10 mm Hg at flows of 150 to 300 ml/min. The flow at this time was usually lower because systemic hypothermia had already been instituted. The prebypass total WBC ( ), neutrophil ( ), and platelet (309 12) counts were within normal limits and not significantly different between acyanotic and cyanotic patients. In the non WBC-filtered infants, the total WBC ( ) and neutrophil ( ) counts were moderately decreased 5 minutes after initiating bypass. However, they were already increasing after 30 minutes of bypass (WBC, ; neutrophil, ) and were increased above prebypass levels on arrival of the infants in the intensive care unit (WBC, ; neutrophil, ; Fig 1). Conversely, if leukocyte depletion was used, the total WBC and neutrophil counts fell abruptly 5 minutes after starting bypass (WBC, ; neutrophil, ; p 0.05 versus nonfiltered patients) and remained low 30 minutes later (WBC, ; neutrophil, ). Although on arrival of these infants in the intensive care unit these levels were not as in high as those in the nonfiltered patients, they had returned to normal levels (WBC, ; neutrophil, ; see Fig 1). Platelet counts followed similar trends, with a greater reduction in the WBC filter patients at 5 minutes (46 13 versus ; p 0.05) and 30 minutes (48 12 versus 98 12; p 0.05) after initiating bypass and a lower level on their arrival in the intensive care unit (93 14 versus ; p 0.05). Results of tissue measurements are summarized in Figures 2 to 4. There was no difference in the prebypass antioxidant reserve capacity between cyanotic and acyanotic hearts ( versus nmol MDA/g protein). However, after cardiopulmonary bypass was initiated, MDA production rose markedly in cyanotic infants without a WBC filter (groups 1 and 2) as compared with the production in acyanotic infants (322% 140% versus 40% 29%; p 0.05) (see Fig 2). In cyanotic infants initiating bypass using 100% oxygen (group 1) Fig 3. Antioxidant reserve capacity in cyanotic infants after reoxygenation with cardiopulmonary bypass using 100% oxygen, 21% oxygen, or white blood cell (WBC) filtration. Note that there is a substantial loss of the antioxidant reserve capacity in infants reoxygenated using 100% oxygen, indicating exposure to a large amount of oxygen free radicals. Although this injury is reduced by using 21% oxygen, there is still a substantial depletion of tissue antioxidants. Conversely, there is almost no change in the antioxidant reserve capacity in infants reoxygenated using leukocyte filtration (see text for description). (p 0.05; MDA malondialdehyde.)

4 1384 ALLEN ET AL Ann Thorac Surg PREVENTING REOXYGENATION INJURY 1997;64: Fig 4. Antioxidant reserve capacity in cyanotic infants reoxygenated using cardiopulmonary bypass with leukocyte filtration at either 100% or 21% oxygen. Note that, although leukocyte filtration results in only a small change in the antioxidant reserve capacity, the generation of oxygen radicals is lowest if white blood cell filtration is combined with 21% oxygen (see text for description). (p 0.05; MDA malondialdehyde.) resulted in a greater MDA production than that in infants in whom bypass was initiated using 21% oxygen (group 2), indicating greater exposure to oxygen free radicals during reoxygenation (see Fig 3), but this did not reach statistical significance (407% 170% versus 227% 78%). Conversely, there was only a minimal increase in MDA production in the 8 of 21 cyanotic infants (group 3) in whom WBCs were filtered (19% 11% versus 322% 140%; p 0.05), indicating a marked reduction in the loss of tissue antioxidants when a leukocyte depletion filter was used. Indeed the antioxidant reserve capacity did not differ from that in the acyanotic infants (19% 11% versus 40% 29%). Although all patients undergoing leukocyte depletion had relatively small changes in their antioxidant reserve capacity, the smallest increase occurred in the 5 infants who also had bypass initiated using 21% oxygen (6% 6% versus 39% 24%; p 0.05) (see Fig 4). Comment Previous clinical and experimental studies have shown that an unintended injury occurs in response to sudden reoxygenation of the cyanotic heart and this injury is mediated by oxygen free radicals and results in significant myocardial depression and pulmonary dysfunction [1 4]. These adverse effects have been modified experimentally by initiating bypass at a lower oxygen concentration or using WBC filtration [4, 5]. The present study supports the validity of these experimental findings, as shown by the finding that cyanotic patients reoxygenated on cardiopulmonary bypass generated abundant oxygen free radicals, this production was reduced by using a lower level of oxygen, and leukocyte depletion resulted in the lowest production of oxygen free radicals. Myocardial tissue was biopsied before and after bypass was initiated to determine the antioxidant reserve capacity, thus allowing the quantification oxygen free radical formation during reoxygenation. The antioxidant reserve capacity is determined by adding a strong oxidant (tbutylhydroperoxide) to myocardial tissue and measures the tissue s ability to scavenge the resulting oxygen radicals and prevent MDA formation (a by-product of lipid peroxidation) [1, 2, 4, 6]. Therefore it tests the endogenous tissue stores of oxygen radical scavengers (eg, glutathione, vitamin E, superoxide dismutase, catalase). The more MDA produced, the lower the levels of these endogenous stores. Tissue antioxidants are lost when oxygen free radicals are produced and need to be scavenged, such as when the hypoxemic heart is abruptly reoxygenated using cardiopulmonary bypass [1, 2, 4]. We chose this test because it has been used in numerous experimental studies of acute hypoxia and therefore allowed us to compare our clinical results with the results of previous experimental studies to determine the validity of our experimental model [1 4]. In addition, there appears to be a close link between antioxidant depletion, oxidant damage, and cardiac and pulmonary dysfunction [1 4]. The antioxidant reserve capacity also predicts the ability of the heart to withstand a subsequent ischemic challenge. Only a minor functional impairment develops after aortic clamping in normal hearts with abundant antioxidants, whereas hearts with a limited antioxidant reserve capacity exhibit marked contractile depression after cardioplegic arrest [1, 5, 7, 8]. There was no difference in the prebypass antioxidant reserve capacity between cyanotic and acyanotic hearts. This parallels experimental findings observed after acute hypoxia [1, 4]. However, abrupt reoxygenation of cyanotic infants without a WBC filter (groups 1 and 2) resulted in a significant depletion of endogenous tissue antioxidants (see Fig 2). Because the prebypass endogenous tissue stores of antioxidants were not different between acyanotic and cyanotic hearts, this suggests that the abrupt reoxygenation of chronically hypoxemic infants causes the generation of abundant oxygen free radicals. Conversely, the initiation of bypass in acyanotic infants caused a minimal change in the antioxidant reserve capacity, implying that, in the absence of hypoxia, only a small quantity of oxygen free radicals is generated. This small increase in the antioxidant reserve capacity of acyanotic hearts may reflect the high levels of oxygen (100%) used in the bypass circuit, as there is recent evidence that even a PO 2 of 185 mm Hg may be detrimental [9]. Alternatively, cardiopulmonary bypass has been shown to produce an inflammatory reaction characterized by the activation of numerous pathways, some of which may result in the generation of oxygen radicals [10]. Cyanotic infants reoxygenated using a PO 2 of 400 to 550 mm Hg (group 1, 100% oxygen) showed the greatest loss of the myocardial antioxidant reserve capacity (highest MDA formation), indicating the greatest exposure to oxygen free radicals (see Fig 3). This phenomenon of oxidant damage in response to reoxygenation has previously been demonstrated in cyanotic patients [8, 11, 12]. However, the same test (antioxidant reserve capacity) was used in the present investigation as that used in

5 Ann Thorac Surg ALLEN ET AL 1997;64: PREVENTING REOXYGENATION INJURY 1385 experimental studies of acute hypoxia to allow comparisons [1, 2, 4]. Although both cyanotic infants and acute hypoxic animals exposed to 100% oxygen show a reoxygenation injury, the quantity of MDA generated is four to six times greater in cyanotic infants [1, 4]. This suggests a greater production of oxygen free radicals in response to reoxygenation after chronic cyanosis. One reason for this may be that cyanotic infants can become ischemic during exercise or stress, subjecting them to not only a hypoxic, but also an ischemic, injury [13, 14]. In contrast, our acute hypoxic experimental model does not result in ischemia [4, 15, 16]. Alternatively, compensatory changes that occur in the cyanotic infant may predispose to the generation of larger amounts of oxygen radicals with the reintroduction of high levels of oxygen. This oxidant injury may explain why ventricular function is often depressed in cyanotic infants undergoing extracorporeal membrane oxygenation or surgical repair [7, 12, 17, 18]. Compared with cyanotic infants reoxygenated using 100% oxygen (group 1), the initiation of bypass using 21% oxygen (group 2) reduced the change in the antioxidant reserve capacity, but this did not reach statistical significance because of the large variability in the patients and the small number of patients (see Fig 3). The improvement in the antioxidant reserve capacity that occurs with lower levels of oxygen also parallels experimental results, but the change was still substantially greater than that seen after acute experimental hypoxia, suggesting once again that the cyanotic patient is more susceptible to an oxygen-mediated injury [5]. However, the use of 21% oxygen to prime the bypass circuit resulted in a PO 2 of 140 to 155 mm Hg, which was substantially higher than the PO 2 of 80 to 100 mm Hg used in the experimental study [5]. This may have at least partially accounted for the lack of improvement seen in response to 21% oxygen, as several investigators have shown that the oxygen free radical production and myocardial injury that occur after the reoxygenation of isolated heart preparations are proportional to the oxygen tension [19, 20]. Because the current membrane oxygenators are so efficient, however, it will take oxygen concentrations of less than 21% to obtain a PO 2 of 80 to 100 mm Hg in the bypass prime. These levels (PO 2, 80 to 100 mm Hg) have also been shown to improve tissue perfusion during cardiopulmonary bypass, and therefore higher oxygen levels are probably never needed, because a PO 2 of greater than 100 to 150 mm Hg confers only a negligible increase in the oxygen content [5, 9]. Activated WBCs have been shown to play a major role in the generation of oxygen free radicals after ischemia [4, 21 23]. It therefore seems likely they are also active in the reoxygenation injury, because both ischemia and hypoxia subject myocardial tissue to low levels of oxygen [1, 3, 11, 21]. Our data support this hypothesis. When the number of neutrophils was reduced in cyanotic infants by a leukocyte-depleting filter (group 3), the detrimental effects of sudden reoxygenation were eliminated, resulting in preservation of the antioxidant reserve capacity (see Fig 3). This is once again precisely what was demonstrated in the experimental setting after acute hypoxia, in which it correlated with an improvement in myocardial and pulmonary function [4]. In addition, other clinical investigators have shown that even partially removing WBCs during cardiopulmonary bypass can cause both oxygen radical formation and postoperative pulmonary vascular resistance to be reduced [24]. It is therefore likely that an improvement in the antioxidant reserve capacity translates into an improvement in myocardial and pulmonary function. Furthermore, although no patient reoxygenated with leukocyte-depleted blood showed a substantial change in the antioxidant reserve capacity, the generation of oxygen free radicals was further suppressed by using 21% oxygen (see Fig 4). Indeed the antioxidant reserve capacity in these infants was unchanged from baseline values and even lower than that in acyanotic patients, suggesting the effects of oxygen and WBC filtration are additive. In the present study, WBC filtration substantially reduced the number of leukocytes, both initially and after 30 minutes of cardiopulmonary bypass (see Fig 1). Besides the injury caused by activated WBCs, they also bind to activated platelets, and platelet counts were substantially lower in the leukocyte-depleted infants. By removing the activated platelets, a WBC filter may also help prevent other adverse effects, such as thromboxane release and vasoconstriction [21, 25 27]. In addition, because activated WBC-platelet complexes are larger, they are more likely to be trapped by a filter. Therefore, even though the neutrophil count was not reduced to 0, it is probable that very few activated WBCs escaped filtration. Unfortunately, no WBC filter currently on the market is ideal. We chose the Pall BC-1 filter because it is the most efficient leukocyte filter available and it worked well in our experimental study [4]. At flows of 500 to 600 ml/min the BC-1 removes almost all WBCs in one pass, and these flows are adequate for 3 to 4-kg newborns. We kept the filter in the bypass circuit for up to 60 minutes without any complications, and the pressure across the filter remained low with no significant change over time. However, because WBC filters are flow dependent, if higher flows are utilized, these filters are less efficient. In addition, the BC-1 can only be used up to flows of approximately 600 to 700 ml/min. Therefore it would not be applicable in larger infants. Instead, a Pall LG-6 filter would be required, because this filter can be used at flows of up to 6 L/min. However, it is less efficient at removing WBCs during the first pass, and instead removes neutrophils slowly over time. However, because this filter is also flow dependent, it may remove a large number of WBCs at lower flow rates. Therefore it may still be effective in infants or small children. In contrast, there is no good arterial leukocyte-depleting filter for larger children. Because the reoxygenation injury probably occurs early, the complete removal of WBCs is important during the initial reintroduction of oxygen. Because of this, we prefiltered all blood added to the cardiopulmonary circuit using a Pall RC-400 leukocyte filter. This resulted in extremely low WBC counts in the prime and has been shown by Komai and associates [24] to reduce oxygen radical formation and improve postop-

6 1386 ALLEN ET AL Ann Thorac Surg PREVENTING REOXYGENATION INJURY 1997;64: erative pulmonary vascular resistance even in acyanotic infants. Therefore, even if an in-line arterial filter is not used, we strongly believe blood for the bypass prime should always be leukocyte depleted. Although there is some concern that leukocyte depletion may cause postoperative infection rates to increase, this has not been noted in more than 21,000 patients; indeed, there is evidence that it may even lower the risk of infection (Ortolano, personal communication, 1997 [28, 29]). In fact, the WBC count had been restored to prebypass levels by the time the patient arrived in the intensive care unit, making an increased risk of infection unlikely. There are potential limitations to this study. The ages and preoperative myocardial state of the acyanotic and cyanotic infants were different. The acyanotic patients were older, hemodynamically more stable, and undergoing elective operations. In contrast, the cyanotic infants were predominantly newborns requiring urgent operations for the management of much more complex cardiac problems. Despite these differences, however, the preoperative antioxidant reserve capacity did not differ between the cyanotic and acyanotic infants nor did it correlate with the age of the patient. Similarly, the postoperative antioxidant reserve capacity seemed to correlate only with the method of reoxygenation, and not with the patient s age. This is evident from the fact that, although the type of congenital anomaly and the age of the patients were similar in the cyanotic infants in groups 1 (100% oxygen) and 2 (21% oxygen), the change in the postoperative antioxidant reserve capacity was markedly different. The specific mechanism responsible for the increase in oxygen free radical formation is unknown and was not addressed. However, it appears that it is at least partially related to the oxygen concentration and the WBC count, as indicated by the finding that modifying either of these factors reduced oxygen radical formation during reoxygenation. Although the effect of chronic cyanosis on neutrophil function is unknown, hypoxia is known to alter the vascular endothelial cell, thereby increasing its ability to bind to activated WBCs. This may partially explain why neutrophils cause adverse effects in cyanotic infants. In summary, our study findings support those of previous investigations and show that cyanotic infants are predisposed to the generation of large quantities of oxygen free radicals in response to the initiation of cardiopulmonary bypass [8, 11]. However, oxygen free radical production can be limited by decreasing the oxygen concentration of the bypass circuit or, more effectively, by leukocyte filtration. Because myocardial necrosis and decreased ventricular function can occur in cyanotic infants after apparently successful surgical procedures, reducing the oxygen level and leukocyte filtration may improve operative results, as experimentally myocardial function has been found to correlate with the antioxidant reserve capacity [1, 4, 5, 7]. Furthermore, and possibly most importantly, these results closely parallel the experimental findings observed after acute hypoxia, suggesting that this model is clinically applicable to the cyanotic infant and can be used for further study of this phenomenon. Supported in part by the Pilesburg Fellowship (Dr Kronon). We thank Ms Kym Montecinos and Ms Christina Green for organizational and secretarial assistance, and Mr Greg Mork and Mr James Murray for technical support. References 1. Buckberg GD. Studies of hypoxemic/rexoygenation injury: I. Linkage between cardiac function and oxidant damage. J Thorac Cardiovasc 1995;110: Ihnken K, Morita K, Buckberg GD, et al. Studies of hypoxemic/reoxygenation injury: without aortic clamping. II. Evidence for reoxygenation damage. J Thorac Cardiovasc Surg 1995;110: Ihnken K, Morita K, Buckberg GD, Sherman MP, Young HH. Studies of hypoxemic/reoxygenation injury: without aortic clamping. III. Comparison of the magnitude of damage by hypoxemia/reoxygentation versus ischemia/reperfusion. J Thorac Cardiovasc Surg 1995;110: Bolling KS, Halldorsson A, Allen BS, et al. Prevention of the hypoxic/reoxygenation injury using a leukocyte depleting filter. J Thorac and Cardiovasc Surg (in press). 5. Morita K, Ihnken K, Buckberg G, Sherman MP, Young HH. Studies of hypoxemic/reoxygenation injury: without aortic cross clamping. IX. Importance of avoiding perioperative hyperoxemia in the setting of previous cyanosis. J Thorac and Cardiovasc Surg 1995;110: Godin D, Garnett M. Altered antioxidant status in the ischemic/reperfused rabbit myocardium: effects of allopurinol. Can J Cardiol 1989;5: Del Nido PJ, Mickle DAG, Wilson GJ, et al. Inadequate myocardial protection with cold cardioplegic arrest during repair of tetralogy of Fallot. J Thorac Cardiovasc Surg 1988; 95: Teoh KH, Mickle DAG, Weisel RD, Li R, Tumiati L, Coles JG. Effect of oxygen tension and cardiovascular operations on the myocardial antioxidant enzyme activities in patients with tetralogy of Fallot and aorta-coronary bypass. J Thorac Cardiovasc Surg 1992;104: Joachimsson P, Sjoberg F, Forsman M, Johansson M, Casimir Ahn H, Rutberg H. Adverse effects of hyperoxemia during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1996;112: Castañeda AR, Jonas RA, Mayer JE Jr, Hanley FL. Myocardial preservation in the immature heart. In: Castañeda AR, Jonas RA, Mayer JE Jr, Hanley FL, eds. Cardiac surgery of the neonate and infant. Philadelphia: Saunders, 1994: Del Nido PJ, Mickle DAG, Wilson G, Benson LN, Coles JG, Trusler GA. Evidence of myocardial free radical injury during elective repair of of tetralogy of Fallot. Circulation 1987;76: Martin G, Short B, Abbott C, O Brien A. Cardiac stun in infants undergoing extracorporeal membrane oxygenation. J Thorac Cardiovasc Surg 1991;101: Boucek RJ Jr., Kasselberg AG, Boerth RC, Parrish MD, Graham TP Jr. Myocardial injury in infants with congenital heart disease: evaluation by creatine kinase MB isoenzyme analysis. Am J Cardiol 1982;50: Graham J, Erath HG Jr, Buckspan GS, Fisher RD. Myocardial anaerobic metabolism during isoprenaline infusion in a cyanotic animal model: possible cause of myocardial dysfunction in cyanotic congenital heart disease. Cardiovasc Res 1979;13: Bolling KS, Kronon M, Allen BS, et al. Myocardial protection in normal and hypoxically stressed neonatal hearts: the superiority of hypocalcemic versus normocalcemic blood cardioplegia. J Thorac Cardiovasc Surg 1996;112:

7 Ann Thorac Surg ALLEN ET AL 1997;64: PREVENTING REOXYGENATION INJURY Bolling KS, Kronon M, Allen BS, Wang T, Ramon S, Feinberg H. Myocardial protection in normal and hypoxically stressed neonatal hearts: the superiority of blood versus crystalloid cardioplegia. J Thorac and Cardiovasc Surg (in press). 17. Hirschl R, Heiss K, Bartlett R. Severe myocardial dysfunction during extracorporeal membrane oxygenation. J Pediatr Surg 1992;27: Rocchini AP, Keane JF, Castaneda AR, Nadas AS. Left ventricular function following attempted surgical repair of tetralogy of Fallot. Circulation 1978;57: Gauduel Y, Menasché P, Duvelleroy M. Enzyme release and mitochondrial activity in reoxygenated cardiac muscle: relationship with oxygen-induced lipid perioxidation. Gen Physiol Biophys 1989;8: Hearse D, Humphrey S, Bullock G. The oxygen paradox and the calcium paradox: two facets of the same problem? J Mol Cell Cardiol 1978;10: Byrne J, Appleyard R, Lee C, et al. Controlled reperfusion of the regionally ischemic myocardium with leukocyte-depleted blood reduces stunning, the no-reflow phenomenon, and infarct size. J Thorac Cardiovasc Surg 1992;103: Fantone J, Ward P. Role of oxygen-derived free radicals and metabolites in leukocyte-dependent inflammatory reactions. Am J Pathol 1982;107: Breda MA, Drinkwater DC, Laks H, Bhuta S, Corno A. Prevention of reperfusion injury in the neonatal heart with leukocyte-depleted blood. J Thorac Cardiovasc Surg 1989;97: Komai H, Yamamoto F, Tanaka K, Yagihara T, Kawashima Y. Prevention of lung injury during open heart operations for congenital heart defects. Ann Thorac Surg 1994;57: Mullane K, Read N, Salmon J, Moncada S. Role of leukocyte in acute myocardial infarction in anesthetized dogs: relationship to myocardial salvage by anti-inflammatory drugs. Circ Res 1983;53: Harlan J. Leukocyte-endothelial interactions. Blood 1985;65: Engler R, Schmid-Schoebein G, Pavelec R. Role of leukocyte capillary plugging in preventing myocardial reperfusion. Circulation 1981;64:IV Jewett PH, Kissmeyer-Nielsen P, Wolff B, Qvist N. Randomised comparison of leucocyte-depleted versus buffy-coatpoor blood transfusion and complications after colorectal surgery. Lancet 1996;348: Bowden AR, Slichter JS, Sayer M, et al. A comparison of filtered leukocyte-reduced and cytomegalovirus (CMV) seronegative blood products for the prevention of transfusionassociated CMV infection after marrow transplant. Blood 1995;86: DISCUSSION DR WILLIAM A. BAUMGARTNER (Baltimore, MD): I congratulate you on a terrific study and for getting the leukocyte filters to work. We have had an interest in this for many years. In a clinical adult setting, we have never been able to reduce leukocytes to the level that I think you need to demonstrate a reduction in oxygen free radical release. I have a couple questions. We saw that there was a correlation with a rise in white blood cell counts during filtration once rewarming took place. I wonder if you saw that in your clinical study? Second, how long were your patients on bypass, and what were your mean bypass runs? Do you think that the reduction in white blood cells in your model was a result of the size of the patients, as opposed to some other mechanism? My last question is, did you look at any other evidence that led to a decrease in the overall inflammatory state usually initiated by cardiopulmonary bypass, such as cytokine production or lung function, which at least experimentally has enjoyed a really beneficial effect of reducing the number of white blood cells? I enjoyed your paper very much. DR ALLEN: Thank you for your comments. We are well aware of your previous work, and it helped stimulate our interest in this subject. The question of how to obtain low white blood cell counts in patients is an excellent one, and it is extensively addressed in the manuscript. We all recognize that leukocytes are deleterious, but because most filters are unable to effectively reduce the white blood cell count, they have only been sporadically used. However, we did several things in this study that allowed us to obtain low white blood cell counts. Because we used blood to prime the bypass circuit, we prefiltered the blood using a Pall RC-400 filter in all patients. This resulted in white blood cell counts in the cardiopulmonary bypass circuit of almost zero. Because we believe the reoxygenation injury occurs early, initiating cardiopulmonary bypass with very low white blood cell counts is very important. Next, instead of using one of the standard arterial white blood cell filters, we used one designed for the cardioplegia circuit (Pall-BC-1) but placed it in the arterial line. The company says these filters can be used at flows of up to 500 ml/min, but we used them at flows of up to 700 ml/min. However, this meant they could only be used in patients weighing less than 4 kg. The advantage of this filter is that it is very effective and removes almost all white blood cells during one pass. The problem is that it gets filled up over time. In contrast, the Pall LG-6 leukocyte filter is specifically designed to be used in the arterial line of the bypass circuit and it can accommodate flows of up to 6 L/min. However, it is not very effective at higher flow rates, removing white blood cells very slowly over time. This probably explains why most investigators have not been able to significantly reduce white blood cell counts in adult patients. We measured the pressure across the BC-1 filter during cardiopulmonary bypass, and it did not increase during the 30 to 60 minutes it was in place. We then removed it from the bypass circuit by redirecting the flow through a standard arterial filter for fear of increased resistance. However, we never had any problems in this study, or in our laboratory, where the filter was left in place for up to 90 minutes. I should mention that although the BC-1 filter cannot be used at higher flow rates, we still leukodeplete the bypass prime in all children, as this has been shown to be beneficial, even when used without an arterial line filter. We have considered using a Pall LG-6 leukocyte-depleting filter in smaller children, say those weighing between 5 and 15 kg, because the removal of white blood cells by this filter is flow dependent. Therefore, at low flow rates, this filter may be more effective than it is in adults. As you have stated, the white blood cell counts do rise over time, and we found that they were back to normal by the time the patients arrive in the intensive care unit (see Figures 1A and 1B in the article). We did not assay the myocardial tissue for other evidence of inflammation, because most of our patients were infants or neonates and the amount of tissue that could be sampled was limited. This is especially true in our patients because we took biopsy specimens both before and after bypass to make comparisons. However, as you have mentioned, there is abundant

8 1388 ALLEN ET AL Ann Thorac Surg PREVENTING REOXYGENATION INJURY 1997;64: evidence that many inflammatory substances are reduced if the white blood cells are effectively filtered out. DR JOHN E. MAYER, JR (Boston, MA): This makes me wonder a little bit about what the effects of chronic hypoxia or cyanosis are on neutrophil function and on neutrophils in general. Is there any information about that that you know of? DR ALLEN: I am not currently aware of the effects of chronic hypoxia on neutrophil function. Several studies have shown that neutrophils can be deleterious after ischemia, and this study shows that neutrophils may also be detrimental after chronic hypoxia. Chronic hypoxia does, however, affect the immune system as well as increase neutrophil adherence to endothelial cells. This would imply that hypoxia increases the ability of the neutrophil to cause injury. DR MAYER: The effect of chronic cyanosis on endothelial cells would be another interesting question. DR ALLEN: Experimental studies have shown that endothelial function may be preserved, but I am unaware of any clinical studies that have specifically examined endothelial cell function after chronic hypoxia. However, the vascular endothelial cell does increase neutrophil adherence during hypoxia, thereby increasing the chance for injury. DR MAYER: And one last question. Did you have any clinical impressions or do you have any data on what the impact was on myocardial or other organ system function? DR ALLEN: I do not have any firm data, but I do have several clinical impressions, both my own and those of our pediatric intensivists. This group of infants had a wide variety of diagnoses, and most were very sick with severe congenital abnormalities. For instance, 6 patients had hypoplastic left heart syndrome. This would be very difficult to compare with, say, tetralogy of Fallot. It was therefore impossible for us to compare the hemodynamics or other organ system function in such a small, diverse group of patients. However, our intensivists always seemed to be able to tell us which patients had undergone white blood cell filtration. They would say This must be a white blood cell filtered patient, because the oxygenation and pulmonary compliance are excellent. We also seemed to have fewer problems with postoperative pulmonary hypertension and lower requirements for inotropic support. So, anecdotally we noticed a significant difference, but I have no hard data to support these observations. DR DUKE E. CAMERON (Baltimore, MD): I think this subject of the modification of perfusion for the cyanosed infant is fascinating. And many of the things we have been doing for years may not be right, the most obvious being hyperoxic reperfusion after a period of aortic cross-clamping. We have known for years that oxygen on reperfusion is detrimental, yet we still have the arterial blood PO 2 up to 300 or 400 mm Hg when the cross-clamp comes off. I would like to take a quick survey, if I could, among those who do operations on cyanosed infants, on whether you modify your perfusion in terms of the oxygen concentration or PO 2 level on bypass or use leukocyte filtration. Let me ask first if you could raise your hand to show if you do pediatric heart operations? (Audience responds) DR CAMERON: And how many do not modify perfusion for cyanosed infants? (Audience responds) DR CAMERON: And those who do? (Audience responds) DR CAMERON: I think there is a slight majority of those who, at least up to today, have not modified their perfusion practice. Thank you very much, Dr Allen.