Regional High-Flow Cerebral Perfusion Improves Both Cerebral and Somatic Tissue Oxygenation in Aortic Arch Repair

Regional High-Flow Cerebral Perfusion Improves Both Cerebral and Somatic Tissue Oxygenation in Aortic Arch Repair Kagami Miyaji, MD, PhD, Takashi Miyamoto, MD, PhD, Satoshi Kohira, CCP, Kei-ichi Itatani, MD, Takahiro Tomoyasu, MD, Nobuyuki Inoue, MD, and Kuniyoshi Ohara, MD, PhD Department of Cardiovascular Surgery, Kitasato University, School of Medicine, Sagamihara, Japan Background. Regional cerebral perfusion provides cerebral circulatory support during aortic arch reconstruction. We report the effectiveness of high-flow regional cerebral perfusion (HFRCP) from the right innominate artery to maintain sufficient cerebral and somatic oxygen delivery through collateral vessels. Methods. Frontal cerebral and thoracolumbar probes to measure somatic regional oxygen saturation (rso 2 ) were used to continuously measure oxygenation during cardiopulmonary bypass in 18 patients (weight, 2.1 to 4.3 kg) who underwent arch reconstruction using HFRCP (mean flow, 82; range, 43 to 108 ml/kg/min). Procedures included 9 Norwood procedures, 5 coarctation of aorta/interruption of aorta complex repairs, and 4 aortic arch repairs for a single ventricle. Mean HFRCP duration was 51 17 minutes under moderate hypothermia. Mean radial arterial pressure was kept at less than 50 mm Hg during HFRCP, and chlorpromazine (mean dose, 2.8 mg/kg) was given to all patients before and during HFRCP to increase regional cerebral perfusion flow. Plasma lactate concentration was measured before and after HFRCP. Results. During HFRCP, mean cerebral rso 2 was 78.8% 9.5%, somatic rso 2 was 65.4% 12.1%, and lactate concentration increased from 3.8 2.2 to 5.5 2.1 mmol/l. There was significant correlation between regional cerebral perfusion flow and somatic rso 2. Significant inverse correlations were noted between regional cerebral perfusion flow and the increase of lactate concentration and between somatic rso 2 and the increase of lactate concentration. Conclusions. High-flow regional cerebral perfusion preserved sufficient cerebral and somatic tissue oxygenation during aortic arch repair. The reduction of vascular resistance of collateral vessels increased both cerebral and somatic blood flow, resulting in improved tissue oxygen delivery. (Ann Thorac Surg 2010;90:593 9) 2010 by The Society of Thoracic Surgeons In aortic arch repair, including Norwood stage I palliation, deep hypothermic circulatory arrest (DHCA), which provides surgeons a bloodless and uncluttered operative field, has been used as an adjunctive therapy with cardiopulmonary bypass (CPB). Because of concerns about the effect of DHCA on neurologic complications, alternatives have been sought [1 3]. Regional cerebral perfusion (RCP) has been shown to provide cerebral circulatory support during aortic arch reconstruction. The maintenance of cerebral oxygen saturation during RCP has been demonstrated with near-infrared spectroscopy [2, 4, 5]. In 2001, Pigula and colleagues [5] reported that low-flow RCP (LFRCP) provided somatic circulatory support during neonatal arch surgical procedures and that support of the subdiaphragmatic viscera should improve the ability of neonates to survive the postoperative period [5]. Accepted for publication March 26, 2010. Presented at the Forty-sixth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 25 27, 2010. Address correspondence to Dr Miyaji, Department of Cardiovascular Surgery, Kitasato University, School of Medicine, Kitasato 1-15-1, Sagamihara 228-8555, Japan; e-mail: kagami111@aol.com. In our institution, high-flow RCP (HFRCP) from the right innominate artery has been induced to maintain sufficient cerebral and somatic oxygen delivery through collateral vessels. We used near-infrared spectroscopy devices to assess changes in oxyhemoglobin saturation in cerebral [2, 4, 5] and somatic [6, 7] regional circulations, before, during, and immediately after aortic arch repair to estimate the effectiveness of HFRCP for cerebral and somatic tissue oxygenation. Material and Methods This study received Institutional Review Board approval. Parental consent was obtained for all patients to participate in this study. Patients and Surgical Technique The study comprised 18 patients with a mean weight of 3.0 kg (range, 2.1 to 4.3 kg) who underwent arch reconstruction using HFRCP and whose clinical physiologic data were reviewed. Their mean age was 28 days (Table 1). The types of procedures included 9 Norwood stage I palliations, 5 coarctation of aorta/interruption of aorta 2010 by The Society of Thoracic Surgeons 0003-4975/$36.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2010.03.113

594 MIYAJI ET AL Ann Thorac Surg REGIONAL HIGH-FLOW CEREBRAL PERFUSION 2010;90:593 9 Abbreviations and Acronyms CPB cardiopulmonary bypass DHCA deep hypothermic circulatory arrest HFRCP high-flow regional cerebral perfusion LFRCP low-flow regional cerebral perfusion RCP regional cerebral perfusion rso 2 regional oxygen saturation SD standard deviation complex repair, and 4 aortic arch repairs for a single ventricle (Table 1). Among these 18 patients, 6 (33.3%) needed preoperative mechanical ventilation, and 9 (50%) needed inotropic support because of unstable hemodynamics. All 18 patients underwent RCP through the right innominate artery. Norwood stage I palliation was performed using techniques previously described [8]. A 3.5-mm polytetrafluoroethylene tube graft was anastomosed to the innominate artery as an arterial catheter. During CPB, the patient was cooled to 25 C. The ascending aorta was transected, and a cardioplegic solution was given. After cardiac arrest, the ascending aorta was incised vertically down to the sinus level and anastomosed to the main pulmonary artery in a side-to-side fashion to maintain sufficient coronary blood flow. Under RCP, the descending aorta was clamped, and the main pulmonary artery was anastomosed directly to the aortic arch. The right ventricular pulmonary artery conduit was contracted through the right side of the neoaorta using a 5-mm expanded polytetrafluoroethylene graft. For aortic arch repair, under RCP, the descending aorta was directly anastomosed to the ascending aorta or the aortic arch after complete resection of ductal tissue. For patients with a coarctation of aorta/interruption of aorta complex, the ventricular septal defect was closed under cardiac arrest after completion of the aortic arch repair. Table 1. Patient Demographic Data Variable Mean SD or No. (%) Age, d 28 47 Weight, kg 3.0 0.6 Type of procedure Norwood 9 CoA/IAA complex 5 Arch repair for SV 4 Pre-op mechanical ventilation 6/18 (33.3) Pre-op inotropic support 9/18 (50.0) CPB time, min 171 51 RCP time, min 51 17 RCP flow, ml/min/kg 82 20 RAP during RCP, mmhg 40 10 CoA coarctation of aorta; CPB cardiopulmonary bypass; IAA interruption of aorta; RAP radial arterial pressure; RCP regional cerebral perfusion; SD standard deviation; SV single ventricle. CPB Systems and Techniques Our miniaturized CPB system was reported previously [9, 10]. To achieve a CPB system with a low priming volume, a low prime oxygenator and reservoir with a priming volume of 40 ml (Baby RX, Terumo Inc, Tokyo, Japan), 15-mL arterial filters (Filtia, JMS Inc, Hiroshima, Japan), and a smaller and shortened extracorporeal circuit was needed. To shorten the circuit, the CPB roller pump was placed close enough to the operative field to minimize the tubing length. Our CPB system consisted of a distant roller-pump head, a remote-controlled unit, and a sterilized sheet. The distant roller-pump and remotecontrolled unit (Tonokura Compo III, Tonokura Medical Inc, Tokyo, Japan) allows maximal proximity to the operative field. The sterilized sheet (50 100 mm, Steri- Sheet, Tonokura Medical Inc, Tokyo, Japan), made of polyvinyl chloride, acts as a protective barrier between the first assistant and the CPB unit [9]. The minimum priming volume of this system is currently 140 ml, with 15 ml in the reservoir level. The biocompatible surface coating can reduce the inflammatory response and improve the outcome of the operation [9]. Poly-2- methoxyethyl acrylate is one of the potential coating materials, and poly-2-methoxyethyl acrylate-coated circuits have already been reported to suppress the inflammatory response in clinical settings [11]. High-flow (200 ml/kg/min) moderate hypothermic (25 C) CPB was used. Blood-gas management was performed using the ph-stat strategy. The arterial oxygen pressure was kept at more than 300 mm Hg with oxygen saturation of 100% in all patients. Before HFRCP, mannitol (5 mg/kg) was given to all patients as deoxygenation to prevent the brain injury due to hyperoxygenation. The hematocrit level was kept at more than 25% during CPB using transfusion of red blood cells. DHCA was not used. Crystalloid cardioplegic solution (10 ml/kg) was given every 20 minutes. After termination of CPB, modified ultrafiltration was performed with a polymethylmethacrylate hemofilter for all patients. Modified ultrafiltration was started with an ultrafiltration rate of 20 ml/kg/min for 10 minutes. Heparinization was neutralized by protamine sulfate until the activated coagulation time had normalized. Solu-Medrol (30 mg/kg; Pfizer, New York, NY) was routinely given to all the patients before CPB. Aprotinin was not used in this study. Monitoring and Data Acquisition Blood pressure was invasively monitored at the right radial artery and the femoral artery. The central venous pressure was monitored with an inferior venous catheter from the right femoral vein. Arterial oxygen saturation was monitored continuously in the upper and lower extremities (Nellcor N200; Pleasanton, CA). Systemic venous oxygenation was monitored continuously during CPB from the venous drainage (Terumo CDI-500; Tokyo, Japan). Near-infrared spectroscopy probes were placed on the patient s midline forehead (cerebral) and midline back

Ann Thorac Surg MIYAJI ET AL 2010;90:593 9 REGIONAL HIGH-FLOW CEREBRAL PERFUSION 595 on the T10-L2 posterior flank (somatic) after entry into the operating room. The probes were monitored by a dual-detector device (Somanetics INVOS 5100, Troy, MI) and trended at 1-minute intervals. Cardiopulmonary bypass flows, temperatures, and pressures were recorded at 5-minute intervals. Arterial and venous blood gases were obtained at clinically appropriate intervals, with tensions reported at 37 C (Radiometer ABL, Copenhagen, Denmark). The mean radial arterial pressure was kept less than 50 mm Hg during RCP, and chlorpromazine was given to all the patients before and during HFRCP to increase RCP flow. The plasma lactate level (mmol/l) was also monitored during CPB. Prospective, manually recorded data were inserted into a common statistical database. Each patient s clinical data were divided into five major time periods consisting of: 1. surgical incision and preparation of CPB, 2. initiation of CPB and cooling, 3. initiation of myocardial ischemia and RCP, 4. declamping of the descending aorta and warming, and 5. separation from CPB and surgical closure. Within each period for each patient, data were collapsed into 5 epochs of equal duration to create 25 distinct epochs for subsequent statistical analysis. Data were summarized within intervals and periods and are presented as mean SD. Paired t tests were selected in comparison between time periods with significance cutoff at p 0.05. Study Protocol Changes of cerebral and somatic regional oxygen saturation (rso 2 ) were collected from all 18 patients. The plasma lactate level was measured at the beginning of the operation, at the initiation of CPB, the initiation of RCP, the end of RCP, the separation from CPB, and at the end of the operation. Changes in the lactate level were also examined for all the patients. In HFRCP, somatic blood flow, such as renal blood flow, increased in comparison with LFRCP due to collateral vessels. Correlations between RCP flow and urine output, between RCP flow and postoperative creatinine levels (12 hours postoperatively), and between RCP flow and blood urea nitrogen level (12 hours postoperatively) were estimated. To clarify the effectiveness of HFRCP for cerebral and somatic tissue oxygenation, correlations between RCP flow and somatic rso 2, between RCP flow and lactate increase during RCP, and between somatic rso 2 and lactate increase were examined. Finally, a brain computed tomography scan was performed 2 months after the operation. Results Demographic data in the 18 patients are summarized in Table 1. There was one early death. A patient with hypoplastic left heart syndrome and pulmonary artery sling who underwent the Norwood procedure and left pulmonary artery plasty died of respiratory failure due to tracheobronchial stenosis. Hypothermic HFRCP was accomplished in all patients at a flow rate of 82 20 ml/kg/min. Regional saturation data differed markedly during the operation. Regional saturations at major time periods are summarized in Table 2. The pre-cpb baseline cerebral rso 2 was 57.9% 10.5%, and baseline somatic rso 2 was 61.7% 13.3%. During cooling on CPB, cerebral rso 2 increased to 68.6% 9.9%, and somatic rso 2 increased to 76.9% 9.8%. During HFRCP, cerebral rso 2 increased to 78.8% 9.5%, and somatic rso 2 decreased but was maintained at 65.4% 12.1% (greater than 50%). During rewarming on CPB, cerebral rso 2 decreased to 66.8% 8.8%, whereas somatic rso 2 increased to 86.7% 6.6%. After separation from CPB, regional saturation in both measured beds decreased; both cerebral rso 2 and somatic rso 2 decreased, to 54.7% 10.0% and 77.5% 9.1%, respectively, but remained above or almost the same at baseline. Cerebral and somatic rso 2 were different from each other and from baseline values at all time points (p 0.01). The changes of both cerebral and somatic regional saturations are shown in Figures 1 and 2. The change of plasma lactate level during procedure is shown in Figure 3. During HFRCP, the lactate level significantly increased from 3.8 2.2 to 5.5 2.1 mmol/l (p 0.0001). There were no correlations between RCP flow and urine output, between RCP flow and postoperative creatinine level, and between RCP flow and blood urea nitrogen level during HFRCP. There was significant correlation between RCP flow and somatic rso 2 (r 0.547, p 0.0001) in all 18 patients (Fig 4). Significant inverse correlations were noted between RCP flow and the increase of lactate concentration (r 0.648, p 0.01, Fig 5) and between somatic rso 2 and the increase of lactate concentration (r 0.795, p 0.0001, Fig 6). These results indicated that the reduction of vascular resistance of collateral vessels increased both cerebral and somatic blood flow, resulting in improved tissue oxygen delivery. Finally, a brain computed tomography scan was performed in 12 patients (66.7%) at 2 months after the operation. No patients had any cerebral bleeding or infarction. The mean follow-up was 28 months. No major Table 2. Regional Saturation Data Summarized Within Major Operative Time Periods a Time Period Cerebral rso 2,% Mean SD Somatic rso 2,% Mean SD Pre-CPB 57.9 10.5 61.7 13.3 CPB cooling 68.6 9.9 b 76.9 9.8 bc RCP 78.8 9.5 b 65.4 12.1 c CPB warming 66.8 8.8 b 86.7 6.6 bc Post-CPB 54.7 10.0 77.5 9.1 bc a All differences are significant at p 0.01. with pre-cpb. c Differences between sites. RCP regional cerebral perfu- CPB cardiopulmonary bypass; sion; SD standard deviation. b Differences compared

596 MIYAJI ET AL Ann Thorac Surg REGIONAL HIGH-FLOW CEREBRAL PERFUSION 2010;90:593 9 Fig 1. Cerebral somatic regional oxygen saturation (rso 2 ) during the operation. The surgical procedure was divided into five major periods. Data are shown as mean and standard deviation for each period. (CPB cardiopulmonary bypass; HFRCP high-flow regional cerebral perfusion.) neurologic complications were documented in any patient. Comment Various techniques have been reported to avoid neurologic complications after an aortic arch repair. The following techniques have been used to successfully accomplish an aortic arch repair, including Norwood stage I palliation, in many institutions. Deep Hypothermic Circulatory Arrest Deep hypothermic circulatory arrest has been widely used as an easy and effective technique to perform aortic arch repair in pediatric patients. This technique provides a bloodless field and adequate brain protection. As survival rates after complex operations such as Norwood stage I palliation have increased, significant neurodevelopmental problems are now observed at long-term follow-up [12 14]. Because surgical results have improved, the avoidance of neurologic complications should be addressed. Recently, Goldberg and associates [15] reported the result of a randomized clinical trial of RCP vs DHCA investigating neurodevelopment for patients undergoing Fig 3. The change of plasma lactate level during procedure. During high-flow regional cerebral perfusion (HFRCP), lactate level significantly increased. Data are shown as mean and standard deviation. the Norwood procedure [15]. Infant development is delayed after the Norwood procedure, and their pilot data do not suggest that LFRCP improves infant development. However, the RCP flow rate in their study, which was considerably lower than the rate reported here, was initiated at 5 ml/kg/min, gradually increasing to 20 ml/kg/min. The cerebral rso 2 was measured during the procedure, but no precise data of cerebral rso 2 were reported. Because neurologic injury in children with congenital heart disease is now recognized as having a multifactorial origin, there has been no evidence that DHCA is superior to RCP. Low-Flow RCP The effectiveness of LFRCP has been reported by a number of centers during the past decade [1 5]. Although a number of studies recommend targeting RCP flow to either cerebral saturations or perfusion pressure, it remains unclear what actually constitutes optimal or even adequate cerebral oxygenation during hypothermia. Hoffman and colleagues [6] reported changes in cerebral and somatic oxygenation during Norwood stage I palliation using LFRCP [6]. They presented under continuous RCP with a flow rate from 30 to 70 ml/kg/ min, while cerebral oxygenation was maintained during RCP at pre-cpb levels with DHCA. However, after re- Fig 2. Intraoperative somatic regional oxygen saturation (rso 2 ). The surgical procedure was divided into five major periods. Data are shown as mean and standard deviation for each period. (CPB cardiopulmonary bypass; HFRCP high-flow regional cerebral perfusion.) Fig 4. Correlation between regional cerebral perfusion (RCP) flow and somatic regional oxygen saturation (rso 2 ) in 18 patients. There was significant correlation between RCP flow and somatic rso 2 (y 0.39x 34.0).

Ann Thorac Surg MIYAJI ET AL 2010;90:593 9 REGIONAL HIGH-FLOW CEREBRAL PERFUSION 597 Fig 5. Correlation between increase of lactate and regional cerebral perfusion (RCP) flow in 18 patients. There was significant inverse correlation between RCP flow and the increase of lactate concentration (y.051x 5.7). warming and separation from CPB, cerebral oxygenation was lower compared with pre-cpb or somatic values. These results indicate that cerebrovascular resistance is increased after DHCA, even with continuous perfusion techniques, placing patients on cerebral circulation at risk postoperatively. Somatic saturation also decreased less than 50% during RCP in 75% of the patients. In the current study, we used a vasodilator to reduce the cerebral and somatic vascular resistance before and during RCP, resulting in increased cerebral and somatic oxygen saturation during and after CPB. Descending Aorta Cannulation Combined With RCP In 1999, Imoto and colleagues [16] reported a special technique of descending aorta cannulation combined with RCP for Norwood stage I palliation. They reported that the cannulation of the distal descending thoracic aorta through a median sternotomy seems a safe, easy, and effective technique for perfusion of the lower body in the Norwood procedure. The combination of this technique and cerebral perfusion through the innominate artery enables complete avoidance of DHCA throughout the operation. The CPB system is complicated, however, and CPB circuits and cannulas are crowded in small operative fields. Another issue is how much cerebral perfusion flow is provided. In this technique, extracorporeal circulation with a single pump and double arterial catheters is established. Therefore, the cerebral perfusion flow depends on both cerebral and somatic vascular resistance. The flow measurement of cerebral perfusion circuits should be performed, or both cerebral and somatic rso 2 monitoring would be needed. High-Flow RCP Regional cerebral perfusion has been shown to provide cerebral circulatory support during aortic arch reconstruction. In our institution, HFRCP from the right innominate artery has been induced to maintain sufficient cerebral and somatic oxygen delivery through collateral vessels. HFRCP is the same surgical technique as LFRCP. A vasodilator is needed, to increase RCP. To avoid brain edema or bleeding due to excessive cerebral blood flow, as much as possible, intraoperative pressure monitoring of the left superficial temporal artery or the right radial artery is recommended for the sake of safety. In the present study, cerebral flow was limited by the right radial arterial pressure, which was maintained at less than 50 mm Hg. The mean cerebral flow was 82 ml/kg/ min (range, 43 to 108 ml/kg/min). In our high-flow CPB (150 to 200 ml/kg/ min) strategy, the cerebral perfusion flow is not excessive if the adequate blood supply to the upper body is half that of the total body in neonates and small infants. We began this prospective study in 2006. Our protocol regulated the dose of chlorpromazine to 3 mg/kg given to the patients during CPB. As we showed, 9 patients required preoperative inotropic support because of unstable hemodynamics. The cerebral and collateral vessels were both constricted in some of these patients, and sufficient vasodilation could not be obtained. Therefore, chlorpromazine at 3 mg/kg in these patients achieved a maximum RCP flow of about 50 ml/kg with less than 50 mm Hg of radial arterial pressure. For these patients, a higher dosage of chlorpromazine should have been administrated to achieve better flow, up to 80 to 100 ml/kg. Chlorpromazine was given to all the patients as a vasodilator as well as a sedative in the present study. Phenoxybenzamine can be used during CPB as a vasodilator. There were no correlations between RCP flow and urine output, between RCP flow and postoperative creatinine levels, and between RCP flow and blood urea nitrogen level during HFRCP, although there were significant inverse correlations between RCP flow and the increase of lactate concentration and between somatic rso 2 and the increase of lactate concentration. Among 18 patients, 6 patients required preoperative mechanical ventilation, and 9 required inotropic support because of unstable hemodynamics. Six patients (33.3%) had no urine output preoperatively or during CPB; therefore, RCP flow did not correlate with urine output during CPB. Fig 6. Correlation between increase of lactate and somatic regional oxygen saturation (rso 2 ) in 18 patients. There was significant inverse correlation between somatic rso 2 and the increase of lactate concentration (y 0.083x 7.04).

598 MIYAJI ET AL Ann Thorac Surg REGIONAL HIGH-FLOW CEREBRAL PERFUSION 2010;90:593 9 Study Limitations The present study is a prospective study; however, the number of patients was small (N 18). The patient group included various diagnoses and procedures, such as the Norwood procedure and biventricular repair for the coarctation of aorta/interruption of aorta complex. The postoperative course depends on the patient s preoperative condition and the procedure performed. Nevertheless, there was a significant correlation between RCP flow and the increase of lactate concentration, and between somatic rso 2 and the increase of lactate concentration. Conclusions The present study revealed that HFRCP preserved sufficient cerebral and somatic tissue oxygenation during aortic arch repair. The reduction of vascular resistance of collateral vessels increased both cerebral and somatic blood flow, resulting in improved tissue oxygen delivery. HFRCP is a safe and easy technique for aortic arch repair in neonates and small infants. References 1. Pigula FA, Siewers RD, Nemoto EM. Regional perfusion of the brain during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 1999;117:1023 4. 2. Pigula FA, Nemoto EM, Griffith BP, Siewers RD. Regional low-flow perfusion provides cerebral circulatory support during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 2000;119:331 9. 3. Hannan RL, Ybarra MA, Ojito JW, Alonso FA, Rossi AF, Burke RP. Complex neonatal single ventricle palliation using antegrade cerebral perfusion. Ann Thorac Surg 2006;82: 1278 85. 4. Andropoulos DB, Stayer SA, McKenzie ED, Fraser CD Jr. Novel cerebral physiologic monitoring to guide lowflow cerebral perfusion during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 2003;125:491 9. 5. Pigula FA, Gandhi SK, Siewers RD, Davis PJ, Webber SA, Nemoto EM. Regional low-flow perfusion provides somatic circulatory support during neonatal aortic arch surgery. Ann Thorac Surg 2001;72:401 6. 6. Hoffman GM, Stuth EA, Jaquiss RD, et al. Changes in cerebral and somatic oxygenation during stage 1 palliation of hypoplastic left heart syndrome using continuous regional cerebral perfusion. J Thorac Cardiovasc Surg 2004; 127:223 33. 7. Pigula FA. Arch reconstruction without circulatory arrest: scientific basis for continued use and application to patients with arch anomalies. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2002;5:104 15. 8. Tomoyasu T, Miyaji K, Miyamoto T, Inoue N. The bilateral pulmonary artery banding for hypoplastic left heart syndrome with a diminutive ascending aorta. Interact Cardiovasc Thorac Surg 2009;8:479 81. 9. Miyaji K, Kohira S, Miyamoto T, et al. Pediatric cardiac surgery without homologous blood transfusion, using a miniaturized bypass system in infants with lower body weight. J Thorac Cardiovasc Surg 2007;134:284 9. 10. Miyaji K, Miyamoto T, Kohira S, et al. Miniaturized cardiopulmonary bypass system in neonates and small infants. Interact Cardiovasc Thorac Surg 2008;7:75 8. 11. Ninomiya M, Miyaji K, Takamoto S. Influence of PMEAcoated bypass circuits on perioperative inflammatory response. Ann Thorac Surg 2003;75:913 8. 12. Bellinger DC, Jonas RA, Rappaport LA, et al. Developmental and neurologic status of children after heart surgery with hypothermic circulatory arrest or low-flow cardiopulmonary bypass. N Engl J Med 1995;332:549 55. 13. Mahle WT, Clancy RR, Moss EM, Gerdes M, Jobes DR, Wernovsky G. Neurodevelopmental outcome and lifestyle assessment in school-aged and adolescent children with hypoplastic left heart syndrome. Pediatrics 2000;105:1082 9. 14. Wernovsky G. Current insights regarding neurological and developmental abnormalities in children and young adults with complex congenital cardiac disease. Cardiol Young 2006;16(suppl 1):92 104. 15. Goldberg CS, Bove EL, Devaney EJ, et al. A randomized clinical trial of regional cerebral perfusion versus deep hypothermic circulatory arrest: outcomes for infants with functional single ventricle. J Thorac Cardiovasc Surg 2007; 133:880 7. 16. Imoto Y, Kado H, Shiokawa Y, Fukae K, Yasui H. Norwood procedure without circulatory arrest. Ann Thorac Surg 1999; 68:559 61. DISCUSSION DR JAMES GANGEMI (Charlottesville, VA): I noticed that 2 months following the surgery you performed a brain computed tomography (CT) scan. Was there any thought to doing an magnetic resonance imaging (MRI) instead, thinking that an MRI might be a little more sensitive to small changes in the brain? DR MIYAJI: We didn t have any MRI data in this study. And just after surgery, we routinely use the echography through the anterior fontanel for neonates, and then there were no data of brain edema and bleeding. DR ADITYA KAZA (Salt Lake City, UT): I just want to ask the audience about their flow patterns during regional perfusion. Our perfusionist is very particular about any flows over about 50 cc/kg when we are perfusing the innominate artery. I just want to get an idea from the audience about what they do when they do regional perfusion, or from the panel members. DR MIYAJI: You mean perfusion flowing at 50 ml? DR KAZA: Yes. You said you flow at about 80 cc, correct? DR MIYAJI: Yes. DR KAZA: Do you have any idea about how their fontanels are doing in the intensive care unit in these babies? DR MIYAJI: No, we don t have any of that. DR KAZA: I know this neurologic imaging was remote from the time of perfusion, but I am concerned that it is a lot of flow for the brain and if that is going to result in a lot of neurologic edema. DR MIYAJI: Yes, that is what I am saying about using an echography to check the brain edema. There are no data about the brain edema just after surgery. DR KAZA: Dr Plunkett, can you ask the audience about the flow rates that they use?

Ann Thorac Surg MIYAJI ET AL 2010;90:593 9 REGIONAL HIGH-FLOW CEREBRAL PERFUSION 599 DR MARK PLUNKETT (Lexington, KY): Yes. Can we see a show of hands of those that flow 50 or those that are flowing higher. Is that the general rule? (Show of hands.) DR MIYAJI: Thank you. DR FRANK PIGULA (Boston, MA): That was a very nice presentation. I think it points out that there is a great variability in how this technique is practiced. Do you have any concerns about some of the experimental data that suggests that at higher flow rates you may run the risk of cerebral edema? I will limit flow 30 to 40 ml/kg/min, and I have little experience with anything above deep hypothermic temperatures with this technique because of concerns. One of the things we have learned the last couple of years is that surgeons use this technique with great variability in practice. So I am interested in finding out that you can flow these high rates without having any overt evidence of injury. So did you have concerns about those high flow rates? DR MIYAJI: Well, before this study, we usually use low flow rate, maybe around 30 or 40 ml/kg/min. In these patients, we checked in all the patients lactate concentrations during bypass. In those patients, the plasma lactate level significantly increased during regional cerebral perfusion (RCP). Then I discussed with our perfusionists, how we will increase flow? And then we began to use vasodilator, such as chlorpromazine or isosorbide dinitrate (ISDN). After that we can achieve high flow RCP, around 80 ml/kg/min with less than 50 mmhg radial artery pressure. Actually, I was very anxious about brain edema after surgery. Therefore, we checked all the patients echography through the fontanel just after surgery. Because we couldn t do the CT scan in some of the patients with their chests open. Now we re doing the high-flow RCP with flow rate of around 100 ml/kg/min without any problems. DR PIGULA: What temperature do you use? DR MIYAJI: Under 28 or around 30 centigrade. DR PIGULA: Thank you. DR WILLIAM DECAMPLI (Orlando, FL): Again, your abstract gives quite a range of flow rates from 42 to 124 ml/kg/min. And you state that you try to maintain the perfusion pressure below 50 mm Hg. But given that wide range in flow rates, exactly what parameter were you measuring in your protocol to determine what flow rate you actually used? Also, did your dose of vasodilators vary? DR MIYAJI: It was just regulated by radial artery pressure less than 50. DR DECAMPLI: So less than 50 and greater than what? I mean 50 is a maximum, what would be DR MIYAJI: Fifty is the maximum. Maximum pressure of the radial artery pressure is 50. DR DECAMPLI: And how low would you allow it to be? DR MIYAJI: We don t make a limit of the low limit of the radial artery pressure. When the patient s radial arterial pressure was around 20, we increased the RCP flow until the radial arterial pressure reached around 40, less than 50. We began this prospective study in 2006, and our protocol regulated dose of the vasodilators, such as chlorpromazine and ISDN. In some patients, preoperative hemodynamics were very unstable, those patients had inotropic supports before surgery. In some of these patients, both cerebral and collateral vessels were constricted, and sufficient vasodilation could not be obtained. Therefore, using 3 mg/kg of chlorpromazine, the maximum RCP flow was around 50 ml/kg with less than 50 mm Hg of radial arterial pressure. That is the reason why there was the wide range of the RCP flow in this study.