Aortic arch hypoplasia is a common constituent of

Transcription

1 Regional Low-Flow Perfusion Provides Somatic Circulatory Support During Neonatal Aortic Arch Surgery Frank A. Pigula, MD, Sanjiv K. Gandhi, MD, Ralph D. Siewers, MD, Peter J. Davis, MD, Steven A. Webber, MD, and Edwin M. Nemoto, PhD Division of Pediatric Cardiothoracic Surgery, Children s Hospital of Pittsburgh, University of Pittsburgh, Pittsburgh, Pennsylvania Background. Regional low-flow perfusion has been shown to provide cerebral circulatory support during neonatal aortic arch operations. However, its ability to provide somatic circulatory support remains unknown. Methods. Fifteen neonates undergoing arch reconstruction with regional perfusion were studied. Three techniques were used to assess somatic perfusion: abdominal aortic blood pressure, quadriceps blood flow (nearinfrared spectroscopy), and gastric tonometry. Results. Twelve patients required operation for hypoplastic left heart syndrome, and 3 required arch reconstruction with a biventricular repair. There was one death (7%). Abdominal aortic blood pressure was higher (12 3 mm Hg versus 0 0 mm Hg), and quadriceps blood volumes (5 24 versus 17 26) and oxygen saturations (57 25 versus 33 12) were greater during regional perfusion than during deep hypothermic circulatory arrest (p < 0.05). During rewarming, the arterial gastric mucosal carbon dioxide tension difference was lower after circulatory arrest than after regional perfusion ( mm Hg versus mm Hg, p < 0.05). Conclusions. Regional low-flow perfusion provides somatic circulatory support during neonatal arch surgical procedures. Support of the subdiaphragmatic viscera should improve the ability of neonates to survive the postoperative period. (Ann Thorac Surg 2001;72:401 7) 2001 by The Society of Thoracic Surgeons Aortic arch hypoplasia is a common constituent of congenital heart disease, repair of which is usually performed during a period of deep hypothermic circulatory arrest (DHCA). Because of concerns about the impact of DHCA on outcome in terms of neurologic morbidity and mortality, alternatives have been sought [1, 2]. We have previously described a technique of regional low-flow perfusion (RLFP) that provides cerebral circulatory support during surgical reconstruction of the aortic arch in neonates. Because of our observation that major backbleeding occurs from the descending thoracic aorta during RLFP, we have hypothesized that somatic circulatory support is provided as well. This study extends our clinical investigations by exploring the ability of RLFP to provide subdiaphragmatic somatic circulatory support in the neonate undergoing aortic arch operation. Material and Methods Fifteen consecutive neonates with hypoplasia of the ascending aorta, aortic arch, or both underwent repair (Table 1). Twelve were seen with classic hypoplastic left heart syndrome and 3, with biventricular anatomy with associated aortic pathologic conditions. Presented at the Thirty-seventh Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 29 31, Address reprint requests to Dr Pigula, Pediatric Cardiothoracic Surgery, Room 2820, 2 Main, Children s Hospital of Pittsburgh, Pittsburgh, PA 15213; pigulaf@heart.chp.edu. The operative technique has previously been described in detail [1]. Briefly, in the case of children with hypoplastic left heart syndrome, standard cannulation (through the ductus by way of pulmonary artery) is performed and the preparatory dissection completed. The proximal anastomosis of the anticipated Blalock- Taussig shunt is accomplished. The graft is left long, out of the operative field. In patients undergoing a Norwood operation, a brief period of DHCA is used to perform the atrial septectomy. The arterial cannula is then inserted into the graft, and after careful deairing, RLFP is initiated. With control of the brachiocephalic vessels and the descending thoracic aorta, exposure comparable to that obtained during DHCA is obtained. In the case of biventricular repairs, a graft (3.5-mm Gore-Tex; W. L. Gore and Associates, Flagstaff, AZ) is anastomosed to the innominate artery in a manner identical to that for placement of a Blalock-Taussig shunt. However, the shunt serves as the primary cannulation site and is unchanged for the duration of the operation. After separation of the patient from cardiopulmonary bypass, the graft is clipped flush with the innominate artery and the stump, oversewn. All operations were performed using deep hypothermia (18 C) and alpha-stat management. The RLFP rate was guided by near-infrared spectroscopy (NIRS) with the goal of maintaining cerebral blood volumes at baseline levels (obtained on full-flow hypothermic cardiopulmonary bypass). Assessment of somatic perfusion was 2001 by The Society of Thoracic Surgeons /01/$20.00 Published by Elsevier Science Inc PII S (01)

2 402 PIGULA ET AL Ann Thorac Surg SOMATIC CIRCULATORY SUPPORT IN NEONATES 2001;72:401 7 Table 1. Summary of Clinical Characteristics Patient No. Diagnosis Age (d) DHCA (min) RLFP (min) Bypass Time Cardiopulmonary (min) 1 HLHS HLHS HLHS Biventricular anatomy a HLHS HLHS HLHS HLHS HLHS HLHS Biventricular anatomy a HLHS Biventricular anatomy a HLHS HLHS Mean SD Range Median a Biventricular repairs included Yasui operation for aortic atresia and ventricular septal defect (1 patient), Taussig-Bing heart with arch hypoplasia (1 patient), and interrupted aortic arch (type B) with ventricular septal defect (1 patient). DHCA deep hypothermic circulatory asset; HLHS hypoplastic left heart syndrome; RLFP regional low-flow perfusion; SD standard deviation. performed in three ways: abdominal aortic blood pressure was measured with an umbilical artery catheter; quadriceps muscle blood volumes and saturations were measured with NIRS; and visceral perfusion was assessed by measuring the arterial-mucosal carbon dioxide tension (Pco 2 ) gradient (Pco 2 gap) in the stomach using gastric tonometry. Monitoring and Data Acquisition Standard hemodynamic monitoring was applied to all children. Arterial blood pressure was measured simultaneously by a left radial artery catheter and an umbilical artery catheter (placed preoperatively for resuscitation and monitoring). After informed consent was obtained from the parents, all children underwent intraoperative monitoring of relative cerebral blood volumes and oxygen saturations using NIRS (INVOS 5100; Somanetics Corp, Troy, MI) [2]. These data have been reported previously and were used to guide the RLFP rate [2]. A NIRS sensor was placed on the quadriceps muscle to measure relative muscle oxygenation and relative quadriceps muscle blood volume index. Data were downloaded and stored for later analysis. After the induction of anesthesia, a gastrointestinal tonometer catheter (Instrumentarium Corp, Helsinki, Finland) was placed in the stomach and attached to the tonometer (Tonometrics Corp, Helsinki, Finland). Gastric mucosal Pco 2 was automatically measured at 10-minute intervals and recorded. Simultaneous arterial blood gases were obtained from the radial artery, and the arterial gastric mucosal Pco 2 gap was calculated. Statistical Analysis Data are presented as the mean the standard deviation unless otherwise stated. Evaluation of continuous variables was performed using analysis of variance for repeated measures, and significance was assumed when the value of P was less than Results Three patients had biventricular repair, and 12 underwent the Norwood operation for hypoplastic left heart syndrome. The mean DHCA time was 12 9 minutes, with a mean duration of RLFP of minutes. There were no observed neurologic sequelae, and there were no instances of renal or hepatic dysfunction. A child with classic hypoplastic left heart syndrome died. During RLFP, the mean left radial artery blood pressure was 29 5 mm Hg, and the mean abdominal aortic blood pressure was 12 3 mm Hg. Paired arterial blood gas data (obtained simultaneously from the radial and umbilical artery catheters) were comparable (radial: ph 7.4; Pco 2, 29 mm Hg; oxygen tension, 194 mm Hg; the bicarbonate radical, 19.2; and lactate, 5.1 mmol/l; and umbilical: ph 7.39; Pco 2, 29 mm Hg; oxygen tension, 133 mm Hg; the bicarbonate radical, 19.1; and lactate, 5.4 mmol/l). Quadriceps data were obtained by NIRS for all 15 patients. Quadriceps muscle blood volumes and quadriceps saturations were measured on full-flow hypothermic cardiopulmonary bypass, during the brief period of

3 Ann Thorac Surg PIGULA ET AL 2001;72:401 7 SOMATIC CIRCULATORY SUPPORT IN NEONATES 403 Fig 1. Quadriceps muscle near-infrared spectroscopy in neonate undergoing Norwood operation for hypoplastic left heart syndrome. With the initiation of cardiopulmonary bypass (CPB), there is an increase in quadriceps muscle oxygen saturation (RQrSO2) with a stable quadriceps muscle blood volume. During a brief period (6 minutes) of circulatory arrest (CIRC ARREST), there is a sharp decline in muscle saturations and blood volumes. Immediately after the initiation of regional low-flow perfusion (RLFP) (20 ml/min), there is an increase in muscle blood volumes that continues as RLFP rate increases to 30 and then 40 ml/min. After approximately 5 minutes, there is a corresponding increase in muscle saturations. With completion of the neo-aorta, RLFP is stopped momentarily to allow for central recannulation, and standard CPB is resumed. DHCA required for atrial septectomy and shunt cannulation (for those children undergoing a Norwood operation), and during RLFP (Fig 1). The data showed a significant increase in both relative blood volume index and oxygenation in the quadriceps muscle during RLFP compared with values obtained during the short period of DHCA (mean duration, 9 4 minutes). In fact they approximated the values obtained during full-flow hypothermic cardiopulmonary bypass (Table 2). To evaluate the degree of splanchnic flow provided by RLFP, gastric tonometry was performed in 9 neonates during RLFP and compared with the data from 3 neonates undergoing cardiac repair during DHCA (mean duration, 34 minutes). The results of gastric tonometry showed no significant differences in the arterial gastric mucosal Pco 2 gradient before or during RLFP or DHCA, but there was a significant gradient observed during rewarming (Fig 2). The children undergoing repair during DHCA demonstrated an arterial gastric mucosal Pco 2 gradient (Pco 2 gap) of mm Hg compared with mm Hg in the neonates undergoing repair during RLFP (p 0.03). Because arterial blood gases were unobtainable during DHCA, the Pco 2 gap could not be calculated; thus, comparisons during RLFP and DHCA were impossible. Table 2. Abdominal Aortic Blood Pressure and Quadriceps Near-Infrared Spectroscopic Data a Variable CPB DHCA RLFP p Value b Abdominal aortic blood pressure (mm Hg) QrBVI QrSO a Data are shown as the mean the standard deviation (analysis of variance). b The p value shows the difference between DHCA and RLFP data. CPB cardiopulmonary bypass; DHCA deep hypothermic circulatory arrest; QrBVI quadriceps muscle relative blood volume index; QrSO 2 quadriceps muscle relative blood saturation; RLFP regional low-flow perfusion.

4 404 PIGULA ET AL Ann Thorac Surg SOMATIC CIRCULATORY SUPPORT IN NEONATES 2001;72:401 7 Fig 2. Gastric tonometry for 11 patients undergoing arch repair with regional low-flow perfusion (RLFP) versus 3 patients undergoing cardiac repair during deep hypothermic circulatory arrest (CA). Data are presented as the difference between the arterial and gastric mucosal carbon dioxide tensions (pco 2 ), the Pco 2 gap (arterial Pco 2 gastric Pco 2 ). There were no significant differences before cardiopulmonary bypass (Pre-CPB), during cooling on bypass (Cooling), or after separation from bypass (Post-CPB). During rewarming, gastric mucosal Pco 2 increased relative to arterial Pco 2, thus creating a negative Pco 2 gap, suggesting ischemia. Differences in the Pco 2 gap between the two groups were significant only during rewarming (p 0.03 by analysis of variance [ANOVA]). Finally, blood urea nitrogen and creatinine values were measured preoperatively and for 3 days postoperatively (Fig 3). There were no instances of oliguria and no requirement of renal replacement therapy in any patient. arterial cannulation. Second, whereas conventional low-flow perfusion maintains a cardiac index of 0.75 L min 1 m 2, the flow rates required to support the neonate using RLFP are substantially lower (cardiac index of 0.45 L min 1 m 2 ). This assertion is based on our earlier studies [2] showing baseline cerebral blood saturations and blood volumes are obtained at this flow rate. Because NIRS allows real-time assessment of the cerebral circulation, flows can be optimized to maintain those obtained at baseline (defined as full-flow cardiopulmonary bypass). Thus, the incomplete ischemia postulated to occur during conventional low-flow perfusion may be avoided [8]. During RLFP (30 to 40 ml kg 1 min 1 for a cardiac index of 0.45 L min 1 m 2 ), the mean left radial artery blood pressure was 29 mm Hg with a mean abdominal aortic blood pressure of 12 mm Hg. These findings support the speculation that in the neonate, there is an extensive network of vascular collaterals traversing the relatively short spatial distance between the supradiaphragmatic and subdiaphragmatic vasculature. These collaterals may include the internal mammary arteries and the intercostal arteries, as well as unnamed muscular and cutaneous connections. This degree of collateralization is clinically apparent by the extremely low incidence of paraplegia among neonates undergoing repair of aortic coarctation. Comment Although DHCA has been an indispensable technique in congenital heart surgery, the safe duration of DHCA remains uncertain. Most clinical and experimental studies suggest that 30 to 45 minutes is probably acceptable, but beyond that time, the risk of neurologic morbidity increases [3, 4]. The use of DHCA can also have an impact on survival. Clancy and coauthors [5] have recently shown that the duration of DHCA is an independent risk factor for mortality in neonatal heart surgical procedures. Because of these issues, alternatives to DHCA have been sought. Low-flow perfusion is one such alternative. Its advantages, identified in the Boston Circulatory Arrest Study [6], include better neurologic outcomes in neonates compared with the results in patients undergoing repair during DHCA. Low-flow perfusion, a well-described technique, has been applied during operations for other forms of congenital heart disease, such as tetralogy of Fallot, and has been important in the efforts to perform earlier anatomic cardiac repairs [7]. Regional low-flow perfusion differs from conventional low-flow perfusion, as described in the Boston study [6], in two important ways. First, cannulation is regional rather than central, and is accomplished through a perfusion conduit (shunt) anastomosed to the innominate artery. This is required in children with aortic arch hypoplasia, as the diminutive aorta precludes standard Fig 3. Mean blood urea nitrogen (BUN) (A) and creatinine (B) levels of 15 neonates undergoing cardiac repair during regional lowflow perfusion preoperatively (preop) and on postoperative days (pod) 1,2,and3.

5 Ann Thorac Surg PIGULA ET AL 2001;72:401 7 SOMATIC CIRCULATORY SUPPORT IN NEONATES 405 Near-infrared spectroscopy, a recent noninvasive technology capable of measuring tissue chromophores, exploits the differences in absorption peaks between oxygenated and deoxygenated hemoglobin and provides information on the changes in these compounds over time [9]. Thus, relative changes in oxygen saturations and blood volumes are obtainable [10]. In this study, we used NIRS to provide information on the relative blood volumes and saturations in the quadriceps muscle. Validation of this technique in skeletal muscle has been pursued clinically and experimentally. Edwards and colleagues [11] reported a good correlation between NIRS and venous occlusion plethysmography in the human forearm. Experimentally, Tran and associates [12] performed a comparative analysis of nuclear magnetic resonance and NIRS measurements of intracellular oxygen tension in human skeletal muscle and reported that the NIRS signals closely match the desaturation kinetics of myoglobin. They concluded that skeletal muscle (gastrocnemius) NIRS largely reflects the change in oxymyoglobin and deoxymyoglobin rather than hemoglobin. This conclusion would do little to alter the interpretation of limb ischemia. Indeed, an assessment of tissue, rather than hemoglobin oxygenation, would provide a superior assessment of the adequacy of blood flow. With this experience, the use of NIRS in the assessment of limb blood flow is increasingly common in the clinical setting [13]. Kooijman and coworkers [14] found that NIRS is an effective noninvasive method for assessing the oxygen debt in the lower extremities of patients with peripheral vascular disease. Komiyama and associates [15] recently suggested that NIRS grades the severity of intermittent claudication in diabetics more accurately than the ankle brachial pressure index. In this study, quadriceps NIRS recorded increasing muscle blood volumes and saturations about 5 minutes after the initiation of RLFP (see Fig 1). This delay probably represents the time required for blood to traverse the collateral vascular beds between the point of delivery (supradiaphragmatic) to the subdiaphragmatic aorta. These NIRS data are consistent with our observations of aortic backbleeding and a mean blood pressure of 12 mm Hg in the abdominal aorta during RLFP. Finally, because the role of gastrointestinal ischemia in the development of sepsis and multiorgan dysfunction is well established, we thought it important to assess the physiological importance of splanchnic blood flow provided by RLFP. Gastric tonometry is a technique that measures the gastric mucosal Pco 2 by way of a balloontipped catheter inserted into the stomach. There is evidence to suggest that increases in gastric mucosal Pco 2 relative to arterial Pco 2 (the Pco 2 gap) are both sensitive and specific for the detection of gut ischemia [16, 17]. This phenomenon is thought to reflect two separate but related pathophysiological processes; carbon dioxide accumulation as a consequence of impaired blood flow or as a result of increased CO 2 production under anaerobic conditions [18]. Although studies defining the precise Pco 2 gap indicative of mucosal ischemia are lacking, Duke and associates [19] reported that among patients receiving pediatric extracorporeal membrane oxygenation, survivors had a significantly smaller Pco 2 gap ( 4.7) than did nonsurvivors ( 24) (p 0.003). Furthermore, animal studies [18] examining the linkage between blood flow reduction and Pco 2 have shown that there appears to be a critical lower limit of blood flow (approximately a 60% reduction) below which there is a sudden rise in gastric mucosal Pco 2. Thus, we employed gastric tonometry to detect differences in gastric mucosal Pco 2 after repair during DHCA or RLFP. Although the 3 patients having repair during DHCA presented with a slightly negative (gastric greater than arterial) Pco 2 gap, it was mild ( 0.5 mm Hg) and may reflect the fact that these tended to be older patients with medically refractory congestive heart failure. During cooling on bypass, the Pco 2 gap was similar between the two groups. During rewarming, patients exposed to DHCA experienced a negative Pco 2 gap (gastric mucosal Pco 2 higher than arterial Pco 2 ), whereas those having operation during RLFP did not (see Fig 2). Gastric mucosal hypercarbia identified in children undergoing repair during DHCA suggests an ischemic event not seen in patients supported with RLFP. Although the value of statistical conclusions drawn from small samples is uncertain, we prefer to perform neonatal repairs during RLFP whenever possible. In practical terms, DHCA is now generally confined to the repair of obstructed total anomalous pulmonary venous return. Despite the small sample size, our data are consistent with the blood pressure and NIRS data and support the contention that RLFP provides significant subdiaphragmatic circulatory support during arch reconstruction. Support of the subdiaphragmatic viscera would be noteworthy because substantial morbidity and mortality resulting from noncardiac-related organ failure or sepsis occur in the postoperative period. Poirier and coauthors [20] reported their experience with 59 neonates undergoing the Norwood operation for hypoplastic left heart syndrome or its variants. The median DHCA time was 37 minutes, and the early postoperative survival rate was 83%. Postoperative sepsis (not otherwise specified) occurred in 13 patients, necrotizing enterocolitis in 1 patient, and seizures in 1 patient. Clancy and colleagues [5] reported a survival rate of 84% among 318 neonates undergoing a variety of one- and two-ventricle repairs during DHCA. Among the 52 deaths, 24 (46%) occurred between 3 and 35 days postoperatively, when mortality is generally secondary to causes other than overt inadequacy of the hemodynamic repair. These studies would support the position that any postoperative organ dysfunction, be it renal, gut, or hepatic, can erode the already precarious clinical situation of a recovering neonate. Overall survival (30-day and hospital discharge) for patients undergoing repair with RLFP was 93% (14/15). The subgroup of 12 neonates undergoing a Norwood operation with RLFP had a 92% survival rate (11/12), and all 3 children undergoing biventricular repair survived. The results in this small cohort compare favorably with those in published series, but factors other than the use

6 406 PIGULA ET AL Ann Thorac Surg SOMATIC CIRCULATORY SUPPORT IN NEONATES 2001;72:401 7 of RLFP bear consideration [21, 22]. Postoperative management, directed at reducing fluctuations in pulmonary vascular resistance, may also be important. From these data, we conclude that RLFP provides significant somatic circulatory support during neonatal aortic arch surgical procedures. Circulatory support of the subdiaphragmatic viscera should improve the ability of neonates to survive the postoperative period. These results complement those in our previous report [2] documenting cerebral circulatory support in neonates and justify the conclusion that RLFP truly reduces DHCA time during neonatal aortic arch operations. Furthermore, this modification of low-flow perfusion is applicable to most, if not all, forms of congenital heart disease. References 1. Pigula FA, Siewers RD, Nemoto EM. Regional perfusion of the brain during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 1999;117: 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: Kirklin JW, Barratt-Boyes BG. Hypothermia, circulatory arrest, and cardiopulmonary bypass. In: Kirklin JW, Barratt- Boyes BG, eds. Cardiac surgery. 2nd ed. New York: Wiley, 1993: Wragg RE, Dimsdale SE, Moser KM, Daily PO, Dembitsky WP, Archibald C. Operative predictors of delirium after pulmonary thromboendarterectomy. J Thorac Cardiovas: Surg 1988;96: Clancy RR, McGaurn SA, Weinovsky G, et al. Preoperative risk-of-death prediction model in heart surgery with deep hypothermic circulatory arrest in the neonate. J Thorac Cardiovasc Surg 2000;119: 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: Pigula FA, Khalil PN, Mayer JE, Del Nido PJ, Jonas RA. Repair of tetralogy of Fallot in neonates and young infants. Circulation 1999;100(Suppl 2): Du Plessis AJ. Mechanisms of brain injury during infant cardiac surgery. Semin Pediatr Neurol 1999;6: Jobsis FF. Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science 1977;198: Wray S, Cope M, Delpy DT, Wyatt JS, Reynolds EO. Characterization of the near infrared absorption spectra of cytochrome aa3 and haemoglobin for the non-invasive monitoring of cerebral oxygenation. Biochim Biophys Acta 1988;933: Edwards AD, Richardson C, van der Zee P, et al. Measurement of hemoglobin flow and blood flow by near-infrared spectroscopy. J Appl Physiol 1993;75: Tran TK, Sailasuta N, Kreutzer U, et al. Comparative analysis of NMR and NIRS measurements of intracellular PO 2 in human skeletal muscle. Am J Physiol 1999;276 (6 Pt2):R Boushel R, Piantadosi CA. Near-infrared spectroscopy for monitoring muscle oxygenation. Acta Physiol Scand 2000: 168: Kooijman HM, Hopman MT, Colier WN, van der Vliet JA, Oeseburg B. Near infrared spectroscopy for noninvasive assessment of claudication. J Surg Res 1997;72: Komiyama T, Shigematsu H, Yasuhara H, Muto T. Nearinfrared spectroscopy grades the severity of intermittent claudication in diabetics more accurately than ankle pressure measurement. Br J Surg 2000;87: Montgomery A, Hartman M, Jonsson K, et al. Intramucosal ph measurement with tonometers for detecting gastrointestinal ischemia in porcine hemorrhagic shock. Circ Shock 1998;29: Tang W, Weil MH, Sun S, Nol M, Gazmuri RJ, Biseja J. Gastric intramural PCO2 as monitor of perfusion failure during hemorrhagic and anaphylactic shock. J Appl Physiol 1994;76: Knichwitz G, Van Aken H, Brussel T. Gastrointestinal monitoring using measurement of intramucosal PCO2. Anesth Analg 1998;87: Duke T, Butt W, South M, Shann F. The DCO2 measured by gastric tonometry predicts survival in children receiving extracorporeal life support. Comparison with other hemodynamic and biochemical information. Royal Children s Hospital ECMO Nursing Team. Chest 1997;111: Poirier NC, Drummond-Webb JJ, Hisamochi K, Imamura M, Harrison AM, Mee RB. Modified Norwood procedure with a high-flow cardiopulmonary bypass strategy results in low mortality without late arch obstruction. J Thorac Cardiovasc Surg 2000;120: Daebritz SH, Nollert GD, Zurakowski D, et al. Results of Norwood stage I operation: comparison of hypoplastic left heart syndrome with other malformations. J Thorac Cardiovasc Surg 2000;119: Bove EL. Current status of staged reconstruction for hypoplastic left heart syndrome. Pediatr Cardiol 1998;19: DISCUSSION DR CHRISTO I. TCHERVENKOV (Montreal, Quebec, Canada): This is an excellent study, and I congratulate you for adding objective data to the similar surgical techniques my colleagues and I are using in Montreal. Although we have not used that kind of sophistication, your findings are very consistent with ours. Tomorrow at the moderated poster session, we outline our experience with 18 patients, half of whom had the Norwood procedure using several techniques of selective low-flow cerebral perfusion during neonatal aortic arch repair. What is very interesting in support of your findings is that when you remove the clamp from the descending thoracic aorta, you literally get flooded with backbleeding from the lower-body circulation. My question is whether you have tried to quantify how much of the flow in the innominate artery reaches the cerebral circulation and how much ends up perfusing the lower body? DR PIGULA: No, we have not done that, and actually we were talking about it last night. I suppose it is possible to get some idea if you can quantify the return from the inferior vera cava versus, the superior vera cava. That is manageable; we just have not pursued it. DR TCHERVENKOV: The flow rates we have used clinically are somewhat higher than yours. They range from 0.23 to 1 L min 1 m 2, which works out to somewhere between 20 and 60 ml kg 1 min 1. I wonder if there is an upper limit of perfusion above which one might actually be doing damage to the brain.

7 Ann Thorac Surg PIGULA ET AL 2001;72:401 7 SOMATIC CIRCULATORY SUPPORT IN NEONATES 407 Do you have a sense of what the upper limit of regional perfusion flow is within the safe margin? DR PIGULA: I do not have a sense of what the upper limit is, but I admit we have liberalized our flows a little bit. With flow at 30 to 40 ml kg 1 min 1, we check the left radial artery blood pressure often. I did not show those data. With this technique, when there is a descending aortic of 12 mm Hg, there is a left radial artery pressure of 25 mm Hg. So I usually try to maintain a mean left radial artery pressure of 25 to 30 mm Hg. MR JAMES L. MONRO (Southampton, UK): Congratulations on a very nice study. It is important to know that the perfusion of the lower part of the body is good, but it is the brain about which we are particularly concerned. If I understood correctly, 3 of your patients did not have a hypoplastic heart, and they had rather longer circulatory arrest times. Were you perfusing the right innominate artery? A trick my associates and I have used is to slide the aortic cannula up and snug it. When doing a hypoplastic correction, you use the shunt you have already placed. I do not understand how you did it in the 3 patients without a hypoplastic left heart. DR PIGULA: With the biventricular repairs, I sewed a shunt to the innominate artery and cannulated that directly as the primary cannulation site. These 3 patients had longer circulatory arrest times because, at some points, it was more trouble technically doing the intracardiac portions of the repair and the ventricular septal defect closures in the operative field than I thought it was worth. So we did resort to some circulatory arrest time in those patients. MR MONRO: As I understood it, you cannulated the innominate artery directly? DR PIGULA: No, I sewed in a Gore-Tex graft, as in anticipation of a Blalock-Taussig shunt, and cannulate the Gore-Tex graft. MR MONRO: But if you are not doing a shunt, my point is that you can use the DLP cannulas. You put one in the ascending aorta, and then when you want to do the anastomosis, you just slide it up and snug it. DR ERLE H. AUSTIN (Louisville, KY): I enjoyed your presentation, and I thank you for helping introduce this technique of regional low-flow perfusion that allows us to do arch reconstruction without circulatory arrest. We have been using your technique for the past 2 years, and it has virtually eliminated the use of circulatory arrest in my practice. However, a note of caution, based on my own anecdotal experience with this technique, is warranted. This note refers to the point of your presentation that suggests that during regional flow to the cerebral circulation, adequate flow is occurring to the subdiaphragmatic organs. In our experience, by placing one end of the shunt on the innominate artery prior to initiating bypass and using the shunt for all the arterial inflow, we were able to perform the operation without any period of circulatory arrest. We were comfortable with the technique but wondered whether or not we needed to cool the infant to 18 C. Therefore, for several patients, we decided to cool to no lower than 25 C to shorten the time it takes to rewarm. Doing this resulted in a relatively expeditious operation with good cerebral protection. However, we also noted that these patients had impaired, postoperative renal function with several days of oligúria that was unresponsive to diuretics, a situation that can be a problem postoperatively in such children. After seeing this in 3 consecutive patients, we hypothesized that failing to cool the patient to less than 20 C may have contributed to the renal dysfunction. Subsequently, we have gone back to using deep hypothermic levels until reconstruction is completed and flow is restored to the descending aorta. Since then, we have noted no problems with postoperative renal function. I recognize this is not very scientific, but I thought our experience might indicate that temperature is also an important factor in preserving the integrity of the subdiaphragmatic organs such as the kidneys. DR PIGULA: I think that is a very important observation. Clearly, we are not providing physiologic blood flow necessarily, and we have to bear in mind that it is a supply demand, issue. The demands at lower temperatures are much less than they are at the higher temperatures. DR TOM R. KARL (Philadelphia, PA): I really enjoyed the presentation. This was a nicely thought-out strategy. As more of us adopt this technique, in your opinion, what sort of data should we be collecting prospectively to prove that it is better than circulatory arrest? DR PIGULA: I have thought a bit about that. Mortality is a very crude assessment, and I think if we make a difference between an 80% survival rate and a 90% survival rate, it will take a lot of patients to try to prove a difference. In my estimation, the most important thing is to assess the developmental outcomes in these patients, even though new developmental assessments are fraught with difficulties.