Right Ventricle to Pulmonary Artery Conduit Versus Blalock-Taussig Shunt: A Hemodynamic Comparison

Transcription

1 Right Ventricle to Pulmonary Artery Conduit Versus Blalock-Taussig Shunt: A Hemodynamic Comparison Nancy S. Ghanayem, MD, Robert D. B. Jaquiss, MD, Joseph R. Cava, MD, Peter C. Frommelt, MD, Kathleen A. Mussatto, RN, George M. Hoffman, MD, and James S. Tweddell, MD Herma Heart Center at the Children s Hospital of Wisconsin, and the Divisions of Critical Care and Cardiology, Department of Pediatrics, Department of Anesthesia, and the Division of Cardiothoracic Surgery, Department of Surgery, Medical College of Wisconsin, Milwaukee, Wisconsin CARDIOVASCULAR Background. A comprehensive assessment of 48 hour postoperative hemodynamics in neonates randomized to the right ventricle to pulmonary artery (RV-PA) conduit or modified Blalock-Taussig (BT) shunt for stage 1 palliation of hypoplastic left heart syndrome was performed to determine the potential benefits of the modified technique. Methods. Randomization to either RV-PA conduit or BT shunt was stratified by surgeon and the presence of aortic atresia. The designated procedure was performed by using hypothermic cardiopulmonary bypass with phenoxybenzamine, continuous cerebral perfusion, phstat blood gas management, and continuous postoperative venous oximetry. Differences between treatments were analyzed by time-series generalized least-squares regression, 2 tests, two-way repeated measures analysis of variance, and the Levene variance ratio test for variability in parameters, as appropriate. Results. All patients underwent the procedure to which they were randomized. There were no differences in age, weight, deep hypothermic circulatory arrest, or cardiopulmonary bypass times between patients receiving the BT shunt (n 8) or the RV-PA conduit (n 9). There was one early and one late death in the RV-PA conduit group, and one interstage death in the BT shunt group. Other than diastolic blood pressure (39 mm Hg in BT shunt versus 46 mm Hg in RV-PA conduit, p < 0.001), there were no differences in the mean values of arterial saturation, venous oximetry, mean arterial blood pressure, pulmonary-to-systemic flow ratio (Qp/Qs), or any other physiologic or inotropic support variable between groups. The variability of physiologic values related to pulmonary blood flow was greater in the RV-PA group (Qp/Qs coefficient of variation, 0.91 versus 2.50, p < 0.001). Conclusions. In this randomized prospective study, no hemodynamic benefits of the RV-PA modification for stage 1 palliation of hypoplastic left heart syndrome were found. Pulmonary blood flow was more variable, and the diastolic blood pressure was higher. These findings did not influence indicators of systemic oxygen delivery with our afterload reduction strategy. (Ann Thorac Surg 2006;82: ) 2006 by The Society of Thoracic Surgeons Two decades after Norwood s description of initial palliation for hypoplastic left heart syndrome, the right ventricle to pulmonary artery (RV-PA) conduit for provision of pulmonary blood flow was reintroduced as part of stage 1 palliation [1, 2]. Since then, several centers have reported improved results using the RV-PA conduit, but these studies have compared noncontemporary groups of patients in which the surgical treatment strategy changed over time from the modified Blalock- Taussig (BT) shunt to the RV-PA conduit [3 5]. To date, studies comparing the results between these two techniques have not been prospective or randomized. Accepted for publication May 22, Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30 Feb 1, Address correspondence to Dr Ghanayem, Children s Hospital of Wisconsin, 9000 W Wisconsin Ave, Milwaukee, WI 53201; nancyg@mcw.edu. Because the risk factors for early mortality are similar to risk factors for impaired early systemic oxygen delivery [6, 7], we questioned whether the purported survival advantage of the RV-PA conduit might be a reflection of more robust hemodynamics as measured by variables of systemic oxygen economy. We hypothesized patients with hypoplastic left heart syndrome who received the RV-PA conduit as part of the initial palliation would have more favorable hemodynamics compared with similar patients who had stage 1 palliation with a BT shunt. Specifically, the primary outcomes included comparisons of invasive and noninvasive measures of oxygen delivery, derived variables of arteriovenous oxygen saturation and content difference, systemic and pulmonary flow indices, and pulmonary/systemic blood flow ratio. This was a randomized, prospective study of the postoperative hemodynamics in patients undergoing placement of either the RV-PA conduit or BT shunt as part of initial palliation for hypoplastic left heart syndrome by The Society of Thoracic Surgeons /06/$32.00 Published by Elsevier Inc doi: /j.athoracsur

2 CARDIOVASCULAR 1604 GHANAYEM ET AL Ann Thorac Surg EARLY HEMODYNAMICS AT STAGE 1 PALLIATION 2006;82: Patients and Methods Patients At Children s Hospital of Wisconsin from January 2004 through April 2005, 21 patients with hypoplastic left heart syndrome and planned staged palliation were screened for participation in a randomized, prospective trial comparing the RV-PA conduit and BT shunt, with specific focus on early postoperative hemodynamics. Three patients with cardiac or extracardiac anomalies thought to increase the risk of morbidity or mortality independent of shunt type met exclusion criteria and were not included in the trial. One of these patients had heterotaxy with interrupted inferior vena cava and intestinal malrotation, another patient underwent attempted biventricular repair but then progressed to staged palliation 1 week later, and a third patient had unresolved multiorgan dysfunction with a large hemorrhagic stroke preoperatively. A fourth patient was excluded owing to the absence of informed consent. After informed consent was obtained from parents, 17 patients eligible for the trial were randomized to receive either the RV-PA conduit or BT shunt. Randomization was performed the night before surgery and stratified by the presence of aortic atresia and surgeon. Two surgeons experienced in staged palliation for hypoplastic left heart syndrome participated in this trial. Treatment assignments were made according to a randomized blocks design. Randomization schemes were developed in advance and placed in sealed, numbered envelopes. This study was approved by the Human Studies Committee at Children s Hospital of Wisconsin. Intraoperative Management The operative approach for this cohort was identical other than shunt type. The surgical repair has been described previously and included relief of aortic arch obstruction by means of coarctectomy, augmentation of the aortic arch with allograft pulmonary artery, and side-to-side anastomosis of the aortic and pulmonary roots with a proximal cut-back below the sinotubular ridge of the pulmonary root [7]. The diameter of the BT shunts was determined by patient weight, whereas all RV-PA conduits were 6.0- mm, nonvalved, polytetrafluoroethylene grafts extending from the right ventricular infundibulum to the central pulmonary artery confluence, which was patched in all cases with a pulmonary allograft. The 6.0-mm conduit was initially selected as a result of anecdotal reports from centers that reported progressive hypoxemia and the need for early cavopulmonary anastomosis after placement of 4.0-mm and 5.0-mm conduits. Continued use of the 6.0-mm conduit stems from assessment of detailed postoperative hemodynamics that have not suggested shunt-size mismatch. Hypothermic cardiopulmonary bypass with continuous cerebral perfusion through a shunt sewn on the innominate artery was used in all cases. Circulatory arrest was used during the atrial septectomy in all cases and during arterial cannula repositioning in BT shunt cases. Modified ultrafiltration was used in all patients. All patients received methylprednisolone (10 mg/kg) at 8 hours and 2 hours before surgery. Aprotinin KIU/m 2 was administered intravenously before the skin incision and in the bypass circuit prime and as a continuous infusion of KIU m 2 hr during the procedure. Phenoxybenzamine (0.25 mg/kg) was added to the pump prime according to an Institutional Review Board and Food and Drug Administration approved protocol and investigational device exemption. Oximetric catheters were surgically placed in the superior vena cava before separation from bypass [8]. Milrinone at 0.5 g kg min was initiated during rewarming, and norepinephrine was titrated before separation from bypass to achieve an approximate systemic vascular resistance (SVR) index of 12 woods units. Epinephrine was used for inotropic support. All patients underwent delayed sternal closure in the intensive care unit when they demonstrated hemodynamic stability and resolution of chest wall edema [9]. Postoperative Management Standardized postoperative monitoring included continuous invasive arterial blood pressure, central venous pressure (CVP), arterial saturation (SaO 2, Masimo, Irvine, CA), and venous oximetry (SvO 2, Abbott Laboratories, Abbott Park, IL) from the superior vena cava. Twosite regional oximetry with near infrared spectroscopy (INVOS, Somanetics 5100A, Troy, MI) was used to assess regional tissue perfusion and oxyhemoglobin saturation (rso 2 ) of cerebral and somatic (renal) vascular beds [10]. Postoperative management targets of SaO 2 80%, SvO 2 50%, cerebral rso 2 50%, somatic rso 2 60%, mean arterial blood pressure (MABP) 45 mm Hg, diastolic blood pressure (DBP) 30 mm Hg, and hematocrit 45% were achieved by titration of vasoactive drugs, red blood cell transfusion, and controlled ventilation. Fentanyl infusion was initiated in the operating room and continued until after sternal closure. For the first postoperative night, patients also received neuromuscular blockade and scheduled benzodiazepine sedation. Hemodynamic Assessment Perioperative hemodynamic indices were recorded prospectively for the first 48 hours after neonatal stage 1 palliation, including SaO 2, SvO 2, MABP, systolic blood pressure (SBP), DBP, cerebral rso 2, somatic rso 2, CVP, heart rate, hemoglobin level, arterial partial pressure of carbon dioxide (PaCO 2 ), base excess, inspired oxygen fraction (fio 2 ), ventilatory index (the product of the ventilator rate times the difference between peak and end-expiratory pressure pco 2 / 1000), derived values of arteriovenous oxygen saturation difference (Sa-vO 2 ) and content difference (CavO 2 ), pulmonary/systemic blood flow ratio (Qp/Qs, assuming a pulmonary vein saturation of 97%), systemic (Qs) and pulmonary (Qp) flow indices (assuming an oxygen consumption of 160 ml/[m 2 min]), and stroke volume index [(Qs Qp)/heart rate, L m 2

3 Ann Thorac Surg GHANAYEM ET AL 2006;82: EARLY HEMODYNAMICS AT STAGE 1 PALLIATION beat]. Intraoperative variables recorded included duration of deep hypothermic circulatory arrest and cardiopulmonary bypass and use of phenoxybenzamine and inotropic support. Table 1. Patient and Operative Characteristics Characteristics BT Shunt a n RV-PA Conduit a n 9 CARDIOVASCULAR Statistical Analysis Data were expressed as mean standard deviation for descriptive statistics and as median with range between 5% to 95% confidence intervals, when appropriate. The difference in mean values of hemodynamic variables and support levels between patients in the RV-PA conduit or BT shunt groups was assessed by generalized least squares time-series regression with correction for autocorrelation. The degree of variability in physiologic and support variables between treatment groups was assessed by the Levene conservative test of homogeneity of variances, with adjustment of the degrees of freedom for the number of patients in each group. Continuous values for hemodynamic parameters were divided into clinically appropriate strata to test for nonlinear differences between treatment groups at different physiologic conditions by two-way repeated-measures analysis of variance. The cutoff for significance was p 0.05 after multiple comparison correction using the Tukey honestly significant difference, or the Bonferroni method, when applicable. All calculations were performed with STATA 8 (STATA Corp, College Station, TX). Results The clinical trial enrolled 17 patients. All patients received the shunt they were randomized to, with no intraoperative or postoperative crossover to the alternative shunt. Nine patients received the RV-PA conduit, and 8 received the BT shunt. All RV-PA conduits were 6.0-mm nonvalved expanded polytetrafluoroethylene shunts. In the BT shunt group, 3.0-mm shunts were placed in 2 patients, 3.5-mm shunts were placed in 5 patients, and a 4.0-mm shunt was placed in 1 patient. Five of the 9 patients with RV-PA conduits and 4 of the 8 patients with BT shunts had aortic atresia. A diminutive ascending aorta ( 2.5 mm) was present in 4 patients with RV-PA conduits and in 3 patients with BT shunts. Of the remaining 4 patients without aortic atresia who received RV-PA conduits, 1 had aortic stenosis and mitral stenosis, 1 had aortic stenosis, mitral atresia, and ventricular septal defect, 1 had aortic stenosis, mitral stenosis, and a ventricular septal defect, and 1 had double outlet right ventricle, mitral atresia, and coarctation. In the BT shunt patients without aortic atresia, 3 patients had aortic stenosis and mitral stenosis, and 1 had aortic stenosis, mitral stenosis, and a ventricular septal defect. Moderate atrioventricular valve insufficiency was present on the preoperative echocardiogram in 3 of 17 patients, all of whom received RV-PA conduits (p 0.25). Antenatal diagnosis of hypoplastic left heart syndrome was made in 10 (59%) of 17 of the cohort. In this group, 3 patients were randomized to receive RV-PA conduits and 7 received BT shunts. Of the 7 patients without prenatal Male 6 (75%) 5 (56%) Birth weight (kg) Aortic atresia 4 (50%) 5 (56%) Ascending aorta 2.5 mm 3 (38%) 4 (44%) Ascending aorta diameter (mm) Atrioventricular valve insufficieny 0 3 (33%) Age at operation (days) Cardiopulmonary bypass time (min) Circulatory arrest time (min) a Value of p were not significant. BT Blalock-Taussig; RV-PA right ventricle-to-pulmonary artery. diagnosis, 6 received RV-PA conduits and 1 received a BT shunt. Of the 7 patients with postnatal diagnosis, 3 presented in extremis with multiorgan failure and 2 of these patients received RV-PA conduits and 1 received a BT shunt. Each of these patients had recovered organ function before surgery. Three patients (2 with RV-PA conduits and 1 with a BT shunt) were premature ( 37 weeks gestational age), one of which was diagnosed with severe respiratory distress syndrome. This patient required oscillating mechanical ventilation and received a RV-PA conduit after recovery of pulmonary function. One patient who received a RV-PA conduit had symptomatic supraglottic laryngomalacia and left mainstem bronchomalacia. These extracardiac anomalies have not proven to increase morbidity or mortality in our experience and were included in this series. Patient characteristics and intraoperative data are summarized in Table 1. Birth weight, age of operation, ascending aorta diameter, and intraoperative support times were similar between patient groups. Preoperative moderate atrioventricular valve insufficiency was present in 3 patients in the RV-PA conduit group and absent in all patients who received a BT shunt. Complete postoperative hemodynamic variables are listed in Table 2 and shown as mean values standard deviation over 48 hours. Systolic and mean arterial blood pressures were not different between groups, but DBP was lower in the BT shunt group than in the RV-PA conduit group (39 4 mm Hg versus 46 6mm Hg, p 0.001; Fig 1). Mean heart rate, CVP, SaO 2, SvO 2, cerebral rso 2 and somatic rso 2, Sa-vO 2, Ca-vO 2,Qp, Qs, Qp/Qs, and stroke volume index did not differ between groups during the first 48 postoperative hours (Fig 2). However, the variability of physiologic values related to pulmonary blood flow (SaO 2, Qp, and Qp/ Qs) was greater in the RV-PA group (Fig 3). There was greater variability in Qp, but not Qs, over time in the RV-PA group, with resulting variability in Qp/Qs (Fig 3). The relationship between MABP and Qp, Qs, and Qp/Qs are shown in Figure 4. At high MABP, the RV-PA conduit group exhibited greater variability in

4 CARDIOVASCULAR 1606 GHANAYEM ET AL Ann Thorac Surg EARLY HEMODYNAMICS AT STAGE 1 PALLIATION 2006;82: Table 2. Hemodynamics and Support Variables During First 48 Postoperative Hours Variable BT Shunt (n 8) RV-PA Conduit (n 9) Mean SD Range Mean SD Range p Means p Variances Blood pressure (mmhg) Diastolic Systolic Mean SaO 2 (%) SvO 2 (%) CVP (mmhg) Heart rate (1/min) Qp/Qs Qp (l/min/m 2 ) L/(min m 2 ) Qs (l/min/m 2 ) L/(min m 2 ) Ca-vO 2 (ml/dl) Sa-vO 2 (%) Cerebral rso Somatic rso PaCO 2 (torr) Ventilatory index fio Base excess (meq/l) Stroke volume index (L m 2 beat) Epinephrine ( g kg min) Norepinephrine ( g kg min) ) Differences between means were tested by time-series generalized least squares regression. Differences between variances were tested by Levene s test of equality of variances. Ventilatory index [(PIP peep)*imv rate* pco 2 ]/1000. PIP PEEP IMV rate pco 2 /1000. Stroke volume index (Qs Qp)/heart rate. BT Blalock-Taussig; RV-PA right ventricle to pulmonary artery; CVP central venous pressure; Qp pulmonary blood flow; Qs systemic blood flow; PIP peak inspiratory pressure; PEEP positive end expiration pressure; IMV intermittent mechanical ventilation. pulmonary blood flow; however, Qs remained relatively invariant of MABP in both groups. Support indices were similar between groups. There were no differences in vasodilator or potent inotrope drug doses. Other than a higher fio 2 in the BT shunt group ( versus , p 0.01), there was no difference in the ventilatory support, particularly in ventilatory index and pco 2 (Table 2). Within this cohort, no patient required cardiopulmonary resuscitation or mechanical cardiac support in the early postoperative period. One patient in the BT shunt group died at home 42 days after stage 1 palliation. Two patients in the RV-PA conduit group died. One patient delivered prematurely with preoperative, yet resolved, severe respiratory distress syndrome died 68 days after stage 1 palliation from recurrent respiratory failure. The second patient had preoperative moderate atrioventricular valve insufficiency that progressed to severe valve insufficiency during the interstage period. Support was withdrawn after stage 2 palliation with atrioventricular valvuloplasty (age 105 days) at the request of the family, who declined further intervention. This patient required extracorporeal mechanical support during the interstage period. Fig 1. Arterial (AP) and central venous (CVP) pressures are shown as mean standard deviation in mm Hg between groups. The Blalock-Taussig (BT) shunt group is on the left panel and the right ventricle-to-pulmonary artery (RV-PA) conduit group is on the right panel. Systolic (gray line with circles), mean arterial (black line with circles), and central venous pressures (black line with ovals) were similar between groups. Diastolic blood pressure was lower in the BT shunt group compared with the RV-PA conduit group (39 4 versus 46 6 mm Hg, p 0.001).

5 Ann Thorac Surg GHANAYEM ET AL 2006;82: EARLY HEMODYNAMICS AT STAGE 1 PALLIATION 1607 CARDIOVASCULAR Fig 2. Arterial saturation (SaO 2, solid line), superior vena cava venous oximetry (SvO 2, gray line), and near infrared spectroscopy cerebral (dashed black line) and somatic oxyhemoglobin saturations (rso 2, dashed gray line) displayed in 4-hour intervals over the first 48 hours in the Blalock-Taussig (BT) shunt and right ventricle to pulmonary artery (RV-PA) conduit groups. Each point represents the group 4-hour mean and 95% confidence interval. There was no significant difference in mean values or in variation of parameters between groups. Comment Operative survival after stage 1 palliation for hypoplastic left heart syndrome has greatly improved since the Norwood and colleagues [11] description of a reliable operative strategy for initial palliation and in the past decade has ranged from 80% to 95% at many centers [7, 12 14]. Innovative perioperative strategies that include modifications to cardiopulmonary bypass, introduction of -adrenergic blockade for SVR reduction, technologic advances in perioperative monitoring, and adoption of novel surgical techniques have all contributed to the Fig 3. Pulmonary blood flow (Qp), systemic blood flow (Qs), and pulmonary/systemic blood flow ratio (Qp/Qs) displayed in 4-hour intervals over the first 48 hours in the Blalock-Taussig (BT) shunt and right ventricle to pulmonary artery (RV-PA) conduit groups. Each point represents the group 4-hour mean, and the 95% confidence interval is displayed for each point. Although the 48 hour means are not different, the variability was greater in the RV-PA conduit group for Qs (p 0.01 by the Levene test) and Qp/Qs (p 0.001). Fig 4. The relationship between mean arterial blood pressure (MABP) and arterial saturation (SaO 2 ), superior vena cava venous oximetry (SvO 2,) (solid lines), and pulmonary blood flow (Qp), and systemic blood flow (Qs) (dashed line) is shown by fractional polynomial modeling in the modified Blalock-Taussig (BT) shunt and right ventricle to pulmonary artery (RV-PA) conduit groups. Qp was higher and more variable at MABP greater than 50 mm Hg in the RV-PA conduit group (p 0.05 by fractional polynomial regression and analysis of variance). Data are shown as mean standard deviation (gray shaded area). increasing success observed in the palliation of hypoplastic left heart syndrome [1, 7, 8, 10, 15 17]. Among the most widely adopted surgical modifications has been the change to the RV-PA conduit rather than the BT shunt as a means of providing pulmonary blood flow during the initial palliation [1 5]. Although initial reports of the RV-PA conduit modification for stage 1 palliation indicated a reduction in early mortality, these results have not been reproduced at all centers, particularly if mortality with the BT shunt was less than 20% [13, 14, 18, 19]. Much of the literature describing experience with the RV-PA conduit has focused on survival, without systematic examination of hemodynamic differences compared with patients with BT shunts. Bradley and colleagues [13] found a higher DBP but no difference in indices of oxygen delivery in patients who received the RV-PA conduit compared with a historical control cohort who received a BT shunt. Our data confirm higher DBP with the RV-PA conduits (46 6 mm Hg versus 39 4 mm Hg in the BT shunt group, p 0.001) but do not show concomitant increased systemic oxygen delivery as evidenced by SvO 2, two-site noninvasive cerebral and somatic oximetry, Sa-vO 2 and Ca-vO 2. Similarly, there was not a statistical difference in the mean Qp/Qs during 48 hours. Because our data are derived from a baseline, resting, minimal oxygen consumption state rather than from a high demand state, as in sepsis for example, some benefit in myocardial reserve may be associated with the RV-PA conduit that our study was not designed to detect. Mechanisms of potential benefits of one shunt over another have been extensively hypothesized but incompletely validated. The BT shunt, which avoids a ventriculotomy in a univentricular heart, is frequently said to confer a significant risk of myocardial ischemia due to

6 CARDIOVASCULAR 1608 GHANAYEM ET AL Ann Thorac Surg EARLY HEMODYNAMICS AT STAGE 1 PALLIATION 2006;82: diastolic hypotension and coronary insufficiency. By separating the systemic and pulmonary circulations during diastole, the RV-PA conduit is said to eliminate the diastolic runoff observed with the BT shunt, which might increase coronary blood flow and, in turn, might improve ventricular function (despite the ventriculotomy). Our study did not find evidence of benefit from the postulated increase in myocardial oxygen reserve conferred by increased DBP; however, our patients were studied while sedated and with significant control of autonomic responses from phenoxybenzamine. The variability of physiologic indices related to pulmonary blood flow was greater in the RV-PA conduit group, which might reflect the uniform size of the RV-PA conduit regardless of patient weight, unlike those patients in the BT shunt group. However, systemic flow indices were well maintained in the RV-PA conduit group as well as the BT shunt group under conditions of increased blood pressure, suggesting both groups had matched shunt size. Blood pressure is determined by systemic flow and vascular resistance. In patients with the BT shunt, use of phenoxybenzamine results in higher Svo 2 and less variable Qp/Qs over a wider blood pressure range, because cardiac output, not systemic vascular resistance, becomes the more important determinant of blood pressure [7, 16]. We have continued to use phenoxybenzamine at our institution, and thus all patients in this series had systemic vascular resistance stabilized with -adrenergic blockade. Although it is curious that the periodically high Qp/Qs was observed without reduction in indices of systemic flow in the RV-PA conduit group, we cannot conclusively presume that the RV-PA conduit modification is more resistant to increases in systemic vascular resistance. The RV-PA conduit modification is seemingly more energy efficient than the BT shunt. Although phasic, BT shunt flow is continuous, and a significant amount of flow occurs during diastole. This diastolic flow depends on energy stored during systole. This energy is a factor of the excess volume to be delivered to the pulmonary arteries and the pressure of that stored blood volume. In essence, the elastic aorta and systemic arteries act as a capacitor, charged during systole, and the aorta discharges the volume-pressure energy to the pulmonary circuit during diastole. Other studies have shown an increase in systolic blood pressure in patients with the BT shunt compared with the RV-PA conduit modification at stage 1 palliation and may reflect this increased energy utilization with the BT shunt [20]. In this study, we did not identify a significant increase in SBP among the BT shunt patients, but we were able to show that the larger RV-PA conduit provides the same amount of pulmonary blood flow entirely during systole without the inefficiencies of storing energy at a higher level in the aorta. This improved energy efficiency comes at the cost of a ventriculotomy and a degree of ventricular volume overload related to regurgitant conduit flow. Ultimately, the potential benefits of the RV-PA conduit will be a function of the positive impact of improved energy utilization and a possible decrease in thrombotic BT shunt complications balanced against the long-term risk of a ventriculotomy and volume overload. Our single-center hemodynamic study is limited by the small sample size and is thus unable to assess the potential benefits of either surgical strategy in patients with additional cardiac or extracardiac morbidity. In particular, preoperative moderate atrioventricular valve insufficiency was seen in 3 patients in the RV-PA conduit group and not observed in the BT shunt group. Our perioperative strategy that emphasizes a reduction in systemic vascular resistance and a minimal oxygen consumption state is optimal for the patient with atrioventricular valve insufficiency. As a result, the impact of atrioventricular valve insufficiency on outcomes after stage 1 palliation might be reflected in late outcomes and not in immediate postoperative hemodynamics. The patients in this report represent our enrollment in a study initially developed as a multicenter controlled clinical protocol created through a collaborative effort between Children s Hospital of Wisconsin, University of Michigan Health System, Children s Hospital of Philadelphia, and Children s Hospital Boston. Ultimately, the importance of this study was recognized by the Pediatric Heart Network and was adopted as the currently active Single Ventricle Reconstruction Trial. Continued efforts in this current multicenter randomized trial will be necessary for accrual of the sample size necessary to generate the statistical power to truly evaluate the differences between groups not only in early postoperative hemodynamics but also in longer-term outcomes such as survival to stage 2, 1 year of age, rate of progression to transplantation, and neurodevelopmental outcomes. References 1. Sano S, Ishino K, Kawada M, et al. Right ventriclepulmonary artery shunt in the first-stage palliation of hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 2003;126: Kishimoto H, Kawahira Y, Kawata H, et al. The modified Norwood palliation on a beating heart. J Thorac Cardiovasc Surg 1999;118: Pizarro C, Malec E, Maher IO, et al. Right ventricle to pulmonary artery conduit improves outcomes after stage I Norwood for hypoplastic left heart syndrome. Circulation 2003;108(Suppl 2): Malec E, Januszewska K, Kolcz J, Mroczek T. Right ventricleto-pulmonary artery shunt versus modified Blalock-Taussig shunt in the Norwood procedure for hypoplastic left heart syndrome influence on early and late haemodynamic status. Eur J Cardiothorac Surg 2003; Mair R, Tulzer G, Sames E, et al. Right ventricular to pulmonary artery conduit instead of modified Blalock- Taussig shunt improves postoperative hemodynamics in newborns after the Norwood operation. J Thorac Cardiovasc Surg 2003;126: Tweddell JS, Hoffman GM, Fedderly RT, et al. Patients at risk for low systemic oxygen delivery after the Norwood procedure. Ann Thorac Surg 2000;69: Tweddell JS, Hoffman GM, Mussatto KA, et al. Improved survival of patients undergoing palliation of hypoplastic left heart syndrome: lessons learned from 115 consecutive patients. Circulation 2002;106:I82 9.

7 Ann Thorac Surg GHANAYEM ET AL 2006;82: EARLY HEMODYNAMICS AT STAGE 1 PALLIATION Tweddell JS, Hoffman GM, Fedderly RT, et al. Phenoxybenzamine improves systemic oxygen delivery after the Norwood procedure. Ann Thorac Surg 1999;67: Tabbutt S, Duncan BW, McLaughlin D, et al. Delayed sternal closure after cardiac operations in a pediatric population. J Thorac Cardiovasc Surg 1997;113: Hoffman GM, Stuth EA, Jaquiss RDJ, 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: Norwood WI, Lang P, Hansen DD. Physiologic repair of aortic atresia-hypoplastic left heart syndrome. N Engl J Med 1983;308: Ungerleider RM, Shen I, Yeh T, et al. Routine mechanical ventricular assist following the Norwood procedure improved neurologic outcome and excellent hospital survival. Ann Thorac Surg 2004;77: Bradley SM, Simsic JM, McQuinn TC, et al. Hemodynamic status after the Norwood procedure: a comparison of right ventricle-to-pulmonary artery connection versus modified Blalock-Taussig shunt. Ann Thorac Surg 2004;78: Tabbutt S, Dominguez TE, Ravishankar C, et al. Outcomes after the stage I reconstruction comparing the right ventricular to pulmonary artery conduit with the modified Blalock-Taussig shunt. Ann Thorac Surg 2005;80: Hoffman GM, Ghanayem NS, Kampine JM, et al. Venous saturation and the anaerobic threshold in neonates after the Norwood procedure for hypoplastic left heart syndrome. Ann Thorac Surg 2000;70: Hoffman GM, Tweddell JS, Ghanayem NS, et al. Alteration in critical arteriovenous oxygen saturation relationship by sustained afterload reduction after the Norwood procedure. J Thorac Cardiovasc Surg 2004;127: Prevention of early sudden circulatory collapse after the Norwood operation. Circulation 2004;110:II Mahle WT, Cuadrado AR, Tam VK. Early experience with a modified Norwood procedure using the right ventricle to pulmonary artery conduit. Ann Thorac Surg 2003;76: Azakie A, Martinez D, Sapru A, et al. Impact of right ventricle to pulmonary artery conduit on outcome of the modified Norwood procedure. Ann Thorac Surg 2004;77: Pizarro C, Mroczek T, Malec E, et al. Right ventricle to pulmonary artery conduit reduces interim mortality after stage 1 Norwood for hypoplastic left heart syndrome. Ann Thorac Surg 2004;78: CARDIOVASCULAR DISCUSSION DR JACEK MOLL (Lodz, Poland): I am surprised that you used 6-mm Gore-Tex tube for RV-PA shunt. I usually use 5 mm for children whose weight is less than 3.5 kg, and it s enough. I used 6 mm in one patient, and I had pulmonary overflow and I had to decrease the size. And from that time, I electively use 5-mm Gore-Tex shunt, and it s quite enough for these patients. We have next postoperative day usually saturation in the range of 75 to 80, and I believe that the 5-mm shunt is enough for those patients. DR GHANAYEM: Our experience with venous oximetry over the past decade has taught us that targeting arterial saturation incompletely reflects systemic oxygen delivery. Through invasive and noninvasive oximetry, we have been able to more reliably tailor shunt size and perioperative management. With a 6.0 mm RV-to-PA conduit, our patients generally have saturations in the 80s and Pao 2 s in the 40s. We don t believe these shunts are too big given that the concomitant venous saturations, regional oximetry, and organ function are all quite acceptable. Others have reported their experience with the RV-to-PA conduit, specifically with the use 4.0- and 5.0-mm shunts, and have remarked on the need for earlier cavopulmonary anastomosis. The 6.0-mm conduits in our series have allowed for sufficient pulmonary artery growth and, generally, avoidance of interstage hypoxemia that prompts intervention on the shunt or earlier second stage palliation. When we looked at these patients in terms of timing of the second operation, we found that patients with the RV to PA conduit had their cavopulmonary anastomosis at nearly 4 months of age versus the modified BT shunt group who underwent stage 2 palliation at just over 3 months of age. DR MOLL: At what age do you do second-stage operation? Because usually we do it at 4 months of age, and until this time the 5-mm shunt is usually enough. DR GHANAYEM: With the interstage home monitoring program, we have identified a reduction in growth velocity between 4 to 5 months of age and, therefore, see little benefit in delaying the second stage operation beyond this age. Our practice has changed over the past 5 years; we now minimize the infant s period of at-risk parallel circulation and generally proceed with the cavopulmonary anastomosis electively at 4 months of age unless physiologic data suggest proceeding earlier. DR EMILE M. BACHA (Boston, MA): I have two questions. Your total N is fairly small, and so my first question would be, do you think you would see differences if you increased the number of your patients from a statistical standpoint? And the other question, that may be more important, is that it s our impression that the RV-to-PA conduit is more beneficial in the more extreme cases, such as the low-birth-weight babies or babies who have other anomalies in addition to HLHS. Do you think that including these kinds of patients your mean weight I believe was in the 3-kg range decreasing that weight for example, would show you a difference? DR GHANAYEM: To answer the first question, you re absolutely correct; the sample size is small in this study. Hopefully, randomization of greater than 400 patients in the ongoing multicenter Single Ventricle Reconstruction Trial will better assess the significant differences between shunt types, if one exists. Your second question does address a very important concept and that is whether the RV-to-PA conduit is more beneficial for low-birth-weight babies or those with extracardiac anomalies and morbidities. I don t believe we have enough experience or understanding of the long-term implications of the RV-to-PA conduit to conclude that one shunt is superior to the other for those patients identified as high risk. As you correctly identified, our study sample size is too small and certainly limits our ability to conclude such a benefit exists for the high risk neonate. This is yet another question that will be better answered with a large multicenter trial. DR FRANÇOIS G. LACOUR-GAYET (Denver, CO): I have a similar question to that of Emile Bacha. If I understand well, you

8 CARDIOVASCULAR 1610 GHANAYEM ET AL Ann Thorac Surg EARLY HEMODYNAMICS AT STAGE 1 PALLIATION 2006;82: have removed from the study the most complex forms. What did you do with these patients? DR GHANAYEM: On the contrary, the most complex patients were not removed from this trial. Within this series, patients with prematurity, low birth weight, preoperative organ failure, and extracardiac anomalies were included. There were, however, 4 patients screened and not included in this trial. One patient had heterotaxy with interrupted IVC and malrotation. She received an RV-to-PA conduit. Another patient had an initial attempt at biventricular repair that failed and underwent stage 1 palliation with a modified BT shunt 1 week later. A third patient presented in shock with multiorgan dysfunction as well as a stroke prior to undergoing initial palliation; this patient was the first RV-to-PA conduit for one of our surgeons and we felt was not suitable for randomization because of experience. The fourth screened patient had no obvious comorbidities and was not consented. This patient received a RV-to-PA conduit at the surgeon s discretion. DR LACOUR-GAYET: Even if you look at low birth weight? DR GHANAYEM: We didn t have extremely low-birth-weight patients in this series. Our smallest patients were 2.5 kg: one received a BT shunt and two received RV-to-PA conduits. DR MARSHALL L. JACOBS (Philadelphia, PA): Let me ask you for a moment to put on your more subjective critical care physician hat. When Dr Sano championed this modification, he said it creates a scenario where the patient is easier to care for in the intensive care unit. Apart from the objective analysis with the doses of inotropes and ventilatory times and whatnot, is there anything subjective about your experience that has led you to assess one strategy as having more merit than the other? conduit and BT shunt. Another important question, which I don t know the answer to as of yet, is whether one shunt is advantageous over the other beyond the immediate postoperative period when sympathetic tone is less well controlled. DR FRANK A. PIGULA (Boston, MA): I have a question for you related to the pulmonary blood flow, the Qp/Qs. There is a fair amount of variability in the data, of course, and my question really pertains to the fact that do you have a sense, or a shunt preference, when you re faced with a child that may have elevated pulmonary vascular resistance, do you have a sense of which shunt may be preferable in those circumstances? And if you do, why? DR GHANAYEM: We actually have just discussed this within the past couple months at our own institution. In the patient with a very restrictive or intact atrial septum, an extreme situation where pulmonary vascular resistance is high, perhaps a modified BT shunt would provide greater pulmonary circulation, given that flow through the shunt is continuous during systole and diastole. Furthermore, during diastole, the aorta is able to continually generate volume pressure-energy to the pulmonary circuit with a BT shunt. This is, however, only theoretical. In our experience, we have placed RV-to-PA conduits in patients with elevated pulmonary vascular resistance with resultant oxygenation that is similar to patients with BT shunts. DR CHARLES D. FRASER, JR (Houston, TX): I wasn t clear, was the intensive care unit team blinded to the operation? DR GHANAYEM: Characteristic physical examination findings for each shunt type precluded blinding the ICU team to the randomization. DR GHANAYEM: To date our perioperative management has not differed between shunt types. I d say subjectively that there is little difference in the early postoperative management. This might very well be attributed to our invasive and noninvasive monitoring protocols as well as the perioperative use of phenoxybenzamine DR JACOBS: And maybe very good surgery. DR GHANAYEM: Oh, yes, and we have excellent surgeons. DR JACOBS: I m sorry, I didn t mean to interrupt. DR GHANAYEM: That s quite all right. Perhaps the advantage of the RV-to-PA conduit in the early postoperative period as proposed by Dr Sano and others would be more apparent in the absence of pharmacologic sympathectomy with alphaadrenergic blockade. However, Dr Bradley and his colleagues in South Carolina do not use phenoxybenzamine and last year reported similar results in their comparison of the RV-to-PA DR CHRISTOPHER KNOTT-CRAIG (Oklahoma City, OK): Do you think the use of phenoxybenzamine in the Sano-Norwood group may have complicated the postoperative care unnecessarily, or perhaps muted the differences in the two groups, since stage 1 Norwood requires a low systemic resistance, while the Sano-Norwood group relies on a normal SVR? And if so, is this a policy which you re planning to continue? DR GHANAYEM: The use of phenoxybenzamine at our institution has not been viewed as complicating care to achieve improved results; in fact, care is seemingly easier and more streamlined. The inherent parallel circulation of HLHS and the potential imbalance in pulmonary-to-systemic blood flow are of physiologic concern in the patient with either the BT shunt or RV-to-PA conduit. Regardless of shunt type, the goal of postoperative management is the same and targeted at balancing the partitioning blood flow and optimizing systemic oxygen delivery. Therefore, in the patient with planned RV-to-PA conduit, we continue to use phenoxybenzamine as part of our perioperative strategy.