Short and Long-Term Outcomes for Bidirectional Glenn Procedure Performed With and Without Cardiopulmonary Bypass

Transcription

1 Short and Long-Term Outcomes for Bidirectional Glenn Procedure Performed With and Without Cardiopulmonary Bypass Damien J. LaPar, MD, MS, Carlos M. Mery, MD, MPH, Benjamin B. Peeler, MD, Irving L. Kron, MD, and James J. Gangemi, MD Department of Surgery, Division of Thoracic and Cardiovascular Surgery, University of Virginia Health System, Charlottesville, Virginia Background. The bidirectional Glenn (BDG) procedure is most commonly used as staged palliation for complex cyanotic congenital heart defects. The benefits of a BDG procedure without the use of cardiopulmonary bypass (CPB) remain mixed within reported series. The purpose of this study was to compare short- and long-term outcomes for performance of a BDG procedure with and without the use of CPB. Methods. From 2001 to 2010, 106 patients underwent a BDG procedure. Patients were stratified into CPB (n 72; age 202 days) and non-cpb (n 34; age 182 days) groups. Primary outcomes included operative mortality and postoperative complications as well as differences in long-term Kaplan-Meier survival. Results. Median follow-up was 30 months. Preoperative patient characteristics were similar among patients despite the use of CPB. The most frequent indications for a BDG procedure were hypoplastic left heart syndrome (HLHS) (35.8%) and tricuspid atresia (TA) (17.9%). Median perfusion time was 73 minutes for CPB patients. Overall mortality was 0.9% and no deaths occurred among non-cpb patients (0.0% versus 1.4%; p > 0.99). Similarly, no significant differences existed between non-cpb patients and CPB patients with respect to overall complication rates (11.8% versus 18.1%; p 0.57) or postoperative length of stay (7.0 [ ] versus 7.0 [ ] days; p 0.38). Furthermore, 1-, 3-, and 5-year survival was high and similar between groups. Conclusions. The BDG procedure can be performed with no significant differences in operative mortality, morbidity, or use of resources, with or without CPB support. Long-term survival after the BDG procedure is high with both strategies. Performance of an off-pump BDG procedure should be considered a safe alternative to the conventional use of CPB for appropriately selected patients. (Ann Thorac Surg 2012;94:164 71) 2012 by The Society of Thoracic Surgeons Superior cavopulmonary anastomoses are commonly performed for complex cyanotic congenital heart defects. Initially thought to have been developed simultaneously by several different surgeons working independently at different institutions, the first clinical use of a unidirectional cavopulmonary anastomosis occurred in 1951 by Carlon [1] at the University of Padua. Subsequent revisions resulted in the classic approach described by William Glenn [2] in 1958, which has since been refined to deliver bidirectional flow through the pulmonary arteries. The bidirectional Glenn (BDG) anastomosis is most frequently used as staged palliation for patients with functionally univentricular hearts in preparation for a later Fontan operation. However it may also be used in patients with hypoplastic right ventricles and selected other lesions, such as Ebstein s anomaly. Accepted for publication March 1, Presented at the Fifty-eighth Annual Meeting of the Southern Thoracic Surgical Association, San Antonio, TX, Nov 9 12, Address correspondence to Dr Gangemi, Department of Surgery, University of Virginia, School of Medicine, PO Box , Charlottesville, VA 22908; jjg5d@virginia.edu. Performance of a BDG procedure has been described with and without the use of cardiopulmonary bypass (CPB) support. The use of CPB has a well-established risk profile, including increased lung injury and consequent prolonged mechanical ventilation requirements, hemolysis, and increased costs. Prolonged CPB has been recently identified as an independent predictor of adverse outcomes after a BDG procedure [3]. Anatomically, to perform the BDG anastomosis without the use of CPB, an alternative source of left pulmonary artery blood flow must exist when the right pulmonary artery is clamped. As a result, it is routine practice at our institution for any patient who requires a BDG procedure and does not have an exclusively right Blalock-Taussig (B-T) shunt to be considered a potential candidate for a non-cpb BDG procedure. It is important to consider, however, that performance of a BDG procedure without CPB may be associated with significant superior vena cava (SVC) pressure elevations that may predispose patients to neurologic injury. Thus the optimal operative approach for performance of a BDG procedure with respect to the use of CPB remains unsettled, and the comparative effective by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc

2 Ann Thorac Surg LAPAR ET AL 2012;94: BIDIRECTIONAL GLENN AND CARDIOPULMONARY BYPASS 165 ness of a BDG procedure with or without CPB remains ill defined. The purpose of the present study was to compare short- and long-term outcomes for patients undergoing a BDG procedure with and without the use of CPB. Although few reports exist on general outcomes after a BDG procedure without CPB, even fewer provide longterm survival data for these patients. Patients and Methods Patients Approval for this investigation was obtained by the Institutional Review Board and Human Investigation Committee of the University of Virginia Health System, including a waiver for the need to obtain patient consent. All consecutive patients undergoing performance of bidirectional superior cavopulmonary anastomoses were reviewed from 2001 to Patient demographics, principal surgical diagnosis, and operative strategy for the performance of a BDG procedure with or without CPB were used to define the study cohort. Patients were then stratified based on the use of CPB into 2 study groups: CPB and non-cpb. Perioperative indications for performance of a non-cpb BDG procedure were detailed, as were circumstances surrounding conversion from non- CPB to CPB. All operations were performed by 3 independent surgeons (J.J.G., B.B.P., and I.L.K.) at the University of Virginia. Primary outcomes of interest were short- and long-term mortality differences between study groups. Secondary outcomes included observed differences in patient characteristics, postoperative events, and use of resources. Accepted definitions of all patient- and operation-related variables as well as outcomes as defined by the Society of Thoracic Surgeons Congenital Heart Database were used in this study. Prolonged ventilation occurred with continuous mechanical ventilation longer than 7 days. Total hospital length of stay refers to each patient s entire inpatient admission, including both preoperative and postoperative admissions. Operations BDG PROCEDURE WITHOUT CARDIOPULMONARY BYPASS. For non- CPB patients, circumferential control was gained around the entire length of the SVC. The azygos vein was doubly ligated and divided. The entire length of the right pulmonary artery was dissected out, and circumferential control was gained around the right main pulmonary artery as well as the hilar branches. For patients with bilateral SVC involvement, the left pulmonary artery was dissected out in a similar manner. The patient was systemically heparinized with 300 units/kg heparin. A 12F right-angle cannula was placed high on the SVC or innominate vein. A 12F straight cannula was placed in the right atrium. These cannulas had the air removed and were hooked together to create a venoatrial shunt and allow drainage of the upper body while the proximal SVC was occluded. The pulmonary artery was temporarily occluded to ensure acceptable oxygen saturations and hemodynamic stability. For patients with bilateral SVC involvement, decompression cannulas were not used. The midportion of the SVC was occluded either with a clamp or snared with a caval tape and the SVC was divided from the right atrium. The right pulmonary artery was occluded proximally and distally, and a superior pulmonary arteriotomy was made and extended with Potts scissors. A standard end-to-side BDG anastomosis was then performed for a unilateral BDG procedure. Similar procedures were then performed for the left pulmonary artery for a bilateral BDG procedure. All anastomoses were performed using a continuous suture technique with 7-0 polypropylene. At the conclusion of the anastomoses, the clamps were released after air was removed from the anastomosis, all cannulas were removed, and the previously placed purse-string sutures were tied down. The systemic-pulmonary shunt (right ventricle to pulmonary artery conduit, modified B-T shunt, central shunt, or patent ductus arteriosus [PDA]) was then doubly ligated and divided. BDG PROCEDURE WITH CARDIOPULMONARY BYPASS. For patients requiring CPB, the venous cannulation strategy was the same, whereas an aortic cannula was placed in the ascending aorta. Each patient was then placed on standard CPB. All operations in this group were performed with a beating heart except in patients requiring a brief period of cardioplegic arrest to remove the atrial septum. Once on bypass, the B-T shunt was ligated and divided between sutures. The graft/pulmonary artery anastomosis was taken down, and the existing arteriotomy was extended. Subsequent end-to-side Glenn anastomoses were then performed in the same manner as already described. Statistical Analysis All statistical methodology was designed to test the null hypothesis that outcomes after performance of BDG shunts wound not be significantly different either with the use of CPB or without. Standard statistical significance was set to All study outcomes and data comparisons were established a priori before data collection. Descriptive statistics for all variable comparisons were calculated using appropriate univariate hypothesis tests. Categorical variables are expressed as within-group percentages and were compared for independent samples using either the Pearson s 2 test or Fisher s exact test. Continuous variables are expressed as either mean standard deviation or median [interquartile range] depending on overall variable distribution. Independent sample, single-factor analysis of variance was used for parametric data comparisons, whereas the Mann Whitney U test was used for all nonparametric data comparisons. Differences in longterm survival were analyzed by Kaplan-Meier analysis and compared using the log-rank test. Calculated test statistics were used to derive all reported 2-tailed p values. All statistical analyses were performed using

3 166 LAPAR ET AL Ann Thorac Surg BIDIRECTIONAL GLENN AND CARDIOPULMONARY BYPASS 2012;94: Predictive Analytics SoftWare (PASW), version 18 (IBM Corporation, Somers, NY). Results Patient Characteristics, Operative Details, and Surgical Diagnoses A total of 106 consecutive patients were included in this study, including 34 non-cpb patients undergoing a BDG procedure and 72 CPB patients undergoing a BDG procedure. Table 1 presents descriptive statistics for patient demographics, characteristics, and operative details. Overall, performance of operations was more common among males (67%) in both study groups. Average patient age within the non-cpb group was days compared with days in the CPB group (p 0.11). Similarly, no differences were observed in median patient weight (p 0.20), whereas CPB patients had slightly higher median heights (p 0.01). With respect to operative details, 91.5% of operations were unilateral BDG procedures and 8.5% were bilateral BDG procedures, circulatory arrest was performed in 4 (3.8%) patients, median CPB time for those undergoing a BDG procedure with CPB was 73 minutes ( minutes), and a total of 4 patients originally intended to undergo non-cpb BDG procedures underwent intraoperative conversion to CPB procedures. None of the patients undergoing either CPB or non-cpb BDG procedures required pulmonary arterioplasty or patch repairs. Regarding primary surgical diagnoses (Table 2), hypoplastic left heart syndrome (HLHS) and tricuspid atresia (TA) were the 2 most common indications for a BDG procedure in both study cohorts. HLHS was present in 52.9% of non-cpb patients and in 33.3% of CPB patients. TA was present in 11.8% of non-cpb patients and in a slightly higher proportion of CPB patients (20.8%). In addition, double-inlet left ventricle, double-outlet right Table 2. Principal Diagnoses for Performance of BDG Procedure with (CPB) or Without (non-cpb) Cardiopulmonary Bypass Support Diagnosis Non-CPB (n 34) CPB (n 72) No. % No. % HLHS TA DORV DILV UAVC PA/IVS LTGA/PA Ebstein s anomaly CPB cardiopulmonary bypass; DILV double-inlet left ventricle; DORV double-outlet right ventricle; HLHS hypoplastic left heart syndrome; IVS intact ventricular septum; LTGA L transposition of great arteries; PA pulmonary atresia; TA tricuspid atresia; UAVC unbalanced atrioventricular canal. ventricle, and pulmonary atresia were present in 4% to 12% of all patients. Patients with mitral atresia, transposition of the great arteries, hypoplastic right ventricle, and pulmonary stenosis all underwent a BDG procedure with CPB. The frequency of preoperative surgical indications that allowed the performance of a non-cpb BDG procedure are detailed in Table 3. Postoperative Events and Resource Use Overall and study group incidences of postoperative events and the extent resource use are listed in Table 4. Overall, the composite incidence of any complication was 16.0%. The most common complication after a BDG procedure was phrenic nerve paralysis (4.7%). Other noteworthy events included pleural effusion (2.8%), prolonged mechanical ventilation in excess of 7 days (2.8%), Table 1. Patient Characteristics and Perioperative Events Overall (n 106) Non-CPB (n 34) CPB (n 72) Perioperative Variable No. Mean Median % (SD) [IQR] No. Mean Median % (SD) [IQR] No. Mean Median % (SD) [IQR] p Sex 0.12 Female % % % Male % % % Age (days) (84.9) (87) (83.7) 0.11 Height (cm) 61.0 [61 67] 62.0 (60 64) 64.0 [61 68] 0.01 Weight (kg) 5.6 [ ] 6.2 [ ] 6.5 [ ] 0.20 Unilateral BDG % % % 0.99 Bilateral BDG 9 8.5% % 5 6.9% 0.99 Cardiopulmonary 73 [55 107] [55 107] 0.43 bypass time (min) Intraoperative non-cpb to CPB conversion 4 3.8% BDG Bidirectional Glenn; CPB cardiopulmonary bypass; IQR interquartile range; SD standard deviation.

4 Ann Thorac Surg LAPAR ET AL 2012;94: BIDIRECTIONAL GLENN AND CARDIOPULMONARY BYPASS 167 Table 3. Principal Surgical Indication for Performance of Non-CPB BDG Indication Non-CPB (n 34) No. % RV to PA shunt Forward PA flow Left B-T shunt Right B-T shunt with double SVC Central shunt Left B-T shunt with double SVC PA flow through patent PDA BDG bidirectional Glenn; B-T Blalock-Taussig; CPB cardiopulmonary bypass; PA pulmonary artery; PDA patent ductus arteriosus; RV right ventricle; SVC superior vena cava. need for reintubation (1.9%), reoperation for bleeding (1.9%), cardiac arrest (1.9%), chylothorax (1.9%), and need for postoperative extracorporeal membrane oxygenation (ECMO, 1.9%). Importantly, no significant differences in postoperative events were observed between patients in either study group. Median total hospital length of stay was 7.5 days [5 13 days], whereas postoperative length of stay was 7 days [5 12 days]; neither were significantly different between study groups (p 0.23 and p 0.38, respectively). Except for 2 patients requiring ECMO, all patients were hemodynamically stable with acceptable arterial blood gas analyses on transfer to the pediatric intensive care unit postoperatively. Short- and Long-Term Mortality Overall, operative mortality was 0.9%. No deaths occurred among non-cpb patients, whereas 1 patient in the CPB group died in the hospital (0.0% versus 1.4%; p 0.99). With respect to patient survival functions, patients undergoing BDG procedures demonstrated high survival with and without the use of CPB. Among non-cpb patients, 2 deaths occurred at 2 months and 5 months postoperatively, resulting in 93% 1-, 3-, and 5-year Kaplan-Meier survival. For those undergoing a BDG procedure with CPB, a total of 7 deaths occurred, resulting in 94%, 90%, and 87% 1-, 3-, and 5-year Kaplan-Meier survival, respectively. Among the 9 deaths, 1 mortality occurred from congestive heart failure, 1 mortality occurred from sepsis and multiorgan failure, 1 mortality resulted from intraoperative injury, and 6 deaths were a result of unknown cause after discharge from the hospital. Among these decedents, 2 patients survived to undergo a Fontan completion. Importantly, survival functions were not different between study groups as determined by the log-rank test (log-rank p 0.68) (Fig 1). Among CPB patients, 46 (63.8%) underwent Fontan completion at a median of 33 months, whereas 14 (43.8%) non-cpb patients underwent Fontan operations at a median of 32 months. Table 4. Incidence of Postoperative Events for Patients Undergoing Bidirectional Glenn Operations With (CPB) or Without (Non-CPB) Cardiopulmonary Bypass Support Overall (n 106) Non-CPB (n 34) CPB (n 72) Postoperative Event No. % No. % No. % p Total hospital length of stay (d) 7.5 (5 13) 9 (5 16) 7 (5 12) 0.23 Postoperative length of stay (d) 7.0 (5 12) 7 (5 12) 7 (5 11) 0.38 Operative mortality Any complication Systemic venous obstruction NA Reoperation (bleeding) Sternum left open Persistent neurologic deficit Infection (wound infection) Infection (mediastinitis) Pneumothorax Pleural effusion Pneumonia NA Prolonged mechanical ventilation ( 7 d) Need for reintubation Tracheostomy Phrenic nerve paralysis Cardiac arrest ECMO Arrhythmia Chylothorax CPB cardiopulmonary bypass; ECMO extracorporeal membrane oxygenation; NA not available.

5 168 LAPAR ET AL Ann Thorac Surg BIDIRECTIONAL GLENN AND CARDIOPULMONARY BYPASS 2012;94: Fig 1. Kaplan-Meier survival functions for patients undergoing BDG operations with (CBP) or without (non-cpb) cardiopulmonary bypass (CPB) support. (BDG bidirectional Glenn.) Comment The present study provides a contemporary analysis of outcomes after BDG procedures for staged palliation of functional single-ventricle congenital heart defects and demonstrates the effect of performance of a BDG procedure without the use of CPB. Based on these analyses, performance of a non-cpb BDG procedure is not associated with increased morbidity, operative mortality, or use of resources. More importantly, these data provide important short- and long-term survival data comparable to published BDG procedure literature. This study demonstrates that a BDG procedure is associated with high long-term survival and that survival after a BDG procedure is no different with or without the use of CPB. Although a paucity of reported outcomes for non-cpb BDG procedures exists, these findings support the use of this approach to surgical palliation for appropriately selected patients with single-ventricle anomalies. Outcomes after the performance of BDG procedures remain underreported in recent literature, and the principal findings of this report are in agreement with other reported series. In a recent 2008 report of outcomes and risk factors related to the performance of a BDG procedure, Kogon and colleagues [3] reported on their experience with 270 consecutive BDG operations between 2001 and Serving as 1 of the largest reports on BDG procedure outcomes, they documented a 0.7% mortality rate and 27.0% overall complication rate among patients with median hospital lengths of stay of 5 days. Other series have reported similar results in an attempt to define the optimal timing for stage II BDG procedure palliation of functionally univentricular hearts. In 1997, Reddy and colleagues [4] reported operative mortality of 4.8%, a reoperation rate of 17%, and 2-year actuarial survival of 86% in 42 infants less than 6 months of age, whereas Jaquiss and coworkers [5] later documented 0.0% operative morality, 96% 1-year Kaplan-Meier survival, and median hospitalizations of 8.0 and 6.0 days for infants less than 4 months and greater than 4 months of age undergoing a BDG procedure, respectively. These results are commensurate with those reported in the present series. In the present report, the documented overall operative mortality of 0.9%, composite complication rate of 16%, and reoperation for bleeding rate of 1.9% are favorable, and the overall median postoperative hospital length of stay (7.0 days) is comparable to the aforementioned series. Even more underreported are outcomes after performance of a BDG procedure without CPB. In a 2004 report of 20 patients undergoing BDG procedures, 19 patients had non-cpb BDG procedures, whereas 1 patient (5.0%) underwent intraoperative conversion to CPB as a result of hemodynamic instability secondary to arrhythmia [6]. In this report, operative mortality was 0.0%, and mean intensive care unit length of stay was days. In a separate report of 36 patients that year, Luo and colleagues [7] documented a similar mortality rate as well as the incidence of postoperative chylothorax in 1 patient (2.8%). Another more recent series included 22 patients undergoing non-cpb BDG procedures [8] and documented no operative mortalities or neurologic complications, with mean intensive care unit lengths of stay of days. The results of these 3 series are in agreement with the principal findings of the present study. In the present cohort of 34 patients who underwent non-cpb BDG procedures, excellent outcomes were achieved with no operative mortalities and low morbidity, including a 2.9% reoperation rate for bleeding and 0.0% rates of prolonged mechanical ventilation, need for reintubation, persistent neurologic deficits, arrhythmia, pleural effusion, and chylothorax. Moreover, these data extend those of other series to report on high 1-, 3-, and 5-year long-term survival. The performance of the BDG shunt without the use of CPB has been described using several different approaches. Originally described by Lamberti and colleagues [9] in 1990, the temporary use of a venoatrial shunt between the SVC and right atrium allowed for the avoidance of CPB support [9]. Alternatively, the use of a venopulmonary shunt has been described [10]. Still others have described a non-cpb approach without the use of temporary shunts [11]. In a series of 7 patients undergoing non-cpb BDG procedures without decompression shunts, Jahangiri and associates [11] argued that a temporary shunt may not be needed with cerebral perfusion pressures maintained at greater than or equal to 30 mm Hg through augmentation of systemic blood pressure using dopamine or alpha-agonist infusions, or both. Although such reports have documented no significant differences in neurologic sequelae after clamping the SVC without the use of a proximal shunt, standard venoatrial shunts were routinely used in all patients

6 Ann Thorac Surg LAPAR ET AL 2012;94: BIDIRECTIONAL GLENN AND CARDIOPULMONARY BYPASS 169 undergoing non-cpb BDG procedures in this series in an attempt to protect the brain from high upper body venous pressures, resulting in the absence of any persistent neurologic deficits (0.0%). In fact, clamping of the SVC, without use of a shunt, has been demonstrated to result in up to a 50% reduction in cerebral blood flow [12], and it has been shown to result in significant electroencephalographic changes after SVC clamping [13]. Thus the use of temporary shunts should be considered for the routine performance of non-cpb BDG procedures. Patient selection with respect to the performance of a BDG procedure without CPB in this series was critical to achieving successful outcomes. In the present series, the presence of a preexisting Sano right ventricle to pulmonary artery shunt (52.9%) was the most common surgical condition allowing the performance of non-cpb BDG procedures, followed by the presence of adequate forward pulmonary artery flow (17.6%) and the presence of a left B-T shunt (14.7%). Two (5.9%) patients with right B-T shunts were able to undergo non-cpb BDG procedures because of the presence of a double SVC. Two (5.9%) patients also had central shunts. One (2.9%) patient was able to tolerate a non-cpb BDG procedure because of adequate pulmonary artery flow through a PDA secondary to prostaglandin E1 administration. Other important factors to consider in the selection of patients for a BDG procedure without CPB include appropriate preoperative catheterization indices, including mean pulmonary artery pressure and pulmonary vascular resistance, surgeon experience to minimize SVC clamp times and upper body venous hypertension, patient age, and a low McGoon index [6]. At our institution, we generally adhere to standard catheterization guidelines, including end-diastolic pressure 10 to 15 mm Hg, pulmonary vascular resistance less than 2 Wood units/ m 2, mean pulmonary artery pressure less than 18 mm Hg, transpulmonary gradient less than 10 mm Hg, and no significant ventricular outflow obstruction ( 15 mm Hg). Importantly, the results of the catheterization do not impact our decision as to whether or not we perform an on- or off-pump BDG procedure. Select perioperative events in this series warrant further discussion. Among those patients meeting preoperative criteria for a non-cpb BDG procedure (n 38), 4 patients (10.5%) underwent intraoperative conversion to a BDG procedure with CPB support. One patient had a stroke after stage I Norwood palliation secondary to a hypercoaguable state and was found to have a diffusely small-caliber Sano shunt possibly due to extensive clot burden. Intraoperatively, this patient suffered from low blood oxygen saturation leading to bradycardia, requiring CPB support. One patient with a preoperative history of supraventricular tachycardia experienced similar cardiac arrhythmias during initial mediastinal dissection rendering her too hemodynamically unstable to proceed without CPB. A third patient with a right B-T shunt had undergone previous pulmonary banding resulting in forward pulmonary flow initially felt to be substantial enough to support an off-pump approach; however the degree of forward flow was insufficient intraoperatively to complete the operation without CPB. The fourth patient underwent a bilateral BDG procedure and became hemodynamically unstable after completion of the right Glenn anastomosis for unknown reasons and ultimately required conversion to CPB support. Importantly, except for this patient, none of the aforementioned conversions were performed emergently. Careful consideration of the circumstances surrounding these conversions is important to emphasize the role of preoperative catheterization and associated risk-factor analysis in the preoperative planning and patient selection processes for the performance of a BDG procedure without CPB. The present study has select limitations and important clinical implications. First, the influence of patient selection bias and surgeon choice for the performance of a BDG procedure with or without CPB must be considered in any retrospective study design. Similarly the impact of changing practice patterns and surgeon experience over time with the performance of non-cpb BDG procedures may have impacted the reported results by biasing them toward the null hypothesis. Changes in CPB pump and surgical technology as well as advances in anesthesia and postoperative management over time are also unable to be accounted for in these analyses. Importantly, however, these limitations are not specific to the present study and impact any longitudinal study reporting on outcomes after a BDG procedure. Conversely, this study has important clinical implications because it provides a recent analysis not only of short-term operative outcomes but also of long-term survival data to complement existing literature on the performance of BDG procedures. Moreover, the extensive reporting of postoperative outcomes in this series provides important point estimates of various event frequencies for future therapeutic comparisons. In conclusion, the results of the present study demonstrate that the BDG procedure can be performed with no significant differences in operative mortality, morbidity, or use of resources either with or without the use of CPB support. Long-term survival after a BDG procedure is high and is similar with both strategies. Thus the performance of off-pump BDG procedures should be considered a safe alternative to the conventional use of CPB for appropriately selected patients with candidate preoperative anatomy and catheterization results. The authors would like to thank Rebecca Chamberlain, RN, for her assistance in data collection and management related to this study. This study was supported by Award Number 2T32HL A1 (ILK) from the National Heart, Lung, and Blood Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute or the National Institutes of Health. References 1. Karl TR, Stellin G. Early Italian contributions to cavopulmonary shunt surgery. Ann Thorac Surg 1999;67:1175.

7 170 LAPAR ET AL Ann Thorac Surg BIDIRECTIONAL GLENN AND CARDIOPULMONARY BYPASS 2012;94: Glenn WW. Circulatory bypass of the right side of the heart. IV. Shunt between superior vena cava and distal right pulmonary artery; report of clinical application. N Engl J Med 1958;259: Kogon BE, Plattner C, Leong T, Simsic J, Kirshbom PM, Kanter KR. The bidirectional Glenn operation: a risk factor analysis for morbidity and mortality. J Thorac Cardiovasc Surg 2008;136: Reddy VM, McElhinney DB, Moore P, Haas GS, Hanley FL. Outcomes after bidirectional cavopulmonary shunt in infants less than 6 months old. J Am Coll Cardiol 1997;29: Jaquiss RD, Ghanayem NS, Hoffman GM, et al. Early cavopulmonary anastomosis in very young infants after the Norwood procedure: impact on oxygenation, resource utilization, and mortality. J Thorac Cardiovasc Surg 2004;127: Liu J, Lu Y, Chen H, Shi Z, Su Z, Ding W. Bidirectional Glenn procedure without cardiopulmonary bypass. Ann Thorac Surg 2004;77: Luo XJ, Yan J, Wu QY, Yang KM, Xu JP, Liu YL. Clinical application of bidirectional Glenn shunt with off-pump technique. Asian Cardiovasc Thorac Ann 2004;12: Hussain ST, Bhan A, Sapra S, Juneja R, Das S, Sharma S. The bidirectional cavopulmonary (Glenn) shunt without cardiopulmonary bypass: is it a safe option? Interact Cardiovasc Thorac Surg 2007;6: Lamberti JJ, Spicer RL, Waldman JD, et al. The bidirectional cavopulmonary shunt. J Thorac Cardiovasc Surg 1990;100: 22 9; discussion Murthy KS, Coelho R, Naik SK, Punnoose A, Thomas W, Cherian KM. Novel techniques of bidirectional Glenn shunt without cardiopulmonary bypass. Ann Thorac Surg 1999;67: Jahangiri M, Keogh B, Shinebourne EA, Lincoln C. Should the bidirectional Glenn procedure be performed through a thoracotomy without cardiopulmonary bypass? J Thorac Cardiovasc Surg 1999;118: Rodriguez RA, Weerasena NA, Cornel G. Should the bidirectional Glenn procedure be better performed through the support of cardiopulmonary bypass? J Thorac Cardiovasc Surg 2000;119: Rodriguez RA, Cornel G, Semelhago L, Splinter WM, Weerasena NA. Cerebral effects in superior vena caval cannula obstruction: the role of brain monitoring. Ann Thorac Surg 1997;64: DISCUSSION DR HENRY WALTERS (Detroit, MI): Dr LaPar, I congratulate you and your colleagues on an excellent presentation and thank you for providing me the manuscript in advance of this meeting. I certainly agree that avoiding unnecessary cardiopulmonary bypass is a laudable goal. As you point out in your manuscript, the following factors, including the retrospective study design, the obvious effect of era, the surgeon preference, the patient clinical condition, and the anatomical constraints that precluded consideration of the non-bypass approach all could, indeed, have made your comparison groups different in important ways that could, in turn, have an important effect upon your results. Nonetheless, those are the liabilities of any retrospective study. I believe that I am correct in saying, though, that the nonbypass technique is not yet common practice in our specialty. In congenital heart surgery we sadly do not often have or, perhaps, take the opportunity to perform prospective, randomized studies. Often, a given technique becomes so engrained in our practices that we unwittingly pass the point of no return before which we might have considered a prospective trial. With the non-bypass bidirectional Glenn I do not think that we have passed that point. So, now that you have generated excellent retrospective data that this non-bypass approach has a comparable safety profile to the traditional approach, I would first ask you what are the contraindications to performing the nonbypass technique and how would you feel about prospectively randomizing a group of patients to bypass and non-bypass, all of whom are suitable for the non-bypass approach, and report those comparative results to us at a future meeting of this society? In our institution we feel that it is important to suture closed the pulmonary valve leaflets, as part of the process of dividing the main pulmonary artery, to reduce the chances of clot formation in the pulmonary artery stump and subsequent embolization. At your institution how do you deal with the patient with forward flow at the level of the pulmonary valve and does it differ between the bypass and non-bypass techniques? In your title you mention long-term follow-up and I am glad of that because the results of the bidirectional Glenn live on up to and well beyond the point of converting the patient to a full cavopulmonary connection. One of the important things that we look for just prior to the completion Fontan, for example, is the pulmonary artery anatomy. Did you compare, between the bypass and non-bypass groups, the anatomical features that could reasonably be attributed to the BDG at the time of their pre-fontan evaluation? DR LAPAR: Thank you very much for your comments and your questions. First, with respect to contraindications for the use of cardiopulmonary bypass, for the most part, it is general practice at our institution that we approach almost every patient requiring a Glenn with the potential for an off-pump approach, unless they have a clear indication for the use of cardiopulmonary bypass, say, eg, if they need an atrial septectomy or there is an indication for immediate arch reconstruction, but by and large, we prefer now an off-pump technique if we can do so. With respect to the need for a prospective study and whether or not I feel that is feasible, I certainly think that that s something that is warranted now. Although there is currently a paucity of literature looking at this topic, and even though they are generally small series, I think there is enough out there now to make the case for more rigorous clinical investigation. This is certainly something that we have talked about at our institution. Your third question was related to what exactly? DR WALTERS: Let s say that there is forward flow at the level of the pulmonary valve. I suppose you could leave that as a supplemental source of pulmonary blood flow (in addition to the cavopulmonary connection); some centers report doing that on a regular basis. Some, alternatively, I suppose just ligate the pulmonary artery. We, eg, will first deliberately suture closed the pulmonary valve leaflets and then close the stump, because we feel pretty certain that this is otherwise a source of embolism. DR LAPAR: We have had experience with both practices, and I think in general it is safe to say that we prefer a subtotal ligation of the proximal main PA, allowing a little bit of forward flow through there for 2 reasons: first, because we believe it decreases the risk of embolism and clot formation there, as well as it also allows circulation of hepatic factors, which we feel are protective.

8 Ann Thorac Surg LAPAR ET AL 2012;94: BIDIRECTIONAL GLENN AND CARDIOPULMONARY BYPASS 171 In terms of your last question, although we did specifically look at the incidence or degree of PA stenosis pre-fontan, in my extensive review of the patient charts I can t recall any cases that required angioplasty or stent placement before they got completion Fontans. DR WALTERS: Thank you. I have 1 follow-up question. I didn t hear that the presence of a right-sided shunt is a contraindication to this technique. I thought, from the manuscript, that it might, indeed, be a relative contraindication. In your bypass technique, eg, you do say that you routinely use the shunt insertion site for the site of implantation of the SVC. DR LAPAR: You are correct, a right-sided BT shunt would be a contraindication to an off-pump technique. DR WALTERS: Are you able to sort of wiggle your way around the shunt if there is a right-sided shunt? DR LAPAR: No. DR WALTERS: Thank you. DR CONSTANTINE MAVROUDIS (Orlando, FL): This was a very nice presentation. Did you have enough comparative data to show if either 1 group or the other had a higher lactate level or more acidosis? I don t think I saw it in your presentation. DR LAPAR: Thank you very much. We didn t collect serum lactate levels for inclusion in this study. DR MAVROUDIS: I hope you can get it before you publish the paper. It will add credence to your hypothesis. DR LAPAR: None of the patients undergoing off-pump Glenn in this series were hemodynamically unstable, as those who were had conversion to an on-pump approach, so I m not sure that I would expect that much variation in lactate levels between our study groups. DR ERLE AUSTIN (Louisville, KY): That was a nice presentation. When the anatomy allows it, we also try to do the Glenn off pump, but I take some exception to your conclusion that implies that you can apply an off-pump technique to every patient. You just indicated that if you had done a Norwood you would have used a right-sided Blalock-Taussig shunt. In that case I don t think an off-pump technique is safe. First off, you must assure oxygenation, so if you are off pump there must be pulmonary blood flow, and you also must decompress the venous system in the head. Well, you can decompress the head off-pump with cannulas, but you will not have pulmonary blood flow when you change your right-sided BT shunt to a bidirectional Glenn anastomosis. I think you should revise your conclusion statement. I do think an off-pump Glenn may be advisable if you can do it, but I don t think it can be done in every situation where a bidirectional Glenn is the goal. DR LAPAR: I completely agree with you that an off-pump approach can be applied to appropriately selected patients. DR JORGE SALAZAR (Jackson, MS): What kind of cerebral monitoring do you utilize during your bidirectional Glenn s, whether on or off bypass? DR LAPAR: We do use the NIRS intraoperatively. However, we are not necessarily sure at this point of its utility during performance of an off-pump BDG. It is something that we are looking very closely at, though. DR SALAZAR: I believe there are an increasing number of congenital heart surgeons whose practice is significantly influenced by the NIRS, in particular for superior vena caval cannulation. With regard to performing the Glenn operation off bypass, we want to avoid compromising the cerebral drainage, let alone changing the operation that we would otherwise perform if done on pump.