The first-ever successful superior vena cava (SVC) to. Half a Century s Experience With the Superior Cavopulmonary (Classic Glenn) Shunt

Transcription

1 Half a Century s Experience With the Superior Cavopulmonary (Classic Glenn) Shunt Riad Abou Zahr, MD, Paul M. Kirshbom, MD, Gary S. Kopf, MD, Sandeep Sainathan, MD, Margaret M. Steele, PA, Robert W. Elder, MD, and Mohsen Karimi, MD Department of Pediatrics, Section of Pediatric Cardiology, and Department of Surgery, Section of Pediatric Cardiothoracic Surgery, Yale University School of Medicine, New Haven, Connecticut Background. Five decades after its introduction, the Glenn shunt remains an integral step for patients undergoing single-ventricle palliation. We performed a longitudinal follow-up of the original cohort of patients who underwent Glenn shunt. Methods. We performed a retrospective study of the original cohort of patients who underwent Glenn shunt at Yale between 1958 and Electronic medical records and chart review up to current era were used to collect data. Results. Ninety-one patients underwent a Glenn shunt at an average age of 6.6 ± 2.5 years, of which 89 were classic Glenn shunts. Median overall survival was 43 years (range, 2 to 56; 95% confidence interval [CI], 39.5 to 46.5) while median survival from the Glenn shunt was 31.4 years (range, 0 to 45; 95% CI, 23.9 to 38.9). Forty-six patients died, 7 in the early postoperative period and 39 late deaths. Twenty-six patients were lost to follow-up. Nineteen patients remain alive with active clinical follow-up, 6 of whom still live with their classic Glenn shunt without conversion to bidirectional Glenn. Twenty-six patients (31%) developed pulmonary arteriovenous fistula with 11 patients (42%) requiring coil embolization. No patient developed thrombosis of the Glenn shunt. There were a total of 28 patients who developed arrhythmias, mostly in the tricuspid atresia group (n [ 16), with the majority being atrial tachyarrhythmias (48%). Sixteen patients required permanent pacemaker placement for sinus node dysfunction. Conclusions. The Glenn shunt continues to provide excellent staged palliation in single-ventricle patients and a bridge to two-ventricle repair. Arrhythmias and pulmonary arteriovenous fistulas were common among single-ventricle cohort. Quality-of-life evaluation of the surviving patients would be an important outcome measure for future investigation. (Ann Thorac Surg 2016;101:177 82) Ó 2016 by The Society of Thoracic Surgeons The first-ever successful superior vena cava (SVC) to right pulmonary artery (PA) shunt (classic Glenn shunt) was performed in 1958 at Yale by Dr William Glenn [1]. Five and a half decades later, the Glenn shunt remains an integral stage in palliation for single-ventricle congenital heart diseases. The superior cavopulmonary shunt introduced the concept of partial circulatory bypass of the right heart, which evolved into complete right heart bypass, namely the Fontan procedure. The Glenn shunt provides a stable means of providing low-pressure pulmonary blood flow in cyanotic congenital heart disease and volume unloads a single ventricle as compared to systemic shunts. Over time, the classic Glenn shunt has evolved from a unidirectional to a bidirectional cavopulmonary anastomosis. This provided less distortion to the right PA and allowed for some blood flow to the left lung. Despite this evolution, there are a number of adult Accepted for publication Aug 17, Address correspondence to Dr Karimi, Yale University School of Medicine, Department of Surgery, Section of Cardiac Surgery, Pediatric Cardiothoracic Surgery, 330 Cedar St, Boardman Bldg 204, New Haven, CT 06510; mohsen.karimi@yale.edu. congenital heart disease patients who retain classic Glenn shunts to date. Medium- and long-term clinical outcomes of the original experience was previously reported [2, 3]. The aim of this study was to evaluate the very long-term outcome of the original cohort of Glenn shunt patients. Patients and Methods We performed a single-center retrospective review of the original cohort of patients who underwent an SVC-PA (classic Glenn) shunt at Yale University. Between 1958 and 1988, Glenn shunts were performed on 91 patients. The most recent clinical and surgical medical records were reviewed through Data collection included demographics, age and date of Glenn shunt, surgical and interventional procedures, arrhythmias, complications, and mortality. The study was approved by Yale University Institutional Review Board. Statistical analysis was performed using SPSS version 20.0 for Windows (SPSS, Chicago, IL). Data was described as mean SD or median and range with confidence intervals for normally and nonnormally Ó 2016 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier

2 178 ZAHR ET AL Ann Thorac Surg CLASSIC GLENN OUTCOMES OVER FIVE DECADES 2016;101: distributed continuous variables, respectively. Analyses of categorical variables were done using Pearson s chisquare test or Fisher s exact test when appropriate. Continuous variables with a normal distribution and continuous variables with nonnormal distributions were analyzed using the Student t test and Mann-Whitney U test, respectively. A p value less than 0.05 was considered statistically significant. Results A total of 91 patients underwent Glenn shunts between 1958 and Eighty-nine patients underwent a classic Glenn shunt consisting of an SVC to RPA anastomosis in an end-to-end fashion while only 2 patients underwent a bidirectional Glenn shunt (BDG). Most surgeries were performed through an anterolateral thoracotomy while eight cases were done through a median sternotomy. The mean age at the time of surgery was years. Tetralogy of Fallot patients underwent surgery at an older age (mean, years) compared to the rest of the cohort. Of the 91 patients, 65 (71%) had available clinical follow-up while 26 (29%) were lost to follow-up. Of the lost to follow group, clinical information was available for the first 3 to 4 decades. Table 1 shows the distribution of diagnoses with the corresponding patient status and mean age at Glenn. Of note a third of the patients had tricuspid atresia and half had functionally univentricular hearts. A total of 46 (50%) patients died; 7 (7%) were early hospital deaths within 30 days of surgery. There were 39 (43%) late deaths; 8 were secondary to surgical complications while 31 were attributed to cardiac disease (Table 2). Arrhythmias, heart failure, and Fontan failure or thrombosis were the most common causes of late death. Median overall survival was 43 years (95% confidence interval [CI], 39.5 to 46.5) while median survival from the time of the Glenn shunt was 31.4 years (95% CI, 23.9 to 38.9) as demonstrated in Figure 1. The 30-, 40-, and 50-year survival rates for all patients were 52%, 34%, and 21%, respectively. Univariate analysis between the groups of survivors and deceased was performed. No statistically significant differences were found based on gender, history of pulmonary arteriovenous fistulas (AVF), placement of Table 1. Original Cohort of Classic Glenn Shunt axillary AVF, diagnosis, or age at Glenn. Kaplan-Meier survival curve for single- and two-ventricle groups demonstrated favorable survival of single-ventricle cohort in the first 40 years of follow-up, which was not statistically significant (Fig 2). For the single-ventricle group, median overall survival from the time of the Glenn shunt was 32.4 years (95% CI, 26.9 to 37.9) and for the two-ventricle group was 28.6 years (95% CI, 11.7 to 45.5). No survival difference was found in the periods before 1970, during the 1970s, or in the 1980s. Nineteen patients are alive with active clinical follow-up, 11 (58%) of whom had tricuspid atresia. The mean overall survival of this group was years while their mean survival post Glenn is years. Their median oxygen saturation is 92% with a range of 85% to 99%. Six patients continue to have a functional classic Glenn shunt without revision. These patients are now in their fifth or sixth decades of life (Table 3). One has tricuspid atresia with a Fontan conduit to the left PA. The second patient has tricuspid atresia with a Potts shunt. The third has tetralogy of Fallot with a right ventricle that pumps to the left PA through a recently placed Melody valve. The fourth patient has transposition of the great arteries and a hemi-mustard baffle. The fifth patient has transposition of the great arteries, declined Fontan palliation and has a left Glenn with a Waterston shunt. The last patient has Ebstein s anomaly and underwent tricuspid valve replacement. Approximately two-thirds of patients in this cohort underwent cardiac catheterizations. Thirty-one percent of patients were found to have pulmonary arteriovenous malformation. These were more prevalent among univentricular patients. Forty-two percent required coil embolization in the catheterization laboratory at the side of the Glenn shunt (Table 4). Among patients with available follow-up to the modern era, 28 (44%) developed arrhythmias. Most of these patients had tricuspid atresia (n ¼ 16; 57%) with the majority (48%) having atrial tachyarrhythmias. Sixteen (57%) patients developed sinus node dysfunction requiring pacemaker implantation. Prior to the Glenn shunt, a total of 59 patients underwent palliative procedures in the form of Blalock-Taussig shunt (n ¼ 25), central shunt (n ¼ 14), PA banding (n ¼ 12), and atrial septectomy (n ¼ 8) with 32 patients (35%) requiring no prior palliation. Most patients (95%) Diagnosis No. Age at Glenn (years) Alive Deceased Unknown TA (12) 16 (17) 6 (7) TOF (2) 7 (8) 6 (7) TGA/VSD/PS (2) 12 (13) 5 (5) PA (2) 1 (1) 2 (2) Complex SV (1) 7 (8) 7 (8) Ebstein (1) 3 (3) 0 Total (21) 46 (50) 26 (29) Values are mean SD or n (%). PA ¼ pulmonary atresia; SV ¼ single ventricle; TA ¼ tricuspid atresia; TGA/VSD/PS ¼ transposition of great arteries/ventricular septal defect/pulmonary stenosis; TOF ¼ tetralogy of Fallot.

3 Ann Thorac Surg ZAHR ET AL 2016;101: CLASSIC GLENN OUTCOMES OVER FIVE DECADES 179 Table 2. Causes of Early and Late Deaths Etiology of Death Early Late Surgical 7 8 Arrhythmias 7 Cardiac failure 4 Pulmonary complications 2 Stroke 2 Fontan complications 5 Sepsis 1 Unknown 10 Total 7 39 Total early and late deaths 46 with functionally univentricular hearts needed additional sources of pulmonary blood flow. The surgical procedures after the Glenn shunt were Fontan palliation (n ¼ 35), Blalock-Taussig shunt (n ¼ 14), axillary AVF (n ¼ 8), biventricular repair (n ¼ 8), central shunts (n ¼ 3), tricuspid valve repair or replacement (n ¼ 2), and heart transplantation (n ¼ 2). Many classic Glenn shunts were converted to a bidirectional shunt at the time of the Fontan. Fontan procedures were classic atriopulmonary anastomosis in the single-ventricle group (n ¼ 15) with 10 patients subsequently undergoing conversion to lateral tunnel or extracardiac Fontan. There were 8 patients in the single-ventricle group who underwent lateral tunnel or extracardiac Fontan from the original Glenn or bidirectional Glenn shunt. The majority of two-ventricle group underwent a biventricular repair (Table 5). On average, univentricular patients required more surgical procedures after their Glenn shunt compared to biventricular patients (surgical load, 1.4 vs 1.0). Eight patients underwent axillary AVF to augment pulmonary blood flow. Only 1 of those patients is alive. None of the patients developed thrombosis within the Glenn shunt itself. Fig 2. Kaplan-Meier curve for overall survival after the classic Glenn for single- and two-ventricle cohorts. Comment This study represents the longest follow-up of the oldest cohort of Glenn patients, including the original patients operated on by Dr William Glenn himself [1 3]. The first patient in this series had complex single ventricle and underwent Glenn shunt at 7 years of age followed by a Blalock-Taussig shunt and later Fontan palliation. The Glenn shunt was originally performed as a unidirectional end-to-end SVC to right PA anastomosis. The main purpose of this shunt was to establish a low-pressure source of pulmonary blood flow and to volume unload a functionally single-ventricle heart [4]. It has evolved to a bidirectional end-to-side SVC to the right PA connection. The bidirectional Glenn provides blood flow to the left lung and makes subsequent third-stage palliation (Fontan procedure) less complicated [4, 5]. Currently, the bidirectional Glenn shunt is the standard second stage palliation for many cyanotic congenital heart diseases including tricuspid atresia, complex single ventricles, and severely unbalanced atrioventricular canals. A classic shunt may still be performed in exceedingly rare cases such as a single-ventricle patient with one functional Table 3. Patients Alive With Classic Glenn in 2014 Diagnosis Glenn Date Age at Glenn (years) Age (years) Tricuspid atresia Tricuspid atresia Tetralogy of Fallot TGA TGA Ebstein s anomaly Fig 1. Kaplan-Meier survival curve from the classic Glenn shunt. TGA ¼ transposition of great arteries.

4 180 ZAHR ET AL Ann Thorac Surg CLASSIC GLENN OUTCOMES OVER FIVE DECADES 2016;101: Table 4. Coil Embolization of Pulmonary Arteriovenous Fistulas Following Classic Glenn Shunt Diagnosis No. AVF Coil Embolization TA (13) 5 (19) TOF 15 4 (5) 2 (8) TGA/VSD/PS 19 2 (2) 2 (8) PA 5 1 (1) 0 Complex SV 15 8 (10) 2 (8) Ebstein Total (31) a 11 (13) a ; (42) b a Percentage of patients excluding early deaths. b Percentage of total with AVF. Values are n (%). AVF ¼ arteriovenous fistulas; PA ¼ pulmonary atresia; SV ¼ single ventricle; TA ¼ tricuspid atresia; TGA/VSD/PS ¼ transposition of great arteries/ventricular septal defect/pulmonary stenosis; TOF ¼ tetralogy of Fallot. lung. In the late 1980s, more Fontan operations were performed on younger children without previous Glenn palliations owing to the fact that short-term outcomes were similar [3]. However, long-term outcomes of performing a Fontan procedure without a preceding Glenn shunt were not encouraging and this approach has therefore been mostly abandoned [6, 7]. Currently, bidirectional Glenn shunts are routinely performed via a median sternotomy. This provides better exposure for the surgeon to anastomose the SVC closer to the branch PA bifurcation, less compression of the right lung, one less scar, and lower risk of scoliosis. While this procedure may be done off cardiopulmonary bypass, performing it on bypass is technically easier and allows for resection of any residual atrial septum. Interestingly in this study, the survival rate of the single-ventricle group was better than the biventricular group for the first four decades. This finding was previously noted and continued to be valid in this report [3]. Multiple factors may explain this result. The twoventricle group was older at the time of surgery and half of the cases were performed before 1968 when no corrective surgeries were available for many congenital heart defects. Moreover, most of the two-ventricle patients (88%) underwent Glenn procedure before 1970 while less than half of the single-ventricle patients (43%) underwent Glenn procedure in the same period of time. This raised the possibility of worse outcomes in the first decade of the experience compared to subsequent years, which may be partially explained by an improved learning curve in surgical era. Nonetheless, when accounting for an era effect over three different decades, no significant difference was found. Patients with Glenn shunts placed in the first few weeks of life had very poor outcome [3]. These findings support the current physiological understanding and practice of performing Glenn shunts after pulmonary vascular resistance had significantly dropped, typically sometime after 2 months of age. Nicolas and colleagues [8] observed that performing Glenn shunts before 120 days of life carried significantly higher mortality. PA distortion and pulmonary vascular resistance in excess of 3 Wood units m 2 were found to be risk factors for mortality in one study [9]. Patients who had systemic to pulmonary shunts for prolonged periods of time may not be good candidates for a Glenn shunt and should undergo hemodynamic evaluation in the catheterization laboratory prior to shunt placement. When pulmonary hypertension is found, the pulmonary vascular bed should be tested with vasodilators for reactivity. If pulmonary vascular resistance drops below 3.5 Wood units m 2, some evidence suggests that it is safe to perform a Glenn shunt [10]. In patients with low risk for pulmonary hypertension and no known branch PA stenosis, pre-glenn evaluation with cardiac magnetic resonance imaging was shown to be an acceptable Table 5. Operative Procedures Following Classic Glenn Shunt in Patients With Available Follow-Up Diagnosis Operation TA (n ¼ 21) TOF (n ¼ 6) TGA/VSD/ PS (n ¼ 3) PA (n ¼ 4) Complex SV (n ¼ 7) Ebstein (n ¼ 1) A-P Fontan Fontan 8 a a 2 a 0 Conversion LT/EC Fontan Two ventricle Transplant Axillary AVF Waterston shunt TV replacement No intervention a Fontan conversion to either LT/EC Fontan from classic Fontan. A-P ¼ aortopulmonary; AVF ¼ arteriovenous fistula; LT/EC ¼ lateral tunnel/extracardiac; PA ¼ pulmonary atresia; SV ¼ single ventricle; TA ¼ tricuspid atresia; TGA/VSD/PS ¼ transposition of great arteries/ventricular septal defect/pulmonary stenosis; TOF ¼ tetralogy of Fallot; TV ¼ tricuspid valve.

5 Ann Thorac Surg ZAHR ET AL 2016;101: CLASSIC GLENN OUTCOMES OVER FIVE DECADES 181 alternative to cardiac catheterization both in the shortterm post-glenn and up to 8 years afterward [11, 12]. Patients with pulmonary AVF presented with worsening cyanosis and lower oxygen saturations. The rate of angiographic AVF formation remained similar to our previously reported incidence of 30% [3]. Most of the fistulas formed in the right lower lobe of the lung at the side of the Glenn shunt. In our early experience, length of time from the Glenn and the presence of pulmonary hypertension in the contralateral lung had significant correlation with fistula formation [3]. Recent data, however, suggests that absence of hepatic venous effluent, which likely contains a biochemical agent, is involved in formation of pulmonary AVF [13]. This is supported by the fact that collaterals and fistulas regress after inclusion of hepatic venous blood in the pulmonary circulation [14]. Interestingly, increased right to left intrapulmonary shunting may be a universal phenomenon in patients with cavopulmonary anastomoses compared to healthy controls [15]. This intrapulmonary shunting is higher in patients whose Glenn is the only source of pulmonary blood flow. Although, right to left intrapulmonary shunting may occur in all patients with a Glenn shunt at a microvascular level, our data shows that it manifests clinically and angiographically in only a third of patients. In addition to pulmonary AVF, systemic to pulmonary arterial collaterals frequently occur in patients with Glenn shunts with an incidence of 59% to 71% [16]. In our cohort, pulmonary AVF were of more clinical relevance because they caused worsening symptoms such as cyanosis and had potential to cause paradoxical emboli and strokes. It has been suggested that the presence of systemic to pulmonary arterial collaterals may cause prolonged effusions after Fontan palliation. This, however, has not been a consistent observation [16]. Although in situ thrombosis in Fontan patients is well known, in our series no Glenn thrombosis occurred. One explanation is that a Fontan conduit is made of prosthetic material with a higher likelihood of thrombosis. In addition, due to its length, the resistance to flow is higher. Alternatively, gravity or some other postural factor may enhance flow through the Glenn shunt but impedes that through the Fontan conduit. Ravn and colleagues [17], found an increase in platelet aggregation and hence reactivity after Glenn shunts. The reason for this observation is unclear. Furthermore, they demonstrated a significant reduction in the naturally occurring anticoagulants protein C, protein S, and antithrombin III in this patient population. All of these abnormalities contribute to an increased risk of thrombosis. This data supports the rationale for using antiplatelet agents such as aspirin to prevent clot formation. Most of the patients in our cohort, however, did not receive any form of antiplatelet or anticoagulant agents after the Glenn shunt. It was only after Fontan palliation that aspirin was used. Anticoagulation with warfarin was added for those who developed Fontan thrombosis or strokes. Right axillary AVF were performed in 8 patients because of worsening hypoxia. Oxygen saturations improved in the short term, but these patients developed complications such as arm edema and ecchymosis requiring fistula ligation. Overall, this subset of patients did poorly with no long-term survival advantage. The average age of the Glenn shunt among the 19 survivors is about 40 years. Some of these patients continue to be functional and able to perform many of their activities of daily living. Study Limitations First, this study has the inherent limitation of retrospective nature, which is prone to administrative errors of documentation. Second, some patients were completely lost to follow-up for the past two decades representing less than a third of the original cohort. Conclusion In summary, we report half a century s experience with the classic Glenn shunt. This procedure has indeed withstood the test of time. In its bidirectional form, it continues to provide excellent staged palliation in singleventricle patients and has allowed for long-term survival in many patients with complex congenital heart disease. The clinical usefulness of the contemporary bidirectional Glenn shunt has been well demonstrated over the years as a stage palliation in single-ventricle pathway, one-anda-half ventricle physiology, and bridge to two-ventricle repair. Although some of the patients experienced AVF, arrhythmia, and multiple interventions during the course of their lives, some remain alive with reasonable functional status into adulthood from their original operation. A comprehensive functional status and quality-of-life study would be an important adjunct to this work to compare with general population. In 1984 Dr Glenn wrote, Someone once said that the clinical usefulness of a new operation cannot be properly assessed for 25 years. After 25 years experience with the clinical application of cava-pulmonary artery anastomosis, I would extend that period to 50 years. Certainly, an epitaph for the cava-pulmonary artery shunt is not appropriate at this time [18]. The authors wish to thank Yale University, Yale New Haven Hospital and the patients for facilitating this work. References 1. 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: Mathur M, Glenn WW. Long-term evaluation of cavapulmonary artery anastomosis. Surgery 1973;74: Kopf GS, Laks H, Stansel HC, Hellenbrand WE, Kleinman CS, Talner NS. Thirty-year follow-up of superior vena cava-pulmonary artery (Glenn) shunts. J Thorac Cardiovasc Surg 1990;100:662 70; discussion Hopkins RA, Armstrong BE, Serwer GA, Peterson RJ, Oldham HN Jr. Physiological rationale for a bidirectional cavopulmonary shunt. A versatile complement to the Fontan principle. J Thorac Cardiovasc Surg 1985;90:391 8.

6 182 ZAHR ET AL Ann Thorac Surg CLASSIC GLENN OUTCOMES OVER FIVE DECADES 2016;101: Bridges ND, Jonas RA, Mayer JE, Flanagan MF, Keane JF, Castaneda AR. Bidirectional cavopulmonary anastomosis as interim palliation for high-risk Fontan candidates. Early results. Circulation 1990;82:IV Jonas RA. Indications and timing for the bidirectional Glenn shunt versus the fenestrated Fontan circulation. J Thorac Cardiovasc Surg 1994;108: Duncan BW, Desai S. Pulmonary arteriovenous malformations after cavopulmonary anastomosis. Ann Thorac Surg 2003;76: Nicolas RT, Hills C, Moller JH, Huddleston CB, Johnson MC. Early outcome after Glenn shunt and Fontan palliation and the impact of operation during viral respiratory season: analysis of a 19-year multi-institutional experience. Ann Thorac Surg 2005;79:613 7; discussion Pridjian AK, Mendelsohn AM, Lupinetti FM, et al. Usefulness of the bidirectional Glenn procedure as staged reconstruction for the functional single ventricle. Am J Cardiol 1993;71: Hussain A, Arfi AM, Hussamuddin M, et al. Comparative outcome of bidirectional Glenn shunt in patients with pulmonary vascular resistance > or ¼ 3.5 woods units versus < 3.5 woods units. Am J Cardiol 2008;102: Brown DW, Gauvreau K, Powell AJ, et al. Cardiac magnetic resonance versus routine cardiac catheterization before bidirectional glenn anastomosis in infants with functional single ventricle: a prospective randomized trial. Circulation 2007;116: Brown DW, Gauvreau K, Powell AJ, et al. Cardiac magnetic resonance versus routine cardiac catheterization before bidirectional Glenn anastomosis: long-term follow-up of a prospective randomized trial. J Thorac Cardiovasc Surg 2013;146: Knight WB, Mee RB. A cure for pulmonary arteriovenous fistulas? Ann Thorac Surg 1995;59: Shah MJ, Rychik J, Fogel MA, Murphy JD, Jacobs ML. Pulmonary AV malformations after superior cavopulmonary connection: resolution after inclusion of hepatic veins in the pulmonary circulation. Ann Thorac Surg 1997;63: Vettukattil JJ, Slavik Z, Lamb RK, et al. Intrapulmonary arteriovenous shunting may be a universal phenomenon in patients with the superior cavopulmonary anastomosis: a radionuclide study. Heart 2000;83: McElhinney DB, Reddy VM, Tworetzky W, Petrossian E, Hanley FL, Moore P. Incidence and implications of systemic to pulmonary collaterals after bidirectional cavopulmonary anastomosis. Ann Thorac Surg 2000;69: Ravn HB, Hjortdal VE, Stenbog EV, et al. Increased platelet reactivity and significant changes in coagulation markers after cavopulmonary connection. Heart 2001;85: Glenn W, Hellenbrand WE, Henisz A, et al. Superior vena cava-right pulmonary artery anastomosis: present status., in Obstructive lesions of the right heart. 1984, Baltimore University Press: Baltimore. p