- Email: [email protected]
Right Ventricular Unloading for Heart Failure Related to Ebstein Malformation
Right Ventricular Unloading for Heart Failure Related to Ebstein Malformation Vijayakumar Raju, MS, MCh, Joseph A. Dearani, MD, Harold M. Burkhart, MD, Martha Grogan, MD, MPH, Sabrina D. Phillips, MD, Naser Ammash, MD, Roxann P. Pike, MD, Jonathan N. Johnson, MD, and Patrick W. O’Leary, MD Divisions of Cardiovascular Surgery, Cardiovascular Diseases, Pediatric Cardiology, and Department of Anesthesiology, Mayo Clinic, Rochester, Minnesota
Background. Patients with Ebstein malformation (EM) and severe RV dilatation and dysfunction have increased operative risk. Early results with right ventricular unloading utilizing the bidirectional cavopulmonary shunt (BCPS) during repair of EM have been encouraging. We report our experience of the 1.5-ventricle repair strategy for this difficult group of patients. Methods. Between July 1999 and January 2013, 62 patients with severe EM underwent BCPS at the time of tricuspid repair. Median age was 21.5 years (range, 9 months to 57 years), 51.6% were male, and 72.5% were children. Severe RV dilatation was present in all patients; severe RV dysfunction was present in 72.5% (n [ 45) and moderate to severe RV dysfunction in 22.5% (n [ 14). Mean RV systolic pressure was 32.7 ± 0.7 mm Hg and mean PA pressure was 15.6 ± 2.1 mm Hg. Mean preoperative left ventricular ejection fraction (LVEF) was 0.536 ± 0.071; it was less than 40% in 10 patients (16.1%). Fatigue or reduced exercise tolerance was the most common symptom (61.2%), followed by cyanosis (27.4%). New York Heart Association class III/IV heart failure was present in 43 patients (69.3%) preoperatively and 20 patients (32.2%) were initially referred for heart transplant evaluation. Prior EM surgery occurred in 35.4% (n [ 22; 8 prior valve repair, 8 prior valve replacement, BlalockTaussig shunt in 4, atrial septal defect (ASD) closure in 2). Results. Tricuspid repair was performed in 51.6% (n [ 32, 5 had re-repair). Bioprosthetic valve replacement was performed in 48.4% (n [ 30, 8 had rereplacement). The
BCPS was a planned procedure in 53 patients (85.5%) because of RV dysfunction; BCPS was added after unsuccessful weaning from bypass in 7 (11.2%), and in the early postoperative period due to hemodynamic instability in 2. Concomitant procedures included ASD closure in 48.3%, maze in 38.7%, and mitral valve repair in 6.4%. Postoperative extracorporeal membrane oxygenation support was needed in 8 patients. Delayed chest closure was performed in 25.8%. Early mortality was 1.6% (n [ 1). Mean mechanical ventilation time was 69.7 hours. Mean intensive care unit and hospital stays were 5.4 ± 3.5 and 10.7 ± 3.5 days, respectively. Mean follow-up was 3.6 ± 2.6 years (maximum, 12.8 years). Patients (n [ 10) with low preoperative LVEF (0.362 ± 0.035) improved to 0.517 ± 0.042 postoperatively (p [ 0.001). There was 1 late death in a patient with cystic fibrosis. Late reintervention was needed in 5 patients (8%). Late follow-up was available in 95% (n [ 59); all were acyanotic and 88% were in New York Heart Association functional class I/II. Conclusions. Concomitant BCPS is a useful adjunct in repair of advanced EM with severe RV dilatation and dysfunction. Operation can be performed with low early mortality. Intermediate-term survival and quality of life is good to excellent, and transplantation can be delayed or avoided in the majority.
E
of the bidirectional cavopulmonary shunt (BCPS) or “1.5 ventricle repair” improves hemodynamics by RV unloading and increasing LV preload. The BCPS was first applied in EM in 1980 [1] and since then other studies have shown improved outcome with its use [2–4]. We have previously reported our early results with 1.5 ventricle repair for EM in a small number of patients (n ¼ 14) [2]. The objective of this paper is to review early and midterm outcome of the BCPS in a larger number of high-risk patients with EM.
bstein malformation (EM) with right ventricular (RV) failure remains a challenge with high operative risk. Establishing tricuspid valve competence in EM increases RV afterload and increases the chance of RV failure, which in turn reduces the left ventricular (LV) preload and also produces the pancake effect on the LV. These physiologic alterations contribute to higher operative morbidity and mortality in EM. The addition
Accepted for publication March 4, 2014. Presented at the Sixtieth Annual Meeting of the Southern Thoracic Surgical Association, Scottsdale, AZ, Oct 30–Nov 2, 2013. Address correspondence to Dr Dearani, Division of Cardiovascular Surgery, Mayo Clinic, 200 First St SW, Rochester, MN 55905; e-mail: [email protected].
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2014;-:-–-) Ó 2014 by The Society of Thoracic Surgeons
Patients and Methods The Mayo Clinic Institutional Review Board approved this study. Between July 1999 and December 2012, 319 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.03.009
2
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
patients underwent surgery for EM. Among these, 62 (19.4%) with EM and severe RV dilatation or RV dysfunction underwent EM repair with concomitant BCPS; this group forms the cohort for this review. Of this 62 patient cohort, 45 (72.5%) were children 18 or less years of age; 8 were less than1 year old, 14 were age 1 to 5 years, 7 age 6 to 10 years, and 16 were 11 to 18 years old. Among the entire 62 patients, 20 patients (32.2%) were initially referred for the heart transplantation. The preoperative demographics are listed in Table 1. Fatigue and poor exercise tolerance were the most common presenting symptoms (61.2%), followed by cyanosis (27%). Right-sided heart failure (ascites) was present in 22.5%. Preoperative renal insufficiency or hepatic dysfunction was present in 2 patients (3.2%) each. Cardiomegaly on chest X-ray was present in all; mean cardiothoracic ratio was 0.8 0.04. All patients had echocardiographic evidence of severe RV dilatation; assessment of RV function and tricuspid regurgitation is shown in Table 1. Assessments of RV function included the entire anatomic RV (both atrialized and functional zones). We used this approach as postoperatively both zones are beyond the functional tricuspid valve and we feel that it reduces the variability between examiners. Parameters assessed included the following: fractional area change of the RV cavity in the 4-chamber and short axis views (both subjectively and quantitatively); systolic descent of the tricuspid annular plane systolic excursion; myocardial performance index; and a subjective, visual assessment of systolic RV myocardial thickening. These parameters were combined by the interpreting cardiologist in order to classify the degree of RV dysfunction as mild, moderate, or severe as outlined in Table 1. No patient was judged to have normal RV function. Grading of the degree of tricuspid regurgitation reflected the amount of systolic reversed flow crossing the Table 1. Patient Demographics Demographic
Number
Variable Median age Male Cardiac rhythm Sinus Atrial fibrillation Paroxysmal atrial fibrillation WPW syndrome Tricuspid regurgitation grade Moderate Moderately severe Severe RV dysfunction Mild Moderate Severe RV ¼ right ventricular;
Patient number (n ¼ 62) 21.5 years (range, 9 months–57 years) n ¼ 32 (51.6%) 35 14 7 6
(56.4%) (22.5%) (11.2%) (9.6%)
2 (3.2%) 5 (8%) 55 (88.7%) 3 (4.8%) 14 (22.5%) 45 (72.5%)
WPW ¼ Wolff-Parkinson-White.
Ann Thorac Surg 2014;-:-–-
plane of the functional tricuspid orifice (from the functional to the atrialized RV). A combination of anatomic leaflet coaptation (or the lack thereof), regurgitant jet width, Doppler regurgitant signal strength, and systolic distortion of hepatic venous Doppler flow profile was used by the interpreting cardiologist to classify regurgitant severity. In cases with a single regurgitant orifice, the width of the regurgitant jet at its origin (the vena contracta) was used to define jet width. In the mature patient, a vena contracta width of less than 3 mm is associated with mild regurgitation, widths of 8 to 10 mm represented severe transvalvar regurgitation. In young patients, vena contracta diameters which are less than 10% of the normal annulus were considered mild. Jet origins measuring more than 25% to 30% of the normal annular dimension were classified as severe. When a combination of multiple jets existed, the examiner mentally combined the size of the regurgitant orifices in order to determine the overall grade. Preoperative hemodynamic variables included the following: mean RV systolic pressure 32.7 0.7 mm Hg; mean pulmonary artery pressure 15.6 2.1 mm Hg; and left ventricular ejection fraction (LVEF) 0.536 0.071. Biventricular dysfunction was seen in 10 patients (16.1%); mean LVEF was 0.362 0.035. Preoperative magnetic resonance imaging (MRI) was available in 21 (33.8%) patients. The MRI derived mean RVEF, RV end systolic and diastolic volumes, indexed to body surface area were 0.295 0.325, and 267.7 185.3 and 388.3 133.6 mL/m2, respectively. Twenty-two patients (35.4%) had previous surgery. Eight patients had recurrent tricuspid regurgitation after previous valve repair (7 had repair elsewhere) and 8 had bioprosthetic tricuspid valve failure (2 had multiple previous valve replacements); 4 had modified BlalockTaussig shunt and 2 had isolated atrial septal defect (ASD) closure. The average time interval between the previous tricuspid operations and the current operation was 2.6 years (repair) and 10.3 years with prior tricuspid replacement. All patients with prior tricuspid repair had severe recurrent tricuspid regurgitation. In those patients with prior replacement, the mean prosthetic gradient was 13.2 0.7mm Hg. Severe bioprosthetic regurgitation was present in 4 patients, moderate regurgitation in 3, and mild regurgitation in 1. All patients underwent tricuspid repair or replacement and BCPS; concomitant procedures are shown in Table 2. Patients requiring reoperation for EM with BCPS performed elsewhere, and EM with pulmonary atresia with a RV-to-pulmonary artery conduit were excluded from the study.
Operative Technique and Perioperative Management Median sternotomy with aortic and bicaval cannulation was utilized; the superior vena cava was cannulated percutaneously using the Seldinger technique through the right internal jugular vein in older children and adults. This facilitated construction of the BCPS and it also provided a source of venous access for cardiopulmonary bypass in the event that peripheral cannulation
Ann Thorac Surg 2014;-:-–-
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
Table 2. Operative Details Surgical procedure Tricuspid repair (n ¼ 32) Monocusp repair (n ¼ 4) Cone repair (n ¼ 28) Tricuspid replacement (n ¼ 30) Indication for BCPS (n ¼ 62)
Concomitant procedures Right atrial reduction Right ventricular plication ASD/PFO closure Right-sided maze Biatrial maze Ablation WPW Mitral valve repair Epicardial ICD
Number (N¼62) TV repair with BCPS, n ¼ 27 TV re-repair with BCPS, n ¼ 5 (bilateral BCPS in one) TVR with BCPS, n ¼ 20 Redo TVR with BCPS, n ¼ 8 TVR with bilateral BCPS, n ¼ 2 Planned procedure n ¼ 53 (85.4%) Unable to separate from bypass n ¼ 7 (11.2%) Early postoperative n ¼ 2 (3.2%) n, patients n ¼ 53 n ¼ 52 n ¼ 30 n ¼ 24 n¼8 n¼4 n¼4 n¼4
ASD ¼ atrial septal defect; BCPS ¼ bidirectional cavopulmonary shunt; ICD ¼ implantable cardioverter defibrillator; PFO ¼ patent foramen ovale; TV ¼ tricuspid valve; TVR ¼ tricuspid valve replacement; WPW ¼ Wolff-Parkinson-White.
was needed during resternotomy. The conduct of operation included normothermic or mildly hypothermic (34 C) bypass and cold blood antegrade cardioplegia with subsequent doses at 20-minute intervals during the cross-clamp period. Iced slush saline was applied topically to the RV. Tricuspid repair, RV plication, and closure of intracardiac shunts were performed with cardioplegic arrest. In patients undergoing tricuspid replacement, the cross-clamp was removed after intracardiac shunt closure and RV plication. Valve suture placement and seating of the prosthesis was performed with the heart beating and perfused to insure maintenance of sinus rhythm and to minimize myocardial ischemic time. Valve repair techniques are described in detail in prior publications and were monocusp version in the early portion of this series [5–7] and the cone reconstruction in the recent era (2007 to present) [8–10]. In general, anatomy favorable for valve repair included a large anterior leaflet and some (even diminutive) septal leaflet. Complete surgical delamination and mobilization of leaflet tissue was performed so that 360 degrees of rotated, reconstructed leaflet tissue was anchored at the true tricuspid annulus. When tricuspid valve replacement was performed, a porcine bioprosthesis was utilized; technique of insertion has been previously described [11, 12]. Valve anatomy that made anatomic cone repair unsuccessful included significantly muscularized or deformed leaflet tissue that resulted in reduced pliability and poor coaptation, or absent septal leaflet.
3
The BCPS was performed after intracardiac repair and right reduction atrioplasty was complete. The anastomosis was continuous when somatic growth was complete and the anterior suture line was interrupted when it was not. The azygos vein was routinely ligated. All patients were separated from bypass on epinephrine 0.05 mcg $ kg1 min1, milrinone 0.25 $ mcg $ kg1 $ min1. Vasopressin 0.02 to 0.04 mcg $ kg1 $ min1 was utilized selectively to maintain a mean blood pressure of greater than 50 mm Hg. Inhaled nitric oxide was utilized to offset the pulmonary vasoconstrictive effects of inotropic support; this was applied when epinephrine was greater than 0.05 mcg $ kg1 $ min1 or when vasopressin was used. Temporary atrial pacing at 110 to 120 was done routinely for the first 24 hours if the intrinsic heart rate was less than 100. Extracorporeal membrane oxygenation (ECMO) support was accomplished with central aortic and right atrial cannulation. Inadequate venous return was not an issue despite the presence of BCPS. The ECMO flows were maintained with a cardiac index of 1.5 to 2.0 L $ min1 $ m2. Discharge medications included ace-inhibiter and beta-blocker therapy. Sildenafil was used for 4 weeks postoperatively in most patients and routinely in those who underwent valve repair and lived at high altitude. Amiodarone therapy was utilized for 3 to 6 months in those patients who had RV plication or had episodes of atrial or ventricular tachyarrhythmias postoperatively. Warfarin therapy was used for tricuspid replacement and was maintained for 3 to 6 months and discontinued when there was echocardiographic confirmation that prosthetic valve leaflets were moving normally. Aspirin therapy (81 mg) was used indefinitely in patients undergoing valve replacement.
Results Early Results Tricuspid valve repair (n ¼ 32; 51.6%) or porcine bioprosthetic tricuspid replacement (n ¼ 30; 48.4%) was performed with BCPS in all patients (Table 2). Intraoperative and early postoperative data are shown in Table 3. An intraaortic balloon pump was utilized in 4 adults for unstable hemodynamics on optimal inotropic support when separating from bypass; all had LV dysfunction preoperatively. Extracorporeal membrane oxygenation was utilized in 8 patients (12.9%); 5 with valve repair and 3 with valve replacement. Median age of the patients requiring ECMO was 2 years (range, 9 months to 41 years). In the valve repair group, 1 patient (a 15-year-old child) was transferred from an outside institution on ECMO after failure of valve repair and cardiogenic shock; she underwent successful valve re-repair with BCPS and was weaned from ECMO 11 days later. Among 3 patients requiring ECMO who underwent tricuspid replacement, 2 needed it in the operating room for failure to wean from bypass and 1 needed ECMO on the second postoperative
4
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
Ann Thorac Surg 2014;-:-–-
Table 3. Perioperative Details Variable CPB time (minutes) Cross-clamp time (minutes) ECMO IABP Open chest Mean duration of ventilation (hours) Mean ICU stay (days) Mean hospital stay (days) CPB ¼ cardiopulmonary bypass; unit.
Entire Population (n ¼ 62)
Valve Repair (n ¼ 32)
Valve Replacement (n ¼ 30)
112.5 17.7 54.1 34.6 8 patients (12.9%) (mean 5.8 days) 4 patients (6.4%) 16 patients (25.8%) 69.7 84.8 5.4 3.5 10.7 3.5
113.5 26.3 74.9 22.1 5 patients (mean 6.1 days) None 5 patients 71.8 93.3 5.8 3.5 11 4.2
111.4 38.8 31.9 11.3 3 patients (mean 7 days) 4 patients 11 patients 67.6 22.6 5.1 2.1 10.5 2.8
ECMO ¼ extracorporeal membrane oxygenation;
day for poor hemodynamics. Seven of the 8 patients requiring ECMO support were successfully weaned with a mean duration of 5.8 days. Early postoperative morbidity and mortality are shown in Table 4. There was 1 early death (1.6%). This patient had valve replacement (27-mm porcine) and developed worsening of hemodynamics on the second postoperative day. Extracorporeal membrane oxygenation was placed with conversion to an extracardiac Fontan; this patient died 11 days later from multiorgan failure. Early reoperation occurred in 7 patients (11.2%). Four patients initially repaired needed early reoperation for recurrent severe tricuspid regurgitation. Three patients underwent successful re-repair and 1 underwent valve replacement after unsuccessful re-repair. Three patients needed early reoperation after valve replacement. One patient had BCPS added on the second postoperative day for severe RV dysfunction and 2 patients needed surgery for bioprosthetic valve thrombosis (1 had valve re-replacement and 1 had thrombus removal at the time of ECMO separation). Dismissal echocardiography in the valve repair group (n ¼ 32) showed no or mild tricuspid regurgitation 28 patients (87.5%), moderate tricuspid regurgitation in 3 (9.3%), and severe tricuspid regurgitation in 1 (3.1%);
IABP ¼ intraaortic balloon pump;
ICU ¼ intensive care
this patient underwent early reoperation and valve replacement after unsuccessful valve re-repair. In the bioprosthetic valve replacement group (n ¼ 30), all patients had no regurgitation and the mean gradient was 6.1 2.1 mm Hg.
Late Results Complete follow-up was available in 95% (n ¼ 59) with a mean follow-up of 3.6 2.6 years (maximum, 12.8 years). All patients were acyanotic and 88% of patients (n ¼ 52) were in New York Heart Association functional class I/II. Recurrent atrial or ventricular arrhythmias were common during the follow-up period (n ¼ 7; 11.2%). Among the 20 patients who were initially referred for heart transplant, only 3 were listed for transplantation postoperatively. One patient (57-years old) underwent transplant 6 years after operation (valve replacement with BCPS), and 1 patient is currently listed for heart transplant. The third patient died 3 years after surgery awaiting transplantation. This patient was a 32-year-old woman with cystic fibrosis who had tricuspid valve replacement and BCPS. She died from multiorgan failure while awaiting combined heart and liver transplant. The remaining 17 patients were able to avoid transplantation.
Table 4. Early Morbidity and Mortality Variable Morbidity Arrhythmias (atrial and ventricular) Valve thrombosis Stroke Reintubation Pneumonia Heart block Chylous Pericardium Mortality (early) Early reoperation
Entire Population (n ¼ 62) 7 2 2 2 2 1 1 1 7
(11.2%) (3.2%) (3.2%) (3.2%) (3.2%) (1.6%) (1.6%) (1.6%) (11.2%)
Valve Repair (n ¼ 32)
Valve Replacement (n ¼ 30)
3 (9.3%) None 2 (6.2%) 2 None None None None 4 (12.5 %)
4 (13.3%) 2 (6.6%) None None 2 1 1 1 3 (10%)
Ann Thorac Surg 2014;-:-–-
Overall survival after1.5 ventricle repair at 1, 3, and 5 years was 98.3%, 98.3%, and 95.4%, respectively (Fig 1). Late reintervention occurred in 5 (8%) patients, including 2 in the valve repair group; 1 patient needed tricuspid valve replacement for recurrent tricuspid regurgitation 3 years after valve repair and the other patient needed implantable cardioverter defibrillator placement for recurrent ventricular arrhythmias. In the valve replacement group, 1 patient underwent heart transplant 6 years after operation, 1 patient had placement of an epicardial pacing system for heart block, and 1 needed stenting of a stenotic BCPS connection 2 years after operation. Overall freedom from reintervention at 1, 3, and 5 years was 96.6%, 96.6%, and 74.3%, respectively (Fig 2). Echocardiography at latest follow-up in the valve repair group showed mild tricuspid regurgitation in 84.3% and mild-to-moderate regurgitation in 6.2%. In the tricuspid valve replacement group (n ¼ 30), 8 patients (26.6%) had a prosthesis-related abnormality. Moderate-to-severe bioprosthetic stenosis (mean gradient >10 mm Hg) or moderate prosthetic regurgitation was seen in 4 patients each. In those patients with depressed LVEF preoperatively (n ¼ 10), LVEF increased from 0.362 0.03 to 0.517 0.042 (p ¼ 0.001). Chest radiography at latest follow-up showed a significant reduction in cardiothoracic ratio from 0.8 0.04 to 0.59 0.03 (p ¼ 0.001).
Comment Tricuspid valve repair in EM has been challenging and the recent application of cone reconstruction has facilitated successful repair for most patients. A right ventricle dysfunction has been identified as a risk factor for early mortality in EM [13]. Symptoms are often inconspicuous until adulthood, particularly when the atrial septum is intact or the patent foramen ovale is small because cyanosis or arterial desaturation is minimized or absent. This can result in late referral to surgery. The most common symptom is fatigue or decreased exercise tolerance; frank right-sided heart failure is uncommon unless marked RV dysfunction is present. Progressive RV
Fig 1. Kaplan-Meier survival curve after 1.5 ventricle repair.
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
5
Fig 2. Overall freedom from reintervention after 1.5 ventricle repair.
enlargement and dysfunction is very common in EM resulting in increased operative risk. This review demonstrates that operation for EM in the setting of advanced RV enlargement or dysfunction can be safely performed when the BCPS is applied to EM repair. In this advanced group, factors that facilitate the decision to offer operation are the absence of overt signs of heart failure on physical exam, the presence of normal LV EF, and valve anatomy favorable for repair [9, 10]. The BCPS is helpful in the setting of EM because it reduces the volume load to the enlarged dysfunctional RV and it improves preload to the chronically under filled and compressed LV. However, a BCPS is not applicable to all patients with EM and RV dysfunction. In the adult patient, or when there is depressed LV function, a preoperative cardiac catheterization is performed. The following hemodynamics are required for BCPS: LVEDP less than12 mm Hg; transpulmonary gradient less than10 mm Hg; and mean pulmonary artery pressure less than18 mm Hg. Overall indications and contraindications for BCPS are shown in Table 5 [2, 3, 9, 10]. The decision to repair versus replace the tricuspid valve is critical in this advanced EM group. While valve repair is preferred, it is often reserved for children and those patients with favorable valve anatomy, which may not be the case in the advanced, adult patient group. Complex intracardiac repair (tricuspid valve repair, RV plication, and ASD closure) with massive RV and annular enlargement can result in a long cross-clamp time, which can further worsen the already dysfunctional right ventricle. The decision to proceed with valve replacement in this setting may be more prudent as the cross-clamp time is shorter by limiting it to patent foramen ovale or ASD closure and RV plication. This allows the majority of the procedure to be performed with the heart beating and perfused. Furthermore, it assures tricuspid competence at the end of the operation and avoids the need for repeated bypass runs to revise an unacceptable valve repair and then subsequent replacement. Finally, the decision to proceed with replacement should be made early and preferably before the BCPS has been performed. This is important as the prosthesis is intentionally downsized because the
6
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
Table 5. Indications and Contraindications for Bidirectional Cavopulmonary Shunt Heart failure indications Cardiothoracic ratio > 65% RVEDV > 250 mL/m2 RVEF < 0.25 Small, squashed, D-shaped left ventricle Postrepair RAP:LAP > 1.5:1 Postrepair low cardiac output: persistent metabolic acidosis, low urine output, increasing creatinine, poor peripheral perfusion, low mean arterial blood pressure < 50 mm Hg Non-heart failure indications Reduce tension on complex tricuspid repair Postrepair tricuspid stenosis (mean gradient > 8–10 mm Hg) Contraindications Mean PA pressure > 20 mm Hg Pulmonary arteriolar resistance > 4 Woods units LVEDP or LA pressure >12 mm Hg Significant pulmonary artery hypoplasia LA ¼ left atrial; LVEDP ¼ left ventricular end-diastolic pressure; PA ¼ pulmonary artery; RAP:LAP ¼ right atrial pressure to left atrial pressure; RVEDV ¼ right ventricular end-diastolic volume; RVEF ¼ right ventricular ejection fraction.
amount of blood flow across the tricuspid valve is reduced by a factor of 30% to 40% with the BCPS. In our experience it is preferable to have a gradient of 5 to 6 mm Hg across the prosthesis so that all leaflets open and close. When the prosthesis is too large (easy to do given marked annular dilatation), a gradient of only 1 to 2 mm Hg usually results in 1 or 2 of the prosthetic leaflets not opening normally. This results in the development of thrombus and accelerated bioprosthetic deterioration. This finding was the stimulus for warfarin therapy for 3 to 6 months postoperatively. Mechanical valves should not be used in this setting as there can be poor disc motion due to low right atrial pressure and poor RV function that predisposes to valve thrombosis even in the face of therapeutic anticoagulation. The role of atrial septal fenestration can also be considered. As mentioned above, the BCPS significantly unloads the RV and provides preload to the LV. This combination is particularly helpful when the hemodynamic situation is poor. Furthermore, the right-to-left shunting across the atrial level shunt can result in profound arterial desaturation in the postoperative period. In patients with LV hypertrophy or dysfunction it is important to determine left atrial and LV end diastolic pressures preoperatively as they may be elevated and would then result in left-to-right shunting across the atrial level shunt. Finally, in the adult EM population, a persistent atrial level shunt is associated with increased risk of paradoxical embolism [14]. In our practice, subtotal closure of the ASD is reserved for neonates and children with lesser degrees of RV dysfunction. Limitations to BCPS include reduced right atrial or RV access for electrophysiologic intervention, mild facial suffusion, or development of upper torso venous
Ann Thorac Surg 2014;-:-–-
collaterals or pulmonary arteriovenous fistulae. Only 1 patient (1.6%) in our series developed a stenosis of the BCPS anastomosis that was relieved with percutaneous stenting. This low incidence of problems related to BCPS in EM has been demonstrated by others [15, 16]. We had a low threshold for delayed sternal closure and a low threshold for temporary ECMO support. Our preference for ECMO over RV assist device therapy is based on our extensive ECMO experience with other high-risk postcardiotomy patients. The assurance of adequate oxygenation and the ability to rapidly wean inotropic agents on ECMO support has been the rule in our experience. We generally begin with ECMO in the immediate postoperative period, achieve aggressive diuresis, and plan ECMO separation 4 to 7 days later. If ECMO separation is unsuccessful, we make a transition to RV assist device if mechanical support is needed for a longer duration; fortunately, this has rarely been the case. The role of MRI in EM is important. In general, we prefer it in all patients preoperatively and postoperatively. The MRI was not done in many of the younger children because of the need for general anesthesia and many of the older patients had hemodynamic cardiac catheterization performed preoperatively. At late follow-up, arrhythmias (atrial and ventricular) were the most common reason for hospitalization. This emphasizes the importance of a long-term surveillance protocol for late arrhythmias that includes routine exercise testing at 6 months postoperatively and periodic Holter monitoring to determine need for arrhythmia therapy. While most are atrial tachyarrhythmias (flutter and fibrillation), ventricular arrhythmias are also important. The decision to offer conventional operation (1.5 ventricle repair) in this high-risk group is difficult when LV dysfunction is present. Our decision to offer surgery when LVEF was below normal was based on the presence of a small LV cavity and normal anterior and lateral LV free wall motion. Specifically, the ventricular septal motion abnormality was the predominant explanation for LV dysfunction, which we thought would improve after RV decompression with the BCPS. We would not advise operation with severe LV dysfunction or a dilated, dysfunctional LV; these patients would be referred to transplantation.
Limitations of the Study We did not have a protocol for application of the BCPS in the early part of this review. As we learned from our own experience and others [3], indications for BCPS evolved and are outlined above. Objective analysis of RV size and function at late follow-up with MRI would be preferred and was limited to the minority of this cohort. Longer follow-up is required to examine late outcome and to see if transplantation can be further delayed in this high-risk group.
Conclusions Concomitant BCPS is a useful adjunct in EM repair with severe RV dilatation or dysfunction. Operation can be
Ann Thorac Surg 2014;-:-–-
performed safely but postoperative care is prolonged and morbidity is increased. The LV function improves in those with low EF preoperatively. Intermediate-term survival and quality of life is good to excellent, and transplantation can be delayed or avoided in the majority.
References 1. Marcelletti C, D€ uren DR, Schuilenburg RM, Becker AE. Fontan’s operation for Ebstein anomaly. J Thorac Cardiovasc Surg 1980;79:63–6. 2. Quinonez LG, Dearani JA, Puga FJ, et al. Results of the 1.5ventricle repair for Ebstein anomaly and the failing right ventricle. J Thorac Cardiovasc Surg 2007;133:1303–10. 3. Malhotra SP, Petrossian E, Reddy VM, et al. Selective right ventricular unloading and novel technical concepts in Ebstein’s anomaly. Ann Thorac Surg 2009;88:1975–81. 4. Chauvaud S, Fuzellier JF, Berrebi A, et al. Bi-directional cavopulmonary shunt associated with ventriculo and valvuloplasty in Ebstein’s anomaly: benefits in high risk patients. Eur J Cardiothorac Surg 1998;13:514–9. 5. Danielson GK, Maloney J, Devloo R. Surgical repair of Ebstein’s anomaly. Mayo Clin Proc 1979;54:185–92. 6. Dearani JA, Danielson GK. Tricuspid valve repair for Ebstein’s anomaly. Op Tech Thorac Cardiovasc Surg 2003;8: 188–92. 7. Dearani JA, O’Leary PW, Danielson GK. Surgical treatment of Ebstein’s malformation: state of the art 2006. Cardiol Young 2006;16(Suppl 3):12–20.
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
7
8. Dearani JA, Bacha E, da Silva JP. Cone reconstruction of tricuspid valve for Ebstein’s anomaly: anatomic repair. Oper Tech Thoracic Cardiovasc Surg 2008;13:109–25. 9. Dearani JA, Said SM, O’Leary PW, et al. Anatomic repair of Ebstein’s malformation: lessons learned with cone reconstruction. Ann Thorac Surg 2013;95:220–6. 10. Dearani JA, Said SM, Burkhart HM, Pike RB, O’Leary PW, Cetta F. Strategies for tricuspid re-repair in Ebstein malformation using the cone technique. Ann Thorac Surg 2013;96: 202–8. 11. Dearani JA, Danielson GK. Ebstein’s anomaly. In: Sellke FW, del Nido PJ, Swanson SW, eds. Sabiston & Spencer Surgery of the Chest. Philadelphia, PA: Elsevier; 2005:2223–35. 12. Kiziltan HT, Theodoro DA, Warnes CA, O’Leary PW, Anderson BJ, Danielson GK. Late results of bioprosthetic tricuspid valve replacement for Ebstein’s anomaly. Ann Thorac Surg 1998;66:1539–45. 13. Brown ML, Dearani JA, Danielson GK, et al. The outcomes of operations for 539 patients with Ebstein anomaly. J Thorac Cardiovasc Surg 2008;135:1120–36.e1-7. 14. Attenhofer Jost CH, Connolly HM, Dearani JA, Edwards WD, Danielson GK. Ebstein’s anomaly. Circulation 2007;115: 277–85. 15. Kim S, Al-Radi O, Mark KF, et al. Superior vena cava to pulmonary artery anastomosis as an adjunct to biventricular repair: 38-year follow-up. Ann Thorac Surg 2009;87: 1475–83. 16. Liu J, Qiu L, Zhu Z, Chen H, Hong H. Cone reconstruction of the tricuspid valve in Ebstein anomaly with or without one and a half ventricle repair. J Thorac Cardiovasc Surg 2011;141:1178–83.
DISCUSSION DR EMMETT MCKENZIE (Houston, TX): That was very nicely presented, Dr Raju. Thank you. Thank you also for getting me the manuscript. I was able the review that before the meeting. You and your colleagues are to be congratulated for outstanding results in a very sick group of patients with very advanced Ebstein’s disease. This experience that you present today is only the latest iteration from the very large series at Mayo Clinic that has been so helpful in our understanding of Ebstein’s anomaly. So thank you again for putting that together. I have a couple of questions. You note that a third of the patients included in this report had been referred initially for transplant. I would submit, however, that it is quite common for patients with congenital heart disease to be referred for a transplant and in fact to have reparable residual heart disease. The fact that the majority of the patients experienced durable relief of their symptoms following operation suggests that the right heart dysfunction, not necessarily the right ventricle but the global right heart dysfunction, is actually treatable and reversible. In an earlier report from your colleagues at the Mayo Clinic, low cardiac output was the main predictor of early mortality. And the fact that seven of the eight patients that required ECMO in this series recovered and survived suggests again that the low cardiac output and the right heart dysfunction resolved and resolved quickly. So my question is, is it possible that highly selected use of a right ventricular assist device or some form of temporary mechanical right heart support would achieve the same result without the long-term consequences of a bidirectional Glenn shunt? DR RAJU: Thank you, Dr McKenzie, for your kind remarks. Ebstein’s anomaly is a disease of the myocardium; basically it is an RV [right ventricular] myopathy. The abnormality with the leaflets are secondary to the myopathy. We had a limited number of
patients who had a postoperative MRI [magnetic resonance imaging], and in those patients we observed reduction in RV size and some improvement of the RV function. Since this is a disease of the myocardium and results in a myopathy, RV function improvement may be limited. This is the reason we opted to proceed with RV unloading with the bidirectional cavopulmonary shunt. DR MCKENZIE: My other question relates to the perioperative management. In this group, 25% of the patients had delayed sternal closure and the time to extubation was three days, on average. There is fairly compelling evidence that with a Fontan operation early extubation is beneficial, and these patients have no right ventricle. Even patients undergoing Fontan conversion, a much more lengthy operation in a much sicker group of patients, can benefit from early extubation; if not in the OR [operating room], in the first couple of hours after surgery. Is the decision to delay the sternal closure related to the size of the heart, pulmonary hypoplasia, or high airway pressures? And do you think that this group of patients would benefit from early extubation, also? DR RAJU: Dr Dearani’s approach here is anybody who needs bidirectional cavopulmonary shunt as an unplanned or a salvage procedure, his preference is to leave the chest open in order to avoid further hemodynamic instability. Conversely, if they had RV unloading as a planned procedure, we are less likely to leave the chest open; almost all were closed primarily at the time of initial repair when the bidirectional cavopulmonary shunt was planned. DR MCKENZIE: And one last quick question. What are your recommendations for timing of surgery, recognizing that many of these patients aren’t symptomatic and aren’t diagnosed until
8
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
late in their disease? You mentioned that the right ventricle suffers from a myopathy, and with our getting over the learning curve of the cone operation and it seemingly being applicable to more variants, are you recommending earlier repair now? DR RAJU: The decision regarding the timing of the surgery is determined by the experience of the surgeon and the surgical team. If the surgeon is able to produce a reliable and reproducible valve repair, we recommend surgery between two to five years of age in the asymptomatic child with severe tricuspid regurgitation and RV enlargement. The learning curve for an anatomic cone-type repair is steep, so if the surgeon, or the team, is not confident or comfortable in repairing and producing a good result, then we think it is more prudent to follow the traditional indications for surgery for tricuspid regurgitation. This includes the presence of symptoms, progressive RV enlargement, onset of atrial tachyarrhythmias, and cyanosis. Ebstein anomaly is a serious disease and a reproducible, durable repair is difficult to obtain when experience is limited. For these reasons, consideration should be given to referral to a center with documented experience so outcome can be optimized, particularly when the patient is asymptomatic. Thank you. DR TARA KARAMLOU (San Francisco, CA): Two quick questions. One, did you notice a detriment in those patients who you performed the cavopulmonary shunts as a salvage operation versus the patients that it was done a priori? And then my second question is: Do you have long-term exercise data that can support your contention that cardiac output is improved? At the CHSS [Congenital Heart Surgeons’ Society], we published a paper looking at exercise data in patients with pulmonary atresia with intact septum, showing that patients with a 1.5 ventricle repair, including a bidirectional Glenn, had a distinct benefit compared to the 1 or 2 ventricle patients in that the unloading actually recovered their cardiac output and they did better. Thanks. DR RAJU: I think Dr Dearani can respond. I can answer your second question. We did notice about 18% of the patients in this group had a good functional capacity at follow-up and their exercise tolerance improved, and they are in class 1 and 2. Doctor Dearani? DR DEARANI: What was the first part of the question?
Ann Thorac Surg 2014;-:-–-
DR KARAMLOU: So, Joe, the first question was, is there a downside to doing this as a salvage? We know in congenital heart surgery many times if you make the “wrong decision” up front that you pay a price down the line. And, two, do you have any formal cardiopulmonary exercise testing to support your contention about cardiac output, especially in the setting of chronotropic incompetence? DR DEARANI: There were no downsides. Actually, in the early part of this series it was resorting to the bidirectional Glenn that was the eye-opening finding that the patients turned around almost immediately. There was no detriment from a hemodynamic standpoint in the long term. In the older patients, there could be some concern about the development of prominent varicosities in the shoulders and the neck, facial suffusion, and, of course, venous access for arrhythmia analysis down the road may be more challenging from the femoral vein below. The majority of patients have had late exercise testing, and there is an improvement in exercise performance. It is due in part to increased cardiac output, and also probably related to the elimination of right to left shunting with ASD [atrial septal defect] closure. Even isolated ASD closure with elimination of arterial desaturation results in improvement in exercise tolerance, too. So it’s probably a combination of both things. DR ANDREW C. FIORE (St. Louis, MO): Tell us your thoughts about the atrial septal defect. How do you manage it in these patients? DR DEARANI: We believe the bidirectional Glenn is a better pathway when you have severe advanced right-sided heart failure. The atrial septal fenestration does not unload the right ventricle, it just provides a little bit of extra preload to the left ventricle. So when you have a compressed, pancaked left ventricle, the atrial septal fenestration does not unload and decompress the RV Furthermore, cyanosis is the indication for surgery in many of these patients, and so intentionally leaving a fenestrated atrial septum may result in an element of desaturation, which is often the indication for surgery. And finally, in the adult patients an atrial septal fenestration in Ebstein’s anomaly is a risk factor for paradoxical embolism. We believe the role of routine subtotal closure of the ASD includes Ebstein repair in infancy and in young children who have lesser degrees of RV dysfunction that is not quite bad enough to proceed with a Glenn.
Background. Patients with Ebstein malformation (EM) and severe RV dilatation and dysfunction have increased operative risk. Early results with right ventricular unloading utilizing the bidirectional cavopulmonary shunt (BCPS) during repair of EM have been encouraging. We report our experience of the 1.5-ventricle repair strategy for this difficult group of patients. Methods. Between July 1999 and January 2013, 62 patients with severe EM underwent BCPS at the time of tricuspid repair. Median age was 21.5 years (range, 9 months to 57 years), 51.6% were male, and 72.5% were children. Severe RV dilatation was present in all patients; severe RV dysfunction was present in 72.5% (n [ 45) and moderate to severe RV dysfunction in 22.5% (n [ 14). Mean RV systolic pressure was 32.7 ± 0.7 mm Hg and mean PA pressure was 15.6 ± 2.1 mm Hg. Mean preoperative left ventricular ejection fraction (LVEF) was 0.536 ± 0.071; it was less than 40% in 10 patients (16.1%). Fatigue or reduced exercise tolerance was the most common symptom (61.2%), followed by cyanosis (27.4%). New York Heart Association class III/IV heart failure was present in 43 patients (69.3%) preoperatively and 20 patients (32.2%) were initially referred for heart transplant evaluation. Prior EM surgery occurred in 35.4% (n [ 22; 8 prior valve repair, 8 prior valve replacement, BlalockTaussig shunt in 4, atrial septal defect (ASD) closure in 2). Results. Tricuspid repair was performed in 51.6% (n [ 32, 5 had re-repair). Bioprosthetic valve replacement was performed in 48.4% (n [ 30, 8 had rereplacement). The
BCPS was a planned procedure in 53 patients (85.5%) because of RV dysfunction; BCPS was added after unsuccessful weaning from bypass in 7 (11.2%), and in the early postoperative period due to hemodynamic instability in 2. Concomitant procedures included ASD closure in 48.3%, maze in 38.7%, and mitral valve repair in 6.4%. Postoperative extracorporeal membrane oxygenation support was needed in 8 patients. Delayed chest closure was performed in 25.8%. Early mortality was 1.6% (n [ 1). Mean mechanical ventilation time was 69.7 hours. Mean intensive care unit and hospital stays were 5.4 ± 3.5 and 10.7 ± 3.5 days, respectively. Mean follow-up was 3.6 ± 2.6 years (maximum, 12.8 years). Patients (n [ 10) with low preoperative LVEF (0.362 ± 0.035) improved to 0.517 ± 0.042 postoperatively (p [ 0.001). There was 1 late death in a patient with cystic fibrosis. Late reintervention was needed in 5 patients (8%). Late follow-up was available in 95% (n [ 59); all were acyanotic and 88% were in New York Heart Association functional class I/II. Conclusions. Concomitant BCPS is a useful adjunct in repair of advanced EM with severe RV dilatation and dysfunction. Operation can be performed with low early mortality. Intermediate-term survival and quality of life is good to excellent, and transplantation can be delayed or avoided in the majority.
E
of the bidirectional cavopulmonary shunt (BCPS) or “1.5 ventricle repair” improves hemodynamics by RV unloading and increasing LV preload. The BCPS was first applied in EM in 1980 [1] and since then other studies have shown improved outcome with its use [2–4]. We have previously reported our early results with 1.5 ventricle repair for EM in a small number of patients (n ¼ 14) [2]. The objective of this paper is to review early and midterm outcome of the BCPS in a larger number of high-risk patients with EM.
bstein malformation (EM) with right ventricular (RV) failure remains a challenge with high operative risk. Establishing tricuspid valve competence in EM increases RV afterload and increases the chance of RV failure, which in turn reduces the left ventricular (LV) preload and also produces the pancake effect on the LV. These physiologic alterations contribute to higher operative morbidity and mortality in EM. The addition
Accepted for publication March 4, 2014. Presented at the Sixtieth Annual Meeting of the Southern Thoracic Surgical Association, Scottsdale, AZ, Oct 30–Nov 2, 2013. Address correspondence to Dr Dearani, Division of Cardiovascular Surgery, Mayo Clinic, 200 First St SW, Rochester, MN 55905; e-mail: [email protected].
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2014;-:-–-) Ó 2014 by The Society of Thoracic Surgeons
Patients and Methods The Mayo Clinic Institutional Review Board approved this study. Between July 1999 and December 2012, 319 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.03.009
2
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
patients underwent surgery for EM. Among these, 62 (19.4%) with EM and severe RV dilatation or RV dysfunction underwent EM repair with concomitant BCPS; this group forms the cohort for this review. Of this 62 patient cohort, 45 (72.5%) were children 18 or less years of age; 8 were less than1 year old, 14 were age 1 to 5 years, 7 age 6 to 10 years, and 16 were 11 to 18 years old. Among the entire 62 patients, 20 patients (32.2%) were initially referred for the heart transplantation. The preoperative demographics are listed in Table 1. Fatigue and poor exercise tolerance were the most common presenting symptoms (61.2%), followed by cyanosis (27%). Right-sided heart failure (ascites) was present in 22.5%. Preoperative renal insufficiency or hepatic dysfunction was present in 2 patients (3.2%) each. Cardiomegaly on chest X-ray was present in all; mean cardiothoracic ratio was 0.8 0.04. All patients had echocardiographic evidence of severe RV dilatation; assessment of RV function and tricuspid regurgitation is shown in Table 1. Assessments of RV function included the entire anatomic RV (both atrialized and functional zones). We used this approach as postoperatively both zones are beyond the functional tricuspid valve and we feel that it reduces the variability between examiners. Parameters assessed included the following: fractional area change of the RV cavity in the 4-chamber and short axis views (both subjectively and quantitatively); systolic descent of the tricuspid annular plane systolic excursion; myocardial performance index; and a subjective, visual assessment of systolic RV myocardial thickening. These parameters were combined by the interpreting cardiologist in order to classify the degree of RV dysfunction as mild, moderate, or severe as outlined in Table 1. No patient was judged to have normal RV function. Grading of the degree of tricuspid regurgitation reflected the amount of systolic reversed flow crossing the Table 1. Patient Demographics Demographic
Number
Variable Median age Male Cardiac rhythm Sinus Atrial fibrillation Paroxysmal atrial fibrillation WPW syndrome Tricuspid regurgitation grade Moderate Moderately severe Severe RV dysfunction Mild Moderate Severe RV ¼ right ventricular;
Patient number (n ¼ 62) 21.5 years (range, 9 months–57 years) n ¼ 32 (51.6%) 35 14 7 6
(56.4%) (22.5%) (11.2%) (9.6%)
2 (3.2%) 5 (8%) 55 (88.7%) 3 (4.8%) 14 (22.5%) 45 (72.5%)
WPW ¼ Wolff-Parkinson-White.
Ann Thorac Surg 2014;-:-–-
plane of the functional tricuspid orifice (from the functional to the atrialized RV). A combination of anatomic leaflet coaptation (or the lack thereof), regurgitant jet width, Doppler regurgitant signal strength, and systolic distortion of hepatic venous Doppler flow profile was used by the interpreting cardiologist to classify regurgitant severity. In cases with a single regurgitant orifice, the width of the regurgitant jet at its origin (the vena contracta) was used to define jet width. In the mature patient, a vena contracta width of less than 3 mm is associated with mild regurgitation, widths of 8 to 10 mm represented severe transvalvar regurgitation. In young patients, vena contracta diameters which are less than 10% of the normal annulus were considered mild. Jet origins measuring more than 25% to 30% of the normal annular dimension were classified as severe. When a combination of multiple jets existed, the examiner mentally combined the size of the regurgitant orifices in order to determine the overall grade. Preoperative hemodynamic variables included the following: mean RV systolic pressure 32.7 0.7 mm Hg; mean pulmonary artery pressure 15.6 2.1 mm Hg; and left ventricular ejection fraction (LVEF) 0.536 0.071. Biventricular dysfunction was seen in 10 patients (16.1%); mean LVEF was 0.362 0.035. Preoperative magnetic resonance imaging (MRI) was available in 21 (33.8%) patients. The MRI derived mean RVEF, RV end systolic and diastolic volumes, indexed to body surface area were 0.295 0.325, and 267.7 185.3 and 388.3 133.6 mL/m2, respectively. Twenty-two patients (35.4%) had previous surgery. Eight patients had recurrent tricuspid regurgitation after previous valve repair (7 had repair elsewhere) and 8 had bioprosthetic tricuspid valve failure (2 had multiple previous valve replacements); 4 had modified BlalockTaussig shunt and 2 had isolated atrial septal defect (ASD) closure. The average time interval between the previous tricuspid operations and the current operation was 2.6 years (repair) and 10.3 years with prior tricuspid replacement. All patients with prior tricuspid repair had severe recurrent tricuspid regurgitation. In those patients with prior replacement, the mean prosthetic gradient was 13.2 0.7mm Hg. Severe bioprosthetic regurgitation was present in 4 patients, moderate regurgitation in 3, and mild regurgitation in 1. All patients underwent tricuspid repair or replacement and BCPS; concomitant procedures are shown in Table 2. Patients requiring reoperation for EM with BCPS performed elsewhere, and EM with pulmonary atresia with a RV-to-pulmonary artery conduit were excluded from the study.
Operative Technique and Perioperative Management Median sternotomy with aortic and bicaval cannulation was utilized; the superior vena cava was cannulated percutaneously using the Seldinger technique through the right internal jugular vein in older children and adults. This facilitated construction of the BCPS and it also provided a source of venous access for cardiopulmonary bypass in the event that peripheral cannulation
Ann Thorac Surg 2014;-:-–-
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
Table 2. Operative Details Surgical procedure Tricuspid repair (n ¼ 32) Monocusp repair (n ¼ 4) Cone repair (n ¼ 28) Tricuspid replacement (n ¼ 30) Indication for BCPS (n ¼ 62)
Concomitant procedures Right atrial reduction Right ventricular plication ASD/PFO closure Right-sided maze Biatrial maze Ablation WPW Mitral valve repair Epicardial ICD
Number (N¼62) TV repair with BCPS, n ¼ 27 TV re-repair with BCPS, n ¼ 5 (bilateral BCPS in one) TVR with BCPS, n ¼ 20 Redo TVR with BCPS, n ¼ 8 TVR with bilateral BCPS, n ¼ 2 Planned procedure n ¼ 53 (85.4%) Unable to separate from bypass n ¼ 7 (11.2%) Early postoperative n ¼ 2 (3.2%) n, patients n ¼ 53 n ¼ 52 n ¼ 30 n ¼ 24 n¼8 n¼4 n¼4 n¼4
ASD ¼ atrial septal defect; BCPS ¼ bidirectional cavopulmonary shunt; ICD ¼ implantable cardioverter defibrillator; PFO ¼ patent foramen ovale; TV ¼ tricuspid valve; TVR ¼ tricuspid valve replacement; WPW ¼ Wolff-Parkinson-White.
was needed during resternotomy. The conduct of operation included normothermic or mildly hypothermic (34 C) bypass and cold blood antegrade cardioplegia with subsequent doses at 20-minute intervals during the cross-clamp period. Iced slush saline was applied topically to the RV. Tricuspid repair, RV plication, and closure of intracardiac shunts were performed with cardioplegic arrest. In patients undergoing tricuspid replacement, the cross-clamp was removed after intracardiac shunt closure and RV plication. Valve suture placement and seating of the prosthesis was performed with the heart beating and perfused to insure maintenance of sinus rhythm and to minimize myocardial ischemic time. Valve repair techniques are described in detail in prior publications and were monocusp version in the early portion of this series [5–7] and the cone reconstruction in the recent era (2007 to present) [8–10]. In general, anatomy favorable for valve repair included a large anterior leaflet and some (even diminutive) septal leaflet. Complete surgical delamination and mobilization of leaflet tissue was performed so that 360 degrees of rotated, reconstructed leaflet tissue was anchored at the true tricuspid annulus. When tricuspid valve replacement was performed, a porcine bioprosthesis was utilized; technique of insertion has been previously described [11, 12]. Valve anatomy that made anatomic cone repair unsuccessful included significantly muscularized or deformed leaflet tissue that resulted in reduced pliability and poor coaptation, or absent septal leaflet.
3
The BCPS was performed after intracardiac repair and right reduction atrioplasty was complete. The anastomosis was continuous when somatic growth was complete and the anterior suture line was interrupted when it was not. The azygos vein was routinely ligated. All patients were separated from bypass on epinephrine 0.05 mcg $ kg1 min1, milrinone 0.25 $ mcg $ kg1 $ min1. Vasopressin 0.02 to 0.04 mcg $ kg1 $ min1 was utilized selectively to maintain a mean blood pressure of greater than 50 mm Hg. Inhaled nitric oxide was utilized to offset the pulmonary vasoconstrictive effects of inotropic support; this was applied when epinephrine was greater than 0.05 mcg $ kg1 $ min1 or when vasopressin was used. Temporary atrial pacing at 110 to 120 was done routinely for the first 24 hours if the intrinsic heart rate was less than 100. Extracorporeal membrane oxygenation (ECMO) support was accomplished with central aortic and right atrial cannulation. Inadequate venous return was not an issue despite the presence of BCPS. The ECMO flows were maintained with a cardiac index of 1.5 to 2.0 L $ min1 $ m2. Discharge medications included ace-inhibiter and beta-blocker therapy. Sildenafil was used for 4 weeks postoperatively in most patients and routinely in those who underwent valve repair and lived at high altitude. Amiodarone therapy was utilized for 3 to 6 months in those patients who had RV plication or had episodes of atrial or ventricular tachyarrhythmias postoperatively. Warfarin therapy was used for tricuspid replacement and was maintained for 3 to 6 months and discontinued when there was echocardiographic confirmation that prosthetic valve leaflets were moving normally. Aspirin therapy (81 mg) was used indefinitely in patients undergoing valve replacement.
Results Early Results Tricuspid valve repair (n ¼ 32; 51.6%) or porcine bioprosthetic tricuspid replacement (n ¼ 30; 48.4%) was performed with BCPS in all patients (Table 2). Intraoperative and early postoperative data are shown in Table 3. An intraaortic balloon pump was utilized in 4 adults for unstable hemodynamics on optimal inotropic support when separating from bypass; all had LV dysfunction preoperatively. Extracorporeal membrane oxygenation was utilized in 8 patients (12.9%); 5 with valve repair and 3 with valve replacement. Median age of the patients requiring ECMO was 2 years (range, 9 months to 41 years). In the valve repair group, 1 patient (a 15-year-old child) was transferred from an outside institution on ECMO after failure of valve repair and cardiogenic shock; she underwent successful valve re-repair with BCPS and was weaned from ECMO 11 days later. Among 3 patients requiring ECMO who underwent tricuspid replacement, 2 needed it in the operating room for failure to wean from bypass and 1 needed ECMO on the second postoperative
4
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
Ann Thorac Surg 2014;-:-–-
Table 3. Perioperative Details Variable CPB time (minutes) Cross-clamp time (minutes) ECMO IABP Open chest Mean duration of ventilation (hours) Mean ICU stay (days) Mean hospital stay (days) CPB ¼ cardiopulmonary bypass; unit.
Entire Population (n ¼ 62)
Valve Repair (n ¼ 32)
Valve Replacement (n ¼ 30)
112.5 17.7 54.1 34.6 8 patients (12.9%) (mean 5.8 days) 4 patients (6.4%) 16 patients (25.8%) 69.7 84.8 5.4 3.5 10.7 3.5
113.5 26.3 74.9 22.1 5 patients (mean 6.1 days) None 5 patients 71.8 93.3 5.8 3.5 11 4.2
111.4 38.8 31.9 11.3 3 patients (mean 7 days) 4 patients 11 patients 67.6 22.6 5.1 2.1 10.5 2.8
ECMO ¼ extracorporeal membrane oxygenation;
day for poor hemodynamics. Seven of the 8 patients requiring ECMO support were successfully weaned with a mean duration of 5.8 days. Early postoperative morbidity and mortality are shown in Table 4. There was 1 early death (1.6%). This patient had valve replacement (27-mm porcine) and developed worsening of hemodynamics on the second postoperative day. Extracorporeal membrane oxygenation was placed with conversion to an extracardiac Fontan; this patient died 11 days later from multiorgan failure. Early reoperation occurred in 7 patients (11.2%). Four patients initially repaired needed early reoperation for recurrent severe tricuspid regurgitation. Three patients underwent successful re-repair and 1 underwent valve replacement after unsuccessful re-repair. Three patients needed early reoperation after valve replacement. One patient had BCPS added on the second postoperative day for severe RV dysfunction and 2 patients needed surgery for bioprosthetic valve thrombosis (1 had valve re-replacement and 1 had thrombus removal at the time of ECMO separation). Dismissal echocardiography in the valve repair group (n ¼ 32) showed no or mild tricuspid regurgitation 28 patients (87.5%), moderate tricuspid regurgitation in 3 (9.3%), and severe tricuspid regurgitation in 1 (3.1%);
IABP ¼ intraaortic balloon pump;
ICU ¼ intensive care
this patient underwent early reoperation and valve replacement after unsuccessful valve re-repair. In the bioprosthetic valve replacement group (n ¼ 30), all patients had no regurgitation and the mean gradient was 6.1 2.1 mm Hg.
Late Results Complete follow-up was available in 95% (n ¼ 59) with a mean follow-up of 3.6 2.6 years (maximum, 12.8 years). All patients were acyanotic and 88% of patients (n ¼ 52) were in New York Heart Association functional class I/II. Recurrent atrial or ventricular arrhythmias were common during the follow-up period (n ¼ 7; 11.2%). Among the 20 patients who were initially referred for heart transplant, only 3 were listed for transplantation postoperatively. One patient (57-years old) underwent transplant 6 years after operation (valve replacement with BCPS), and 1 patient is currently listed for heart transplant. The third patient died 3 years after surgery awaiting transplantation. This patient was a 32-year-old woman with cystic fibrosis who had tricuspid valve replacement and BCPS. She died from multiorgan failure while awaiting combined heart and liver transplant. The remaining 17 patients were able to avoid transplantation.
Table 4. Early Morbidity and Mortality Variable Morbidity Arrhythmias (atrial and ventricular) Valve thrombosis Stroke Reintubation Pneumonia Heart block Chylous Pericardium Mortality (early) Early reoperation
Entire Population (n ¼ 62) 7 2 2 2 2 1 1 1 7
(11.2%) (3.2%) (3.2%) (3.2%) (3.2%) (1.6%) (1.6%) (1.6%) (11.2%)
Valve Repair (n ¼ 32)
Valve Replacement (n ¼ 30)
3 (9.3%) None 2 (6.2%) 2 None None None None 4 (12.5 %)
4 (13.3%) 2 (6.6%) None None 2 1 1 1 3 (10%)
Ann Thorac Surg 2014;-:-–-
Overall survival after1.5 ventricle repair at 1, 3, and 5 years was 98.3%, 98.3%, and 95.4%, respectively (Fig 1). Late reintervention occurred in 5 (8%) patients, including 2 in the valve repair group; 1 patient needed tricuspid valve replacement for recurrent tricuspid regurgitation 3 years after valve repair and the other patient needed implantable cardioverter defibrillator placement for recurrent ventricular arrhythmias. In the valve replacement group, 1 patient underwent heart transplant 6 years after operation, 1 patient had placement of an epicardial pacing system for heart block, and 1 needed stenting of a stenotic BCPS connection 2 years after operation. Overall freedom from reintervention at 1, 3, and 5 years was 96.6%, 96.6%, and 74.3%, respectively (Fig 2). Echocardiography at latest follow-up in the valve repair group showed mild tricuspid regurgitation in 84.3% and mild-to-moderate regurgitation in 6.2%. In the tricuspid valve replacement group (n ¼ 30), 8 patients (26.6%) had a prosthesis-related abnormality. Moderate-to-severe bioprosthetic stenosis (mean gradient >10 mm Hg) or moderate prosthetic regurgitation was seen in 4 patients each. In those patients with depressed LVEF preoperatively (n ¼ 10), LVEF increased from 0.362 0.03 to 0.517 0.042 (p ¼ 0.001). Chest radiography at latest follow-up showed a significant reduction in cardiothoracic ratio from 0.8 0.04 to 0.59 0.03 (p ¼ 0.001).
Comment Tricuspid valve repair in EM has been challenging and the recent application of cone reconstruction has facilitated successful repair for most patients. A right ventricle dysfunction has been identified as a risk factor for early mortality in EM [13]. Symptoms are often inconspicuous until adulthood, particularly when the atrial septum is intact or the patent foramen ovale is small because cyanosis or arterial desaturation is minimized or absent. This can result in late referral to surgery. The most common symptom is fatigue or decreased exercise tolerance; frank right-sided heart failure is uncommon unless marked RV dysfunction is present. Progressive RV
Fig 1. Kaplan-Meier survival curve after 1.5 ventricle repair.
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
5
Fig 2. Overall freedom from reintervention after 1.5 ventricle repair.
enlargement and dysfunction is very common in EM resulting in increased operative risk. This review demonstrates that operation for EM in the setting of advanced RV enlargement or dysfunction can be safely performed when the BCPS is applied to EM repair. In this advanced group, factors that facilitate the decision to offer operation are the absence of overt signs of heart failure on physical exam, the presence of normal LV EF, and valve anatomy favorable for repair [9, 10]. The BCPS is helpful in the setting of EM because it reduces the volume load to the enlarged dysfunctional RV and it improves preload to the chronically under filled and compressed LV. However, a BCPS is not applicable to all patients with EM and RV dysfunction. In the adult patient, or when there is depressed LV function, a preoperative cardiac catheterization is performed. The following hemodynamics are required for BCPS: LVEDP less than12 mm Hg; transpulmonary gradient less than10 mm Hg; and mean pulmonary artery pressure less than18 mm Hg. Overall indications and contraindications for BCPS are shown in Table 5 [2, 3, 9, 10]. The decision to repair versus replace the tricuspid valve is critical in this advanced EM group. While valve repair is preferred, it is often reserved for children and those patients with favorable valve anatomy, which may not be the case in the advanced, adult patient group. Complex intracardiac repair (tricuspid valve repair, RV plication, and ASD closure) with massive RV and annular enlargement can result in a long cross-clamp time, which can further worsen the already dysfunctional right ventricle. The decision to proceed with valve replacement in this setting may be more prudent as the cross-clamp time is shorter by limiting it to patent foramen ovale or ASD closure and RV plication. This allows the majority of the procedure to be performed with the heart beating and perfused. Furthermore, it assures tricuspid competence at the end of the operation and avoids the need for repeated bypass runs to revise an unacceptable valve repair and then subsequent replacement. Finally, the decision to proceed with replacement should be made early and preferably before the BCPS has been performed. This is important as the prosthesis is intentionally downsized because the
6
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
Table 5. Indications and Contraindications for Bidirectional Cavopulmonary Shunt Heart failure indications Cardiothoracic ratio > 65% RVEDV > 250 mL/m2 RVEF < 0.25 Small, squashed, D-shaped left ventricle Postrepair RAP:LAP > 1.5:1 Postrepair low cardiac output: persistent metabolic acidosis, low urine output, increasing creatinine, poor peripheral perfusion, low mean arterial blood pressure < 50 mm Hg Non-heart failure indications Reduce tension on complex tricuspid repair Postrepair tricuspid stenosis (mean gradient > 8–10 mm Hg) Contraindications Mean PA pressure > 20 mm Hg Pulmonary arteriolar resistance > 4 Woods units LVEDP or LA pressure >12 mm Hg Significant pulmonary artery hypoplasia LA ¼ left atrial; LVEDP ¼ left ventricular end-diastolic pressure; PA ¼ pulmonary artery; RAP:LAP ¼ right atrial pressure to left atrial pressure; RVEDV ¼ right ventricular end-diastolic volume; RVEF ¼ right ventricular ejection fraction.
amount of blood flow across the tricuspid valve is reduced by a factor of 30% to 40% with the BCPS. In our experience it is preferable to have a gradient of 5 to 6 mm Hg across the prosthesis so that all leaflets open and close. When the prosthesis is too large (easy to do given marked annular dilatation), a gradient of only 1 to 2 mm Hg usually results in 1 or 2 of the prosthetic leaflets not opening normally. This results in the development of thrombus and accelerated bioprosthetic deterioration. This finding was the stimulus for warfarin therapy for 3 to 6 months postoperatively. Mechanical valves should not be used in this setting as there can be poor disc motion due to low right atrial pressure and poor RV function that predisposes to valve thrombosis even in the face of therapeutic anticoagulation. The role of atrial septal fenestration can also be considered. As mentioned above, the BCPS significantly unloads the RV and provides preload to the LV. This combination is particularly helpful when the hemodynamic situation is poor. Furthermore, the right-to-left shunting across the atrial level shunt can result in profound arterial desaturation in the postoperative period. In patients with LV hypertrophy or dysfunction it is important to determine left atrial and LV end diastolic pressures preoperatively as they may be elevated and would then result in left-to-right shunting across the atrial level shunt. Finally, in the adult EM population, a persistent atrial level shunt is associated with increased risk of paradoxical embolism [14]. In our practice, subtotal closure of the ASD is reserved for neonates and children with lesser degrees of RV dysfunction. Limitations to BCPS include reduced right atrial or RV access for electrophysiologic intervention, mild facial suffusion, or development of upper torso venous
Ann Thorac Surg 2014;-:-–-
collaterals or pulmonary arteriovenous fistulae. Only 1 patient (1.6%) in our series developed a stenosis of the BCPS anastomosis that was relieved with percutaneous stenting. This low incidence of problems related to BCPS in EM has been demonstrated by others [15, 16]. We had a low threshold for delayed sternal closure and a low threshold for temporary ECMO support. Our preference for ECMO over RV assist device therapy is based on our extensive ECMO experience with other high-risk postcardiotomy patients. The assurance of adequate oxygenation and the ability to rapidly wean inotropic agents on ECMO support has been the rule in our experience. We generally begin with ECMO in the immediate postoperative period, achieve aggressive diuresis, and plan ECMO separation 4 to 7 days later. If ECMO separation is unsuccessful, we make a transition to RV assist device if mechanical support is needed for a longer duration; fortunately, this has rarely been the case. The role of MRI in EM is important. In general, we prefer it in all patients preoperatively and postoperatively. The MRI was not done in many of the younger children because of the need for general anesthesia and many of the older patients had hemodynamic cardiac catheterization performed preoperatively. At late follow-up, arrhythmias (atrial and ventricular) were the most common reason for hospitalization. This emphasizes the importance of a long-term surveillance protocol for late arrhythmias that includes routine exercise testing at 6 months postoperatively and periodic Holter monitoring to determine need for arrhythmia therapy. While most are atrial tachyarrhythmias (flutter and fibrillation), ventricular arrhythmias are also important. The decision to offer conventional operation (1.5 ventricle repair) in this high-risk group is difficult when LV dysfunction is present. Our decision to offer surgery when LVEF was below normal was based on the presence of a small LV cavity and normal anterior and lateral LV free wall motion. Specifically, the ventricular septal motion abnormality was the predominant explanation for LV dysfunction, which we thought would improve after RV decompression with the BCPS. We would not advise operation with severe LV dysfunction or a dilated, dysfunctional LV; these patients would be referred to transplantation.
Limitations of the Study We did not have a protocol for application of the BCPS in the early part of this review. As we learned from our own experience and others [3], indications for BCPS evolved and are outlined above. Objective analysis of RV size and function at late follow-up with MRI would be preferred and was limited to the minority of this cohort. Longer follow-up is required to examine late outcome and to see if transplantation can be further delayed in this high-risk group.
Conclusions Concomitant BCPS is a useful adjunct in EM repair with severe RV dilatation or dysfunction. Operation can be
Ann Thorac Surg 2014;-:-–-
performed safely but postoperative care is prolonged and morbidity is increased. The LV function improves in those with low EF preoperatively. Intermediate-term survival and quality of life is good to excellent, and transplantation can be delayed or avoided in the majority.
References 1. Marcelletti C, D€ uren DR, Schuilenburg RM, Becker AE. Fontan’s operation for Ebstein anomaly. J Thorac Cardiovasc Surg 1980;79:63–6. 2. Quinonez LG, Dearani JA, Puga FJ, et al. Results of the 1.5ventricle repair for Ebstein anomaly and the failing right ventricle. J Thorac Cardiovasc Surg 2007;133:1303–10. 3. Malhotra SP, Petrossian E, Reddy VM, et al. Selective right ventricular unloading and novel technical concepts in Ebstein’s anomaly. Ann Thorac Surg 2009;88:1975–81. 4. Chauvaud S, Fuzellier JF, Berrebi A, et al. Bi-directional cavopulmonary shunt associated with ventriculo and valvuloplasty in Ebstein’s anomaly: benefits in high risk patients. Eur J Cardiothorac Surg 1998;13:514–9. 5. Danielson GK, Maloney J, Devloo R. Surgical repair of Ebstein’s anomaly. Mayo Clin Proc 1979;54:185–92. 6. Dearani JA, Danielson GK. Tricuspid valve repair for Ebstein’s anomaly. Op Tech Thorac Cardiovasc Surg 2003;8: 188–92. 7. Dearani JA, O’Leary PW, Danielson GK. Surgical treatment of Ebstein’s malformation: state of the art 2006. Cardiol Young 2006;16(Suppl 3):12–20.
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
7
8. Dearani JA, Bacha E, da Silva JP. Cone reconstruction of tricuspid valve for Ebstein’s anomaly: anatomic repair. Oper Tech Thoracic Cardiovasc Surg 2008;13:109–25. 9. Dearani JA, Said SM, O’Leary PW, et al. Anatomic repair of Ebstein’s malformation: lessons learned with cone reconstruction. Ann Thorac Surg 2013;95:220–6. 10. Dearani JA, Said SM, Burkhart HM, Pike RB, O’Leary PW, Cetta F. Strategies for tricuspid re-repair in Ebstein malformation using the cone technique. Ann Thorac Surg 2013;96: 202–8. 11. Dearani JA, Danielson GK. Ebstein’s anomaly. In: Sellke FW, del Nido PJ, Swanson SW, eds. Sabiston & Spencer Surgery of the Chest. Philadelphia, PA: Elsevier; 2005:2223–35. 12. Kiziltan HT, Theodoro DA, Warnes CA, O’Leary PW, Anderson BJ, Danielson GK. Late results of bioprosthetic tricuspid valve replacement for Ebstein’s anomaly. Ann Thorac Surg 1998;66:1539–45. 13. Brown ML, Dearani JA, Danielson GK, et al. The outcomes of operations for 539 patients with Ebstein anomaly. J Thorac Cardiovasc Surg 2008;135:1120–36.e1-7. 14. Attenhofer Jost CH, Connolly HM, Dearani JA, Edwards WD, Danielson GK. Ebstein’s anomaly. Circulation 2007;115: 277–85. 15. Kim S, Al-Radi O, Mark KF, et al. Superior vena cava to pulmonary artery anastomosis as an adjunct to biventricular repair: 38-year follow-up. Ann Thorac Surg 2009;87: 1475–83. 16. Liu J, Qiu L, Zhu Z, Chen H, Hong H. Cone reconstruction of the tricuspid valve in Ebstein anomaly with or without one and a half ventricle repair. J Thorac Cardiovasc Surg 2011;141:1178–83.
DISCUSSION DR EMMETT MCKENZIE (Houston, TX): That was very nicely presented, Dr Raju. Thank you. Thank you also for getting me the manuscript. I was able the review that before the meeting. You and your colleagues are to be congratulated for outstanding results in a very sick group of patients with very advanced Ebstein’s disease. This experience that you present today is only the latest iteration from the very large series at Mayo Clinic that has been so helpful in our understanding of Ebstein’s anomaly. So thank you again for putting that together. I have a couple of questions. You note that a third of the patients included in this report had been referred initially for transplant. I would submit, however, that it is quite common for patients with congenital heart disease to be referred for a transplant and in fact to have reparable residual heart disease. The fact that the majority of the patients experienced durable relief of their symptoms following operation suggests that the right heart dysfunction, not necessarily the right ventricle but the global right heart dysfunction, is actually treatable and reversible. In an earlier report from your colleagues at the Mayo Clinic, low cardiac output was the main predictor of early mortality. And the fact that seven of the eight patients that required ECMO in this series recovered and survived suggests again that the low cardiac output and the right heart dysfunction resolved and resolved quickly. So my question is, is it possible that highly selected use of a right ventricular assist device or some form of temporary mechanical right heart support would achieve the same result without the long-term consequences of a bidirectional Glenn shunt? DR RAJU: Thank you, Dr McKenzie, for your kind remarks. Ebstein’s anomaly is a disease of the myocardium; basically it is an RV [right ventricular] myopathy. The abnormality with the leaflets are secondary to the myopathy. We had a limited number of
patients who had a postoperative MRI [magnetic resonance imaging], and in those patients we observed reduction in RV size and some improvement of the RV function. Since this is a disease of the myocardium and results in a myopathy, RV function improvement may be limited. This is the reason we opted to proceed with RV unloading with the bidirectional cavopulmonary shunt. DR MCKENZIE: My other question relates to the perioperative management. In this group, 25% of the patients had delayed sternal closure and the time to extubation was three days, on average. There is fairly compelling evidence that with a Fontan operation early extubation is beneficial, and these patients have no right ventricle. Even patients undergoing Fontan conversion, a much more lengthy operation in a much sicker group of patients, can benefit from early extubation; if not in the OR [operating room], in the first couple of hours after surgery. Is the decision to delay the sternal closure related to the size of the heart, pulmonary hypoplasia, or high airway pressures? And do you think that this group of patients would benefit from early extubation, also? DR RAJU: Dr Dearani’s approach here is anybody who needs bidirectional cavopulmonary shunt as an unplanned or a salvage procedure, his preference is to leave the chest open in order to avoid further hemodynamic instability. Conversely, if they had RV unloading as a planned procedure, we are less likely to leave the chest open; almost all were closed primarily at the time of initial repair when the bidirectional cavopulmonary shunt was planned. DR MCKENZIE: And one last quick question. What are your recommendations for timing of surgery, recognizing that many of these patients aren’t symptomatic and aren’t diagnosed until
8
RAJU ET AL RV UNLOADING FOR HEART FAILURE RELATED TO EM
late in their disease? You mentioned that the right ventricle suffers from a myopathy, and with our getting over the learning curve of the cone operation and it seemingly being applicable to more variants, are you recommending earlier repair now? DR RAJU: The decision regarding the timing of the surgery is determined by the experience of the surgeon and the surgical team. If the surgeon is able to produce a reliable and reproducible valve repair, we recommend surgery between two to five years of age in the asymptomatic child with severe tricuspid regurgitation and RV enlargement. The learning curve for an anatomic cone-type repair is steep, so if the surgeon, or the team, is not confident or comfortable in repairing and producing a good result, then we think it is more prudent to follow the traditional indications for surgery for tricuspid regurgitation. This includes the presence of symptoms, progressive RV enlargement, onset of atrial tachyarrhythmias, and cyanosis. Ebstein anomaly is a serious disease and a reproducible, durable repair is difficult to obtain when experience is limited. For these reasons, consideration should be given to referral to a center with documented experience so outcome can be optimized, particularly when the patient is asymptomatic. Thank you. DR TARA KARAMLOU (San Francisco, CA): Two quick questions. One, did you notice a detriment in those patients who you performed the cavopulmonary shunts as a salvage operation versus the patients that it was done a priori? And then my second question is: Do you have long-term exercise data that can support your contention that cardiac output is improved? At the CHSS [Congenital Heart Surgeons’ Society], we published a paper looking at exercise data in patients with pulmonary atresia with intact septum, showing that patients with a 1.5 ventricle repair, including a bidirectional Glenn, had a distinct benefit compared to the 1 or 2 ventricle patients in that the unloading actually recovered their cardiac output and they did better. Thanks. DR RAJU: I think Dr Dearani can respond. I can answer your second question. We did notice about 18% of the patients in this group had a good functional capacity at follow-up and their exercise tolerance improved, and they are in class 1 and 2. Doctor Dearani? DR DEARANI: What was the first part of the question?
Ann Thorac Surg 2014;-:-–-
DR KARAMLOU: So, Joe, the first question was, is there a downside to doing this as a salvage? We know in congenital heart surgery many times if you make the “wrong decision” up front that you pay a price down the line. And, two, do you have any formal cardiopulmonary exercise testing to support your contention about cardiac output, especially in the setting of chronotropic incompetence? DR DEARANI: There were no downsides. Actually, in the early part of this series it was resorting to the bidirectional Glenn that was the eye-opening finding that the patients turned around almost immediately. There was no detriment from a hemodynamic standpoint in the long term. In the older patients, there could be some concern about the development of prominent varicosities in the shoulders and the neck, facial suffusion, and, of course, venous access for arrhythmia analysis down the road may be more challenging from the femoral vein below. The majority of patients have had late exercise testing, and there is an improvement in exercise performance. It is due in part to increased cardiac output, and also probably related to the elimination of right to left shunting with ASD [atrial septal defect] closure. Even isolated ASD closure with elimination of arterial desaturation results in improvement in exercise tolerance, too. So it’s probably a combination of both things. DR ANDREW C. FIORE (St. Louis, MO): Tell us your thoughts about the atrial septal defect. How do you manage it in these patients? DR DEARANI: We believe the bidirectional Glenn is a better pathway when you have severe advanced right-sided heart failure. The atrial septal fenestration does not unload the right ventricle, it just provides a little bit of extra preload to the left ventricle. So when you have a compressed, pancaked left ventricle, the atrial septal fenestration does not unload and decompress the RV Furthermore, cyanosis is the indication for surgery in many of these patients, and so intentionally leaving a fenestrated atrial septum may result in an element of desaturation, which is often the indication for surgery. And finally, in the adult patients an atrial septal fenestration in Ebstein’s anomaly is a risk factor for paradoxical embolism. We believe the role of routine subtotal closure of the ASD includes Ebstein repair in infancy and in young children who have lesser degrees of RV dysfunction that is not quite bad enough to proceed with a Glenn.