The presence of a bicuspid aortic valve is associated

Neoaortic Bicuspid Valve in Arterial Switch Operation: Mid-Term Follow-Up CARDIOVASCULAR Shahid M. Khan, FRCS, Ahmad Bin Sallehuddin, FRCS, Ziad R. Al-Bulbul, MD, and Zohair Y. Al-Halees, MD King Faisal Heart Institute, King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia Background. We sought to identify the prevalence of bicuspid pulmonary valve among patients with transposition of the great arteries undergoing the arterial switch operation and evaluate functional integrity of that valve in the neoaortic position. Methods. Between October 1985 and December 2001, 391 patients had an arterial switch operation for transposition and its variants. Perioperative information and follow-up data were available for 342 patients. The serial echocardiograms of patients with bicuspid pulmonary valve were reviewed. The neoaortic valve was serially assessed, focusing on aortic insufficiency, annulus diameter, and pressure gradients. Results. Twenty-four patients (7%) had a bicuspid pulmonary valve. Age at operation was 5 days to 12 years. Two patients were lost to follow up, and 22 patients had mean follow-up of 5.3 years (range, 2 months to 13 years), of which 21 patients were alive and 1 died late. At least two postoperative echocardiogram reports were available on 19 patients. Seven patients had no neoaortic regurgitation, and 10 had trivial regurgitation. Severe aortic regurgitation developed in 1 patient with endocarditis and in another with repair of Taussig-Bing anomaly. Neoaortic valve size indexed to body surface area showed an increase in annular diameter over time proportional to somatic growth. No significant valve stenosis developed. Conclusions. Encountering a bicuspid pulmonary valve at the time of an arterial switch operation is not uncommon. The integrity of a bicuspid pulmonary valve in the neoaortic position is maintained at a mean follow-up of 5.3 years. We believe that the presence of a bicuspid pulmonary valve is not a contraindication to an arterial switch operation. (Ann Thorac Surg 2008;85:179 85) 2008 by The Society of Thoracic Surgeons The presence of a bicuspid aortic valve is associated with aortic root dilatation, annuloaortic ectasia, and aortic dissection [1, 2]. It is more susceptible to fibrosis and calcification, leading to aortic stenosis and mixed aortic valve disease in otherwise normal hearts [3, 4]. A bicuspid pulmonary valve is more prone to degenerative changes in its native position. Hence, there has been a concern about the long-term performance of a bicuspid aortic valve used in the systemic circulation (arterial switch, Damus Kaye Stansel, and Norwood operations) [5]. A bicuspid aortic valve is not generally considered a contraindication to the arterial switch operation (ASO); however, there are still limited data on its medium-term to long-term function in the neoaortic aortic position. We sought to examine this particular issue in a cohort of patients undergoing the ASO at our institution. Accepted for publication July 6, 2007. Presented at the Poster Session of the Forty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 29 31, 2007. Address correspondence to Dr Al-Halees, King Faisal Heart Institute, King Faisal Specialist Hospital and Research Centre, MBC-16, PO Box 3354, Riyadh, 11211, Saudi Arabia; e-mail: alhalees@kfshrc.edu.sa. Patients and Methods Patients Between October 1985 and December 2001, 391 patients with simple and complex transposition of the great arteries underwent an ASO. We obtained approval from the Research Advisory Council and Ethics Committee to conduct this retrospective data collection and analysis. The records were reviewed to determine the cardiac anatomy, coronary pattern, presence of bicuspid pulmonary valve, and operative details. Complete data were available for 342 patients from clinical records and the hospital-based database (Apollo Advance LUMEDX Corp 2060, Oakland, CA). A bicuspid pulmonary valve was present in 24 patients (7.0%; 19 boys, 5 girls). The median age at the time of operation was 1 month (range, 5 days to 12 years). The median body surface area (BSA) was 0.25 m 2, and the average weight was 4.8 kg (range, 2.4 to 22.2 kg). All patients had dextrotransposition of the great arteries except one, who had levotransposition of the great arteries. This patient had both right and left atrioventricular valve regurgitation and pulmonary hypertension. In 15 patients the ventricular septum was intact, and 9 had a ventricular septal defect. The arrangement of the great arteries was anteroposterior in 23 patients and side by side in 1. Coronary anatomy was normal in 18 patients, 2008 by The Society of Thoracic Surgeons 0003-4975/08/$34.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2007.07.012

CARDIOVASCULAR 180 KHAN ET AL Ann Thorac Surg NEOAORTIC BICUSPID VALVE IN ASO 2008;85:179 85 Table 1. Diagnosis Diagnosis No. of Patients D-TGA bicuspid pulmonary valve 23 L-TGA bicuspid pulmonary valve 1 D-TGA Intact ventricular septum 15 D-TGA VSD 9 L-TGA with left AV valve regurgitation 1 D-TGA isolated PS 1 D-TGA PS TAPVD 1 D-TGA PS LVOTO 1 D-TGA isolated LVOTO 1 Great arteries Anteroposterior 23 Side by side 1 Coronary anatomy 1 LAD, Cx, 2 RCA 18 1 LAD, 2Cx, RCA 4 Not recorded 2 Cx circumflex coronary artery; D-TGA dextro-transposition of the great arteries; L-TGA levo-transposition of the great arteries; LAD left anterior descending coronary artery; LVOTO left ventricular outflow tract obstruction; PS pulmonary stenosis; RCA right coronary artery; TAPVD total anomalous pulmonary venous drainage; VSD ventricular septal defect. ascending aorta was placed posterior to the pulmonary artery (Lecompte maneuver) in all but 1 patient. Coronary arteries were translocated and incorporated in the neoaorta by using a modified trapdoor method. A pantaloon-shaped baffle of fresh autologous pericardium was used to reconstruct the neopulmonary artery. The primary requirement for a patient to undergo the ASO in relation to the pulmonary valve was suitable valve morphology and annulus size. Patients with a normal pulmonary valve annulus and thin and pliable leaflets with no calcification and normal commissures were considered suitable for ASO. In 1 patient the aortic and pulmonary valves were both bicuspid, and subvalvular myomectomy was required to relieve the LVOT obstruction. One patient required excision of accessory mitral valve tissue that was causing LVOT obstruction. Another patient required pulmonary valvotomy, and associated subvalvular obstruction had to be surgically corrected with subvalvular myomectomy. Eleven patients needed preoperative balloon atrial septostomy, and because of late presentation, 2 patients had pulmonary artery banding with a Blalock-Taussig shunt to prepare the left ventricle for the switch procedure. Concomitant surgical procedures are summarized in Table 2. The chest incision was electively left open in 6 patients at the time of operation. with both left anterior descending and circumflex arteries arising from sinus 1 and right coronary artery arising from sinus 2. Four patients had both circumflex and right coronary arteries arising from sinus 2. Coronary anatomy was not recorded for 2 patients. Three patients had pulmonary valve stenosis: 1 with isolated, 1 with total anomalous pulmonary venous drainage, and 1 with left ventricular outflow tract (LVOT) obstruction. Another patient had isolated LVOT obstruction due to accessory mitral valve tissue but without valvular stenosis (Table 1). Two patients with an intact ventricular septum presented late and underwent a twostage repair. No patients in our study had coarctation of the aorta. Statistical Analysis Descriptive statistics were computed that included frequencies and proportion for categoric variables and means and standard deviation for continuous variables. Repeated measures analysis of variance was used to study change over time in aortic valve diameter size. We defined patient freedom from reintervention as the time from when patients had their first operation to the time of reintervention or most recent follow-up. The probabilities for patient freedom from reintervention were estimated by using the Kaplan-Meier method. Surgical Procedure The ASO was performed with standard aortic and bicaval venous cannulation, hypothermia to 24 C, and cold blood antegrade cardioplegia. Circulation had to be arrested for variable periods of 2 to 18 minutes in 7 patients. The Follow-Up Patients were regularly monitored by pediatric cardiologists in the outpatient clinic, and clinical and detailed echocardiographic assessments were recorded. A single operator reviewed all available echocardiograms. Neo- Table 2. Procedures Procedures No. of Patients Preoperative Balloon atrial septostomy, total 11 IVS D-TGA 6 VSD D-TGA 5 Atrial septectomy PA band BT shunt 2 Concomitant procedures (ASO ) ASD/PFO 21 VSD patch closure 5 VSD primary closure 4 PA debanding, BT shunt take down 2 Relief LVOTO 2 Pulmonary valvotomy 2 Repair cleft MV 1 Repair TAPVD 1 LTGA with left AV valve repair (double switch) 1 Delayed chest closure 6 ASD atrial septal defect; BT Blalock-Taussig; D-TGA dextrotransposition of the great arteries; IVS intact ventricular septum; L-TGA levotransposition of the great arteries; PA pulmonary artery; PFO patent foramen ovale; TAPVD total anomalous pulmonary venous drainage; VSD ventricular septal defect.

Ann Thorac Surg KHAN ET AL 2008;85:179 85 NEOAORTIC BICUSPID VALVE IN ASO aortic valve annulus diameter, change in the indexed aortic annulus diameter, and valve function were serially assessed. The neoaortic regurgitation was graded on a scale of 0 to 4 (0, no regurgitation; 1, trivial; 2, mild; 3, moderate; 4, severe) according to the width of jet at the LVOT, pressure half time of regurgitant jet measured by continuous wave Doppler, and the presence of flow reversal in the abdominal aorta. All patients included in the assessment had at least two postoperative echocardiograms. In general, echocardiograms were done in all patients at 1 month, at 1 year, and every 2 years subsequently. The echocardiographic data were not digitally stored initially in our hospital; hence, the number of patients with available data for review was limited. The first two postoperative and the last follow-up echocardiogram were included in the analysis, where available. Results Two of 24 patients were lost to follow-up, and data were complete on 22. One patient died late of an unknown cause, and 21 patients were alive at the last follow-up. There was no hospital mortality. The median intensive care unit (ICU) stay was 6 days (range, 3 to 69 days). The 6 patients who left the operating room with an open chest incision did well, and the incision was closed in the ICU between 1 and 5 days postoperatively. Two patients needed permanent pacemakers. Mean follow-up was 5.3 years (range, 2 months to 13 years). Patients were censored once they were lost to follow-up or died. Echocardiogram results showed similar indexed aortic annular diameter to body surface area at the initial and late follow-up, a marker of growth (Figs 1 and 2). Mean indexed neoaortic valve diameter was 23.4 mm/m 2 at the initial follow-up and 22 mm/m 2 at the latest follow-up. Repeated measure analysis of mean aortic valve diameter against time selected in general at 1, 2, and 4 years during follow-up was done for all patients who had complete data for three follow-up visits. This showed significant linear growth in diameter by time (p 0.001; Fig 2). No significant valve stenosis developed in any patient. The average peak instantaneous gradient across the aortic valve was 9 mm Hg (range, 4 to 30 mm Hg). Seven patients had no neoaortic valve regurgitation, and 10 patients had trivial aortic insufficiency. No significant progression of the aortic regurgitation has been noted in 17 patients during the follow-up period (Fig 3). Severe neoaortic valve regurgitation developed in 2 patients. In 1 patient this occurred after endocarditis in which the valve was repaired, and the other patient had a Taussig-Bing anomaly repair after a previous palliation by a pulmonary artery band. He needed aortic valve replacement with a size 21-mm bileaflet mechanical prosthesis 4 years after the ASO. None of the patients with relief of LVOT obstruction or relief of pulmonary valve stenosis presented with more than trivial neoaortic regurgitation. Supravalvular pulmonary stenosis developed in 3 patients. Balloon dilatation was performed with adequate results in 2 patients, but the third required surgical relief. At 10 years, freedom from reoperation for neoaortic valve dysfunction was 88% (Fig 4), and freedom from reintervention due to any cause was 74.7% (Fig 5). Comment 181 The ASO is considered to be the treatment of choice for transposition of the great arteries. It permits the definitive repair in neonates or during early infancy, thus avoiding the deleterious effect of cyanosis on the systemic organs [5]. It is generally accepted that the basic requirement for the ASO includes competent and unobstructed ventriculoarterial connections [6]. Pulmonary valve and LVOT abnormalities have not been identified as an absolute contraindication to the ASO [6] provided they can be satisfactorily corrected surgically. The anatomic pulmonary valve, which has thin leaflets with little elastic tissue in the normal heart, is expected to function as the neoaortic valve after the ASO [7]. The long-term function of this valve in the systemic circulation is not known. The aortic and pulmonary valves are morphologically and histologically indistinguishable at birth [8, 9]. Distinct histologic and microscopic changes, presumed secondary to the differences in pulmonary artery and systemic pressures, occur postnatally in the normal heart, resulting in a thin, delicate pulmonary valve with decreased collagen and elastic tissues compared with the normal aortic valve [10, 11]. Neoaortic regurgitation has been reported in up to 41% of patients after an ASO [6]. Schmid and colleagues [12] have shown that neoaortic regurgitation is most likely functional in nature and is attributed to the acutely increased hemodynamic demands. Neoaortic regurgitation is probably the result of a multifactorial process, of which sinotubular geometry and surgical technique are some of the mechanics proposed. A previous study from Formigari and colleagues [13] has proposed that the trapdoor method used for coronary reimplantation has a strong predictive factor for neoaortic regurgitation. Other investigators, however, could not establish any association between the method of coronary transfer and the onset of neoaortic regurgitation [14]. The high prevalence of aortic regurgitation raises the question of adequacy of the anatomic pulmonary valve to support the high-pressure systemic circulation during the patient s lifetime [7]. One study has reported that neoaortic regurgitation may indeed develop progressively over time, reaching a 30% incidence 6 years after the ASO [13]. Other studies showed mild regurgitation in about 35% of patients, and moderate-to-severe regurgitation was demonstrated in 5% or fewer patients [15 18]. This was also noted after operations such as the Ross, Damus-Kaye-Stansel, and Norwood procedures, where the pulmonary valve is functioning in the systemic circulation. In our study the prevalence of bicuspid pulmonary valve was 7%, which is higher than the reported inci- CARDIOVASCULAR

CARDIOVASCULAR 182 KHAN ET AL Ann Thorac Surg NEOAORTIC BICUSPID VALVE IN ASO 2008;85:179 85 dence of 1.2% of patients undergoing an ASO [6]. It therefore becomes important for our practice to evaluate the acceptable long-term performance of a bicuspid pulmonary valve in the aortic position. The question here is whether we can extrapolate the known natural history of a bicuspid aortic valve to a bicuspid pulmonary valve function in the systemic circulation. The bicuspid aortic valve is more susceptible to fibrosis and calcification, leading to aortic stenosis and mixed aortic valve disease in hearts with normally related great arteries. It is also known that the bicuspid pulmonary valve is more susceptible to degenerative changes in its native position. Putting these facts together, it is expected that a bicuspid pulmonary valve will function poorly in the aortic position, which was not the case in our experience at medium-term follow-up. Sohn and colleagues [6] found the function of neoaortic bicuspid valves to be similar to that of trileaflet valves after an ASO. Although the number of patients in that study was very small, our study nevertheless confirms these findings. We also concur with Uemura and colleagues [8], who conclude that the ASO remains an option of choice for patients with an initially bicuspid pulmonary valve, provided there is no severe subpulmonary stenosis [8] or that it can be surgically corrected. Gradients across the pulmonary valve can be due to a true anatomic lesion or may be related to excessive flow in the presence of abnormal shunts. Gradients due to hyperdynamic flow should not need intervention and usually resolve once abnormal shunts are repaired. Easily resectable subvalvular obstructions, such as those caused by membranes or accessory mitral valve tissue, Fig. 1. (A) Indexed aortic valve area. Progression in the size of the neoaortic valve indexed to body surface area (BSA). (B) Absolute aortic valve area. Progression in the actual size of the neoaortic valve.

Ann Thorac Surg KHAN ET AL 2008;85:179 85 NEOAORTIC BICUSPID VALVE IN ASO 183 Fig. 2. Change in neoaortic valve size. Repeated measure analysis of mean aortic diameter indicating linear growth in relation to time (p 0.001). Selected follow-up visits at 1, 2, and 4 years after operation. (BSA body surface area.) CARDIOVASCULAR are not contraindications to the ASO. Pulmonary valvotomy in a pulmonary valve that otherwise appears satisfactory also does not preclude the ASO. Four of our patients had anatomic obstructions at the valvular and subvalvular levels, and these were successfully relieved with good outcome. True anatomic obstructions therefore have to be individually evaluated for feasibility of surgical relief and should not imperatively be a deterrent to an ASO. The bicuspid pulmonary valves in our patients have remained satisfactory in their function, with minimal regurgitation that did not appear to progress during the follow-up period. One of the patients who needed reoperation because of severe neoaortic regurgitation had endocarditis, and this was assumed to be responsible for progression of regurgitation. The other patient had a Taussig-Bing anomaly and was palliated initially by a pulmonary artery band that resulted in a much dilated pulmonary root and, hence, regurgitation. This was in concordance with observation made by other investigators. The Taussig-Bing anomaly and pulmonary artery banding are known risk factors for neoaortic regurgitation [19]. It is reassuring that growth of the aortic annulus that is proportional to somatic growth can be demonstrated by this study. Unlike other studies, our patients did not demonstrate progressive neoaortic root dilatation but rather a true growth. This feature was also demonstrated in our ASO series as well as in young patients undergoing the Ross procedure in which the pulmonary autograft was implanted as a full root [20]. This is a retrospective analysis with all its inherent drawbacks. Echocardiographic data were not digitally stored before 1998 in our institute. Retrospective measuring and analysis were not found to be consistent and reproducible and, hence, were not included in the analysis. This has been reflected by having echocardiographic data on a fewer number of patients. In conclusion, the bicuspid pulmonary valve in this small series of patients was quite stable in the neoaortic position after the ASO. At medium-term follow-up, it Fig. 3. Change of neoaortic regurgitation (AR) from immediate postoperatively to the last follow-up visit. (Ao. V aortic valve.) Fig. 4. Freedom from neoaortic valve intervention. (Line shows survival function; tick mark, censored data.)

CARDIOVASCULAR 184 KHAN ET AL Ann Thorac Surg NEOAORTIC BICUSPID VALVE IN ASO 2008;85:179 85 Fig. 5. Freedom from all-cause reintervention. (Line shows survival function; tick mark, censored data.) does not show exaggerated valve dysfunction requiring repeat operation and valve replacement. The change in indexed neoaortic valve annulus correlates with somatic growth. Patients with a bicuspid pulmonary valve requiring the ASO should not be committed to suboptimal surgical alternatives on the basis of valve status. We would like to thank Gamal El Din Hassan, PhD, Department of Biostatistics, Epidemiology and Scientific Computing at King Faisal Specialist Hospital and Research Centre for his contribution in the statistical analysis of the study. References 1. Keane MG, Wiegers SE, Plappert T, Pochettino A, Bavaria JE, Sutton MG. Bicuspid aortic valves are associated with aortic dilatation out of proportion to coexistent valvular lesions. Circulation 2000;102(suppl III):35 9. 2. Nistri S, Sorbo MD, Martin M, et al. Aortic root dilatation in young men with normally functioning bicuspid aortic valve. Heart 1999;82:19 22. 3. Russo CF, Mazzetti S, Garatti A. Aortic complications after bicuspid aortic valve replacement: long-term results. Ann Thorac Surg 2002;74:S1773 6. 4. Fedek P, Verma S, David TE. Clinical and pathophysiological implications of a bicuspid aortic valve. Circulation 2002;106: 900 4. 5. Uemura H, Kawashima T, Yamamoto F. A bicuspid pulmonary valve is not a contraindication for arterial switch operation. Ann Thorac Surg 1995;59:473 6. 6. Sohn YS, Brizard CP, Cochrane AD, Wilkinson JL, Mas C, Karl TR. Arterial switch in hearts with left ventricular outflow and pulmonary valve abnormalities. Ann Thorac Surg 1998;66:842 8. 7. Jenkins KJ, Hanley FL, Colan SD. Function of the anatomic pulmonary valve in the systemic circulation. Circulation 1991;84(suppl III):III 173 9. 8. Hurle JM, Colvee E. Changes in the endothelial morphology of the developing semilunar heart valves. Anat Embryol (Berl) 1983;167:67 83. 9. Maron BJ, Hutchins GM. The development of semilunar valves in the human heart. Am J Pathol 1974;74:331 44. 10. Gikonyo BM, Lucas RV, Edwards JE. Anatomic features of congenital pulmonary valvar stenosis. Pediatr Cardiol 1987; 8:109 15. 11. Gross L, Kugel MA. Topographic anatomy and histology of the valves in the human heart. Am J Pathol 1931;7:445 81. 12. Schmid FX, Hilker M, Kampann C. Clinical performance of the native pulmonary valve in the systemic circulation. J Heart Valve Dis 1998;7:620 5. 13. Formigari R, Toscano A, Giardini A. Prevalence and predictors of neoaortic regurgitation after arterial switch operation for transposition of great arteries. J Thorac Cardiovasc Surg 2003;126:1753 9. 14. Hwang HY, Kim WH, Kwak JG. Mid-term follow up of neoaortic regurgitation after the arterial switch operation for transposition of the great arteries. Eur J Cardiothorac Surg 2006;29:162 7. 15. Raja SG, Shauq A, Kaarne M. Outcomes after arterial switch operation for simple transposition. Asian Cardiovasc Thorac Ann 2005;13:190 8. 16. Imamura M, Drummond-Webb JJ, McCarthy JF, Mee RB. Aortic valve repair after arterial switch operation. Ann Thorac Surg 2000;69:607 8. 17. Yoshizumi K, Yagihara T, Uemura H. Approach to the neoaortic valve for replacement after the arterial switch procedure in patients with complete transposition. Cardiol Young 2001;11:666 9. 18. Formigari R, Toscano A, Giardini A, et al. Prevalence and predictors of neoaortic regurgitation after arterial switch operation for transposition of the great arteries. J Thorac Cardiovasc Surg 2003;126:1753 9. 19. Losay J, Touchot A, Capderou A, et al. Aortic valve regurgitation after arterial switch operation for transportation of the great arteries. J Am Coll Cardiol 2006;47:2057 62. 20. Al Halees Z, Pieters F, Qadoura F, Shahid M, Al Amri M, Al Fadley F. Ross procedure: the procedure of choice for congenital aortic valve disease. J Thorac Cardiovasc Surg 2002;123:437 42. INVITED COMMENTARY A bicuspid pulmonary valve is sometimes found in association with transposition of the great arteries (TGA). After the arterial switch operation the pulmonary valve will become the new aortic valve; therefore it is important to know the fate of a bicuspid pulmonary valve in the systemic circulation. Khan and colleagues [1] observed a 7% incidence of bicuspid pulmonary valve in their series of 391 arterial switch operations. For unknown reasons this is higher than the prevalence that is reported in other series (ie, 3.2% in Leiden, 1.2% in Melbourne, and 1.7% in Boston [2 4]). It is reassuring for the moment that in all these series the majority of bicuspid pulmonary valves remains with good function after the arterial switch procedure. Insufficiency is usually no more than trivial or mild and does not seem to progress [1 5]. However, maximal follow-up durations that are reported are no longer than 11 to 20 years, and it is not unthinkable that the bicuspid pulmonary valve may show loss of function at a later stage. 2008 by The Society of Thoracic Surgeons 0003-4975/08/$34.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2007.08.065