Surgical Management of Aortopulmonary Window and its Associated Cardiac Lesions

Introduction

Aortopulmonary window (APW) is a congenital cardiac anomaly that is the result of incomplete development of the conotruncal septum, characterized by a communication between the main pulmonary artery (PA) and the ascending aorta in the presence of 2 separate semilunar valves. ¹ Prevalence is 0.2% to 0.6% of all congenital heart diseases. ² APW can occur in association with other cardiac lesions in 50% of the cases. ³ The most common associated lesion is interrupted aortic arch (IAA), followed by ventricular septal defect. ¹ The literature is limited to cohort studies with small number of patients due to rarity of this disease.^4–8 Presence of an additional cardiac anomaly was associated with early death and reoperation in these studies. The aim of this study was to review our institutional experience with patients who underwent surgical repair of APW either isolated or in association with other cardiac anomalies.

Materials and Methods

Results

From January 2006 to December 2020, 183 patients underwent surgical repair of APW at our institute. Sixty-three patients were included in Group 1 (APW with associated anomalies), 120 patients were included in Group 2 (isolated APW). Median ages in Group 1 and 2 were 10 months and 6 months, respectively. Preoperative baseline variables are shown in Table 1. Most common type of APW was type 1, followed by type 2 and then type 3 in both the groups. CT images showing types of APW are shown in Figure 1A to C. Right aortic arch was seen in 9.8% patients. The most common associated lesion was VSD (in 29 patients). A total of 72 associated lesions were seen in 63 patients. IAA was seen in 8 patients (IAA type A in 6 patients and type C in 2 patients). Subaortic membrane was seen in 11 patients, left pulmonary artery (LPA) origin stenosis and anomalous origin of the right coronary artery (RCA) from the pulmonary artery (ARCAPA) in 3 patients each. Severe mitral regurgitation, double outlet right ventricle, coarctation of the aorta, pulmonary atresia, and absent LPA were seen in 2 patients each. Non-confluent branch pulmonary arteries were seen in two cases (main PA continuing as right pulmonary artery [RPA] with the LPA arising from the distal ascending aorta in 1 case and from the distal aortic arch in another, where APW was present between RPA and ascending aorta). Anomalous origin of RCA, tricuspid atresia with hypoplastic right ventricle, tetralogy of Fallot with hypoplastic pulmonary arteries, cortriatrium, aberrant left subclavian artery from main PA, and D-transposition of great arteries were seen in 1 patient each.

OPEN IN VIEWER

Figure 1.

(A) Axial CT image showing type I aortopulmonary window (AP: aortopulmonary; MPA : main pulmonary artery); (B) coronal CT image showing type 1 aortopulmonary window; and (C) coronal CT image showing type 2 aortopulmonary window.

OPEN IN VIEWER

Table 1.

Baseline Variables.

Baseline variables			Group 1 (N = 63)	Group 2 (N = 120)	P value
Age (months)		Median (IQR)	10 (4-30)	6 (3-24)	.393
Weight (kg)			6 (4-9)	5.05 (4-7.85)	.617
Height (cm)		Mean ± SD	74.54 ± 22.17	72.18 ± 23.06	.49
Gender	Male	Number (percentage)	39 (61.9)	83 (69.2)	.328
	Female		24 (38.1)	37 (30.8)
Type of aorto-pulmonary window	Type 1		35 (55.6)	82 (68.3)	.189
	Type 2		19 (30.2)	23 (19.2)
	Type 3		9 (14.3)	15 (12.5)
Arch	Left		56 (88.9)	109 (90.8)	.675
	Right		7 (11.1)	11 (9.2)

Operative Details

Majority of the patients underwent repair using the sandwich patch technique ¹⁰ with treated autologous pericardial patch. Direct anterior APW approach was most commonly used in both the groups, whereas when trans-aorta approach was required, it was more commonly used in Group 1 as compared to Group 2; (p = 0.003). Mean cardiopulmonary bypass time and aortic cross-clamp time were 70 and 36 min in Group 2, as compared to 112 and 65 min, respectively, in Group 1 (both being significantly shorter in Group 2 [P < .001]). Delayed sternal closure was done in 6.3% patients in Group 1 as compared to 0.8% in Group 2; it was significantly higher in Group 1 as compared to Group 2 (P = .049). Additional procedures were done concomitantly in the single stage, depending on associated anomalies. Operative details are given in Table 2.

OPEN IN VIEWER

Table 2.

Intraoperative Variables.

Intraoperative variables			Group 1 (N = 63)	Group 2 (N = 120)	P value
Approach	Aortopulmonary window		48 (76.2)	114 (95)	.003
	Aorta		9 (14.3)	5 (4.2)
	Pulmonary artery		3 (4.8)	1 (0.8)
	Aorta and pulmonary artery		1 (1.6)	0 (0)
	Other		2 (3.2)	0 (0)
CPB time		Mean ± SD	112.35 ± 44.82	64.6 ± 25.52	<.001
AXC time			70.57 ± 28.36	35.75 ± 15.23	<.001
Delayed sternal closure		Number (percentage)	4 (6.3)	1 (0.8)	.049

Bold font indicates statistical significance.

Concomitant procedures: The most common concomitant procedure performed was ventricular septal defect closure in 29 patients. Other commonly performed procedures included IAA repair with end-to- side anastomosis of the descending thoracic aorta to the arch under SACP in 8 patients, and subaortic membrane excision in 11 patients. Intraventricular tunnel repair, midline coarctation repair with extended resection and end-to-side anastomosis, unifocalization of LPA to main PA, modified Blalock-Taussig shunt, arterial switch operation, LPA origin plasty, and reimplantation of RCA to the aorta were performed in 2 patients each. Mitral valve repairs were done in 3 patients (creation of neochordae in 1 and posterior commisuroplasty in 2 patients). ARCAPA with APW repair was done with 2-patch technique in 1 patient. Cortriatrium repair, ligation of aberrant left subclavian artery, and main PA plasty with central shunt were done in 1 patient each.

Postoperative Details

Early mortality occurred in 9 patients (4.91%) overall, including 8 patients in Group 1 and 1 patient in Group 2 (P = .001). Causes of early mortality were low cardiac output (4 patients), sepsis (2 patients), PAH crisis (2 patients), and acute respiratory distress syndrome (1 patient). The single patient who had an early mortality in Group 2 had congenital tracheal stenosis for which tracheostomy and later tracheal reconstruction was done, however he expired due to sepsis. Duration of ventilation (P = .12), postoperative ICU stay (P = .132), hospital stay (P = .076), and prolonged ventilation requiring tracheostomy (P = .184) were not significantly different between the groups. Respiratory distress requiring reintubation occurred more frequently in Group 1 (15.9%) as compared to Group 2 (4.2%) (P = .006). Two patients had sepsis in Group 1 (both had candida in urine culture) whereas 1 patient in Group 2 had sepsis (E coli in endotracheal aspirate culture) (P = .258). ECMO support was not significantly different in Group 1 (2 patients) as compared to in Group 2 (2 patients) (P = .609). Postoperative details are given in Table 3. On univariable analysis, cardiopulmonary bypass time (P < .001), aortic cross-clamp time (P < .001), delayed chest closure (P = .02), sepsis (P = .006), tracheostomy (P = .002), ECMO support (P < .001), associated lesions (P = .001), and PA hypertensive crisis (P < .001) were associated with early mortality. On multivariable analysis, only PA hypertensive crisis was identified as an independent risk factor for early mortality (P = .03; odds ratio = 24). Details of patients who had an early mortality are given in Table 4. Among all the patients, 74 (40.4%) had presented beyond infancy; however, mortality did not differ significantly as compared to infants (P = .316). Six out of the 9 early mortalities occurred in the first half of the study period; however, it was not significantly lower in the second half of the study period (P = .182).

OPEN IN VIEWER

Table 3.

Postoperative Variables.

Postoperative variables		Group 1 (N = 63)	Group 2 (N = 120)	P value
Duration of ventilation (hrs)	Median (IQR)	48 (30-75)	44 (24-57)	.12
ICU stay (days)		6 (4-10)	5 (4-7.25)	.132
Hospital stay (days)		11 (8-21)	10 (7-14)	.076
Re-intubation	Number (percentage)	10 (15.9)	5 (4.2)	.006
Pulmonary hypertensive crisis		5 (7.9)	1 (0.8)	.019
Tracheostomy		4 (6.3)	2 (1.7)	.184
Sepsis		2 (3.2)	1 (0.8)	.273
ECMO		2 (3.2)	2 (1.7)	.609
Hospital mortality		8 (12.7)	1 (0.8)	.001

Bold font indicates statistical significance.

ECMO, extracorporeal membrane oxygenation.

OPEN IN VIEWER

Table 4.

Early Mortality Details.

Age (months)	Weight (kg)	Associated lesion	Delayed chest closure	Cause of mortality
6	3.5	Interrupted aortic arch	No	Low cardiac output
4	3.5	Ventricular septal defect	Yes	Sepsis
10	7	Tracheal stenosis	No	Sepsis
50	13	Coarctation of aorta	No	Pulmonary hypertensive crisis
5	5.5	Ventricular septal defect	No	Pulmonary hypertensive crisis
1	3.5	Ventricular septal defect with aberrant left subclavian artery	No	Low cardiac output
7	4	DORV, d-Transposition of Great arteries, multiple ventricular septal defect	Yes	Low cardiac output, ECMO
21	12	Cortriatrium, hypoplastic left lung	No	Acute respiratory distress syndrome
11	6.5	Ventricular septal defect	No	Low cardiac output, ECMO

DORV, double outlet right ventricle; ECMO, extracorporeal membrane oxygenation.

Follow-up Details

Follow-up details were available in 69% patients with a mean and median of 60.2 months and 54 months, respectively (range, 2-175 months). Late mortality occurred in 2 patients in Group 1: due to congestive cardiac failure 53 months after primary surgery in 1 patient, and in another due to infective endocarditis with mediastinitis within 1 month of primary surgery. Survival at both 5 years and 8 years was 77% ± 6.5 in Group 1 and 98.8% ± 1.2 in Group 2. Overall survival was significantly higher in Group 2 as compared to Group 1 (P ≤ .001) (Figure 2A). Among the patients who had IAA and were discharged (7 patients), 1 was lost to follow-up and the remaining 6 did not have aortic obstruction or supravalvar aortic/pulmonary stenosis requiring reintervention at a mean duration of 75.3 months after the primary surgery. Two patients in Group 1 underwent subsequent surgical reintervention; one for relief of biventricular outflow tract obstruction and one for intracardiac repair with takedown of palliative central shunt, at 20 months and 27 months, respectively, after their initial surgery that included APW repair. None of the patients from Group 2 required reintervention. Freedom from reintervention at both 5 years and 8 years was 92.4% ± 5.2 in Group 1 and 100% in Group 2 (P = .055) (Figure 2B). Analysis comparing the characteristics of patients who were followed-up and those who were lost to follow-up showed no significant differences (Supplemental Tables 1, 2, 3).

OPEN IN VIEWER

Figure 2.

(A) Survival in both the groups and (B) freedom from reintervention in both the groups.

Comment

APW is a rare congenital cardiac anomaly. It was first described in an autopsy study by Elliotson in 1830. ¹¹ Clinical presentation of APW is similar to large patent ductus arteriosus with excessive left-to-right shunt. Most patients present early in life. Previously, cardiac catheterization was an integral part of the diagnostic evaluation for APW. Currently, echocardiography has become the mainstay in diagnosing APW. Early diagnosis and surgical repair have become the mainstay of treatment, as these patients can develop pulmonary vascular disease rapidly. APW has been known to occur either as an isolated lesion or in association with other cardiac anomalies.^7,8 In our study, 34.4% of the patients had an associated cardiac anomaly. Over the years, there have been reports of several single institution series patients undergoing APW repair.^4–8 To our knowledge, our series is the largest single institution study of APW repair to date. In older series, the mortality was linked to the late presentation of patients with development of pulmonary vascular obstructive disease and presence of associated cardiac anomalies. Late presentation continues to be an issue in parts of the developing world, although much less so than was true during the previous 2 decades. The median age of our patients was 7 months with a range of 10 days to 19 years. 40.4% of our patients were beyond infancy (age >1 year). However, presenting beyond infancy was not associated with early mortality (P = .311); as early mortality was seen only in 2.7% of the patients presenting beyond infancy. This is likely related to many of these patients having smaller defects which were associated with a lesser initial degree of left-to-right shunting that allowed them to survive and present later for surgery, as compared to those who presented early in life. These patients, however, may persist in having some degree of PA hypertension and/or pulmonary vascular reactivity despite abolition of the shunt, thus we managed them with oral pulmonary vasodilators postoperatively for 6 months to 2 years, depending on the findings on serial echocardiography.

The other cardiac anomalies that have most commonly been reported in association with APW are ventricular septal defect and IAA. ⁸ In our study, we found VSD as the most common associated anomaly (29 patients), and 8 patients had IAA. A study from the Congenital Heart Surgeons’ Society examining outcomes of 472 patients with IAA identified concomitant APW in 20 patients. ³ Among those patients, 19 underwent surgical repair, with 3 operative deaths and 10-year survival of 84%. We experienced 1 death out of 8 patients with IAA in our series, and 1 patient was lost to follow-up. However, none of the remaining 6 patients had a late mortality nor did they require any reintervention, with a mean follow-up duration of 75.3 months.

We experienced an overall early mortality of 4.9% (9 out of 183 patients). Early mortality for the simple APW group was 0.8% (1 out of 122). The single mortality occurred in a patient who had congenital tracheal stenosis and required tracheostomy and then underwent tracheal resection, who eventually expired due to sepsis. Our results remain consistent with recent literature that suggests very low rates of early mortality for repair of simple APW, ¹² whereas operative mortality for complex APW is 10% to 15%. ¹³ A recent study by Alsoufi and colleagues ⁸ described their experience of 40 patients with both simple and complex APW, with 0% early mortality. We identified CPB time (P < .001), aortic cross-clamp time (P < .001), delayed chest closure (P = .02), sepsis (P = .006), tracheostomy (P = .002), ECMO support (P < .001), associated lesions (P < .001), and PA hypertensive crisis (P < .001) as factors associated with early mortality on univariable analysis. On multivariable analysis, only PAH crisis was identified as an independent risk factor for mortality within 30 days following surgery (P = .03; odds ratio = 24).

Overall survival of our patients was 90.7% at 8 years with 97.3% freedom from reintervention with a mean follow-up duration of 60 months. Other reports have shown an overall survival of 81% to 100% at 10 years.^4–8,12,13 Persisting pulmonary hypertension may be a risk factor for late death.^4,5 Concomitant IAA has also been associated with increased risk of late deaths with a reported survival of 84% after initial admission. ³ A few other studies reported no late deaths associated with IAA.^8,14 IAA associated with APW has been associated with late reinterventions, predominantly for aortic obstruction. ⁸ Regarding the surgical technique for APW with IAA, we perform the repair with direct end-to-side anastomosis of the descending aorta to the proximal arch, relying on extensive dissection of the descending aorta, the arch and its vessels, and the ascending aorta to achieve the arch reconstruction without requiring a patch. We did not experience any late mortality or reintervention in patients with IAA, and would attribute it to our use of this technique. In view of the low number of events (only 1 late mortality), univariable analysis alone was used to assess factors potentially associated with late mortality. None were found to be significant.

Surgical techniques to accomplish repair of APW include simple ligation, ¹⁵ division and suturing, ¹⁶ transpulmonary approach, ¹⁷ transaortic direct closure, ¹⁸ polyester patch closure via the transaortic approach, ¹⁹ and the sandwich technique. ¹⁰ Simple ligation or division was performed earlier for small APW in our institute. Currently, however, for all APWs we perform patch closure on CPB. Thus, there were no patients treated with simple ligation or division in the cohort that is the subject of this study. With regards to the sandwich technique usage, the incision should be made precisely on the aortopulmonary groove over the APW covering half the circumference of APW. The patch is stitched at the exact margin between aorta and PA on the PA side of the APW, taking care not to narrow the RPA and also not to injure the left coronary artery ostium which can be perilously close to the margin especially in cases of a large APW. We did not experience any instances of ascending aortic or PA stenosis in our series, attesting to the efficacy of this technique. Complications that were more prevalent in the complex APW group as compared to the simple APW group include postoperative respiratory distress requiring reintubation (P = .008) and PAH crisis (P = .016). We observed a lower rate of early mortality in the latter half of our study duration (although the difference was not statistically significant). We believe that this may be related in part to the increased use of pulmonary vasodilators (Sildenafil and/or inhaled nitric oxide) in the immediate postoperative period for patients susceptible to PAH crisis. We believe that in general, proceeding with repair at the time of diagnosis is appropriate and is likely to contribute to optimal results.

Limitations

This study is subject to the usual limitations of a retrospective study. In particular, the fact that 31% of patients were lost to follow-up is a major limitation of this study. Notably however, comparative analysis of preoperative, intraoperative, and immediate postoperative variables among patients with follow-up and among patients who were lost to follow-up, revealed no significant differences between these 2 groups with respect to any of these variables.

Conclusion

Early outcomes of APW repair are excellent in cases of isolated APW as compared to those with associated lesions. Long-term outcomes, in terms of survival and freedom from reoperation, are excellent in both the groups. Early repair should be advocated in order to optimize outcomes.

Supplemental Material

Supplemental Material

Supplemental Material