The Effect of Surgical Technique, Age and Trisomy 21 on Early Outcome of Surgical Management of Complete Atrioventricular Canal Defect

Background: The optimal timing, surgical technique and the inuence of Trisomy 21 on outcome of surgical repair of Complete Atrio-Ventricular Canal Defect (CAVC) remain uncertain. We reviewed our experience in repair of CAVC to identify the inuence of these factors on operative outcome. Methods: A prospective study included 70 patients who underwent repair of CAVC at our institute between July 2016 to October 2019. Primary endpoint was mortality and secondary endpoint was degree of left atrio ‐ ventricular valve (LAVV) regurgitation. Results: No signicant difference noted between patients operated on, at the rst 6 months of age versus later, regarding mortality or LAVV regurgitation. Surgical repair by modied single patch technique showed signicant reduction in bypass time (71.13 ± 13.507 min vs 99.19 ± 27.092 min, P value =0.001). Comparison to closure of cleft only, posterior annuloplasty used for repair of LAVV resulted in signicant reduction in the occurrence of post-operative valve regurgitation during the early period (LAVV 2+ 43 % vs 7 %, P value=0.03) and at 6 months of follow up (LAVV 2+ 35.4 % vs 0 %, P value=0.01) respectively. Conclusion: Early intervention, in the rst 6 months in patients with CAVC by surgical repair gives comparable acceptable results to later repair; Trisomy 21 was not found to be a risk factor for early intervention. Repair of common AV valve by cleft closure with posterior LAVV annuloplasty showed better results with signicant decrease in postoperative LAVV regurgitation and early mortality in comparison to closure of cleft only.


Introduction
Complete Atrio-Ventricular Canal Defect (CAVC) is a complex cardiac malformation in which there is a variable de ciency of the atrioventricular (AV) junction in the developing heart (1). This anomaly is characterized by the atrial, ventricular and atrioventricular septa, and the left (LAVV) and right (RAVV) atrioventricular valves (AVV) (2).
During the 1970s and early 1980s, surgeons advocated palliation with pulmonary artery banding to delay surgical repair beyond infancy, aiming to prevent more elevation of pulmonary artery pressure (3). On the other hand, as operative techniques and postoperative care improved and cardiopulmonary bypass in infants became safer (4,5), most cardiac centres abandoned this practice and opted for elective repair before 6 months of age in asymptomatic infants with CAVC (6, 7,8,9). Delay beyond this age has recently been shown to be an incremental risk factor for death and reoperation for LAAV regurgitation (6, 7,8,9). Infants with failure to thrive and with symptoms of heart failure requiring multiple medications should undergo surgery as soon as possible to permit normal growth and development (10).
Two standard methods for repair of CAVC defect are used: rst, is the single-patch technique, where the same patch is used for closure of the atrial septal defect (ASD) and ventricular septal defect (VSD), the AVV lea ets are re-suspended onto this single patch, with a need for lea et division. This allows patch placement for Rastelli type C, which by de nition has a common superior lea et (7,11,12). Second, is the double-patch technique; 2 separate patches are used, one for closure of the VSD and the other for closure of the ASD; the AVV lea ets are typically not divided (13,14,15). In the mid-1990s, the modi ed singlepatch technique was introduced, which is characterized by the obliteration of interventricular

Patients And Methods
A prospective study in which 70 patients with a CAVC underwent de nitive surgical repair from July 2016 to October 2019 at Cardio-thorathic Surgery Department Paediatric Unit, Ain Shams University Hospitals.
Follow-up was done for a duration of 6 months post-surgery through regular outpatient visits with recent echocardiography. The study was approved with waiver of consent by the Institutional Review Boards of the University of Ain-Shams.
All patients were operated on via an elective basis; age ranged between 2 months and 24 months, and cases included Trisomy 21 and Non-Trisomy 21 patients. The selection criteria excluded patients refusing to participate in the study, patients who underwent previous pulmonary artery banding, patients with an incomplete canal defect (partial atrioventricular canal defect), patents with unbalanced complete atrioventricular canal defect not suitable for biventricular repair, patients with cono-truncal complex anomalies (including tetralogy of Fallot, double-outlet right ventricle and pulmonary stenosis) and patients with a single ventricle.
All patients had preoperative echocardiography and multi-slice CT chest to con rm diagnosis and exclude any associated complex anomalies. Surgical repair was performed through median sternotomy, using cardiopulmonary bypass and moderate hypothermia with antegrade cardioplegia to achieve diastolic arrest. This was followed by right atriotomy and performing the repair.
Two techniques were used for repair as follows: The modi ed Single-patch technique (Australian technique), similar to that described by Nunn and colleagues (17), was used in 39 cases (55.7%) and a two-patch technique was used for closure of septal defects in 31 cases (44.3%). Gortex patches were used for the ventricular component and autologous pericardium for the atrial defect. In 40 patients (57.1%) cleft was closed and posterior annuloplasty was added to the repair. Intraoperative testing for competency of the left AVV was done by injecting saline in the left ventricle.
Following completion of the repair, closure of right atriotomy and de-airing of the left side of the heart was performed through the aortic root, followed by weaning from cardiopulmonary bypass, routine haemostasis and closure in layers.
The following data were recorded; age, weight, Trisomy 21 or Non-Trisomy 21 patient, cross-clamp time, total bypass time, technique used for repair: whether double or modi ed single-patch and whether repair was performed with posterior valve annuloplasty or not, and patient rhythm both on going off bypass and on discharge from the operating theatre.
During ICU stay, measurements included postoperative ventilation duration, incidence of re-exploration for bleeding, total intensive care unit stay, incidence of arrhythmia and heart block.
Other operative complications were also recorded including super cial wound infection, mediastinitis, pleural collection, lung collapse, chest infection, early re-operation (within one month of surgery). Inhospital mortality was also recorded.
Postoperative echocardiography was performed to detect the presence and degree of mitral valve regurgitation, left ventricular dimensions, presence of pericardial collection, presence of left ventricular out ow tract (LVOT) obstruction, and presence of residual septal defects.
Following 6 months from hospital discharge, chest X-ray, echocardiography and mortality were also recorded.
Results 37.1% of participants were younger than 6 months old and 62.9 % were older than 6 months. Univariable analysis shows the only signi cant differences between the two groups was regarding ICU stay (3.6 day vs 5.04 day, P value = 0.02) and postoperative complications (53.8 % vs 18.2, P value = 0.01). Modi ed single patch technique was used in 39 cases (55.7%), while in 31 cases (44.3%), double-patch technique was used. Univariate analysis showed that surgical repair by modi ed single patch technique showed signi cant reduction in cross clamp time (mean = 47.6 ± 9.227 min vs 73.55 ± 21.087 min, P value 0.001), shorter bypass time (mean = 71. 13   LAVV dysplasia is de ned as the presence of anomalies including thickened lea ets, elongated chordae, lack of lea et tissue, a double-ori ce LAVV, excessive lea et tissue and fusion of the commissures, and is traditionally diagnosed by echocardiography (10). In our study group, there was a strong association between incidence of LAVV dysplasia and Trisomy 21; (64% vs 13.3% in Trisomy 21 versus Non-Trisomy 21 patients, respectively (P Value = 0.001)).
On multivariate regression analysis, Trisomy 21 was associated with signi cantly less moderate to severe LAVV regurgitation compared to Non-Trisomy 21 patients (P Value = 0.005). Mortality was comparable in Trisomy 21 versus Non-Trisomy 21 patients.
When compared with other variables, Posterior LAV Annuloplasty was found to be signi cantly protective against 6 months moderate to severe LAVVR and short term mortality. The addition of posterior annuloplasty to closure of common AV valve cleft was associated with signi cant reduction in the occurrence of post-operative LAVVR during 6 months of follow up and signi cant reduction in mortality (P value = 0.001, P value = 0.002 respectively).
Mortality was reported in 8.6% of patients (6 patients), 4 early within hospital stay (2 patients died from low COP and 2 from sepsis), whilst 2 died 6 months after discharge; 1 required readmission for a chest infection and died from sepsis, 1 died at home of unknown causes.

Discussion
The optimal age for surgical correction of CAVC defect has been addressed in several studies, most of which showed that successful repair at an early age was associated with competent AV valves at time of diagnosis, even in the presence of a cleft in the LAVV component (18,19). Patients repaired at older ages were found to be more likely presenting with an incompetent AV valve (19,20). This may be due to chronically elevated QP/QS playing an important role in the onset of annular dilatation and secondary LAVV incompetence (9,19). Early correction has been proposed to reduce the incidence of LAVV regurgitation (LAVVR) in the postoperative follow-up period (21,22). This appears to be the most important factor affecting postoperative morbidity and mortality (10). Avoiding irreversible pulmonary hypertension, which may develop before the age of six months particularly in the presence of Trisomy 21, is another reason to justify earlier repair (10,21). Some studies advocate that the optimal timing for surgical intervention in patients with CAVC defect is when symptoms of congestive heart failure cannot be managed by medical therapy (23). Others believe that repair at the age of 4 to 6 months, before developing severe pulmonary hypertension, gives better results with no additional morbidities or mortality (21,22,24).
In our study, we found no statistically signi cant difference between early surgical repair of CAVC defect in the rst 6 months versus later repair, both regarding mortality and postoperative LAVVR. The only signi cant nding was in the ICU duration, which was longer in the younger age group. Stellin et al (10) performed a study assessing surgical treatment of CAVC defects in children before 3 months of age, with ndings demonstrating that repair of CAVC under 3 months of age is the ideal approach to this malformation with a lower mortality rate at operation compared to older patients (10,24). Logistic analysis showed that an operative age after 3 months is compared to an age before or equal to 3 months, an incremental risk factor for hospital mortality with an odds ratio of 4.8 (95% con dence limit 1-23.5) (P value = 0.05) (10). In the long term, freedom from reoperation for LAVV incompetence is higher when compared to children repaired at an older age (10).
Parikh et al (25) found no differences between the early (< 90 days) and late repair (> 90 days) age-groups at surgery for the outcomes of postoperative length of stay, heart block requiring pacemaker placement, rate of moderate or greater LAVVR, reoperation for LAVVR, and mortality (25).
Both modi ed single patch and double patch techniques can be used for the repair of CAVCD, with the former being superior in having shorter bypass time and cross clamp time, making it more popular (11,26), although some surgeons believe that it may have associations with higher incidence of LVOT obstruction (27).
In our study, when compared to the double patch techniques, patients with a modi ed single patch had signi cantly shorter cross clamp time, bypass time and duration of ICU stay. Surgeons' preference in deciding which technique to use in the surgical repair is one important factor, some surgeons tend to use double patch technique with large VSD component, others favour the use of modi ed single patch technique in complete AVSD with large VSD (larger than 0.8 mm and up to 1.5 mm), both, however, were found to have comparable results and outcomes (11,26,27).
Addition of posterior annuloplasty to the common AV valve repair was suggested to prevent late reappearance of LAVVR(26). In this study, it showed a protective effect against moderate to severe LAVVR and short term mortality 6 months postoperatively, when compared to closure of cleft only. We believe that posterior annuloplasty maintains the valve competency and also restricts progressive dilatation of the LAVV, hence decreasing the need for reoperation. In a study performed by Myers et al, 219 patients were included to evaluate the impact of annuloplasty at CAVC repair on postoperative LAVV function(28). The cleft was closed completely in 192 patients (88%). 65 patients had annuloplasty (39 commissural, 32 posterior). There were 5 early deaths (2.3%). At discharge, 4 patients (1.9%) had more than mild regurgitation and no patients had signi cant in ow gradients. During a follow-up of 2.7 ± 2.1 years, there were 6 late deaths (2.8%) and 16 patients (7.3%) required LAVV reoperation. Two of 65 patients (3.1%) with annuloplasty required reoperation, compared to 14 of 148 without annuloplasty (9.5%, p = 0.16). In propensity matched analysis, annuloplasty was signi cantly protective of ≥ moderate LAVVR (OR 0.19, p = 0.008) and non-signi cantly of reoperation (OR 0.28, p = 0.099)(28).
There was a signi cant difference between Trisomy 21 patients and Non-Trisomy 21 patients regarding postoperative, early and 6 months follow-up LAVVR. Additionally, there was a signi cantly higher incidence of common AVV dysplasia and malformation in Non-Trisomy 21 patients when compared to Trisomy 21 patients. The signi cant difference between postoperative LAVVR between the two groups may be due to the difference in AV valve dysplasia and malformations, which makes surgical repair of LAVV and maintenance of valve competency much more di cult, regardless of surgical technique used.
Kaza et al found some preoperative and intraoperative differences between the two groups but of no signi cance, and outcomes were found to be similar (30). Hospital courses were also similar with the exception of a number of in-hospital complications (higher in the Non-Trisomy 21 group). Mortality at 6 months was identical at 4%, but there was a trend towards more reoperations in the rst 6 months in the Non-Trisomy 21 group (13% vs. 2%, p = 0.06) (30). Moderate or severe LAVVR was more likely in the Non-Trisomy 21 group prior to AVSD surgery, however, the prevalence of LAVVR did not differ at 1 month and 6 months post-surgery (30).
We can conclude that Trisomy 21 is no longer a risk factor for adverse outcomes regarding postoperative LAVVR and mortality, so early intervention in the rst 6 months in patients with Trisomy 21 is advocated to provide better results when compared to Non-Trisomy 21 patients, and gives comparable results compared to delaying surgery after 6 months.

STUDY LIMITATIONS
In this study, we ensured that our methodology was coherent. We optimised our selection criteria to make our ndings generalisable to a wider patient population. However, this study still has a number of limitations, namely a limited study sample and a short follow-up. Other limitations, owing to the retrospective nature of the study, included possible unaccounted confounders such as sex, ethnicity and the presence of comorbidities, heterogeneity of operations as well as information bias during data collection.

Conclusion
In our experience, early intervention by surgical repair in CAVC defect patients during the rst 4 to 6 months had good outcomes and comparable results to delayed surgery at older age. Trisomy 21 was not found to be a risk factor for early intervention and had better outcomes regarding postoperative LAVVR   Table 4 Coe cients a required for