Inspired by the well-documented benefits of minimally invasive surgery (MIS) in adults and children, pioneers of pediatric MIS have recognized the potential of thoracoscopic approach to repair the esophageal atresia with tracheoesophageal fistula. During last decade, the thoracotomy approach, refined during almost 100 years (1, 2), is being replaced by thoracoscopic approach for surgical correction of EA (3–8), even some concerns had been raised about thoracoscopy in neonates (9).

This has become the standard approach in many centers, since it allows a better cosmetic result, avoids the musculoskeletal sequelae of thoracotomy, and allows excellent magnified visualization of anatomic structures on the monitor.

Either by thoracotomy or thoracoscopy, the standard technique to correct EA with TEF is performed with the ligation and division of azygos vein.

Recently, preservation of azygos vein has been hypothesized to maintain the mediastinal venous drainage, thereby decreasing postoperative chest congestion and tissue edema, thus working as an additional protective factor against postoperative complications like anastomotic leak (10, 11), but this hypothesis was never hemodynamically confirmed. More recently, Patkowski et al. performed thoracoscopic repair sparing the azygos vein as a strategy to reduce the fistula recurrence by interposing some natural biological tissue between the anastomosis and fistula stump (12). We questioned if the azygos vein would be able to keep its patency in this position. On the other side, we know that patency of the azygos vein is important to preserve to normal hemodynamics in particularly cases (13).

In this line of thought, we hypothesized if fully anatomic thoracoscopic repair of esophageal atresia (ATREA) would be feasible without compromising outcomes. This study presents ATREA technique for the repair of EA with TEF and retrospectively reviews its results comparing with the classic thoracoscopic approach.

Newborns with the diagnosis of EA with TEF, submitted to thoracoscopic repair between September 2012 and September 2016, were operated by the same surgeon (JCP) and included in this study. Patients submitted to open approach or with pure EA were excluded.

This study was approved by the “Comissão de Ética para o Hospital de Braga (ref. CESHB 087/2016)” and “Subcomissão de Ética para as Ciências da Vida e da Saúde (ref. SECVS 015/2016),” and the patients were de-identified to protect their anonymity.

Since 2013, we started to perform the esophageal anastomosis underneath (on the left side) the azygos vein: ATREA. The newborns were divided in two groups according to the surgical technique being performed: ATREA in Group A (patients operated more recently) and the classic thoracoscopic repair sparing the azygos vein in Group B (patients operated until 2013). In the case of technical difficulties, the ATREA was converted in the classical thoracoscopic repair sparing the azygos vein and integrated into the Group B as it occurred in one case when we realized that the anastomosis would be left with moderate to severe tension.

Data were collected regarding the prenatal period (gestational age and prenatal diagnosis), features of the newborn (gender, karyotype changes, associated anomalies, birth weight), surgery (age at the time of surgery, operative time, surgery complications), hospitalization (duration of mechanical ventilation, hospitalization and thoracic drainage, time to first feeding, and early postoperative complications), and follow-up (long-term complications such as fistula recurrence, leak and/or anastomotic stenosis, tracheomalacia, gastroesophageal reflux disease (GERD), and mortality).

Statistical analysis was performed using IMB SPSS Statistics 23. Mann–Whitney U test and Fisher’s exact test were used for the small sample size. Statistical significance was set at p-value of ≤0.05.

Between September 2012 and September 2016, 11 patients with EA and distal TEF were treated by thoracoscopic repair at our center. Seven (4 males) were operated according to ATREA (Group A) and four (1 male) by the classic thoracoscopic repair sparing the azygos vein (Group B).

Data comparing both groups regarding gestational age, gender, karyotype, associated malformations, and birth weight are summarized on Table 1, and there were no statistically significant differences regarding these features.

No karyotype abnormalities identified in both groups. In general, associated anomalies were present in more than 60% of patients (57% of patients from Group A and 75% of patients from Group B). One patient in Group A and two in Group B had VACTERL association. In Group A, there was an infant with a right aortic arc without preoperative diagnosis (Figure 6A), where ATREA was particularly useful as it avoided that the azygos vein and the aortic arch were left compressed in between the esophagus and trachea (Figures 6B,C). All the associated malformations are resumed in Table 2.

Comparative data regarding surgery and immediate postoperative period and long-term complications are presented in Tables 3 and 4, respectively. The median operative time was 115 [54–135] min in Group A and 151 [123–219] min in Group B with no statistically significant difference.

In Group A, the median duration of mechanical ventilation was 3 days, the time until first feeds was 6 days, the duration of chest drainage was 7 days, and the median length of hospitalization was 14 days. In Group B, the median duration of mechanical ventilation was 21 days, the time until first feeds was 8.5 days, the duration of chest drainage 9.5 days, and the median length of hospitalization was 56 days, but the median values in Group B are particularly influenced by one patient with severe tracheomalacia that was hospitalized for 94 days.

Group A had three patients with anastomotic stricture requiring at least one dilation and one patient with mild tracheomalacia with no need for aortopexy. No patient had surgical complications, anastomotic leak, recurrent fistula, GERD, need for new surgical correction, or death. Group B had three patients with anastomotic stricture requiring at least one dilation, three with tracheomalacia, one requiring aortopexy and one with GERD demanding fundoplication. There was a case of anastomotic leak, requiring re-intervention. There were no cases of fistula recurrence or death.

Thoracoscopic operations in neonates such as repair of EA and tracheooesophageal fistula can be associated with intraoperative acidosis and hypercapnia in the absence of hypoxia. These derangements in intraoperative gas exchanges seem to be related to the insufflation and absorption of medical CO₂ with unknown effects on the developing brain (9). While this is not completely clarified, thoracoscopic technique has been adopted worldwide, by several centers, as the first line approach to EA correction.

Thoracoscopic technique was based on the thoracotomy procedure, as it follows its main steps, including azygos vein ligation and division to optimize the approach to the esophageal pouches.

The excellent visualization of the surgical field provided by thoracoscopy turned the division of the azygos vein practically unnecessary and even questionable. As already reported by others, azygos vein is a major vessel of the human body which means that it should be preserved whenever it is possible (10). Previous studies reported the potential benefits of saving azygos vein in the thoracotomy repair. Evans et al. reported a lifesaving case by preserving the azygos vein in a newborn with EA and an interrupted inferior vena cava (IVC) (13). Although this is a rare combination, we know that newborns with EA have increased risk of cardiovascular malformations. In the particular case of an interrupted IVC, dividing the azygos vein would be catastrophic (13). A prospective study of newborns with EA and TEF proposed that preservation of azygos vein offers an additional protection to prevent anastomotic leaks (10). The preservation of the azygos vein avoids early postoperative local edema, offering a protective factor against anastomotic leaks. Another study showed a decreased rate of pneumonitis in the postoperative period when preserving the azygos vein (11). Patkowski et al. reported a series of 23 patients with EA and TEF corrected by thoracoscopy with preservation of azygos vein (12). Differently to ATREA, they separated the anastomosis from the fistula ligation interposing azygos vein between them. With this technique, the authors argued that azygos may prevent recurrent fistula formation (12). This might be true; anyway, the entrapment of the azygos vein between the trachea and esophagus and the friction resulting from the esophagus movements could lead to a variable degree of collapse of the azygos vein, probably making it useless or at least with a major impact on its hemodynamics.

It is our belief that the best approach would be to preserve the normal anatomophysiology of the human body, by reconstructing the esophagus on its anatomic position, medial to the azygos vein. This possibility became feasible in a newborn in whom the TEF was located in a high position of the trachea; therefore, it was almost intuitive to perform the correction by ATREA technique. In this case, we were able to safely finish the anastomosis with no peroperative or postoperative complications. Subsequently, we have attempted to apply the ATREA to every newborn with the diagnosis of EA with distal fistula, and we have succeeded in a total of six patients. The exception was a case in whom the anastomosis was under a great tension and we decided to use the classical thoracoscopic technique for safety reasons. With experience, we could verify that thoracic vessels (azygos vein and aortic branches) are sufficiently elastic to allow dissection around and/or under them. By doing this, we predict we can perform ATREA in the great majority of cases. Actually, our short series could document that ATREA has shown no difference in terms of operative details, hospitalization, and follow-up as compared with the classic thoracoscopic repair sparing azygos vein. Therefore, we can conclude that ATREA is feasible and likely a safe technique.

Additionally, the principles used in ATREA can also be applied in cases of venous return malformations and even in cases of arterial malformations such as aortic arch malformations. In fact, we had a patient in Group A, where we unexpectedly discovered a right aortic arch. In this case, we did not need to do an immediate left thoracoscopy neither to postpone the repair as suggested by a recent systematic review (14). Actually, we could perform dissection and anastomosis in anatomic position. The experience with this patient provides additional evidence that surgeons might be prepared to try anatomic repair in all patients as this experience could be useful when unexpected major aortic malformations appears, including those patients that are still operated by open/thoracotomy access.

The authors recognize the limitations of this study, namely a small sample size and the fact that the all patients were operated on by the same surgeon. It may be arguable that an extensive surgical expertize may influence the results positively; nevertheless, a single surgeon series also eliminates any inter-surgeons variability. So, larger studies or randomized clinical trials are necessary to understand all the benefits or potentially disadvantages, of such a modification. Evaluation of the patency and hemodynamics of the azygos vein after both classic and ATREA repair could also clarify the impact of the surgical approach on such vessel.

This study was submitted and approved by the following ethics committees: “Comissão de Ética para o Hospital de Braga (ref. CESHB 087/2016)” and “Subcomissão de Ética para as Ciências da Vida e da Saúde (ref. SECVS 015/2016).”

Substantial contributions to the conception or design of the work or the acquisition, analysis, or interpretation of data for the work; drafting the work or revising it critically for important intellectual content; final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: JF, CB, RL-P, AS, and JC-P.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.