Thoracoscopy versus thoracotomy for esophageal atresia and tracheoesophageal fistula: Outcomes from the Midwest Pediatric Surgery Consortium.

BACKGROUND/PURPOSE: Controversy persists regarding the ideal surgical approach for repair of esophageal atresia with tracheoesophageal fistula (EA/TEF). We examined complications and outcomes of infants undergoing thoracoscopy and thoracotomy for repair of Type C EA/TEF using propensity score-based overlap weights to minimize the effects of selection bias.

METHODS: Secondary analysis of two databases from multicenter retrospective and prospective studies examining outcomes of infants with proximal EA and distal TEF who underwent repair at 11 institutions was performed based on surgical approach. Regression analysis using propensity score-based overlap weights was utilized to evaluate outcomes of patients undergoing thoracotomy or thoracoscopy for Type C EA/TEF repair.

RESULTS: Of 504 patients included, 448 (89%) underwent thoracotomy and 56 (11%) thoracoscopy. Patients undergoing thoracoscopy were more likely to be full term (37.9 vs. 36.3 weeks estimated gestational age, p < 0.001), have a higher weight at operative repair (2.9 vs. 2.6 kg, p < 0.001), and less likely to have congenital heart disease (16% vs. 39%, p < 0.001). Postoperative stricture rate did not differ by approach, 29 (52%) thoracoscopy and 198 (44%) thoracotomy (p = 0.42). Similarly, there was no significant difference in time from surgery to stricture formation (p > 0.26). Regression analysis using propensity score-based overlap weighting found no significant difference in the odds of vocal cord paresis or paralysis (OR 1.087 p = 0.885), odds of anastomotic leak (OR 1.683 p = 0.123), the hazard of time to anastomotic stricture (HR 1.204 p = 0.378), or the number of dilations (IRR 1.182 p = 0.519) between thoracoscopy and thoracotomy.

CONCLUSION: Infants undergoing thoracoscopic repair of Type C EA/TEF are more commonly full term, with higher weight at repair, and without congenital heart disease as compared to infants repaired via thoracotomy. Utilizing propensity score-based overlap weighting to minimize the effects of selection bias, we found no significant difference in complications based on surgical approach. However, our study may be underpowered to detect such outcome differences owing to the small number of infants undergoing thoracoscopic repair.

LEVEL OF EVIDENCE: Level III.