Risk factors for anastomotic complications following thoracoscopic repair of type III esophageal atresia in neonates: a single-center retrospective cohort study

Background: Although thoracoscopic repair in China has emerged as a preferred and increasingly adopted minimally invasive approach for esophageal atresia (EA), anastomotic complications remain a major concern, with reported incidences ranging from 15% to 80%. However, the independent risk factors for these complications in a homogeneous surgical population remain poorly characterized. Methods: This single-center retrospective cohort study recruited neonates with type III EA who underwent primary thoracoscopic repair. Comprehensive demographic, operative, and anatomic variables were gathered. Univariate and multivariable logistic regression analyses were utilized to identify independent predictors for anastomotic leak (AL), anastomotic stricture (AS), and recurrent tracheoesophageal fistula (recurrent TEF). Results: The cohort had a mean birth weight of 2,597 g and a median operative time of 2.3 h. The incidences of AL, AS, and recurrent TEF were 15.4%, 74.6%, and 6.5%, respectively. Multivariable analysis identified distinct etiological mechanisms: an operative time of ≤2 h independently protected against AL (OR 0.24, 95% CI 0.09-0.63); the institutional learning curve attenuated traditional AS risk factors such as low birth weight; and the presence of a respiratory anomaly was the strongest predictor of recurrent TEF (OR = 8.84; 95% CI, 2.09-33.47). Minimal co-occurrence among the three complications (phi coefficients ≈ 0) confirmed their etiological independence. Conclusion: The three major anastomotic complications of thoracoscopic EA repair stem from distinct mechanisms. Anastomotic leak was independently protected by operative time ≤2 h (OR 0.24, 95% CI 0.09-0.63). The institutional learning curve attenuated traditional risk factors for anastomotic stricture (incidence 74.6%). Recurrent TEF was predominantly associated with respiratory anomalies (OR 8.84, 95% CI 2.09-33.47). This highlights the need for precision-based management. The robustness of these key associations was further confirmed through sensitivity analyses and enhanced methodological rigor, including assessments of predictor correlations and model diagnostics.