Role of Preoperative Transthoracic Echocardiography in Predicting Mortality in Pediatric Patients Diagnosed with Esophageal Atresia with or withoutTracheoesophageal Fistula: A Retrospective Cross-Sectional Study

Background: Esophageal atresia (EA) is a congenital malformation characterized by a discontinuity of the esophagus, with an incidence of approximately 2.99 per 10,000 live births. This condition is often associated with significant comorbidities, particularly cardiac anomalies, which can adversely affect morbidity and mortality rates. The objective of this study is to assess the correlation between preoperative transthoracic echocardiographic (TTE) parameters and surgical outcomes in neonates diagnosed with EA, thereby enhancing risk stratification for improved perioperative decision-making. Methods: This retrospective cross-sectional study was conducted at a tertiary medical facility in southern Iran, analyzing preoperative transthoracic echocardiography (TTE) parameters from patients who underwent surgical repair for esophageal atresia (EA) and tracheoesophageal fistula (TEF) between 2017 and 2023. The inclusion criteria encompassed patients under 18 years of age with a confirmed diagnosis of EA. Data collection was performed using a census-based methodology to minimize selection bias. Ethical approval was obtained, and patient confidentiality was maintained through the de-identification of records. Statistical analyses were conducted in accordance with the STROBE checklist, utilizing descriptive statistics, univariate analysis, and multivariate logistic regression to identify independent predictors of mortality. Results: The analysis revealed that ventricular septal defect (VSD) was a significant predictor of mortality, occurring in 31.8% of deceased patients compared to 6.5% of survivors (p < 0.001). Pulmonary hypertension was observed in 59.1% of deceased patients versus 35.1% of survivors (p = 0.013). Additionally, right ventricular dilation (36.4% vs. 17.6%, p = 0.026) and left ventricular dilation (22.7% vs. 8.8%, p = 0.046) were significantly associated with mortality. Multivariate logistic regression analysis identified VSD as the sole independent predictor of mortality (odds ratio: 0.150; p = 0.002). Machine learning models, particularly XGBoost, demonstrated superior predictive performance (AUC-ROC: 0.87), with ejection fraction percentage, valve pathology, and age recognized as key predictors. Conclusion: Preoperative transthoracic echocardiography (TTE) is essential for predicting postoperative mortality and complications in neonates with esophageal atresia (EA). The findings underscore the necessity of integrating TTE evaluations into preoperative risk stratification frameworks to optimize surgical timing and perioperative management. Future research should focus on validating these predictive models across multicenter cohorts and exploring the impact of emerging echocardiographic technologies on outcome prediction.