R. M. Landisch1, S. Foster3, D. Gregg3, T. Chelius1, L. Cassidy1, D. Lerner4, D. R. Lal1,2 1Medical College Of Wisconsin,Surgery,Milwaukee, WI, USA 2Children’s Hospital Of Wisconsin,Pediatric Surgery,Milwaukee, WI, USA 3Children’s Hospital Of Wisconsin,Radiology,Milwaukee, WI, USA 4Children’s Hospital Of Wisconsin,Gastroenterology,Milwaukee, WI, USA
Introduction: Anastomotic stricture is the most common post-operative complication in infants undergoing repair of esophageal atresia with or without tracheoesophageal fistula (EA/TEF), with an incidence reported as high as 60%. Several stricture indices (SI) based on measurements derived from early post-operative esophagrams have been proposed as a way to predict which infants will develop an anastomotic stricture requiring dilation. We sought to examine which SI more accurately predicted the need for anastomotic dilation.
Methods: An IRB approved retrospective study of infants undergoing primary repair of EA/TEF between 2008-2013 was performed. Digital esophagrams were analyzed to calculate one of four stricture indices: Upper (U-EASI) and Lower (L-EASI) Esophageal Anastomotic Stricture Index, Lateral SI and Anterior/Posterior SI. The primary outcome was stricture resulting in dilation. All esophagrams and dilations in 2-year follow up period were included. Logistic regression analysis was performed to determine if SI was associated with needing anastomotic dilation. A receiver operating characteristic (ROC) curve measured the accuracy of the regression model based on stricture indices to predict dilation. Statistical significance was determined at p-value < 0.05.
Results: Forty-five infants underwent primary repair of their EA/TEF. Anastomotic strictures requiring dilation occurred in 20 (44%) at a median of 95 days post-operatively (range 24-649). Median SI's were calculated from the 1st post-operative esophagram (median 7 days) to categorize the cohort. Only U-EASI as a continuous variable was predictive of need for dilation (p=0.03), however the median U-EASI threshold of ≤ 0.37 as a discrete variable was not significant. Infants with a low U-EASI were dilated at a median of 50.5 days after surgery, with 70% dilated in < 3 months. No association was found between U-EASI and number of dilations required, nor did early dilation decrease the frequency of dilations. In comparison to the 1st post-operative esophagram, the 2nd esophagram (median 28 days, range 10-884) median threshold U-EASI of ≤ 0.39 was significantly associated with dilation (OR 8.51, 95% CI: (1.16, 62.61), p=0.035). The area under the ROC curve of the U-EASI model controlling for days to esophagram demonstrated improved predictive ability from 1st (AUC 0.73) to 2nd esophagram (AUC 0.81).
Conclusion: SI's were not associated with future need for dilation when calculated using the first post-operative esophagram. U-EASI measured on the second post-operative esophagram of ≤ 0.39 should be considered a marker for increased risk of requiring dilation.
