ABSTRACT
BACKGROUND: Leaks of the esophagus and stomach are difficult to manage and associated with significant morbidity and mortality. Endoscopic therapy can manage these leaks without surgical intervention. Our goal is to create a scoring tool to aid in predicting the success of endoscopic therapy in these patients. STUDY DESIGN: An IRB-approved prospectively maintained database was retrospectively reviewed for all patients treated for gastrointestinal leaks from July 2013 to January 2021, including patients treated for esophageal and stomach leaks. Endpoints include success of leak closure for patients treated solely by endoscopic therapy (ET) compared with surgical therapy as failed endoscopic therapy (FET). A multivariable logistic regression model was fitted to identify independent risk factors for predicting success of endoscopic therapy, and a scoring calculator was developed. RESULTS: There were 80 patients (60 females) with a mean age of 50 years. The ET group included 59 patients (74%), whereas the FET group included 21 patients (26%). Patient demographics, comorbidities, surgical history, and timing of leak diagnosis were used. Multivariable analysis resulted in 4 variables associated with higher probability of successful endoscopic leak management without need for additional surgery. These included increased age, lower BMI, lack of previous bariatric surgery, and quicker identification of the leak. Consequently, a scoring nomogram was developed with values from 0 to 22. CONCLUSION: Our data show the development of a scoring calculator capable of quantifying the likelihood of success treating foregut and bariatric leaks with endoscopic therapies. This can be used clinically to guide treatment decisions.
