The utility of radiological upper gastrointestinal series and clinical indicators in detecting leaks after laparoscopic sleeve gastrectomy: a case-controlled study.

Abstract

Leak after laparoscopic sleeve gastrectomy (LSG) often presents after hospital discharge, making timely diagnosis difficult. This study evaluates the utility of radiological upper gastrointestinal (UGI) series and clinical indicators in detecting leak after LSG. A retrospective case-controlled study of 1762 patients who underwent LSG from 2006 to 2014 was performed. All patients with radiographically confirmed leaks were included. Controls consisted of patients who underwent LSG without leak, selected using a 10:1 case-match. Data included baseline patient characteristics, surgical characteristics, and UGI series results. Clinical indicators including vital signs, SIRS criteria, and pain score were compared between patients who developed leak and controls. Of 1762 LSG operations, 20 (1.1%) patients developed leaks and were compared with 200 case-matched controls. Three patients developed leak during their index admission [mean=1.3days, range (1, 2)], while the majority (n=17) were discharged and developed symptoms at a mean of 17.1days [range (4, 63)] postoperatively. Patients diagnosed with leak were similar to controls in baseline and surgical characteristics. Contrast extravasation on routine postoperative UGI identified two patients with early leaks, but was negative in the remainder (89%). Patients with both early and delayed leaks demonstrated significant clinical abnormalities at the time of leak presentation, prior to confirmatory radiographic study. In multiple regression analysis, independent clinical factors associated with leak included fever [OR 16.6, 95% CI (4.04, 68.10), p<0.0001], SIRS criteria [OR 7.0, 95% CI (1.47, 33.26), p=0.014], and pain score ≥9 [OR 19.1, 95% CI (1.38, 263.87), p=0.028]. Contrast extravasation on routine postoperative radiological UGI series may detect early leaks after LSG, but the vast majority of leaks demonstrate normal results and present 2-3weeks after discharge. Therefore, clinical indicators (specifically fever, SIRS criteria, and pain score) are the most useful factors to raise concern for leaks prior to confirmatory radiographic study and may be used as criteria to selectively obtain UGI studies after LSG.