527 ADVERSE EVENTS OF ENDOSCOPIC BALLOON DILATATION FOR ESOPHAGEAL STRICTURE CAUSED BY ENDOSCOPIC SUBMUCOSAL DISSECTION

Endoscopic balloon dilatation (EBD) is useful to resolve esophageal stricture due to endoscopic submucosal dissection (ESD). However, only few reports have described adverse events such as perforation and bleeding caused by EBD. We aimed to evaluate predictive factors of adverse events of EBD for esophageal stricture caused by ESD. This was a retrospective observational study in a single institution. Between October 2006 and November 2016, 638 consecutive patients with 1012 esophageal lesions were treated with ESD. Of these patients, 56 who were treated with EBD for esophageal stricture caused by ESD were enrolled. The primary outcome was the determination of the predictive factors of adverse events of EBD for ESD mediated esophageal stricture. Generalized estimating equations were used to analyze repeated-measures data. A total of 313 EBD procedures were included. The median age of the patients, including 44 men and 12 women, was 70 years. The median diameter of the tumor and circumferential mucosal defect size after ESD were 42 mm and 83.3%, respectively. Prophylactic locoregional steroid was injected for 32 patients after ESD. The incidence of perforation during EBD was 8.9% (5/56) per patient and 1.6% (5/313) per procedure. Wider circumferential mucosal defect after ESD (p=0.006), larger tumor size (p<0.001), higher dysphagia score (p<0.001), and perforation during ESD (p<0.001) were associated with perforation in the crude analysis. The multivariate logistic regression analysis revealed that higher dysphagia score (odds ratio [OR], 5.103; 95% confidence interval [CI], 1.492–17.451; p=0.009) was an independent predictive factor of perforation. The incidence of bleeding during EBD was 8.9% (5/56) per patient and 2.2% (7/313) per procedure. Submucosal fibrosis (p=0.032) and residual ulcer after ESD (p=0.017) were associated with bleeding in the crude analysis. The multivariate logistic regression analysis revealed that residual ulcer after ESD (OR, 17.325; 95% CI, 1.825–164.471; p=0.013) was an independent predictive factor of bleeding. The incidence of all adverse events during EBD was 16.1% (9/56) per patient and 3.8% (12/313) per procedure. Larger tumor size (p<0.001), higher dysphagia score (p<0.001), perforation during ESD (p=0.008), and prophylactic locoregional steroid injections after ESD (p=0.013) were associated with all adverse events in the crude analysis. The multivariate logistic regression analysis revealed that higher dysphagia score (OR, 2.025; 95% CI, 1.217–3.370; p=0.007) and prophylactic locoregional steroid injections (OR, 5.917; 95% CI, 1.127–31.079; p=0.036) were independent predictive factors of all adverse events during EBD. Higher dysphagia score and prophylactic locoregional steroid injections after ESD were significant predictors of adverse events during EBD.