Abstract
A total of 187 patients who underwent ESD for gastric neoplasms between January 2009 and December 2011 were included in this retrospective analysis of risk factors associated with delayed bleeding. Delayed bleeding was defined as clinically evident bleeding that required emergency endoscopy and/or blood transfusion with a decline of more than 2g/dl in serum hemoglobin concentration. Incidental bleeding cases were counted separately when active bleeding corresponding to or beyond Forrest IIa was incidentally found and resulted in additional hemostatic procedures during second-look endoscopy. The following variables were analyzed to determine factors leading to delayed bleeding : patient background (age, and anticoagulants/antiplatelets) , and procedural conditions (specimen size, duration of procedure and location of lesion) , and comorbidities (hypertension, diabetes mellitus, liver cirrhosis and, chronic renal failure with dialysis) . Delayed bleeding occurred in 5 cases (2.7%) . Maximum diameter of the resected specimen was identified as a sole predictor of delayed bleeding by Fisher test while multivariate analysis failed to statistically establish any particular variables predictive of delayed bleeding, although it demonstrated correlation between specimen size and delayed bleeding. Incidental bleeding was observed in 59 cases (31.6%) . In these hypertension, age, and maximum diameter of resected specimen were identified as predictive parameters by chi-squared test, while age and hypertension were identified as predictive parameters by multivariate analysis.─In view of the relatively frequent occurrence of incidental bleeding, additional prophylactic hemostasis immediately following ESD procedures may be useful in preventing delayed bleeding─.
