ABSTRACT
BACKGROUND/AIMS: To investigate the factors contributing to failure of initial hemostasis in patients undergoing endoscopic hemostasis. METHODOLOGY: A total of 316 patients underwent endoscopic hemostasis for bleeding peptic ulcers in a period of 4 years. RESULTS: For hemostatic procedures, application of hemostatic clips, band ligation, injection of hypertonic saline epinephrine solution, soft coagulation, and argon plasma coagulation were employed either singly or in combination. Patients were divided into the following 2 groups for multivariate analysis: durable hemostasis (n = 268) and failed initial (incomplete) hemostasis (n = 48). Hemodialysis was a risk factor of incomplete hemostasis (Odds Ratio [OR] = 2.306, 95% confidence interval [CI] = 1.033-5.147; p = 0.041). Compared with the duodenal 2nd portion, the following bleeding sites had significantly lower risk of incomplete hemostasis (approximately 5 times less likely): The duodenal bulb (D), OR = 0.215, 95% CI = 0.058-0.797 (p = 0.022); the L region, OR = 0.207, 95% CI = 0.046-0.919 (p = 0.038); the M region, OR = 0.132, 95% CI = 0.036-0.482 (p = 0.002); and the U region, OR = 0.164, 95% CI = 0.041-0.649 (p = 0.01). CONCLUSIONn: Hemodialysis and a bleeding site located in the duodenal second portion were the factors strongly associated with incomplete hemostasis in bleeding gastroduodenal ulcers.