Background and study aims: Various clinical and endoscopic factors have been proposed and used as predictors of endoscopic treatment failure in bleeding peptic ulcers. Recently, several endoscopic factors have been considered to be more significant than various clinical factors, except for shock. Detailed knowledge of which endoscopic factors can be classified as predictors of rebleeding following endoscopic hemostasis is needed. This study describes newly defined endoscopic variables and evaluates their usefulness as predictors of endoscopic treatment failure.
Patients and methods: Between January 1996 and April 1999, diagnostic and therapeutic endoscopies were carried out in 143 patients with active bleeding peptic ulcers. Nine clinical and eight endoscopic variables were studied. Endoscopic factors were classified by three types of stigmata bleeding, 14 locations, two ulcer sizes, three ulcer bases, three visible vessel colors, two ulcer depths, two margin shapes, and three endoscopic treatment methods (injection, hemoclipping, and combination).
Results: 36 patients experienced rebleeding (25.2 %), 11 patients needed operations (7.7 %) and five deaths occurred (3.5 %). In univariate analysis, rebleeding was significantly related to: i) presence of spurting activity (odds ratio [OR] = 4.91, P = 0.006), ii) ulcer size larger than 2 cm (OR = 2.78, P = 0.017); and iii) location in stomach (OR = 2.81, P = 0.026). Clinical variables including age, shock, and initial hemoglobin levels were not significantly related to rebleeding. In multiple logistic regression using selected significant variables, presence of spurting activity remained a significant independent predictor of rebleeding (adjusted OR = 6.48, P = 0.003).
Conclusion: Our data support the hypothesis that endoscopic factors are more important than clinical ones when predicting rebleeding of peptic ulcers. Based on statistical analysis of risk factors, the ulcers most likely to rebleed after endoscopic therapy are those which are located in the stomach, are greater than 2 cm in diameter and exhibit oozing or spurting of blood.