RESUMO
BACKGROUND: When patients with advanced gastric cancer experience active bleeding, gastroenterologists normally choose between two treatment modalities, endoscopic hemostasis and transarterial embolization (TAE). In patients with advanced gastric cancer with bleeding, the predictive factors for endoscopic hemostatic failure are still unknown. Thus, the purpose of this study was to evaluate predictive factors for endoscopic hemostasis failure and to differentiate which hemostasis procedure is more effective for advanced gastric cancer with bleeding. METHODS: We reviewed the medical records of patients who were diagnosed with advanced gastric cancer and acute non-variceal gastric bleeding from January 2006 to August 2011. Forty-five patients were enrolled in this study and they were divided into a group of 14 patients who had experienced successful endoscopic hemostasis and a group of 31 patients who had had unsuccessful hemostasis with the first endoscopy and then underwent TAE. RESULTS: Lesion size and bleeding condition of Forrest class 1a or 1b were statistically significant predictive factors for endoscopic hemostatic failure (P = 0.023 and P = 0.017, respectively). On multivariate logistic regression analysis, size (lesion >2 cm) was a significant predictive factor for endoscopic hemostatic failure [adjusted odds ratio (aOR) 8.056; 95% confidence interval (CI) 1.329-48.846]. CONCLUSIONS: We determined that small bleeding lesions (<2 cm) and exposed vessels in the bleeding site with gastric cancer indicated that endoscopic hemostasis would be an effective hemostatic modality to choose. Particularly, in the opposite condition, the presence of large bleeding lesions (>2 cm) and non-exposed vessel bleeding with a tumor, endoscopic hemostasis failure is predicted and TAE could be recommended.