Rebleeding, which occurs in 10-15 % of patients with peptic ulcer bleeding (PUB) [1], is associated with a twoto fivefold mortality increase, depending on the presence of other risk factors [2]. Therefore, identification of the predictors of rebleeding seems meaningful in order to identify high-risk patients needing close observation and rapid treatment in case of the development of rebleeding.According to previous studies, hemodynamic shock, usually defined as a systolic blood pressure \100 mmHg, often combined with tachycardia [100 beats/min, is the most powerful pre-endoscopic predictor of rebleeding [3,4]. In a meta-analysis, hemodynamic shock was associated with an odds ratio (OR) of rebleeding of 3.3 [3]. Conversely, studies on the association between anemia and rebleeding have found conflicting results: Some of the existing data indicate that hemoglobin \10 g/L may be associated with an increased risk of rebleeding [3]. Data concerning the risks of transfusion are even more confounded by differing study protocols (e.g., pre-or postendoscopic transfusion, different categorization of volume) to the point that the rebleeding risk of pre-endoscopic transfusion is unknown. Regarding endoscopic predictors, active bleeding at endoscopy (OR 1.7), ulcer size [2 cm (OR 2.8), posterior duodenal ulcer location (OR 3.8), and high lesser gastric curvature ulcer location (OR 2.9) all predict rebleeding in a meta-analysis [3].The type of endoscopic treatment applied does also affect the risk of rebleeding. A Cochrane analysis reported that combination of epinephrine injection with a second endoscopic treatment modality reduces the relative risk (RR) of rebleeding or persistent bleeding (RR 0.57) compared to endoscopic treatment with epinephrine alone [5]. Therefore, endoscopic monotherapy with injection of epinephrine should be avoided.In a meta-analysis based on eight randomized controlled trials (RCTs) published from 1994 to 2006, performance of second-look endoscopy within 16-48 h was associated with a significant reduction in rebleeding rate (OR 0.55) [6]. Generalization of this finding to current practice standards can be questioned because only one of the included studies used endoscopic combination therapy combined with high-dose infusion of proton-pump inhibitors [7]. Furthermore, detailed review of the fully published component studies revealed that a significant reduction in rebleeding was only evident in two studies that included patients with a very high risk of rebleeding (up to 47 % of included patients had hemodynamic shock) [6]. When these two trials were excluded from the meta-analysis, the association between performance of second-look endoscopy and rebleeding became statistically insignificant [6]. In a cost-effectiveness analysis, performance of secondlook endoscopy was only cost-effective after therapeutic endoscopy if the risk of rebleeding was greater than 31 % [8].In this issue of Digestive Diseases and Sciences, Kim et al. [9] published a prospective multicenter study of risk factors for rebleedi...