Background: Guidelines recommend using prognostic scales for risk stratification in patients with non-variceal upper gastrointestinal bleeding. It remains unclear whether risk scores offer greater accuracy than clinical evaluation. Objective: Compare the diagnostic accuracy of the endoscopist's judgment against different risk-scoring systems (Rockall, Glasgow-Blatchford, Baylor and the Cedars-Sinai scores) for predicting outcomes in peptic ulcer bleeding (PUB). Methods: Between February 2006 and April 2010 we prospectively recruited 401 patients with peptic ulcer bleeding; 225 received endoscopic treatment. The endoscopist recorded his/her subjective assessment ("endoscopist judgment") of the risk of rebleeding and death immediately after endoscopy for each patient. Independent evaluators calculated the different scores. Area under the receiver-operating-characteristics (ROC) curve, sensitivity, specificity, positive and negative predictive values were calculated for rebleeding and mortality. Results: The areas under ROC curve of the endoscopist's clinical judgment for rebleeding (0.67-0.75) and mortality (0.84-0.9) were similar or even superior to the different risk scores in both the whole group and in patients receiving endoscopic therapy. Conclusions: The accuracy of the currently available risk scores for predicting rebleeding and mortality in PUB patients was moderate and not superior to the endoscopist's judgment. More precise prognostic scales are needed.
Keywords: endoscopist’s judgement; peptic ulcer bleeding; risk score.