Clinical risk stratification with scoring systems in peptic ulcer bleeding

Background and Aim: Acute peptic ulcer bleeding is still a serious clinical problem leading to significant morbidity and mortality. To indentify high risk patients clinical and endoscopic scoring systems are available like Baylor and Rockall scores as well as the Forrest classification. The aim of this study was to evaluate the usefulness of those two clinical scores and Forrest classification grades in prediction of clinical outcome in acute ulcer bleeding patients during our daily routine practice. Methods: During a five year period (2004–2008) emergency endoscopy within 24 hours detected gastroduodenal ulcers as bleeding sources in 333 cases. Endoscopic hemostatic therapy (EHT) attempts were performed and patients with primary hemostasis were admitted to our high dependency ward. Rebleeding was defined by repeated appearance of haematemesis or melaena or by Hb drops more than 2g/dl or reappearance of haemodynamic instability and also by the need of repeated EHT during scheduled second look endoscopy. Based on the recorded data and clinical parameters Baylor and Rockall scores were calculated. The sensitivity and specificity to indicate adverse outcomes, including rebleeding, need for surgery, and mortality were analyzed on the basis of clinical scores and the Forrest grades of ulcers during the emergency endoscopy.

Results:

Prediction of clinical outcome measures according to Baylor and Rockall scores and Forrest grades (n=333)
		Forrest Ia, b, IIa (n=135)	Baylor score ≥10 (n=175)	Rockall score ≥8 (n=28)
Rebleeding	sensitivity	64%	67%	19%
	specificity	65%	51%	94%
Surgery	sensitivity	63%	59%	22%
	specificity	63%	48%	93%
Mortality	sensitivity	42%	74%	32%
	specificity	60%	49%	93%

Conclusion: Endoscopy scores indicating high risk stigmata of recent hemorrhage (Forrest Ia,b, IIa) are not inferior to Baylor and Rockall scores to predict adverse outcomes of ulcer bleeding patients.