Introduction: Acute upper gastrointestinal bleeding (AUGIB) is an important gastroenterological emergency. Role of Rockall score in stratification of risk of mortality and rebleeding in Indian patients is uncertain.

Methods: This study was prospectively conducted on 175 consecutive patients admitted with AUGIB who were 18 years or above. The clinical Rockall score and complete scores calculated at admission and after endoscopy were analyzed for association with the outcome (rebleeding and mortality) for both variceal and non-variceal bleeding. 

Results: The mean age of the patients was 48.4±16.4 years and 78.9% were males. History of inake of NSAIDs in patients of AUGIB was present in 13.7% and of antiplatelets in 13.7%. Variceal bleed was the most common etiology (45.7%) followed by peptic ulcer disease (24.0%). Overall rates of surgery, rebleed and mortality were 6.9%, 22.9% and 20% respectively.Both clinical and complete Rockall risk scoring systems were useful tools for stratifying patients with AUGIB into high and low risk categories for mortality. However, for the prediction of rebleed, the complete Rockall scoring system was found to be better than the clinical scoring system. These scoring systems performed much better in predicting the outcomes in non-variceal bleed than in patients with variceal bleed. Multivariate analysis showed that patients with age =65 years and albumin <3 g/dl were at an increased risk of mortality.

Conclusion: Rockall scoring can be used to stratify patients with AUGIB and is more useful in non-variceal bleeding.