Background: Upper gastrointestinal bleeding is a common medical emergency condition which can be very costly with potential risks. Prompt evaluation of the patients’ status and determination of the risks involved are of utmost significance. Rockall score is one of the Methods determining the risk of bleeding. The aim of the present study was to determine sensitivity, specificity and negative and positive predictive values of Rockall scoring system during a 3-month follow-up period in patients with upper gestronintestinal bleeding.

Materials and Methods: 340 hospitalized patients with acute upper gastrointestinal bleeding in Imam Reza hospital were studied from December 2013 to September 2014. The full and clinical Rockall scores, sensitivity, specificity and negative and positive predictive values were specified for all the subjects who were followed for 3 months in relation to complications and mortality.

Results: Of all patients, 204 were males and 136 were females who 15.6% of the subjects had a clinical Rockall score under 3, with 84.4% having a score over 3. In relation to endoscopic Rockall score, 15.6%, 66.8% and 17.6% of the subjects had a score under 3, between 3-8 and over 8, respectively. During the 3-month follow-up period, in the low-risk group, 92.5% did not exhibit hemorrhage again, 3.8% had one recurrent episode of hemorrhage, 1.9% exhibited 2 hemorrhage episodes, with 1.9% having 3 hemorrhage episodes. In the moderate-risk group, 90.9% experienced no recurrence and 9.1% exhibited 1 case of recurrent hemorrhage (p=0.4). There was no mortality in the low-risk group. In the moderate-risk and high-risk groups, the mortality rates were 8.07% and 22.7%, respectively (p<0.001). The sensitivity and specificity of the test were calculated at 11.1% and 81.1%, respectively. The positive and negative predictive values were 5% and 91.4%, respectively.

Conclusion: With due attention to the specificity of 81.8% and negative predictive value of 91.4% for the Rockall score, it might be guaranteed the patient will not suffer from the complications resulting from recurrence of hemorrhage. Given the low sensitivity of the test, the test is not appropriate to rule out disease.