Blatchford score is a useful tool for predicting the need for intervention in cancer patients with upper gastrointestinal bleeding

Background and Aim: The Blatchford score is based on clinical and laboratory variables to predict the need for clinical interventions in upper gastrointestinal bleeding (UGIB). The primary object was to evaluate the Blatchford score with clinical and full Rockall scores in patients with active cancer presenting to the emergency department with UGIB. The secondary object was to assess the accuracy of the Blatchford score at different source of UGIB; cancer bleeding versus non-malignant lesions. Methods: We reviewed and extracted data from electronic medical record on patients with active cancer presenting to the emergency department from January 2009 to December 2011. Clinical interventions included blood transfusion, therapeutic endoscopy, angiographic intervention, and surgery. Results: Of the 225 patients included, 197 (87.6%) received interventions. Comparing the area under receiver-operator curves, the Blatchford score (0.86, 95% confidence interval [CI] 0.77-0.95) was superior to clinical Rockall (0.67, 95% CI 0.55-0.79) and full Rockall score (0.72, 95% CI 0.61-0.83) in predicting interventions. When the score of 2 or less is counted as negative, sensitivity of 0.99 and specificity of 0.54 were calculated. When the patients were separated according to the source of UGIB, sensitivity and specificity were not changed. Conclusions: The Blatchford score outperformed both Rockall scoring system in predicting intervention in patients with active cancer. The source of bleeding was not important factor in the score performance. The Blatchford score has a very good sensitivity. However, suboptimal specificity limits its role as sole means of decision making in cancer patient with UGIB.