Early Risk Score for Predicting Hypotension in Normotensive Patients with Non-Variceal Upper Gastrointestinal Bleeding

Abstract

Risk assessment for upper gastrointestinal bleeding (UGIB) is important; however, current scoring systems are insufficient. We aimed to develop and validate a prediction model for rapidly determining the occurrence of hypotension in non-variceal UGIB patients with normotension (systolic blood pressure 90 mmHg) at emergency department presentation. In this prospective observational cohort study, consecutive non-variceal UGIB patients between January 2012 and April 2017 were enrolled. We developed and validated a new prediction model through logistic regression, with the occurrence of hypotension <24 h as the primary outcome. Among 3363 UGIB patients, 1439 non-variceal UGIB patients were included. The risk factors for the occurrence of hypotension were lactate level, blood in nasogastric tube, and systolic blood pressure. The area under the curve (AUC) of the new scoring model (LBSLactate, Blood in nasogastric tube, Systolic blood pressure) in the development cohort was 0.74, higher than the value of 0.64 of the Glasgow-Blatchford score for predicting the occurrence of hypotension. The AUC of the LBS score in the validation cohort was 0.83. An LBS score of 2 had a negative predictive value of 99.5% and an LBS score of 7 had a specificity of 97.5% in the validation cohort. The new LBS score stratifies normotensive patients with non-variceal UGIB at risk for developing hypotension.