Clinical prediction model may help determine management for acute lower GI bleeding

Factors associated with safe discharge in the development and validation cohorts were age, female sex, no previous admission for lower GI bleeding, no blood on digital rectal exam, heart rate, systolic blood pressure, and hemoglobin level.


A recently developed clinical prediction model may help determine when acute lower GI bleeding can be safely managed on an outpatient basis.

Researchers in the United Kingdom used data from the National Comparative Audit of Lower Gastrointestinal Bleeding, which involved adults with lower GI bleeding at 143 hospitals in 2015, to identify predictors of safe discharge, then converted the model into a risk scoring system. The model was then externally validated in 288 patients admitted to two additional U.K. hospitals with lower GI bleeding, as well as compared with other existing risk scores. The study results were published online June 23 by The Lancet Gastroenterology & Hepatology.

Patients in the derivation cohort had a mean age of 68 years, while those in the validation cohort had a mean age of 66 years. The development cohort involved 2,336 admissions, of which 1,599 (68%) resulted in safe discharge. “Safe discharge” was defined as the absence of rebleeding (additional blood transfusions or additional decrease in hematocrit of ≥20% after 24 hours of clinical stability), red blood cell transfusion, therapeutic intervention to control bleeding (i.e., endoscopic, radiological, or surgical hemostasis), all-cause in-hospital death, and readmission with additional lower GI bleeding within 28 days. Factors associated with safe discharge in the development and validation cohorts were age, female sex, no previous admission for lower GI bleeding, no blood on digital rectal exam, heart rate, systolic blood pressure, and hemoglobin level. Patients with a risk score of 8 or lower were determined to have a 95% chance of safe discharge. The new score was more effective than the Rockall, Blatchford, Strate, BLEED, AIMS65, and NOBLADS scores for predicting safe discharge, with the AIMS65 most discriminative for in-hospital death and BLEED most discriminative for hemostatic intervention.

The researchers noted that the study cohorts included only patients who were admitted to the hospital and that some of the included patients may not have had true lower GI bleeding, among other limitations. They also noted that prospective validation of the new score in patients with different risk profiles, as well as a randomized controlled trial, would be useful to confirm their results. However, they concluded that their score was effective at identifying patients with GI bleeding who could be safely managed as outpatients rather than as inpatients and could lead to economic and resource savings. “[Our score] could be routinely incorporated into triage pathways for acute medical and surgical admissions to identify patients with lower gastrointestinal bleeding who can be safely discharged,” the authors wrote.