Blog → Lower GI Bleed Scores – Which Best Predicts Badness?

Lower GI Bleed Scores – Which Best Predicts Badness?

Listen to this article
0:00 –:–

Written by Shannon Markus

Spoon Feed
For lower GI bleeds, the Oakland and SALGIB risk stratification scores performed well in predicting adverse events and severe bleeding and in identifying patients safe for discharge, but none of the 5 scoring systems reliably predicted need for intervention.

Gee… I… Believe we have a good GIB tool!
Gastrointestinal bleeding (GIB), especially lower GIB (LGIB), remains a common ED presentation with variable severity. Although many LGIB cases resolve without intervention, a significant subset require transfusion, procedures, or lead to severe bleeding or death, making early risk stratification essential. This large retrospective study of 2,051 LGIB patients evaluated five clinical prediction tools (Oakland, SALGIB, CHAMPS, NOBLADS, Strate) for predicting adverse outcomes, severe bleeding, and safe discharge. Adverse events occurred in 38.6% of patients, severe bleeding in 29.8%, and in-hospital mortality was 4.3%. The Oakland and SALGIB scores showed the best overall performance across outcomes, each demonstrating excellent AUROC values for predicting both adverse events and severe bleeding, as well as identifying patients safe for discharge. CHAMPS, NOBLADS, and Strate performed well for predicting mortality but were less effective for broader outcomes. All scoring systems performed poorly in predicting the need for therapeutic intervention. Overall, Oakland and SALGIB were the most reliable tools for ED risk stratification. A major limitation is that this single-center retrospective design limits generalizability, and multicenter prospective studies are needed to confirm whether these findings hold in other settings.

How does this change my practice?
I don’t routinely use formal risk scores for GI bleeds, relying instead on my own key clinical factors (vitals, active bleeding, hemoglobin, orthostasis, anticoagulation). Most patients are obviously high or low risk, but some fall into a gray zone. After reviewing each of the scores’ actual components, the top performers (Oakland and SALGIB) include a few elements I’ll start paying closer attention to (male sex, prior GIB admissions, platelet count). The Oakland Score is already on MDCalc and simple to use, so I plan to try it on my next shift!

Source
Predictive value of clinical risk scores for adverse outcomes and safe discharge in acute lower gastrointestinal bleeding. Am J Emerg Med. 2026 Jan;99:370-375. doi: 10.1016/j.ajem.2025.10.051. Epub 2025 Oct 25. PMID: 41175440.

View Critical Appraisal

Critical Appraisal

Title: Predictive value of clinical risk scores for adverse outcomes and safe discharge in acute lower gastrointestinal bleeding.
Background: Lower gastrointestinal bleeding (LGIB) is a common cause of emergency department visits, with significant morbidity and resource utilization. Identifying patients at risk for adverse outcomes or safe discharge remains a clinical challenge.
Study Question: What is the predictive value of various clinical prediction tools for adverse outcomes and safe discharge in patients with acute LGIB?
Study Design: Retrospective cohort study
Population: Adult patients (≥18 years) with a final diagnosis of acute LGIB in the emergency department.
Intervention: Evaluation of clinical prediction tools (CHAMPS, NOBLADS, Oakland, SALGIB, and Strate scores).
Comparison: Not applicable
Primary Outcome: Predictive performance for adverse outcomes and safe discharge.
Secondary Outcomes: None reported
Key Statistics: Area under the curve (AUC) for predictive performance of clinical scores.
Results: The Oakland score had the highest AUC for adverse outcomes (0.902), severe bleeding (0.925), and safe discharge (0.889). The SALGIB score followed with AUCs of 0.880, 0.901, and 0.867, respectively.
Risk of Bias – Newcastle-Ottawa Scale:
  • Selection (2/4 stars) — Some concerns: Potential selection bias due to retrospective design and use of ICD-10 codes.
  • Comparability (1/2 stars) — Some concerns: Limited control for confounding variables.
  • Outcome (2/3 stars) — Some concerns: Outcome assessment based on electronic health records, with potential for incomplete data.
Take Home: The Oakland and SALGIB scores are effective tools for predicting adverse outcomes and safe discharge in patients with acute LGIB, aiding in clinical decision-making.
This critical appraisal is generated using a structured framework applied to the full text of the original article and is presented separately from the physician’s commentary.

Leave a Reply