TOPLINE

The Oakland score achieved the highest specificity (94.8%), outperforming clinical judgment and other established risk scores for identifying patients with lower gastrointestinal bleeding (LGIB) who could be safely discharged from the emergency department (ED).

METHODOLOGY

• LGIB accounts for more than 100,000 hospital admissions annually in the US. Although it is typically self-limited, a minority of patients experience severe bleeding that may require blood transfusion or surgery or may result in death. Several risk scores to guide safe discharge of patients with LGIB have been developed, but their advantage over clinical judgment remains uncertain.
• Researchers conducted a prospective cohort study involving 344 patients with LGIB (mean age, 59.2 years; 47.7% female) presenting to the ED across eight hospitals in Canada between September 2019 and February 2024. LGIB was defined as the passage of bright red blood per rectum or maroon-colored stool.
• The Oakland score was compared with other established LGIB risk scores (Strate and BLEED), upper gastrointestinal bleeding (UGIB) risk scores (Glasgow-Blatchford score [GBS] and clinical Rockall score [cRockall]), and the clinical judgment of treating physicians regarding patient admission or discharge.
• The primary outcome was safe discharge — defined as the absence of rebleeding; readmission; the need for blood transfusion; the need for endoscopic, radiologic, or surgical hemostasis; and death — within 28 days of presentation.

TAKEAWAY

• Overall, 142 patients were hospitalized. Within 28 days, 16 patients were readmitted, 41 required blood transfusion, 24 needed endoscopic hemostasis, two required radiologic embolization, seven underwent transanal surgery, five needed laparotomy for surgical hemostasis, and five died; none of the deaths were directly attributable to LGIB. Ultimately, 76.7% of the patients met all criteria for safe discharge.
• The Oakland score demonstrated discrimination with an area under the receiver operating characteristic curve (AUROC) of 0.77 for predicting safe discharge; clinical judgment yielded a similar AUROC of 0.76.
• In terms of the LGIB scores, the Oakland score demonstrated significantly better discrimination than the BLEED score (AUROC, 0.62) and discrimination similar to that shown by the Strate score (AUROC, 0.68). In terms of the UGIB scores, GBS showed better discrimination than the Oakland score (AUROC, 0.81 vs 0.77), whereas the Oakland score showed significantly greater discrimination than the cRockall score (AUROC, 0.63).
• Using a cutoff threshold ≤ 8, the Oakland score achieved the highest specificity (94.8%), outperforming the cRockall score (cutoff threshold < 1; specificity, 89.9%), GBS (cutoff threshold ≤ 1; specificity, 87.9%), and clinical judgment (specificity, 81.0%).

IN PRACTICE

“Ultimately, the Oakland score provides a powerful tool to predict who will experience adverse outcomes related to bleeding, but the clinician should still consider nonbleeding factors when deciding who to admit. As such, we would suggest the Oakland score be used in conjunction with clinical judgment to guide discharge decision-making and only when appropriate outpatient follow-up is available in the local jurisdiction,” the authors of the study wrote.

SOURCE

The study was led by Michael Sey, MD, MPH, Western University, London, Ontario, Canada. It was published online in Clinical Gastroenterology and Hepatology.

LIMITATIONS

Digital rectal examination, a component of the Oakland score, was not performed in many patients with LGIB. Milder cases of rebleeding in discharged patients may have been underrecognized. Blood transfusion practices varied, which may have affected the frequency of the observed clinical outcomes.

DISCLOSURES

The study received funding from an Innovation Grant provided by Academic Health Sciences Centres, Ontario, Canada. One author disclosed serving on advisory boards or consulting for various pharmaceutical and healthcare companies.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.