Identifying Emergency Department Patients at Low Risk for a Variceal Source of Upper Gastrointestinal Hemorrhage

Abstract

Background

Assessing the likelihood of a variceal versus nonvariceal source of upper gastrointestinal bleeding (UGIB) guides therapy, but can be difficult to determine on clinical grounds. The objective of this study was to determine if there are easily ascertainable clinical and laboratory findings that can identify a patient as low risk for a variceal source of hemorrhage.

Methods

This was a retrospective cohort study of adult ED patients with UGIB between January 2008 and December 2014 who had upper endoscopy performed during hospitalization. Clinical and laboratory data were abstracted from the medical record. The source of the UGIB was defined as variceal or nonvariceal based on endoscopic reports. Binary recursive partitioning was utilized to create a clinical decision rule. The rule was internally validated and test characteristics were calculated with 1,000 bootstrap replications.

Results

719 patients were identified; mean age was 55 years old and 61% were male. There were 71 (10%) patients with a variceal UGIB identified on endoscopy. Binary recursive partitioning yielded a two-step decision rule (platelet count greater than 200,000/uL, and an INR less than 1.3), which identified patients who were low risk for a variceal source of hemorrhage. For the bootstrapped samples, the rule performed with 97% sensitivity (95% confidence interval [CI] 91-100%) and 49% specificity (95% CI 44- 53%).

Conclusion

Although this derivation study must be externally validated before widespread use, patients presenting to the ED with an acute UGIB with platelets of greater than 200,000/uL and an INR less than 1.3 may be at very low risk for a variceal source of their upper gastrointestinal hemorrhage.

This article is protected by copyright. All rights reserved.

http://ift.tt/2v22mXJ