RHEMITT Score Predicts Rebleed After Capsule Endoscopy: First Validation at a U.S. Tertiary Care Center

Patients with obscure gastrointestinal bleeding undergo small bowel capsule endoscopy (SBCE), but often return for recurrent bleeding or anemia. The RHEMITT score evaluates patients based on 7 variables (heart failure, chronic kidney disease, Saurin P1/P2 lesions, major bleeding, incomplete SBCE, smoking status, and endoscopic treatment) and seeks to predict the risk of rebleeding.

This study aims to perform an external validation of the RHEMITT score in the United States.

SBCEs performed to evaluate anemia or GI bleeding from a tertiary-care center's PillCam database between 1/22/2018 and 7/21/2020 were reviewed. Variables based on the RHEMITT score were collected. The primary outcome was rebleeding, defined as (1) melena or hematochezia or (2) hemoglobin drop of 2 g/dL. Patient were categorized into low, intermediate, and high-risk categories based on RHEMITT score. The accuracy of the RHEMITT score for predicting rebleeding was assessed.

A total of 361 SBCEs were included in the study. Age, indication for SBCE, endoscopic treatment, antiplatelet use, cirrhosis, heart failure, chronic kidney disease, and major bleeding were significantly associated with risk of rebleed (p < 0.05). Each increasing risk category for the RHEMITT score predicted increased probability of this study's primary outcome, rebleeding (p < 0.001). There was a significant association between RHEMITT risk category and rebleeding-free survival (log-rank p < 0.001). An area under the receiver operating characteristic curve for the RHEMITT score was 0.790 (p < 0.001).

Our findings validate the RHEMITT score and confirm acceptable performance for predicting rebleeding at a tertiary referral center in the United States.