The efficacy and comparison of upper gastrointestinal bleeding risk scoring systems on predicting clinical outcomes among emergency unit patients.

BACKGROUND

Gastrointestinal bleeding is a significant cause of morbidity and mortality among emergency unit patients. Several scoring systems are verified for predicting hospitalization and mortality such as Glasgow Blatchford Bleeding Score (GBS), AIMS65 score, Rockall score (RS), and International Bleeding Risk Score (INBS; ABC score). The aim of this study is to evaluate the efficacy and predictive value of these scoring systems.

METHODS

Adult emergency unit patients with gastrointestinal bleeding were retrospectively enrolled. The age, gender, complaints at admission, vitals and examination results, laboratory findings, outcomes, blood transfusion status, and endoscopic interventions were all reported, and GBS, AIMS65, RS, and INBS (ABC) scores were calculated individually for all enrollies.

RESULTS

A total of 311 patients were included. The median age of participants was 70 years (IQR (25-75%): 59-81), and 202 (65%) of them were male. The efficacy of all four scoring systems (GBS, AIMS65, RS, and INBS (ABC)) in predicting hospitalization, need of blood transfusion, determination of high- and low-risk patients, and mortality was found to be statistically significant (p < 0.05 for all). ROC-AUC analysis was revealed that while GBS is the most beneficial in predicting hospitalization, INBS (ABC) has the best predictive value on mortality. Besides, the only scoring model with predictive value in determining the need for endoscopic intervention was RS (p < 0.05).

CONCLUSION

The present study showed that, among adult emergency unit patients with gastrointestinal bleeding, GBS, AIMS65, RS, and INBS (ABC) scores could successfully predict hospitalization, need of blood transfusion, determination of high- and low-risk patients, and mortality. However, the only scoring system that could be used to determine the need of endoscopic intervention is RS. Finally, we believe further studies with prospective enrollment would be beneficial for more accurate conclusions.