Gastrointestinal Bleeding Management and Utility of Clinical Scores in Predicting Disease and Outcome in a Modern European Population

Abstract

Acute gastrointestinal bleeding remains a significant clinical challenge, with early risk stratification playing a critical role in guiding management decisions and optimizing patient outcomes. Risk scores, particularly the Rockall and Glasgow-Blatchford scores, have been widely validated for upper gastrointestinal bleeding (UGIB) over the past few decades, assisting clinicians in differentiating patients who require intensive treatment from those who can be managed conservatively. However, the changing etiology of UGIB, including the decline in Helicobacter pylori infection and increased Nonsteroidal Anti-Inflammatory Drug related peptic ulcers, alongside evolving patient profiles with more comorbidities and frailty, necessitates a re-evaluation of these risk models. This thesis addresses the need for updated risk scores to reflect these changes in patient demographics, bleeding causes, and treatment modalities. In the context of endoscopic treatment, while current therapies are effective, there remains a need for advanced treatments for lesions in difficult-to-reach areas or those with extensive bleeding surfaces. One promising option is the topical haemostatic gel PuraStat®, which has shown potential in clinical trials, especially for lesions where traditional treatments are inadequate. The thesis conducted a retrospective analysis at Tallaght University Hospital (TUH), demonstrating that PuraStat® is both safe and effective, particularly in cases where standard therapies fail. However, further prospective studies are needed to validate these findings and assess PuraStat®'s applicability in different lesion types. The thesis also addresses the evolving landscape of UGIB and the role of clinical scores in predicting outcomes. It highlights that while patient demographics and causes of bleeding have changed, risk scores like Rockall Score and Glasgow-Blatchford Score (GBS) remain valuable for risk stratification, predicting endoscopic findings, and forecasting patient outcomes. A prospective study across hospitals in Ireland and Spain confirmed the continued relevance of these scores, even in a modern clinical setting with younger, more anticoagulant-treated patients. For small bowel bleeding (SBB), the thesis explores the lack of effective clinical scores tailored to this condition. The study showed that the GBS and Rockall scores are also useful in predicting findings on video capsule endoscopy and patient outcomes in overt SBB. A decision-making algorithm incorporating GBS was developed, but further prospective studies are necessary to validate its use in SBB exclusively. Additionally, the RHEMITT score, used to predict rebleeding in SBB, was found to be useful in patients with high scores but less effective for those with lower scores, suggesting a need for refinement.