PO61 | RISK FACTORS FOR INTRACRANIAL HEMORRHAGE: AN UMBRELLA REVIEW TO INFORM MACHINE LEARNING PREDICTION MODELS IN GLIOMA PATIENTS RECEIVING ANTICOAGULATION
T.A. Adeyemo, S. Greenley, F. Ware, W. Jones, A. Maraveyas, F. Haque | 1Centre of Excellence for Data Science, Artificial Intelligence and Modelling DAIM, Faculty of Science and Engineering, University of Hull, UK; 2Hull York Medical School, Hull, UK; 3Academic Library Services, University of Hull, UK; 4NHS Humber Health Partnership, Hull, UK
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.
Authors
Introduction. People with glioma (PWG) requiring anticoagulation face high intracranial hemorrhage (ICH) risk due to tumor neovascularity and blood–brain barrier disruption. Current anticoagulation decisions rely largely on risk models derived from non-cancer populations, with limited validation in brain tumor cohorts.
Aim. To synthesize ICH risk factors across clinical populations, identify cancer-specific evidence gaps, and establish a framework to support algorithm development for predicting hemorrhage risk in anticoagulated PwG.
Methods. Systematic searches of Embase, MEDLINE, and the Cochrane Central Library were conducted up to November 2024 to identify systematic reviews reporting quantified associations between risk factors and intracranial hemorrhage. Following quality assessment using AMSTAR2, extracted associations were classified into nine clinical domains to support structured candidate predictor selection for subsequent validation in glioma populations.
Results. Seventy-nine reviews reporting 324 unique associations were synthesised. The anticoagulant class showed the strongest and most consistent influence on hemorrhage risk. Direct oral anticoagulants were associated with substantially lower intracranial hemorrhage risk than vitamin K antagonists and low-molecular-weight heparin across multiple populations, including brain tumor cohorts. Vitamin K antagonists markedly increased risk in patients with cerebral microbleeds. Dual antiplatelet therapy and polypharmacy further amplified hemorrhagic risk. Several commonly used co-medications were associated with an increased risk, including selective serotonin reuptake inhibitors and statins in acute stroke populations, with genetic modifiers such as APOE variants strongly influencing statin-related hemorrhage risk.
Conclusions. Medication-related factors dominate intracranial hemorrhage risk profiles, with anticoagulant class and drug interactions exerting major effects. These findings support the prioritization of anticoagulant selection and medication burden as core features in future prediction models for anticoagulated glioma patients. Prospective validation in brain tumor cohorts remains essential.
Keywords: Intracranial haemorrhage, glioma, anticoagulation, direct oral anticoagulants, bleeding risk, drug interactions
Downloads
Citations
How to Cite
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
