Stroke: MRI predicts intracranial hemorrhage with long-term oral anticoagulants

  • Martí-Fàbregas J & al.
  • Neurology
  • 19.04.2019

  • von Susan London
  • Clinical Essentials
Der Zugang zum gesamten Inhalt dieser Seite ist nur Angehörigen medizinischer Fachkreise vorbehalten. Der Zugang zum gesamten Inhalt dieser Seite ist nur Angehörigen medizinischer Fachkreise vorbehalten.


  • Among older adults starting long-term oral anticoagulation for recent cardioembolic ischemic stroke, those with advanced cerebral small vessel disease, evidenced by microbleeds or moderate-to-severe white matter hyperintensities, had elevated risk for intracranial hemorrhage (ICH).

Why this matters

  • Concern about ICH risk and lack of reliable predictors limits use of oral anticoagulation.

Key results

  • On MRI, baseline prevalences:
    • Microbleeds: 22.5%.
    • Moderate/severe white matter hyperintensities: 45.1%.
    • Superficial siderosis: 3.0%.
  • Overall, 1.9% of patients experienced intracranial hemorrhage.
  • Multivariate risk for intracranial hemorrhage elevated for patients with:
    • Microbleeds (HR, 2.7; P=.031).
    • Moderate/severe white matter hyperintensities (HR, 5.7; P=.006).
  • Model C index: 0.76.
  • Patients having both microbleeds, moderate/severe white matter hyperintensities had highest ICH rate: 3.76/100 patient-years.

Expert comment

  • In an editorial, Luis F. Maia, MD, PhD, and Duncan Wilson, MD, PhD, write, “Available data do not support changes to the current clinical practice recommendations on anticoagulation after [ischemic stroke] based on imaging evidence for these vasculopathies.”

Study design

  • Multicenter prospective cohort study of 937 patients aged >64 years with recent cardioembolic ischemic stroke who were new users of oral anticoagulation.
  • Baseline MRI used to evaluate microbleeds, white matter hyperintensities, cortical superficial siderosis.
  • Main outcome: ICH during mean 23.1-month follow-up.
  • Funding: Fondo de Investigaciones Sanitarias Instituto de Salud Carlos III; Bristol-Myers Squibb/Pfizer.


  • Possible underestimation of ICH.
  • Some endpoints assessed remotely.
  • Possible selection bias.