Brain MRI scans can inform the choice between OACs and LAA closure for non-valvular AF

M Edip Gurol

M Edip Gurol is a stroke neurologist with particular expertise in the care of patients at high risk for ischaemic (blockage type) strokes and haemorrhages. His research focuses on clarifying the mechanisms of brain small vessel diseases that increase the risk of both types of strokes. In this commentary he discusses how MRI can inform treatment decisions for patients with atrial fibrillation (AF).

Non-valvularatrial fibrillation (NVAF) continues to be a major cause of ischaemic strokes, while oral anticoagulants (OAC) used for prevention of embolism in this setting increase the risk of intracerebral haemorrhages (ICH), a less common but more deadly and disabling type of stroke. Over 50% of patients sustaining a warfarin-related ICH die within the first three months and only one third of survivors achieve some degree of independence in activities of daily living.1 Despite lower rates of ICH in a low risk population demonstrated in phase three trials of non-vitamin K OACs (NOAC), the outcomes of NOAC-ICH are dismal, similar to warfarin-ICH.2 Neurologists see a high volume of both ischaemic strokes and ICHs, realising early during their training the disproportionately high mortality and morbidity of ICH. Most internists who prescribe OACs also encounter at least a few ICHs and easily understand the severity of this stroke type. The fear of OAC related ICH is the main cause of underutilisation of stroke prevention measures in NVAF. The introduction of left atrial appendage closure (LAAC) procedures for stroke prevention in NVAF gave the clinicians an important therapeutic tool that can obviate the need for life-long anticoagulation. As a result, LAAC considerably decreases ICH risk, by over 50% even when compared to NOACs based on the results of a recent network meta-analysis.3 In the light of these advances, it had become even more important to stratify the risk of ICH based on non-invasive imaging markers in order to select the appropriate stroke prevention approaches in NVAF patients.

By far the great majority of the ICHs in older adults occur as a result of cerebral small vessel diseases (cSVD) that are not amenable to endovascular treatment. These SVDs are widespread cerebral pathologies mainly caused by hypertension (HTN-SVD) and cerebral amyloid angiopathy (CAA), the latter being the prototype high ICH risk pathology of brain vessels. Extensive research using brain MRI revealed a number of neuroimaging markers that are associated with CAA and/or HTN-SVD. Many of these age-related lesions are not acutely symptomatic but they represent the severity of the underlying SVD and they are associated with risk of future ICH. A working knowledge of these MRI findings is crucial for treatment decisions involving long-term anticoagulation and their alternatives such as LAAC for NVAF. It is important to note that obtaining a high-resolution susceptibility weighted imaging (3D SWI) MR sequence is important for optimal visualisation of some of the subtle haemorrhagic findings depicted below.

Cerebral microbleeds (CMB) represent tiny amounts of blood extravasated from small vessels of the brain. CMBs found in strictly cortical/superficial parts of the elderly brain are mostly caused by CAA whereas HTN-SVD is the major cause of deep CMBs. In the Rotterdam study, about 19% of clinically healthy adults over 45 years of age were found to have at least one CMB and the presence of one or more CMBs conferred a five to six times higher ICH risk in follow up when compared to similarly aged adults without any CMB.4 Another observational study evaluating the impact of incidentally found CMBs on MRI showed a high 5% risk of first time ICH in these individuals. Older age and warfarin use were associated with significantly higher ICH risk providing support to the view that OACs are not benign medications in patients with CMBs even without past history of brain bleed.5 A large meta-analysis of patients who had brain MRI after an ischemic stroke or TIA found CMBs in one fourth of these patients and having one or more CMBs in this setting was again associated with a 6.3 fold increase in ICH risk during follow-up. Higher number of CMBs were associated with an elevated risk, up to 14 fold in patients who had five or more microbleeds.6 The location of CMBs matter and the ones in more superficial/cortical regions are associated with higher ICH risk than deep-seated microbleeds. There are few and much less common causes for CMBs not related to primary cSVD such as radiotherapy, past infective endocarditis or valve replacement surgery and microbleeds in these patients may not necessarily portend a high risk. A good clinical radiological review is warranted to understand the cause of CMBs and stratify ICH risk whenever these tiny lesions are found on MRI.

Brain MRI showing 2 incidentally found cortical microbleeds (arrows) in an older adult with nonvalvular atrial fibrillation on a Factor Xa inhibitor.
Brain MRI showing 2 incidentally found cortical microbleeds (arrows) in an older adult with nonvalvular atrial fibrillation on a Factor Xa inhibitor. Two years later patient presented with quickly progressive loss of consciousness and head CT shows a large intracerebral haemorrhage probably originating from the site of the cerebral microbleeds

Another subtle haemorrhagic marker that can be found on SWI MRI scans is cortical superficial siderosis (cSS), linear thin regions of bleeding at the most superficial parts of the cortex or even within the brain sulci. They have the typical appearance of “train track”, a sign again noticeable on SWI MRI.2 These lesions are found much less commonly in patients without past ICH, but when found, they are significant risk markers. A different MRI finding, the presence of severe leukoaraiosis or white matter hyperintensities on FLAIR MRI are also associated with higher baseline ICH risk that is exacerbated with long-term anticoagulant use.2

Despite the fact that it is difficult to assess the magnitude of risk in individual patients, it is now well-known that the presence of CMBs, cSS or severe white matter disease on MRI are non-trivial risk markers for future ICH in most situations. Their presence should prompt clinicians caring for NVAF patients to discuss pros and cons of different stroke prevention options such as LAAC, although NOACs can also be considered in select patients at the lower end of the range of ICH risk. While clinical trials in this high-risk patient population are awaited, shared-decision-making is the most appropriate approach to discuss different aspects of the available prevention measures and help the patient make an informed decision.


  1. Rosand J, Eckman MH, Knudsen KA, Singer DE, Greenberg SM. The effect of warfarin and intensity of anticoagulation on outcome of intracerebral hemorrhage. Archives of internal medicine. 2004;164(8):880-884.
  2. Gurol ME. Nonpharmacological Management of Atrial Fibrillation in Patients at High Intracranial Hemorrhage Risk. Stroke. 2018;49(1):247-254.
  3. Koifman E, Lipinski MJ, Escarcega RO, et al. Comparison of Watchman device with new oral anti-coagulants in patients with atrial fibrillation: A network meta-analysis. International journal of cardiology. 2016;205:17-22.
  4. Akoudad S, Portegies ML, Koudstaal PJ, et al. Cerebral Microbleeds Are Associated With an Increased Risk of Stroke: The Rotterdam Study. Circulation. 2015;132(6):509-516.
  5. van Etten ES, Auriel E, Haley KE, et al. Incidence of symptomatic hemorrhage in patients with lobar microbleeds. Stroke. 2014;45(8):2280-2285.
  6. Wilson D, Charidimou A, Ambler G, et al. Recurrent stroke risk and cerebral microbleed burden in ischemic stroke and TIA: A metaanalysis. Neurology. 2016;87(14):1501-1510.


M Edip Gurol is based at Massachusetts General Hospital / Harvard Medical School, Boston, MA, USA


Please enter your comment!
Please enter your name here