Impact of Brain Imaging Findings to Personalise AF Therapy
A framework for how brain MRI may influence treatment decisions is shown in Table 4 and figure 2. However, it is clear that at present the evidence for the predictive value of MRI findings in managing AF in clinical practice remains limited.
(Enlarge Image)
Figure 2.
Atrial fibrillation patients deemed unsuitable for oral anticoagulation (corresponding to table 4). T2 (A) and T2* image (B) at 1.5 T demonstrating confluent white matter hyperintensities and cerebral microbleeds in an atrial fibrillation patient with recent TIA, cognitive impairment, unsteady gait, frequent falls and a HAS-BLED score of 4. T2* images (C/D) at 3.0 T demonstrating an acute lobar haemorrhage and multiple cerebral microbleeds in an atrial fibrillation patient with suspected cerebral amyloid angiopathy.
In general, SBI are associated with an increased risk of overt ischaemic stroke and an increased risk of worsening of cognitive function, which more than doubles the risk of dementia in population cohorts. Comparable data specifically in patients with AF are not yet available despite the fact that cognitive decline in AF cohorts is reported, particularly after ischaemic stroke. Available evidence from clinical studies does not yet prove that optimising treatment of AF reduces the burden of SBI or cognitive decline. CT-detected SBI in patients with AF with clinically evident ischaemic stroke indicated a slightly higher risk for recurrent ischaemic stroke in the European Atrial Fibrillation Trial (EAFT), but this was not statistically significant (HR 1.18 (95% CI 0.79 to 1.77)). However, a recent retrospective MRI study demonstrated a HR of 1.79 (95% CI 1.09 to 2.93) for first time clinically evident stroke during a mean follow-up of 67 months in patients with AF with SBI compared with patients with AF without SBI. Further data in patients with AF with or without prior stroke but with SBI are not yet available.
Therefore, imaging-detected SBI in patients with AF cannot yet be regarded as a proven risk factor for AF-related future stroke or cognitive impairment (Table 4). Furthermore, there is currently no evidence to support the inclusion of SBI as evidence of 'stroke' in clinical risk prediction scores. It is clear that further prospective studies of SBI in AF are required to determine the clinical relevance of these common lesions for prognosis in AF cohorts.
WMH are associated with a higher mortality rate, cognitive dysfunction as well as an increased risk of ischaemic stroke and ICH in population-based studies. Furthermore, a subgroup analysis of the Stroke Prevention In Reversible Ischemia Trial (SPIRIT) and a small case–control study established an association between CT-detected leukoaraiosis and ICH in patients with ischaemic stroke without AF receiving warfarin, but the high target INR (3–4.5) and the specific population included in this study do not allow generalisability to current practice in AF cohorts (Table 4).
The small amount of available evidence does not support stopping or not starting OAC in patients with AF with additional stroke risk factors and WMH. In patients with AF with WMH, treatment of modifiable risk factors for stroke, SVD and ICH (eg, arterial hypertension, alcohol consumption, etc) is strongly encouraged. Whether direct OACs (DOACs) have a lower bleeding risk in patients with AF with WMH remains to be established.
CMBs are associated with an increased risk of ischaemic stroke and spontaneous ICH in both the general population and in ischaemic stroke or transient ischemic attack (TIA) cohorts. In a pooled meta-analysis, the OR for ICH in stroke or TIA cohorts overall was 8.52 while the OR for ischaemic stroke was 1.55. Data in cohorts exposed to OAC are limited but suggest a relationship between CMBs and warfarin-associated ICH in cross-sectional case–case comparisons. A prospective cohort study of 550 patients with ischaemic stroke or TIA in Asia showed that baseline CMBs were related to future ICH risk over 3.1 years, while the CHADS2 and CHA2DS2-VASC scores were not, suggesting that the MRI provided additional prognostic information. A recent study, also in ischaemic stroke and TIA and non-valvular AF, found that during a mean follow-up period of 24.7 months, 29 (10.4%) of 204 patients developed new CMBs, and that increasing age, increased WMH and prior CMBs were significantly associated with development of new CMBs. Thus, anticoagulation may increase the probability that an incident CMB evolves into a 'macrobleed' (symptomatic larger ICH).
Given the known association of CMBs with bleeding-prone arteriopathies (eg, hypertensive arteriopathy, cerebral amyloid angiopathy) (Table 4, figure 2), the presence and number of CMBs may have important implications for increased ICH risk in patients with AF with ischaemic stroke or TIA.
CMBs may also have particular relevance in survivors of ICH with an indication for long-term anticoagulation due to AF; recent reports suggest that this dilemma faces 30% of ICH survivors and that there is a low use of anticoagulants in this cohort due to clinical uncertainty, with a lack of validated risk predictors. Previous studies suggest that the burden of CMBs is a predictor of ICH recurrence in individuals with probable cerebral amyloid angiopathy and that CMBs may interact with antithrombotic treatments (at least for antiplatelet drugs) to increase the risk of further ICH. Thus, a pattern of CMBs suggesting cerebral amyloid angiopathy (CAA) (particularly if the CMB number is large) may present a strong caution against anticoagulation (figure 2C,D), but further prospective data in larger representative ICH cohorts are needed.
Another recent study investigated the risk of future symptomatic ICH in patients with probable CAA with only CMBs (including those patients presenting with cognitive impairment) found a considerable risk of incident ICH, which was increased by use of warfarin after adjusting for other confounders. Nevertheless, data are currently limited and it remains uncertain as to whether CMBs add to existing clinical risk prediction scores. Further prospective studies are needed to assess how CMBs influence OAC-related ICH risk in the different clinical populations including ischaemic stroke/TIA, ICH and those with cognitive impairment.
One recently identified new imaging marker associated with sporadic SVD (in particular CAA) is cortical superficial siderosis (cSS). cSS has been linked to an increased risk of future ICH risk in CAA cohorts, but we are not aware of any specific data on any interaction between cSS and the use of antithrombotic agents, and there are no studies available in AF cohorts.