Author + information
- †Department of Medicine, Columbia University Medical Center, New York, New York
- ‡Department of Epidemiology, Columbia University Medical Center, New York, New York
- ↵∗Reprint requests and correspondence:
Dr. Steven Shea, Columbia University, 535 West 116th Street, New York, New York 10027.
In this issue of the Journal, Gaita et al. (1) report findings from 270 patients, of whom 180 had atrial fibrillation and 90 with similar demographic and clinical characteristics did not have atrial fibrillation. All were free of clinical stroke at the time of the study, and all underwent brain magnetic resonance imaging (MRI). Silent cerebral ischemia (SCI) was defined as a focal, sharply demarcated area on MRI of hyperintensity on T2-weighted fluid-attenuated inversion recovery or isointensity on T1-weighted imaging. The numbers of SCI lesions, and the areas of the brain in which they were observed, were determined. A standardized measure of cognitive function was also obtained. Gaita et al. found a substantially higher prevalence of SCI in those with atrial fibrillation compared with those in sinus rhythm, with an approximately 3-fold greater number of areas of SCI in those with atrial fibrillation. Cognitive performance was worse in those with atrial fibrillation. Two other analyses were also performed. Gaita et al. (1) attempted to distinguish between patients with persistent and those with paroxysmal atrial fibrillation, and they attempted to relate the MRI findings to cognitive performance. We comment below on each of these findings.
Silent cerebral ischemic scars may be caused by small hemorrhagic infarctions, related in part to hypertension, or by emboli, which may originate in the fibrillating atria or from plaque in the cerebrovascular circulation (2). Ezekowitz et al. (3) reported in 1995 that 14.7% of 516 patients with nonrheumatic atrial fibrillation had silent ischemic infarcts on computed tomography, a less sensitive modality than MRI. There were no controls in that study. Kobayashi et al. (4) reported in 2012 that patients with nonvalvular atrial fibrillation had a higher prevalence and greater number of silent infarctions on MRI compared with a well-matched control group. Neither of these reports provided data on cognitive function. Thus, although the association of atrial fibrillation with stroke and transient ischemic attack is firmly established, the study by Gaita et al. (1) adds important confirmatory data supporting an association with SCI. From their imaging description and the example they show in Figure 3, Gaita et al. (1) analyzed 1 of the manifestations of silent brain ischemia, the presence of white matter hyperintensities (WMHs). Surprisingly, they did not report separate results on another less common manifestation of ischemia, namely, the presence of silent brain infarcts (SBI). SBIs, which are focal lesions seen on both computed tomography and MRI with similar intensity as cerebrospinal fluid (2), are present in 8% to 28% of patients in the general population, with most of the difference in prevalence being explained by age (2). While SBI and WMH are strongly associated with each other and share common risk factors (hypertension being the most important ) and probably a common etiology (microvascular disease), they also appear to have differences that go beyond their appearance on brain imaging. In the Framingham Offspring study, both SBI and WMH were independently associated with future stroke occurrence, but only WMH was associated with death and cognitive impairment, independent of vascular risk factors and interim stroke (5). In the population-based Northern Manhattan Study, increased physical activity was associated with a decreased frequency of SBI but had no effect on WMH volume (6). While SBI may be more representative of a vascular etiology (whether microvascular or embolic), WMH may more often have alternative explanations (glyosis, demyelination, degenerative or inflammatory changes, venous disease). An association of atrial fibrillation with SBI in the study of Gaita et al. (1) would have lent stronger support to an embolic etiology of their findings, as opposed to the possible combined effects of other mechanisms that are also more frequent in atrial fibrillation than in sinus rhythm, such as increases in platelet activation, thrombin generation, endothelial dysfunction, and inflammation (7).
To distinguish persistent from paroxysmal atrial fibrillation, Gaita et al. (1) used a clinical definition based on the duration of 1 or more episodes lasting >7 days, on the basis of data from the clinical history and medical record. Gaita et al. (1) did not find important differences between the paroxysmal and persistent groups. Kobayashi et al. (4) also did not find differences for paroxysmal versus persistent atrial fibrillation. However, there are at least 2 serious challenges in these analyses. First, it is difficult to be sure of the validity of the classification, because atrial fibrillation can be asymptomatic. Second, some patients may have had paroxysmal atrial fibrillation before they had persistent atrial fibrillation, but it cannot be determined whether the SCIs observed on MRI occurred during 1 phase or the other. Although inferences about paroxysmal versus persistent atrial fibrillation on the basis of these data are weak, this limitation does not affect the overall comparisons between those with atrial fibrillation versus controls.
Gaita et al. (1) also found measurably lower cognitive performance in patients with atrial fibrillation compared with controls. This finding is consistent with a recent meta-analysis of observational studies in which risk for cognitive impairment in atrial fibrillation was increased in patients with and without stroke before the cognitive assessment (8). As pointed out in that meta-analysis, methods for ascertaining atrial fibrillation and for measuring cognitive function are not standardized across studies. In addition, Gaita et al. (1) analyzed relationships between cognitive performance and SCI density and location. These analyses suggest that location may be more important than the number of lesions, but the analyses were not based on a priori hypotheses, and multiple test statistics were calculated. These findings should therefore be regarded as exploratory. Nonetheless, the observations are provocative. Gaita et al. (1) argue that the finding of bilateral, clustered small lesions, predominantly in the frontal lobes, namely, the pattern observed in the present study, suggests thrombotic emboli arising in the fibrillating atria and can be distinguished from larger ipsilateral (or monolateral) infarctions suggestive of atheroembolic emboli originating in the cerebrovascular circulation. The location of infarctions in the study of Kobayashi et al. (4) was similar to that reported here. Given the aforementioned uncertainty on the possibility of an embolic etiology of WMH, the embolic significance of the preferential (“spotted”) pattern observed by Gaita et al. (1), although an intriguing concept, remains in need of further research.
When analyzing the relationship between atrial fibrillation and silent brain ischemia, the most important question is whether prophylactic antithrombotic treatment may be able to reduce the incidence of silent brain lesions as it has been proved to do for clinical strokes. In the Gaita et al. (1) study, an adequate measurement of patients' exposure to antithrombotic agents during the atrial fibrillation episodes was not available, preventing the investigators from drawing any conclusions regarding treatment effect. This important analysis will require appropriately designed and powered prospective studies.
Finally, paroxysmal, often asymptomatic episodes of atrial fibrillation may be more frequent in the general population than previously thought, and their impact on the attributable risk for stroke, SCI, or cognitive impairment is unknown. The detection, monitoring, and characterization of atrial fibrillation in the population was set as 1 of the future health policy goals by health organizations in the United States (9) and in Europe (10). Devising feasible and cost-effective strategies for detecting asymptomatic atrial fibrillation in subgroups, such as the elderly, that are at high risk for the arrhythmia may become crucial to dealing with a condition that, as the report by Gaita et al. (1) reminds us, can damage the brain in more ways than one.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Drs. Shea and Di Tullio receive funding from the National Institutes of Health.
- 2013 American College of Cardiology Foundation
- Gaita F.,
- Corsinovi L.,
- Anselmino M.,
- et al.
- Ezekowitz M.D.,
- James K.E.,
- Nazarian S.M.,
- et al.
- Kobayashi A.,
- Iguchi M.,
- Shimizu S.,
- Uchiyama S.
- Debette S.,
- Beiser A.,
- DeCarli C.,
- et al.
- Lim H.S.,
- Willoughby S.R.,
- Schultz C.,
- et al.
- Estes N.A. III.,
- Sacco R.L.,
- Al-Khatib S.M.,
- et al.