Author + information
- Received April 20, 2013
- Accepted May 20, 2013
- Published online November 19, 2013.
- Fiorenzo Gaita, MD∗∗ (, )
- Laura Corsinovi, MD, PhD∗,
- Matteo Anselmino, MD, PhD∗,
- Cristina Raimondo, MD∗,
- Martina Pianelli, MD∗,
- Elisabetta Toso, MD∗,
- Laura Bergamasco, Prof†,
- Carlo Boffano, MD‡,
- Maria Consuelo Valentini, MD§,
- Federico Cesarani, MD‖ and
- Marco Scaglione, MD¶
- ∗Cardiology Division, Department of Medical Sciences, University of Turin, Turin, Italy
- †Department of Surgical Sciences, University of Turin, Turin, Italy
- ‡Neuroradiology Department, IRCCS Foundation, Neurological Institute “C. Besta,”, Milan, Italy
- §Division of Neuroradiology, Azienda Ospedaliera Città Della Salute e Della Scienza, Turin, Italy
- ‖Division of Radiology, Cardinal Guglielmo Massaia Hospital, Asti, Italy
- ¶Division of Cardiology, Cardinal Guglielmo Massaia Hospital, Asti, Italy
- ↵∗Reprint requests and correspondence:
Prof. Fiorenzo Gaita, Cardiology Division, Azienda Ospedaliera Città Della Salute e Della Scienza, University of Turin, Corso Bramante 88, 10126 Turin, Italy.
Objectives The aim of this study was to compare the prevalence of silent cerebral ischemia (SCI) and cognitive performance in patients with paroxysmal and persistent atrial fibrillation (AF) and controls in sinus rhythm.
Background Large registries have reported a similar risk for symptomatic stroke in both paroxysmal and persistent AF. The relationship among paroxysmal and persistent AF, SCI, and cognitive impairment has remained uncharted.
Methods Two hundred seventy subjects were enrolled: 180 patients with AF (50% paroxysmal and 50% persistent) and 90 controls. All subjects underwent clinical assessment, neurological examination, cerebral magnetic resonance, and the Repeatable Battery for the Assessment of Neuropsychological Status.
Results At least 1 area of SCI was present in 80 patients (89%) with paroxysmal AF, 83 (92%) with persistent AF (paroxysmal vs. persistent, p = 0.59), and 41 (46%) controls (paroxysmal vs. controls and persistent vs. controls, p < 0.01). The number of areas of SCI per subject was higher in patients with persistent AF than in those with paroxysmal AF (41.1 ± 28.0 vs. 33.2 ± 22.8, p = 0.04), with controls reporting lower figures (12.0 ± 26.7, p < 0.01 for both). Cognitive performance was significantly worse in patients with persistent and paroxysmal AF than in controls (Repeatable Battery for the Assessment of Neuropsychological Status scores 82.9 ± 11.5, 86.2 ± 13.8, and 92.4 ± 15.4 points, respectively, p < 0.01).
Conclusions Patients with paroxysmal and persistent AF had a higher prevalence and number of areas of SCI per patient than controls and worse cognitive performance than subjects in sinus rhythm.
Atrial fibrillation (AF) is known to relate, independently of the presence of other comorbidities, to enhanced mortality and thromboembolism (1), particularly to the brain. To date, large registries have reported a similar risk for symptomatic stroke in both the paroxysmal and the persistent forms of this arrhythmia (2–4). A recent clinical experience in patients with acute cerebral infarctions suggested, however, that patients with persistent AF have significantly worse short-term outcomes in both functional ability and survival compared with those with paroxysmal AF (5).
If symptomatic brain damage is easily diagnosed, the relationship between AF, both paroxysmal and persistent, and silent cerebral ischemia (SCI) has remained uncharted. The presence of SCI is of relevance, because it has been associated with an increased occurrence of stroke (6) and the development of cognitive impairment (7,8).
Few studies (9–13) have investigated the prevalence of SCI in patients with AF. The limited sample sizes, retrospective designs, and heterogeneity of patients and techniques (e.g. computed tomography vs. magnetic resonance [MR]) involved in these studies explain the broad variations reported, from 15% to 86%. In addition, assessment of the relationship between paroxysmal and persistent AF and cognitive decline has, to date, relied only on clinical data not supported by cerebral imaging (14–17).
The aims of the present study were to compare: 1) the prevalence of SCI; and 2) the cognitive performance in patients with paroxysmal and persistent nonvalvular AF and in controls in sinus rhythm (SR), without histories of AF.
From November 2008 to September 2012, consecutive patients referred to our cardiology division or cardiovascular prevention outpatient clinic were screened to generate 3 groups of patients with paroxysmal or persistent AF or without histories of AF, balanced for age, sex, risk factors, and education level.
Exclusion criteria were valvular heart disease, acute coronary syndrome <3 months before, previous transcatheter ablation, pacemaker implantation or other MR contraindications, history of transient ischemic attack or stroke, autoimmune diseases, inflammatory brain diseases, tumors, severe hepatic disorders, severe chronic renal insufficiency, and alcohol abuse.
All patients were screened using the Mini Mental State Examination (18) and the Beck Depression Inventory (19) to rule out dementia and depression. Participants with scores ≤24 points on the Mini Mental State Examination and ≥ 10 on the Beck Depression Inventory were excluded.
A flowchart reporting numbers and reasons for exclusion of the screened cohort down to the final study population is shown in Figure 1.
The final study population consisted of 270 subjects: 180 patients with AF (90 paroxysmal and 90 persistent) and 90 controls in SR without histories of AF. AF was defined paroxysmal if self-terminating within 7 days and persistent if at least 1 AF episode lasted longer than 7 days, on the basis of direct patient interview and/or exhaustive medical record search.
All patients gave written informed consent before enrollment; the study was conducted in accordance to the latest Declaration of Helsinki and approved by the local ethics committee.
All screened subjects underwent extensive clinical assessment, including medical history (targeted to AF type and duration, presence of heart disease, and comorbidities), thromboembolic risk assessment (CHA2DS2-VASc score, consisting of congestive heart failure or left ventricular dysfunction, hypertension, age, diabetes, stroke or transient ischemic attack, thromboembolism, and vascular disease) (20), physical examination, and electrocardiography. Hypertension was defined as blood pressure ≥140/90 mm Hg (grade 1 hypertension, 140/90 to 159/99 mm Hg; grade 2 hypertension, 160/100 to 179/109 mm Hg) (21). Diabetes was defined as glycosylated hemoglobin ≥6.5% (22). Hypercholesterolemia was considered in case of evidence of a total cholesterol level >240 mg/dl (23). Smoking was considered habitual if the patient smoked at least 20 cigarettes a day for at least 1 year.
A standardized neurological examination, according to the National Institutes of Health Stroke Scale, was performed by a certified neurologist. In addition, all patients underwent Doppler echocardiographic sonography to exclude carotid and vertebrobasilar district morphologic and/or functional damage (stenosis of the lumen ≥70%, according to European Carotid Surgery Trial criteria, or peak systolic flow velocity >150 cm/s).
Cognitive function was assessed by a certified neuropsychologist using the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (24,25), exploring 5 domains: immediate memory, visual-spatial abilities, language, attention, and delayed memory.
Cerebral images were obtained using an 8-channel head coil 1.5-T MR imaging system (Magnetom Avanto; Siemens Healthcare, Erlangen, Germany). The imaging protocol included a sagittal T1-weighted spin-echo sequence (repetition time 400 ms, echo time 13 ms, slice thickness 5 mm, field of view 230 mm, matrix size 192 × 256), an axial T2-weighted fluid-attenuated inversion recovery (FLAIR) sequence (repetition time 8,500 ms, echo time 112 ms, inversion time 2,500 ms, slice thickness 5 mm, field of view 240 mm, matrix size 154 × 256), and a diffusion-weighted sequence (repetition time 3,200 ms, echo time 99 ms, slice thickness 5 mm, field of view 230 mm, matrix size 128 × 128, bandwidth 1,502 Hz, gradient strength 22 mT, duration of diffusion gradients 31 ms, gradient separation 42 ms in 3 orthogonal directions, B value 1,000). For each diffusion-weighted sequence, the apparent diffusion coefficient map was obtained. All sequences were centered on the axis defined by a line passing between the anterior and posterior cerebral commissures.
According to anatomic-pathological criteria (26), areas of SCI were defined as focal, sharply demarcated, regularly or irregularly shaped areas hyperintense on T2-weighted FLAIR images or isointense on T1-weighted images. T2-weighted FLAIR images were used to differentiate gliotic ischemic lesions smaller than 3 mm from perivascular spaces and lacunae (hypointense on T2-weighted FLAIR images) (27).
The presence of even only 1 lesion allocated the subject within the SCI group. Each individual area of SCI was recorded, and small lesions closely grouped together, exclusively localized in the subcortical white matter of the frontal lobe, were defined as a “spotted pattern.”
All MR scans were independently analyzed by 2 neuroradiologists, blinded to clinical data; conflict was resolved by common agreement referring to a third expert.
Continuous variables, presented as means and standard deviations, were compared by analysis of variance and subgroup comparisons corrected using the post hoc Bonferroni correction. Categorical variables, presented as counts and percentages, were compared in cross-tabulation tables using chi-square tests (with Yates correction as appropriate), and odds ratio (OR) with their 95% confidence intervals (CIs) were computed. A logistic regression model was performed to determine the correlation between SCI and potential confounders selected on univariate analysis (p < 0.50). All analyses were performed using SPSS for Windows version 18.0 (SPSS, Inc, Chicago, Illinois), and p values <0.05 were considered statistically significant.
Baseline characteristics of the final patient population are summarized in Table 1.
At least 1 area of SCI was present in 80 patients (89%) with paroxysmal AF, 83 (92%) with persistent AF (paroxysmal vs. persistent AF, p = 0.59), and 41 (46%) controls (paroxysmal vs. controls and persistent vs. controls, p < 0.01).
Cerebral lesions were bilateral in 81 patients (90%) with persistent AF and 80 (89%) with paroxysmal AF (p = 1.00), compared with 36 (40%) controls (paroxysmal vs. controls and persistent vs. controls, p < 0.01). Areas of SCI were cortical and subcortical in 75 (83%) and 77 (85%) patients with paroxysmal AF, 79 (88%) and 82 (91%) with persistent AF, and 19 (21%) and 44 (49%) controls, respectively (paroxysmal vs. persistent, p = 0.50 and p = 0.35; patients with AF vs. controls, p < 0.01).
The number of areas of SCI per subject was significantly higher in patients with persistent than those with paroxysmal AF (41.1 ± 28.0 vs. 33.2 ± 22.8, p = 0.04). The number of areas of SCI per subject was significantly higher in patients with paroxysmal and persistent AF than in controls, reporting lower figures (12.0 ± 26.7; paroxysmal vs. controls and persistent vs. controls, p < 0.01). AF showed a higher risk for SCI compared with subjects in SR (OR: 11.2; 95% CI: 6 to 21; p < 0.01), whereas no difference was found between the 2 different forms of the arrhythmia (OR: 1.5; 95% CI: 0.5 to 4.1; p = 0.61). On multivariate analysis, adjusted for age, CHA2DS2-VASc score, and antiplatelet or oral anticoagulation treatment, the presence of AF was strongly independently related to the presence of SCI (OR: 7.2; 95% CI: 2.3 to 22.3; p = 0.001).
The prevalence of the spotted pattern in the frontal lobe was more common in patients with persistent than those with paroxysmal AF (67% vs. 50%, p = 0.03). In controls, the presence of this patterns was negligible (1%).
Cognitive performance, assessed by the RBANS, was significantly worse in patients with persistent and paroxysmal AF than in controls: 82.9 ± 11.5, 86.2 ± 13.8, and 92.4 ± 15.4 points, respectively (paroxysmal vs. persistent AF, p = 0.08; paroxysmal and persistent AF vs. controls, p < 0.01). Specifically, patients with AF obtained lower mean scores than controls in all explored domains: immediate memory, visual-spatial abilities, language, attention, and delayed memory (Fig. 2). Concerning AF type, patients with persistent AF showed a trend toward lower RBANS scores, driven mainly by significantly worse visual-spatial abilities (84.8 ± 14.8 vs. 89.9 ± 18.2, p = 0.04) compared with patients with the paroxysmal form of the arrhythmia.
Among patients with AF, those with the frontal pattern (58%) showed a trend toward a lower mean total score compared with those without (82.9 ± 12.0 vs. 86.5 ± 15, p = 0.06). However, they scored significantly lower on the visual-spatial ability subtest compared with subjects without this pattern (85.1 ± 16.7 vs. 90.7 ± 17, p = 0.02).
The present study is, to the best of our knowledge, the first to investigate the relationship between AF and cognitive decline supported by specifically addressed cerebral MR imaging.
The major findings of this study are as follows: 1) patients with AF (both paroxysmal and persistent) have a higher prevalence but, most of all, number of areas of SCI per subject compared with controls; 2) persistent AF relates to more areas of SCI per subject compared with paroxysmal AF; 3) patients with AF have worse cognitive performance compared with controls; and 4) despite similar cognitive function by the RBANS between patients with paroxysmal and those with persistent AF, visual-spatial abilities were worse in patients with persistent AF.
The incidence of clinical ischemic stroke has been reported as up to 5-fold higher in patients with AF compared with the general population (2,3,28). Similarly, the present study, including subjects with homogenous cardiovascular risk factors by study design, reported a higher risk for SCI in patients with AF compared to subjects in SR.
The high prevalence of SCI in controls (46%), compared with that reported in the general population (29,30), may relate to the baseline characteristics of this group. Because of their selection within patients referring to a cardiovascular prevention outpatient clinic, the control group presented a moderate to high cardiovascular risk profile. Interestingly, the prevalence of SCI in subjects with a similar cardiovascular risk profile in previous experiences ranges from 53% to 58%, similar to that reported in the present study (12,13,31).
Besides the higher prevalence of SCI, by discriminating individual lesions in each patient, we registered a larger number of areas of SCI per person in patients with paroxysmal and persistent AF compared with controls.
SCI may certainly have both ischemic and embolic origins. The design of this study, comparing patients with AF and non-AF controls with similar prevalence of classic cardiovascular factors, should, however, permit a focus on the excess due to AF-related embolic lesions. Emboli of cardiac origin are generally smaller than those due to atherothrombotic material and cause lesions widely distributed, on both sides, of the brain (9,32,33). In contrast, emboli, for instance of carotid artery origin, are usually larger and damage the ipsilateral hemisphere. For these reasons, the peculiar cerebral MR pattern described in 50% and 67% of the patients with paroxysmal and persistent AF, respectively, presenting small, sharply demarcated lesions, often in clusters, with a bilateral distribution, prevalently in the frontal lobe, strongly supports an embolic mechanisms (Fig. 3). This pattern is in fact clearly distinguishable from atherosclerotic damage ipsilateral to a vascular lesion. In our opinion, the spotted pattern described in this study may derive from microembolization of multiple small platelet thrombi in the terminal brain vessels (especially the leptomeningeal arteries).
Furthermore, although patients with paroxysmal and persistent AF did not present differences in SCI prevalence, those with persistent AF showed a larger number of lesions. To date, persistent and paroxysmal AF have been considered to present similar stroke risk (2–4), and studies of symptomatic cerebral damage rarely differentiate between the 2 clinical forms of the arrhythmia.
This result might be related to the longer presence and duration of AF in the persistent form, causing a greater exposure time to AF-related thromboembolic etiology. Considering the high clinical impact this finding may provide, it surely warrants further confirmation.
As shown in Table 1, patients with paroxysmal AF were more commonly prescribed antiaggregant therapy compared with those with persistent AF. This somewhat surprising finding most likely is because patients were enrolled between 2008 and 2012. The evidence that aspirin is not beneficial in patients with AF first emerged in 2011 (34). Before this date, aspirin was commonly, although not on the basis of evidence, chosen for treatment of patients with paroxysmal episodes of AF.
As suggested by previous researches (16,17), AF relates to cognitive decline; in particular, Santangeli et al. (17) found that AF independently increases the risk for incident dementia in elderly patients without acute stroke and in those with normal baseline cognitive function.
Patients with AF performed worse in all investigated cognitive domains compared with subjects without histories of AF. Interestingly, the score difference highlighted between patients with AF and controls relates to a relevant clinical shift from medium to medium-low cognitive performance (24). Although patients with persistent AF had significantly lower RBANS scores for visual-spatial abilities compared with those with paroxysmal AF, no differences were found in all other evaluated subtests. We would argue that this might depend on the sensitivity of the neuropsychological test, which is capable of detecting only a few types of cognitive deficits within general cerebral function. Visual-spatial abilities represent a cognitive function related to the frontal subcortical cerebral circuits (35), which could indeed be damaged by the presence, at this level, of the spotted-pattern lesions. In conclusion, the anatomic brain damage visible by MR did not remain, from a functional point of view, truly “silent.” The present data suggest that initially limited cerebral damage may become overt as the number of cerebral lesions increases because of the persistence of the arrhythmia.
The following general limitations apply to this study. The cross-sectional study design investigating time-dependent risks may represent a source of bias. Future research on this topic should include longitudinal studies with appropriate power and sample sizes. Furthermore, as in most studies of this topic, an accurate measurement of the effective period of anticoagulation or antiaggregant therapy during exposure to the arrhythmia is lacking. For this reason, any correlation between antiaggregant or anticoagulant therapy and cerebral MR findings is avoided. Furthermore, the neuropsychological test used in the present study is a first-level assessment tool, exploring only a limited selection of cognitive functions. We therefore cannot exclude that the cognitive measures used may in certain cases perform above the detection threshold. For these reasons, the results of the present study are to be considered as suggestive, not conclusive, for further research on this clinically relevant topic.
As for symptomatic strokes, patients with AF, both paroxysmal and persistent, have a higher prevalence of AF-related SCI compared with controls in SR. Finally, patients with AF were confirmed to have worse cognitive performance than subjects in SR.
The Hospital of Asti, in collaboration with the University of Turin, supported the present research protocol (magnetic resonance scans). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- confidence interval
- fluid-attenuated inversion recovery
- magnetic resonance
- odds ratio
- Repeatable Battery for the Assessment of Neuropsychological Status
- silent cerebral ischemia
- sinus rhythm
- Received April 20, 2013.
- Accepted May 20, 2013.
- American College of Cardiology Foundation
- Benjamin E.J.,
- Wolf P.A.,
- D'Agostino R.B.,
- Silbershatz H.,
- Kannel W.B.,
- Levy D.
- Friberg L.,
- Hammar N.,
- Rosenqvist M.
- Hohnloser S.H.,
- Pajitnev D.,
- Pogue J.,
- et al.
- Hart R.G.,
- Pearce L.A.,
- Rothbart R.M.,
- McAnulty J.H.,
- Asinger R.W.,
- Halperin J.L.,
- for the Stroke Prevention in Atrial Fibrillation Investigators
- Vermeer S.E.,
- Hollander M.,
- Van Dijk E.J.,
- Hofman A.,
- Koudstaal P.J.,
- Breteler M.M.B.
- Gaita F.,
- Caponi D.,
- Pianelli M.,
- et al.
- Petersen P.,
- Madsen E.B.,
- Brun B.,
- Pedersen F.,
- Gyldensted C.,
- Boysen G.
- Ezekowitz M.D.,
- James K.E.,
- Nazarian S.M.,
- et al.,
- for the Veterans Affairs Stroke Prevention in Nonrheumatic Atrial Fibrillation Investigators
- Miyasaka Y.,
- Barnes M.E.,
- Petersen R.C.,
- et al.
- Beck A.,
- Steer R.,
- Brown G.
- Lip G.Y.H.,
- Nieuwlaat R.,
- Pisters R.,
- Lane D.A.,
- Crijns H.J.G.M.
- Graham I.,
- Atar D.,
- Borch-Johnsen K.,
- et al.
- Rydén L.,
- Standl E.,
- Bartnik M.,
- et al.,
- for the Task Force on Diabetes and Cardiovascular Diseases of the European Society of Cardiology (ESC) and of the European Association for the Study of Diabetes (EASD)
- Reiner Z.,
- Catapano A.L.,
- De Backer G.,
- et al.
- Jokinen H.,
- Kalska H.,
- Mäntylä R.,
- et al.
- Wolf P.A.,
- Abbott R.D.,
- Kannel W.B.
- Vermeer S.E.
- Vermeer S.E.
- Hougaku H.,
- Matsumoto M.,
- Kitagawa K.,
- et al.
- Svensson L.G.,
- Robinson M.F.,
- Esser J.,
- Fritz V.U.,
- Levien L.J.