What's new?

To the best of our knowledge, this is the first study showing an association between silent atrial fibrillation and a higher probability of dementia in a broad, representative population of older adults. The association seems to be independent of other factors, including history of stroke. The above phenomenon requires further research. However, regardless of the nature and direction of the relationship between arrhythmia and cognitive impairment, the results indicate the need for active search for asymptomatic atrial fibrillation (AF) in patients with dementia and screening for cognitive disorders in older individuals with accidentally diagnosed silent AF.

Introduction

Atrial fibrillation (AF) is the most common form of arrhythmia in the older population, and advancing age is the most prominent risk factor both for AF and its complications.¹ In recent years, we have observed not only an increase in the absolute number of AF patients, but also a rise in its prevalence, and mortality rates, which is mainly due to the aging of the general population and the rise in the incidence of age‑related diseases conducive to the development of AF.² It is estimated that individuals aged 60–69, 70–79, and 80–89 years are at a 4.98, 7.35, and 9.33‑fold higher risk of developing AF, respectively, as compared with those aged 50–59 years.³ Most patients with AF are over 65 years old.³ For this reason, proper management of AF patients requires taking into account the complexity of typical geriatric problems.⁴

The data regarding the prevalence of AF in this population appear to be significantly underestimated due to possible oligo- or asymptomatic course of arrhythmia. About one‑third of patients are unaware of AF symptoms.⁵ It seems that older adults are at a high risk of not being diagnosed with AF,^6,7 probably due to the masking influence of comorbidities and their treatment, lack of physical activity that might enhance symptoms, or simply worse medical care. According to current knowledge, silent AF (SAF) is still associated with an increased risk of ischemic stroke, major adverse cardiovascular events, and cardiovascular death, thus requiring assessment of cardioembolic risk and consideration of anticoagulation therapy.⁸ Moreover, in the case of asymptomatic AF, no diagnosis means no treatment, so in the long run, undiagnosed patients are at a higher risk of complications.⁵

Emerging evidence has shown a consistent association between AF and cognitive impairment or dementia, even without a history of stroke.^9-16 Although AF and dementia share many common risk factors, this association appears to be independent of their presence.^11,15 It is unclear, whether silent and symptomatic AF have similar associations with dementia.

Based on the cross‑sectional epidemiological NOMED‑AF (Non‑invasive Monitoring for Early Detection of Atrial Fibrillation) study,¹⁷ we investigated the real AF prevalence in a representative Polish population aged 65 years or more via long‑term continuous electrocardiography (ECG) telemonitoring.^17,18 In the presented ancillary analysis of the NOMED‑AF trial,¹⁷ we aimed to evaluate the relationships between cognitive impairment and both symptomatic and silent AF. To the best of our knowledge, this is the first study to investigate a possible association between SAF and cognitive performance, identified via long‑term ECG monitoring in a large, cross‑sectional population of older adults without implanted devices.

Patients and methods

The study was conducted as a subanalysis of the NOMED‑AF trial,¹⁷ a cross‑sectional observational study aiming to evaluate the prevalence of AF and its associated comorbidities in a representative, community‑dwelling Polish sample of adults aged 65 years or more. The detailed study protocol has been published elsewhere.¹¹ The NOMED‑AF trial¹⁷ used a long‑term wearable noninvasive ECG monitoring system linked with an online platform for data analysis and storage. The studied group comprised of 3014 individuals, and the recruitment was performed between March 15, 2017 and March 10, 2018.

The enrollment was based on a random sample of individuals aged 65 years or more. The entire territory of the country was divided into 59 geographical strata. The regions from each stratum were then randomly selected by the proportional probability. At the end, the participants from the previously chosen areas were selected at random, based on their personal identity number. A similar number of men and women were assigned to each 5‑year age group.

The NOMED‑AF study¹⁷ was approved by the local bioethical committee (26/2015) and registered at clinicaltrials.gov (NCT03243474). Written informed consent was obtained from all participants.

Results

The study included 2203 participants who were assessed for cognitive impairment using the MMSE. Mean (SD) age of the respondents was 77.1 (7.87) years, 51.6% of whom were men.

In the analyzed population, AF was confirmed in 491 patients (22.3%). The respondents with AF were older, more often men, hypertensive, with diabetes mellitus, CAD, HF, CKD, or a history of stroke. They were less likely to suffer from hypercholesterolemia.

The silent form of arrhythmia was recognized in 203 AF patients (41.3%). The participants with SAF were more likely to be men and less likely to have a history of hypertension or HF. We did not observe differences in age, marital status, BMI, and comorbidities between the AF and SAF patients.

Baseline characteristics of the analyzed population are presented in Table 1.

Mean (SD) number of points obtained by the respondents in the MMSE was 25.8 (0.1), and it was lower in patients with AF (25.2 [0.22]), as compared with those free of AF (26 [0.11]; P <⁠0.001). This difference between the participants with and without AF was notable in the subgroup aged 65–79 years (26.8 [0.21] vs 27.4 [0.08]; P = 0.002), and absent in the patients aged 80 years and older (23.6 [0.36] vs 23.3 [0.24]; P = 0.87). Moreover, there was a difference in the number of points between the patients with symptomatic and asymptomatic AF, with the lower score obtained by the latter group (25.8 [0.26] vs 24.3 [0.38]; P = 0.001). This difference was significant in the patients aged 65–69 years (27.4 [0.2] vs 25.9 [0.43]; P = 0.001), but not in those aged 80 years and older (24.1 [0.46] vs 23 [0.56]; P = 0.05).

The number of points in the MMSE questionnaire is presented in Figure 1.

In the group diagnosed with AF, we had treatment data for 449 patients. Of those 261 (58.1%) were taking long‑term oral anticoagulation. There was no difference in the number of points obtained in the MMSE between the patients taking and not taking oral anticoagulants (25.2 [4.8] vs 25.3 [5.64]; P = 0.8).

Finally, in the hierarchic multivariable logistic regression model, the impact of individual factors on the probability of achieving an MMSE score below 24 points was assessed. The association of SAF (but not of symptomatic AF) with the MMSE score indicating the possibility of dementia turned out to be independent of other important demographic and clinical variables, such as age, sex, education, comorbidities, smoking status, the occurrence of falls, and the likelihood of depression (Table 2).

Discussion

In our large cross‑sectional study of an older population, the patients with AF scored lower on a test assessing cognitive abilities. Interestingly, the individuals with SAF seemed to have a higher probability of dementia than those with a symptomatic form of arrhythmia.

Over the last 25 years, the amount of data on the relationship between AF and mild cognitive impairment and dementia, even in the absence of a clinically overt previous stroke, has been growing.^9-16 Furthermore, these data indicate an association between AF and both vascular dementia and Alzheimer disease.^12,13 The vast majority of observational data come from studies that determine the presence of AF based on medical history, medical records, or single ECG records, thus excluding patients with latent and asymptomatic arrhythmia. To the best of our knowledge, this study is the first to investigate a possible association between SAF and cognitive performance assessed with a screening tool.

Interestingly, although the MMSE score was lower among the elderly AF patients, after taking into account other factors (age, sex, marital status, comorbidities), only the silent form of arrhythmia was associated with a result indicating dementia. This may mean that the patient population identified by long‑term ECG monitoring differs from the groups analyzed so far in this respect. It seems that the adopted methodology allowed for the detection of an additional number of patients with paroxysmal arrhythmia. Chen et al¹⁹ indicate a relationship between dementia and only the persistent form of AF. Our results also suggest that age is such a strong prognostic factor for AF and dementia that it is difficult to recognize relationships between them which are independent of it.

The possible link between SAF and poorer performance in cognitive screening can be interpreted in many ways. First, patients with SAF are less likely to benefit from AF treatment protecting against the development of dementia. In fact, there are data on the beneficial effects of anticoagulant treatment²⁰ and AF ablation²¹ on the development of cognitive disorders. Second, patients with SAF are generally likely to receive less medical care, which can lead to a lack of prophylaxis and / or delayed diagnosis and treatment of dementia. An opposite mechanism should also be considered. Patients with dementia less often report symptoms of arrhythmia, so arrhythmia in this group could be diagnosed only with systematic screening. Finally, it cannot be ruled out that SAF has a specific hemodynamic / prothrombotic / inflammatory phenotype, which predisposes a patient to specific neurological damage. This final aspect, of course, requires additional research.

In fact, it is still debatable whether the relationship between AF and cognitive impairment is causal. The basic impediment to resolving this issue is the impact of common risk factors, including, in particular, age and vascular or inflammatory diseases.²² Among the factors responsible for the influence of AF on the development of dementia, are, obviously, microembolisms or silent infarcts,^23,24 but also hemodynamic disturbances related to irregular / fast heart rate,^25,26 a decrease in cardiac output related to the loss of atrioventricular synchrony,²⁷ and damage to the blood‑brain barrier due to the inflammatory process accompanying AF.^28-30 Both the hemodynamic alterations and vascular inflammation could favor amyloid β-protein accumulation and phosphorylation of τ protein leading to Alzheimer disease.^30,31 The possibility of complications resulting from microbleeding within the central nervous system in AF patients who use anticoagulants should also be considered. As already mentioned, stroke prophylaxis is associated with a reduction of the risk of dementia in AF, but interestingly enough, it seems that direct oral anticoagulants carrying a lower risk of intracranial bleeding are more effective in this respect than vitamin K antagonists.³² Our study did not show any differences in the results of the screening test assessing cognitive performance between the patients receiving and not receiving cardioembolic stroke prophylaxis. It should be highlighted, however, that we did not have complete data on this treatment in the study group, which could have influenced the results.

Another interesting aspect indicating a relationship between arrhythmia and cognitive impairment is the influence of AF on the development of cognitive disorders in younger patients. Our study showed that the relationship between the presence of AF in long‑term monitoring and a lower MMSE score was visible in the subgroup of patients aged 65–79 years, but lost relevance in the older subgroup, as confirmed by other studies.^16,33 We might speculate about the additional influence of risk factors of neurodegeneration in the patients aged 80 years and older.

The beneficial effects of education and the relationship between a history of stroke and the likelihood of cognitive disorders are undisputed. However, the protective effect of smoking in the past is interesting. In fact, quitting smoking has been shown to reduce the risk of dementia.³⁴ Perhaps in this case, giving up the addiction was an element of a larger process of health‑promoting changes in the respondents.

This study indicates a substantial frequency of SAF in the group of patients of over 65 years and its relationship with worse performance in the cognitive testing. It seems to point to the need for systematic AF screening in this age group and implementation of efficient treatment. The extent to which this strategy may reduce the development of dementia requires further research.

The study has certain limitations. The cross‑sectional nature of the analysis did not allow for drawing conclusions as to the direction of the causal relationship. The lack of data on the antiocoagulation treatment used did not permit a fully reliable assessment of its impact on the issue under study. Additionally, the selection of patients may have been biased due to the high heterogeneity of the study population and low representation of participants with moderate and severe dementia.