Article Text

Download PDFPDF

Antithrombotic therapy and the risk of new-onset dementia in elderly patients with atrial fibrillation
Free
  1. Chun Ka Wong1,
  2. Duo Huang1,2,
  3. Mi Zhou1,
  4. JoJo Hai1,
  5. Wen Sheng Yue2,
  6. Wen-hua Li3,
  7. Li-Xue Yin3,
  8. Ming-Liang Zuo3,
  9. Ying Qing Feng4,
  10. Ning Tan4,
  11. Ji Yan Chen4,
  12. Joseph Kwan1,
  13. Chung Wah Siu5
  1. 1 Cardiology Division, Department of Medicine, The University of Hong Kong, Hong Kong, China
  2. 2 Medical Imaging Key Laboratory, Affiliated Hospital of North Sichuan Medical College, Nanchong, China
  3. 3 Department of Echocardiography & Non-invasive Cardiology Laboratory, Sichuan Academy of Medical Sciences & Sichuan Provincial People's Hospital, Chengdu, China
  4. 4 Department of Cardiology, Guangdong Cardiovascular Institute, Guangdong Provincial Key Laboratory of Coronary Heart Disease Prevention, Guangdong General Hospital, Guangdong Academic of Medical Sciences, Guangzhou, China
  5. 5 Division of Geriatrics, Department of Medicine, The University of Hong Kong, Hong Kong, China
  1. Correspondence to Chung Wah Siu, Cardiology Division, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, Hong Kong, China; cwdsiu{at}yahoo.com.hk

Abstract

Background Atrial fibrillation (AF) is associated with an increased risk of dementia. Little is known about the relationship of antithrombotic therapy and the risk of dementia in patients with AF without clinical stroke.

Method This was an observational study based on a hospital AF registry. Patients aged 65–85 years at the time of AF diagnosis were identified via the computerised database of the clinical management system. Patients with prior stroke or known cognitive dysfunction were excluded. The primary outcome was newly diagnosed dementia during the follow-up period.

Results 3284 patients (mean age 76.4±5.3 years, 51.6% male) were included for analysis. The mean CHA2DS2-VASc score was 3.94±1.44. 18.5% patients were prescribed warfarin, 39.8% were prescribed aspirin and 41.7% were prescribed no antithrombotic therapy. After a mean follow-up of 3.6 years, 71 patients (2.2%) developed dementia, giving rise to an incidence of 0.61%/year. The incidence of dementia were 1.04%/year, 0.69%/year and 0.14%/year for patients on no therapy, aspirin and warfarin, respectively. Both univariate and multivariate analyses showed that age ≥75 years, female gender and high CHA2DS2-VASc score were associated with significantly higher risk of dementia; warfarin use was associated with significantly lower risk of dementia (HR: 0.14%, 95% CI 0.05 to 0.36, p<0.001). Patients on warfarin with time in therapeutic range (TTR) ≥65% had a non-significant trend towards a lower risk of dementia compared with those with TTR <65%.

Conclusion In elderly AF patients, warfarin therapy was associated with a significantly lower risk of new-onset dementia compared those with no therapy or aspirin.

  • Neurology
  • dementia
  • cardiology

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

INTRODUCTION

Atrial fibrillation (AF) is the most commonly encountered sustained arrhythmia, which causes a fivefold increase in risk of ischaemic stroke.1 AF has been increasingly recognised to be associated with cognitive impairment. One systematic review found that AF was associated with a significantly increased risk of cognitive dysfunction (HR, HR 2.3).2 Another systematic review of 21 prospective and non-prospective studies confirmed that AF was significantly associated with a greater risk of cognitive impairment among patients with a first-ever or recurrent stroke (RR 2.7), as well as in a broader population including patients with or without any prior history of stroke (relative risk, RR 1.4).3 For those patients who have never suffered an acute stroke and with normal cognitive function at baseline, a further systematic review also found that AF was associated with a significantly increased risk of incident dementia (HR 1.42).4 Hence, it seems that the association between AF and cognitive impairment may not be dependent upon a clinically apparent stroke event. However, it is unclear what factors influence this link, and more importantly, whether antithrombotic treatment can reduce the risk of developing new-onset dementia in AF patients. It was demonstrated that warfarin with a target international normalised ratio (INR) of 2.0 to 3.0 reduced risk of ischaemic stroke by 64% in patients with non-valvular AF.5–7 We therefore hypothesised that effective anticoagulation with warfarin may be associated with a lower risk of new-onset dementia, as compared to aspirin or no treatment. This study examined the relationship between the choice of antithrombotic therapy and the risk of developing new-onset dementia among patients with AF and no prior history of stroke or cognitive dysfunction, and no subsequent clinical stroke events.

METHODS

Study design and patients

This was an observational study based on real-world data from AF registry of Queen Mary Hospital, Hong Kong.8–15 The study protocol was approved by Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB). Informed consent was waived due to the registry nature of the study and all clinical data were analysed in an anonymised manner. Patients of Chinese ethnicity aged between 65 and 85 years old at the time of AF diagnosis (made at Queen Mary Hospital, Hong Kong) were identified via the computerised database of the hospital. Patients were excluded if they had previously documented dementia, stroke (ischaemic or haemorrhagic) prior to the diagnosis of AF and/or stroke (ischaemic or haemorrhagic) during the follow-up period. In addition, patients with significant valvular heart disease, prior valvular intervention and/or AF secondary to reversible conditions such as thyroid disease, infection and surgery16–19 were excluded. Patients with incomplete baseline and/or follow-up data were excluded. Demographic data, cardiovascular comorbidities and medications at baseline were documented. The index date was defined as the date when AF was first diagnosed. A blanking period of 14 days after the index date was adopted for the registration of outcomes during follow-up period, as the diagnosis of dementia made within the first few days of AF diagnosis is most likely related to the initial presentation than a new event.20 Stroke risk was estimated using the CHA2DS2-VASc score on index date (C: congestive heart failure (1 point); H: hypertension (1 point); A2: age 65–74 years (1 point) and age ≥75 years (2 points); D: diabetes mellitus (1 point); S: prior stroke or transient ischaemic attack (2 points); VA: vascular disease (1 point); and Sc: sex category (female) (1 point)).21–24 CHA2DS2-VASc score of each individual patient was not continuously updated for new occurrence of risk factors encompassed in the CHA2DS2-VASc score after the index date.

Among those prescribed warfarin, the Rosendaal method was utilised to calculate the time in therapeutic range (TTR).8 11 25–29 Briefly, INR was assumed to vary in a linear manner between intermittent measurements and the assumption was used to interpolate INR values on days without measurement. TTR is defined as the percentage of time during which the measured and/or interpolated INR is within 2.0 to 3.0. Patients were divided into two groups based on their TTR (TTR ≥65% and TTR <65%).

Outcomes, variables and data source

The primary outcome was newly diagnosed dementia during the follow-up period. Data were retrieved from the medical records stored in the territory-wide information network shared by all public hospitals in Hong Kong. Patients with a new coding of ‘dementia’ and/or ‘Alzheimer’s disease’ during the follow-up period were identified as having experienced the primary outcome.

Statistical analysis

Continuous and discrete variables are expressed as mean ± standard derivation and percentages, respectively. Statistical comparison of the baseline clinical characteristics between two groups were performed using Student’s t-test or nonparametric Mann-Whitney test. Comparisons involving more than two groups were performed using one-way ANOVA. Kaplan-Meier survival analyses with the log-rank test were performed for survival analysis. Cox proportional hazards regression model was used to calculate HRs of predictive factors and their 95% CI. Calculations were performed using SPSS software (version 21.0). All tests were two-sided and a p value <0.05 was considered statistically significant.

RESULTS

Between July 1997 and December 2011, 10 195 Chinese patients’ AF from Queen Mary Hospital, Hong Kong, were identified from hospital database. Patients were excluded according to the inclusion and exclusion criteria as stated in the method section. Three thousand two hundred and eighty-four Chinese patients with non-valvular AF and aged between 65 years and 85 years with no history of stroke (ischaemic and/or haemorrhagic) and/or prior dementia were included in this analysis (figure 1). Table 1 summarises their clinical characteristics. Their mean age was 76.4±5.3 years with a male predominance (51.6%); 54.5% patients had hypertension, 22.0% had diabetes mellitus and 22.3% had heart failure. Their mean CHA2DS2-VASc score was 3.94±1.44.

Table 1

Baseline characteristics

Figure 1

Study population. AF, atrial fibrillation; VHD, valvular heart disease.

Of these, 608 patients (18.5%) were prescribed warfarin therapy, 1306 patients (39.8%) were on aspirin and the remaining 1370 patients (41.7%) did not receive any antithrombotic therapy. Compared with patients on aspirin and patients on warfarin, patients on no antithrombotic therapy were less likely to have hypertension (18.2%, vs 25.4% vs 26.2%, p<0.001), diabetes mellitus (46.6%, vs 60.6% vs 58.9%, p=0.001), heart failure (17.7%, vs 25.4% vs 26.2%, p<0.001), coronary artery disease (8.8%, vs 27.5% vs 26.2%, p<0.001) and peripheral artery disease (1.2%, vs 2.5% vs 3.6%, p<0.001) (table 1). Among patients on warfarin, the mean TTR was 40.3±23.4%. There were 94 patients (15.5%) on warfarin achieving TTR ≥65%, and the remaining 514 patients (84.5%) had TTR <65%.

New-onset dementia

After a mean follow-up of 3.6 years (11 605 patient-years in total), there were 71 cases of new-onset dementia (2.2%) with an annual incidence of 0.61%/year. Dementia occurred in 32 patients on no antithrombotic therapy, compared with 34 patients on aspirin, and 5 patients on warfarin. Figure 2A depicts the Kaplan-Meier analysis of dementia among patients on no therapy, aspirin and warfarin (Log-rank: 20.9, p=0.0001). The annualised incidence of dementia among Chinese AF patients on no therapy, on aspirin and on warfarin were 1.04%/year, 0.69%/year and 0.14%/year, respectively (figure 2B). Table 2 summarises the clinical factors that predict new-onset dementia together with the corresponding HRs based on Cox proportional hazard model and 95% CIs. On univariate analysis, age ≥75 years, female gender and high CHA2DS2-VASc score were associated with higher risk of dementia. On the other hand, warfarin therapy was associated with significantly lower risk of dementia (HR: 0.14%, 95% CI 0.05 to 0.36, p<0.001). Nonetheless, in multivariate analysis, these associations remained significant (table 2). To assess the impact of the quality of anticoagulation therapy on the incidence of dementia, patients on warfarin were subdivided into those with TTR ≥65% and TTR <65%. Five out of 514 patients on warfarin with TTR <65% diagnosed dementia during the follow-up period with the annual dementia incidence of 0.16%/year whereas there was no patient with TTR ≥65% diagnosed dementia, with the annual dementia rate essentially 0.0%/year. Although numerically, patients on warfarin with TTR ≥65% had a lower incidence of dementia compared with patients on warfarin with TTR <65%, the Kaplan-Meier analyses did not demonstrate a statistically significant difference in dementia incidence.

Table 2

Association between baseline factors and new-onset dementia

Figure 2

(A) Kaplan-Meier estimates of new-onset dementia-free survival in older AF patients with different antithrombotic regimen; and (B) the annual risk of dementia in older AF patients on no therapy, aspirin and warfarin with time in therapeutic range (TTR) ≥65% and <65%. AF, atrial fibrillation.

DISCUSSION

In a real-world clinical setting with a mean follow-up period of 3.6 years, our prospective study examined the relationship between antithrombotic choice and the incidence of dementia among older Chinese AF patients (aged 65 to 85 years) with no prior history of stroke or cognitive dysfunction, and no subsequent clinical stroke events. The four key findings from our study are: first, older female patients with a higher CHA2DS2-VASc score were at a significantly higher risk of developing new-onset dementia. Second, patients who received no antithrombotic therapy had an annual dementia risk of 1.04% per year. Third, the risk of dementia among patients who received antithrombotic therapy, particularly warfarin, was about 7.5-fold lower than those who received no antithrombotic therapy (0.14%/year vs 1.04%/year). And lastly, AF patients who were on warfarin with TTR ≥65% had a non-significant trend towards a lower risk of new-onset dementia, as compared to those on warfarin with TTR <65% (0.16%/year vs 0%/year).

Due to the rapidly ageing population and improved medical care, life expectancy is improving worldwide. The global disease burden of dementia is immense and dementia is the top contributor to adult disability and morbidity. The prevalence of dementia is projected to double every 20 years with an economic cost of 1% of the global GDP.30 31 In a meta-analysis including 160 studies worldwide,32 the prevalence and incidence of dementia vary widely from 4.6% to 6.9% and from 1.7% per year to 5.3% per year in people aged 60 years or older. It appears that the prevalence and incidence were highest in North America and lowest in Asia. In the Delphi study, the prevalence of dementia in people aged 60 years or older was 5.4–6.4% in the Western countries, and 4.0% in China.33 In a recent systematic review involving 32 116 Chinese people from 13 prospective studies between 1990 and 2010, the incidence of dementia in people aged 60 years or older was 0.99% per year, and the prevalence of dementia increased by 50% over the past 20 years.34 In our study, the highest annual dementia risk for patients who did not receive any antithrombotic therapy was 1.04% per year, and the lowest was 0.14% per year for those who received warfarin. This apparently lower incidence could have resulted from the lack of comprehensive cognitive screening protocol to detect individual patients with early dementia as used in previous study, which means that the true incidence rates were likely to have been higher than reported in our study.

Like dementia, AF is the most common sustained cardiac arrhythmia worldwide with an increasing prevalence with age.35 36 Although the association between AF and dementia risk is no longer in doubt, the underlying mechanism remains unclear. Several mechanisms have been proposed. First, AF and dementia share a number of common risk factors including advanced age, diabetes mellitus, hypertension, heart failure and others.37 Many of these risk factors may lead to micro-vascular and/or macro-vascular diseases resulting in repetitive silent cerebral vascular injuries and white matter lesions.3 38 39 These risk factors also constituted the important components of CHA2DS2-VASc score21; AF patients with higher CHA2DS2-VASc scores tend to be more prone to developing silent or clinically apparent ischaemic stroke, which could also lead to subsequent cerebral injury and cognitive decline.40 41 Logically speaking, one would therefore expect AF to preferentially increase the risk of vascular (over other types of) dementia owing to its thromboembolic tendency to cause silent or clinically apparent subcortical and cortical infarcts.42 However, one large long-term (5 years) follow-up study of 37 025 patients demonstrated that AF was independently associated with all types of incident dementia including senile, vascular and Alzheimer’s dementia.43 In that study, however, the highest risk was found in the younger age group (<70 years), whereas this was opposite to the findings in our study. Importantly, in our study, we excluded patients who have had any prior history of stroke or subsequent clinical stroke events. Hence, the associations found in our study were independent of any clinical stroke events, although we did not examine for the occurrence of silent stroke events (eg, by routine follow-up neuroimaging).

Second, the beat-to-beat variation in stroke volume and resultant reduced cerebral perfusion due to decrease in cardiac output might predispose to cognitive impairment and development of dementia.44 AF has been demonstrated to be associated with reduced brain volume and cognitive function independent of cerebral infarcts,45 as well as exacerbating cerebral perfusion.46 In support of this theory, a study evaluating the impact of AV-nodal ablation in AF patients demonstrated a sustained improvement in memory and cognitive function in patients who had regular stroke volume and normal cerebral perfusion restored after AV nodal ablation followed by permanent pacemaker implantation.47 Lastly, inflammation, endothelial dysfunction and platelet activation may contribute to the prothrombotic pro-inflammatory state associated with AF.48–50 Endothelial activation could lead to the secretion of many pro-inflammatory cytokines and growth factors, such as thrombin, which has in turn been shown to be an inflammatory neurotoxin in AD.51 52 One randomised controlled trial (RCT) of 34 older patients also found that high-intensity cholesterol-lowering therapy provided an anti-inflammatory action—leading to a reduction in plasma concentration of IL-1RA, IL-2, IL-9 and IL-12, and macrophage inflammatory protein-1β (MIP-1β)—that might ameliorate neurocognitive decline and loss of brain volume in amygdala and hippocampus.53 In our study, we did not examine for beat-to-beat variation in stroke volume, cerebral perfusion or blood biomarkers of inflammation or endothelial activation.

This real-world observational study is the first to demonstrate that warfarin therapy may be associated with a significantly lower risk of developing new-onset dementia in AF patients with no prior stroke or cognitive dysfunction, and no subsequent clinical stroke events. Our results should be added to the recent systematic review, which included 19 studies that assessed the association between cognitive decline and AF thrombo-prophylaxis.54 In that review, data from two prospective studies (one RCT) comparing anticoagulation versus antiplatelet therapy revealed that anticoagulation may be associated with a non-significantly lower decline in Mini-Mental Score Examination (MMSE) score between baseline and follow-up.54 However, there was no robust evidence to suggest that anticoagulation may be effective in reducing the risk of new-onset dementia, as compared to antiplatelet agents (two studies) or no treatment (three studies).54 Another ongoing Cochrane systematic review is investigating the effects of antithrombotic therapy in preventing cognitive decline in people with small vessel disease on neuroimaging but without dementia.55 This review plans to include RCTs that have recruited patients with and without AF.

Data from our study also suggested that AF patients who were on warfarin with a higher TTR of ≥65% appeared to have a non-significantly lower risk of new-onset dementia, as compared to those on warfarin but with a lower TTR of <65%. This trend was observed despite none of these patients had a new clinical stroke event. This is consistent with one study of 2605, which showed that, for AF patients on warfarin, time spent inside therapeutic range (TTR) was correlated with the long-term risk of developing dementia.56 This lends further weight in supporting the hypothesis that AF-related cognitive decline may be primarily a result of microvascular occlusive events or silent strokes. The importance of achieving a high TTR in order to maximise the benefit of warfarin therapy for stroke prevention in AF is well known.57 58 However, our study is the first to explore the impact of TTR on the risk of developing new-onset dementia in patients with AF on warfarin but without new clinical stroke events. Due to their non-inferiority and/or superiority of non-vitamin K oral anticoagulants (NOACs) as compared to warfarin for stroke prevention in older people with AF,35 it is reasonable to hypothesise that NOACs might also be effective in preventing cognitive decline and dementia in AF.

Limitations

This study was limited by being single-centre-based and observational in nature. Since the selection of antithrombotic strategies was not randomised on an individual basis, causality of the correlations found in our study cannot be proven. However, real-world databases are useful for generating hypotheses and exploring practice patterns, especially among patient groups that are not usually represented by clinical trials (eg, the older and frailer patients). For this study, there was no institutional guideline restricting choice of antithrombotic therapies and drug regimens were decided by each attending physician. Although stroke risk assessment was on index date using the CHA2DS2-VASc score, patients’ risk profile might have changed over time and continuous update of the CHA2DS2-VASc score was not performed. Data concerning short-term interruption of antithrombotic therapy due to various causes, such as withholding antithrombotic therapy during peri-operative period and/or during occurrence of clinically significant bleeding, were not available. While we ascertained all newly diagnosed dementia and Alzheimer’s disease by careful review of hospitalisation records, we recognised that there were potential problems with documentation and coding, possibly leading to under-diagnosis (more likely than over-diagnosis). Indeed, in the current study, the diagnosis of dementia is often made at the time while the patients exhibited features of dementia instead of through comprehensive cognitive screening protocol, thereby those with early dementia with mild symptoms might not be diagnosed. Moreover, this study was not designed to examine the associations with mild cognitive impairment or subjective cognitive decline. In addition, the study had limited power to detect true difference in dementia incidence between patients on warfarin with different TTR due to the relatively small sample size and low incidence of dementia in patients on warfarin, although there were numerically fewer patients on warfarin with good TTR. Furthermore, in addition to antithrombotic regimen, other cardiovascular risk factors including smoking, and dyslipidemia, use of lipid-lowering agents and certain echocardiographic parameters may have impacts on the development of dementia. Regrettably, those data were not available for analysis. In addition, subjects in the present study were exclusively of Chinese ethnicity living in Hong Kong, thereby potentially limiting the generalisability of our results to other ethnic groups, given the possible variations in disease processes.

CONCLUSIONS

In this prospective observational study based on real-world data, older Chinese AF patients with no prior history of stroke or cognitive dysfunction, and no subsequent clinical stroke events, warfarin therapy were associated with a significantly lower risk of new-onset dementia, when compared with no therapy and/or aspirin. In the future, well-designed large-scale controlled studies are needed to determine the pathophysiological mechanisms of AF-related cognitive decline, and test whether NOACs might be superior to warfarin in preventing cognitive decline and dementia in AF.

Main messages

  • We demonstrated a high incidence of dementia (0.61%/year) in a cohort of 3284 patients with atrial fibrillation.

  • Antithrombotic therapies were associated with lower incidence of dementia with annual incidence of 1.04%/year, 0.69%/year and 0.14%/year in patients no antithrombotic therapy, aspirin, and warfarin, respectively.

  • Among patients on warfarin, there was a non-significant trend towards a lower risk of dementia compared with those with TTR <65%.

Current research questions

  • Does atrial fibrillation cause dementia?

  • What is/are the underlying mechanism(s) for the higher incidence of dementia in patients with atrial fibrillation?

  • Is vitamin K antagonist oral anticoagulant associated with lower risk of dementia?

What is already known on the subject

  • Atrial fibrillation is the most commonly encountered sustained arrhythmia that increases risk of ischaemic stroke.

  • Atrial fibrillation is associated with increased risk of cognitive impairment with or without clinically apparent stroke event.

  • Contemporary antithrombotic therapies reduce risk of ischaemic stroke, nonetheless their effects on cognition remain uncertain.

REFERENCES

Footnotes

  • Contributors CKW, JSK and CWS contributed to the conception and design of the study. CKW, DH, MZ, JH, WSY, WHL, LXY, MLZ, YF, NT, JC, JSK and CWS contributed to the acquisition of data. Data analysis and interpretation were conducted by CKW, JSK and CWS. CKW, JSK and CWS wrote first draft of the protocol. DH and MZ revised the manuscript critically for important intellectual content. All authors have read and approved the final version of the manuscript to be published.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests No, there are no competing interests for any author.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement No data are available.

  • Conflict of interests The authors report no relationships that could be construed as a conflict of interest. 10.1016/S0140-6736(14)61066-7