Clinical characteristics and type of antithrombotic treatment in a Spanish cohort of elderly patients with atrial fibrillation according to dependency, frailty and cognitive impairment.

BACKGROUND: Available data regarding clinical profile and management of elderly patients with atrial fibrillation (AF) according to dependency, fragility and cognitive impairment are scarce. The objective of the study was to analyze the biodemographic data, clinical profile and antithrombotic treatment according to dependency, fragility and cognitive impairment in elderly AF patients.
METHODS: Cross-sectional and multi-center study performed in consecutive AF patients ≥ 75 years treated with oral anticoagulants ≥ 3 months attended in Internal Medicine Departments in Spain.
RESULTS: A total of 837 patients (83.0 ± 5.0 years; CHA2DS2-VASc: 5.0 ± 1.4; HAS-BLED: 2.1 ± 0.9) were included. 44.4% of patients had some degree of dependency, 43.3% were fragile, and 32.3% had cognitive impairment. Patients with any of these conditions were older, had a worse clinical profile, with more comorbidities and higher risks of thromboembolic and bleeding events. All these conditions were independently associated among them. Overall, 70.8% of patients were taking vitamin K antagonists, the remaining 29.2% direct oral anticoagulants and 9.7% oral antiplatelets. This distribution was independent of the presence of dependency or fragility, but there was a trend to a higher prescription of vitamin K antagonists in those patients with cognitive impairment (75.2% vs. 68.8%; P = 0.05).
CONCLUSIONS: Approximately 32%-44% of elderly anticoagulated AF patients attended have some degree of dependency, fragility and/or cognitive impairment. Patients with any of these conditions are older and have a worse clinical profile. Approximately 71% of patients are taking vitamin K antagonists, regardless dependency or frailty, but with a trend to higher prescription in patients with cognitive impairment.

Introduction

Atrial fibrillation (AF) is very common in elderly patients.[1]–[3] Thus, more than 10% of patients aged 70 years or older have AF, number that reaches 18% in those individuals aged 80 years or older.[1] Overall, patients with AF have a high number of comorbidities.[1]–[4] Remarkably, elderly patients have some additional particularities, including cognitive impairment, dependency, falls or frailty that markedly increase the risk of developing complications in this population.[5]–[8] In fact, compared with younger patients, elderly patients with AF have an increased risk of death, stroke and bleeding.[9]

Except when contraindicated, AF patients aged 75 years or older should be chronically anticoagulated.[5],[10],[11] Unfortunately, many elderly patients do not receive oral anticoagulation regardless thromboembolic or bleeding risk, due to a number of causes that include frailty, cognitive impairment, disability, concern about bleeding risk, and classically, the well known limitations of vitamin K antagonists (VKA), such as the narrow therapeutic window, periodic monitoring of anticoagulation activity, frequent dose adjustments, interactions with many drugs, etc.[5],[12] Under anticoagulation leads to a higher risk of stroke and death in these patients.[13]

Direct oral anticoagulants (DOACs) overcome the most relevant limitations of VKA and overall, seem more effective and safer.[14] It has been reported that the efficacy and safety of DOACs compared with VKA are independent of age, and due to the lower risk of intracranial hemorrhage associated with DOACs, they may be a safer choice in elderly patients.[15]

The most relevant information about the efficacy and safety of oral anticoagulants comes from clinical trials in which data regarding elderly patients are limited.[14],[15] Observational studies and registries may allow to determine the clinical profile and management of elderly patients with AF in clinical practice.[16] Unfortunately, the information currently available about this issue according to dependency, frailty and cognitive impairment is clearly insufficient.

The ESPARTA study was performed to determine the degree of adherence to the recommendations performed by the Therapeutic Positioning Report of the Spanish Medicine and Sanitary Products Agency about the treatment with oral anticoagulants, in patients aged ≥ 75 years with nonvalvular atrial fibrillation (NVAF) treated with oral anticoagulants, attended in Spanish Internal Medicine departments in clinical practice.[17] In this study, the bio-demographic data, clinical profile and antithrombotic treatment according to the degree of dependency, fragility and cognitive impairment were analyzed.

Methods

ESPARTA was an observational, cross-sectional and multi-center study that consecutively included patients aged ≥ 75 years old, with a diagnosis of NVAF, attended in Internal Medicine units (hospitalization or outpatients) in Spain, treated with oral anticoagulants ≥ 3 months previous the inclusion and that had provided signed informed consent. Patients with mitral stenosis, prosthetic valve, or valvuloplasty, that had participated in clinical trials involving the use of DOACs in the six months before inclusion, or that were unable to understand or fulfill the questionnaires of the study were excluded. The study was approved by the Clinical Research Ethics Committee of Hospital Universitario de la Princesa de Madrid and endorsed by the local ethic committees of the participating centers.[17]

The study was performed in 63 Internal Medicine units throughout Spain according to clinical practice. The participating sites were representative sample of all such sites in Spain. Patients were included between October 2015 and March 2016. The study consisted in a single visit. No specific diagnostic or therapeutic interventions were performed due to the study. The biodemographic and clinical data were collected from the medical history during the physician interview and recorded by each investigator into an electronic case report form specifically designed for this study.[17]

Data from biodemographic variables (age, sex, place where patients lived, degree of dependency, type of AF, body weight), cardiovascular risk factors (hypertension, diabetes), vascular conditions (heart failure, ischemic stroke, stable angina, peripheral artery disease, renal insufficiency, systemic embolism) and blood analysis (creatinine clearance, hemoglobin) were recorded.[17]

The thromboembolic risk was determined by the CHADS2 and CHA2DS2-VASc scales and the risk of bleeding by the HAS-BLED scale.[10] The type of oral anticoagulant (VKA vs. DOACs) was recorded.

Polymedication was defined as the regular intake of at least five tablets daily at the moment of the visit. Adherence to treatment with oral anticoagulants was determined by using the Morisky-Green test. Good adherence was defined as an average adherence to oral anticoagulation of ≥ 80%.[18] Partial dependency was considered when the patient needed help for at least one of the following activities: personal care, eating, getting dressed, or moving and total dependency when the patient needed help for all previous activities. Cognitive function was assessed by using the Pfeiffer's test, defining the presence of cognitive impairment as having a score ≥ 3.[19] Fragility was determined with the FRAIL scale. The fragile patient was defined as having a score ≥ 3 in the FRAIL scale.[20] Comorbidity was determined according to the Charlson comorbidity index. High comorbidity was defined as having a Charlson comorbidity index ≥ 3.[21]

For the descriptive analysis, quantitative variables were described with measures of central tendency and dispersion (mean ± SD) and qualitative variables were described as absolute (n) and relative (%) frequencies. In the bivariate analysis to compare two means, the nonparametric Mann–Whitney U test was performed. To compare percentages, the chi-square test or Fisher test were used, according to the sample size.

Stepwise binary logistic regression analyses were performed to identify those clinical variables associated with dependency, fragility, cognitive impairment and dependency. Those variables with a P-value < 0.10 in the univariate analysis were included in the multivariate analyses. Odds ratios and their respective 95% confidence intervals were used to describe the results of the multivariate analyses. Statistical significance was set at a P-value < 0.05. The statistical analysis was performed using the SAS statistics package, version 9.4.

Results

A total of 837 patients (mean age 83.0 ± 5.0 years; 48.7% men; 83.3% permanent AF) were included in the study. Patients had a great number of comorbidities (84.3% had hypertension, 62.7% heart failure, 39.1% diabetes and 19.2% prior ischemic stroke). Patients had a high thromboembolic risk (mean CHADS2 score 3.2 ± 1.2; mean CHA2DS2-VASc score 5.0 ± 1.4), but only 28.9% had a high risk of bleeding (Table 1). Overall, 72.1% of patients showed a good adherence to anticoagulant therapy, without significant differences among the type of oral anticoagulant (73.0% in patients taking AVK vs. 70.1% in patients taking DOACs; P = NS).

Table 1.

Clinical profile of patients according to the degree of dependency, fragility and cognitive impairment.

Variables	Overall (n = 837)	Autonomous (n = 465)	Dependent (n = 372)	P	No fragile (n = 472)	Fragile (n = 360)	P	No cognitive impairment (n = 567; 67.7%)	Cognitive impairment (n = 270)	P
Biodemographic data
Age, yrs	83.0 ± 5.0	81.8 ± 4.8	84.5 ± 4.9	< 0.0001	82.4 ± 5.0	83.8 ± 4.9	< 0.0001	82.3 ± 4.9	84.4 ± 5.0	< 0.0001
Gender male	48.7%	57.0%	38.4%	< 0.0001	55.3%	40.3%	< 0.0001	53.8%	38.1%	< 0.0001
Place where patients live
Home	93.8%	97.4%	89.2%	< 0.0001	96.6%	90.0%	< 0.0001	96.5%	88.1%	< 0.0001
Nursing home	6.2%	2.6%	10.8%	3.4%	10.0%	3.5%	11.9%
No dependency	55.6%	100%	0	< 0.0001	72.0%	34.4%	< 0.001	71.3%	22.6%	< 0.0001
Permanent atrial fibrillation	83.3%	80.0%	87.4%	0.01	81.4%	85.8%	NS	80.8%	88.5%	0.006
Body weight, kg	73.7 ± 13.9	74.0 ± 13.8	73.3 ± 14.1	NS	73.4 ± 13.5	74.1 ± 14.5	NS	74.5 ± 14.0	71.9 ± 13.6	0.03
Cardiovascular risk factors
Hypertension	84.3%	81.3%	88.2%	0.006	82.4%	86.9%	NS	83.6%	85.9%	NS
Diabetes mellitus (%)	39.1	35.3	43.8	0.01	34.7%	44.4%	0.004	38.6%	40.0%	NS
Cardiovascular disease
Heart failure	62.7%	52.9%	75.0%	< 0.0001	51.3%	77.5%	< 0.0001	57.0%	74.8%	< 0.0001
Ischemic stroke	19.2%	13.8%	26.1%	< 0.0001	16.1%	23.1%	0.01%	16.4%	25.2%	0.002
Stable angina	9.0%	8.0%	10.2%	NS	6.1%	12.5%	0.001	8.5%	10.0%	NS
Peripheral artery disease	8.1%	8.0%	8.3%	NS	8.5%	7.8%	NS	9.2%	5.9%	NS
Renal insufficiency*	7.9%	7.1%	8.9%	NS	5.7%	10.6%	0.01	6.7%	10.4%	NS
Systemic embolism	2.5%	1.9%	3.2%	NS	1.9%	3.3%	NS	2.3%	3.0%	NS
Creatinine clearance, mL/min	50.7 ± 21.9	53.2 ± 22.0	47.4 ± 21.3	0.0002	52.4 ± 20.9	48.7 ± 22.9	0.001	52.2 ± 21.1	47.4 ± 23.2	0.0002
Thromboembolic and bleeding risk
Mean CHADS2 score	3.2 ± 1.2	3.0 ± 1.1	3.6 ± 1.2	< 0.0001	3.0 ± 1.2	3.6 ± 1.2	< 0.0001	3.1 ± 1.1	3.5 ± 1.3	0.0001
High thromboembolic risk	94.7%	91.8%	98.4%	< 0.0001	92.8%	97.2%	0.004	93.7%	97.0%	0.04
Mean CHA2DS2-VASc score	5.0 ± 1.4	4.7 ± 1.3	5.5 ± 1.4	< 0.0001	4.7 ± 1.3	5.4 ± 1.4	< 0.0001	4.8 ± 1.3	5.4 ± 1.5	< 0.0001
High thromboembolic risk	100%	100%	100%	NS	100%	100%	NS	100%	100%	NS
Mean HAS-BLED score	2.1 ± 0.9	1.9 ± 0.9	2.2 ± 0.9	< 0.0001	1.9 ± 0.9	2.3 ± 1.0	< 0.0001	2.0 ± 0.9	2.2 ± 0.9	0.0006
High bleeding risk	28.9%	22.5%	36.9%	< 0.0001	21.6%	37.4%	< 0.0001	26.2%	34.5%	0.01
Prior bleeding	17.6%	17.6%	17.5%	NS	15.7%	19.4%	NS	19.0%	14.4%	NS
Other conditions
Hemoglobin, g/dL	12.4 ± 2.0	12.9 ± 2.0	11.9 ± 1.8	< 0.0001	12.9 ± 1.8	11.9 ± 2.0	< 0.0001	12.7 ± 2.0	11.9 ± 1.8	< 0.0001
Pfeiffer's test	1.9 ± 2.1	1.0 ± 1.4	3.1 ± 2.3	< 0.0001	1.4 ± 1.9	2.5 ± 2.3	< 0.0001	0.6 ± 0.8	4.5 ± 1.6	< 0.0001
Cognitive impairment	32.3%	13.1%	56.2%	< 0.0001	20.6%	47.5%	< 0.0001	0	100%	< 0.0001
FRAIL scale	2.1 ± 1.4	1.6 ± 1.3	2.7 ± 1.3	< 0.0001	1.0 ± 0.8	3.5 ± 0.6	< 0.0001	1.8 ± 1.4	2.7 ± 1.3	< 0.0001
Fragility	43.3%	26.7%	64.1%	< 0.0001	0	100%	< 0.0001	33.5%	63.8%	< 0.0001
Charlson comorbidity index	2.7 ± 1.9	2.4 ± 1.9	3.1 ± 1.9	< 0.0001	2.3 ± 1.9	3.3 ± 1.9	< 0.0001	2.5 ± 1.9	3.2 ± 2.0	< 0.0001
Falls	25.1%	17.8%	34.1%	< 0.0001	20.6%	30.6%	0.0009	22.4%	30.7%	0.009
Polymedication	91.5%	88.2%	95.7%	0.0001	87.9%	96.4%	< 0.0001	89.6%	95.6%	0.003
Adherence to treatment	72.1%	71.8%	72.5%	NS	74.8%	68.3%	0.03	75.6%	64.8%	0.001

Data are presented as mean ± SD or percent unless other indicated. *Renal insufficiency: chronic dialysis, renal transplantation or serum creatinine ≥ 2.26 mg/dL. NS: not significant.

The clinical profile of patients according to the degree of dependency, fragility and cognitive impairment was analyzed in Table 1. Overall, 465 (55.6%) of patients were completely autonomous, and 372 (44.4%) patients had some degree of dependency (total or partial). Compared with those patients that were autonomous, those patients with dependency were older, lived more commonly at nursing homes, had more comorbidities (hypertension, diabetes, heart failure and ischemic stroke), a worse renal function and a higher thromboembolic and bleeding risk. Anemia and polymedication were more common in patients with dependency. Scores in Pfeiffer's test, FREIL scale and Charlson comorbidity index were also worse in patients with dependency (Table 1).

A total of 360 (43.3%) patients were fragile. Fragile patients were older, had more comorbidities and a higher risk of falls and thromboembolic and bleeding events. The majority (65.6%) of frail patients had some degree of dependency. Cognitive impairment was more common in frail patients (47.5% vs. 20.6%; P < 0.0001) (Table 1). Overall, 270 (32.3%) patients had cognitive impairment. Compared with patients without cognitive impairment, patients with cognitive impairment were older, lived more frequently at nursing homes, the degree of dependency was higher, had more heart failure, ischemic heart disease, anemia and risk of falls, a worse renal function and higher thromboembolic and bleeding risk. Scores in Pfeiffer's test, FREIL scale and Charlson comorbidity index were also worse in patients with cognitive impairment (Table 1).

Overall, falls occurred in 25.1% of patients, being more common among dependent patients (34.1% vs. 17.8%; P < 0.0001), fragile patients (30.6% vs. 20.6%; P = 0.0009) and patients with cognitive impairment (30.7% vs. 22.4%; P = 0.009) (Table 1).

Multivariate analyses were performed to assess those variables that were independently associated with the presence of dependency, fragility, or cognitive impairment (Table 2). A higher thromboembolic risk, falls, higher scores in FRAIL scale and Pfeiffer's test, age and anemia increased the risk of dependency. Dependency, stable angina, heart failure, female sex, cognitive impairment, a higher Charlson comorbidity index and anemia were independently associated with the presence of fragility. Dependency, living at nursing homes, higher scores in FRAIL scale and Charlson comorbidity index, anemia and poor adherence to medication were independently associated with cognitive impairment.

Table 2.

Variables associated with dependency, fragility and cognitive impairment.

	OR	95% confidence interval	P
Variables associated with dependency
CHADS2 (high vs. intermediate risk)	8.197	1.786–38.461	0.0068
Falls (yes vs. no)	1.841	1.188–2.853	0.0063
FRAIL scale	1.644	1.413–1.914	< 0.0001
Pfeiffer's test	1.627	1.451–1.824	< 0.0001
Age, yrs	1.060	1.019–1.103	0.0036
Hemoglobin, g/dL	0.832	0.748-0.924	0.0006
Variables associated with fragility
Dependency
Autonomous	--	--	< 0.0001
Partial dependency	2.830	1.879–4.124
Total dependency	6.813	2.437–19.047
Stable angina (yes vs. no)	2.371	1.285–4.376	0.0058
Heart failure (yes vs. no)	2.080	1.372–3.153	0.0006
Sex (women vs. men)	1.889	1.308–2.728	0.0007
Cognitive impairment (yes vs. no)	1.528	1.042–2.398	0.0309
Charlson comorbidity index	1.156	1.041–1.159	0.0071
Hemoglobin, g/dL	0.840	0.769–0.840	0.0011
Variables associated with cognitive impairment
Dependency
Autonomous	--	--	< 0.0001
Partial dependency	5.001	3.276–7.633
Total dependency	15.890	6.279–40.239
Place where patients live (nursing home vs. home)	2.483	1.095–5.635	0.0295
FRAIL scale	1.174	1.003–1.374	0.0458
Charlson comorbidity index	1.120	1.011–1.241	0.0295
Hemoglobin	0.899	0.810–0.998	0.0459
Adherence (good vs. poor)	0.575	0.376–0.881	0.0109

The use of antithrombotic treatment according to the degree of dependency, frailty and cognitive impairment was also analyzed (Table 3). VKA were the most common oral anticoagulant prescribed in the overall study population, reaching 70.8% of patients. The remaining 29.2% of patients were taking DOACs. In addition to anticoagulant therapy, 9.7% of patients were taking oral antiplatelets. This distribution was independent of the degree of dependency and the presence of fragility, but there was a trend to a higher prescription of VKA in those patients with cognitive impairment (75.2% vs. 68.8%; P = 0.05) and a lower prescription of DOACs (24.8% vs. 31.2%; P = 0.05).

Table 3.

Type of oral anticoagulant according to dependency, frailty and cognitive impairment.

Variable	Overall (n = 837)	Autonomous (n = 465; 55.6%)	Dependent (n = 372; 44.4%)	P	No fragile (n = 472; 56.7%)	Fragile (n = 360; 43.3%)	P	No cognitive impairment (n = 567; 67.7%)	Cognitive impairment (n = 270; 32.3%)	P
VKA	70.8	70.1	71.8	NS	70.1	71.7	NS	68.8	75.2	0.05
DOACs	29.2	29.9	28.2	NS	29.9	28.3	NS	31.2	24.8	0.05
Dabigatran	19.7	19.4	20.0	0.01	22.7	15.7	0.0005	16.9	26.9	0.02
Rivaroxaban	45.1	52.5	35.2		52.5	35.3		50.3	31.3
Apixaban	35.2	28.1	44.8		24.8	49.0		32.8	41.8

DOACs: direct oral anticoagulants; NS: not significant; VKA: vitamin K antagonists.

Discussion

In the ESPARTA study, a wide sample of anticoagulated elderly patients with AF was included. These patients had a high thromboembolic risk (CHA2DS2-VASc 5.0), a high number of comorbidities, but less than 30% of patients had a high bleeding risk. This is in line with previous reports.[22] In the ESPARTA study, approximately 44% of patients had some degree of dependency, 43% of patients were fragile, and 32% had cognitive impairment. Despite some studies have analyzed the clinical profile and management of elderly patients with AF, the majority of these studies have focused only on one of these conditions, but not on all at the same time.[7],[23] In our study, patients with any of these conditions were older and had a worse clinical profile, with more comorbidities and a higher thromboembolic and bleeding risk. In addition, as shown in the multivariate analyses, all these conditions were related, since the presence of one of them promoted the presence of the others. A number of studies have analyzed the impact of fragility in this specific population. Similar to our data, these studies have reported that frail subjects are older, have more comorbidities, and a higher risk of stroke and mortality.[8],[23] Cognitive impairment is more common in this population.[4] Remarkably, the identification of these conditions is mandatory, since it has been shown that the baseline functional status is the strongest predictor for in-hospital mortality in elderly AF patients.[24] In addition, frailty status is positively associated with the CHA2DS2-VASc and the HAS-BLED scales, and it may provide additional prognostic value in this population.[25]

Polymedication is very common in elderly patients with AF, particularly when dependency, frailty or cognitive impairment are present, as our study showed. It has been reported that it may be more important the multifactorial medical profile and functional status rather than cognitive impairment when prescribing long-term oral anticoagulation.[26] Importantly, in our study, medication adherence to anticoagulant therapy was lower in frailty patients and in patients with cognitive impairment. Since drug-drug interactions are more frequent in this population as well as making mistakes while taking the medication,[6] when any of these conditions (dependency, frailty and/or cognitive impairment) are present, additional support may be required for these patients.[5]

Overall, falls were reported in 25% of patients, increasing approximately to one third among dependent patients, fragile patients and patients with cognitive impairment. However, in the multivariate analysis, falls only remained as an independent variable in the case of dependency. Traditionally, one of the causes that have limited the use of oral anticoagulants has been the risk of falls. However, it has been shown that although more caution should be taken in these patients, since the risk of stroke is markedly increased in this population, overall, oral anticoagulation should be considered.[27] In this context, due to the better benefit/risk profile of DOACs over VKA, it seems a better choice the use of DOACs in this setting.[28]

In our study, anemia was independently associated with dependency, frailty and cognitive impairment. Previous studies have shown the same results, not only in individuals with AF, but also in the general elderly population.[29] As a result, low hemoglobin levels among elderly patients may alert clinicians about the increased risk of dependency, frailty or cognitive impairment that these individuals have.

Unless contraindicated, every elderly patient with AF should receive long-term oral anticoagulation.[10] Our study showed that in patients with dependency, frailty or cognitive impairment, the thromboembolic risk was very high, but only 34%–37% of patients had a high risk of bleeding. Similarly, other authors have shown in frail patients with AF that stroke risk is high and bleeding risk levels comparatively low.[30] However, although some studies have suggested an increase of anticoagulation in frail octogenarian patients, many elderly patients do not receive anticoagulant therapy. For example, it has been reported that in more than 40% of older adults with AF after acute ischemic stroke, patients are not discharged with an oral anticoagulant. In this study, fall risk, poor prognosis, older age, and dementia have been shown as predictors of not prescribing oral anticoagulants.[31] Importantly, none of these variables should be considered as an absolute contraindication for anticoagulation by themselves.[5] This is very relevant, since discontinuation of anticoagulation in elderly patients with AF is frequent and it is associated with increased risks of death and adverse events.[32]

In our study, the majority of patients were anticoagulated with VKA, regardless dependency or frailty, and showed a trend to a higher prescription of VKA in those patients with cognitive impairment. However, considering that cognitive disorders in patients with AF significantly reduce the efficacy of VKA therapy[33] and that DOACs exhibit a better beneficial/risk ratio than VKA in elderly patients, not only in clinical trials but also in “real-life” studies,[15],[34]–[37] the prescription of VKA instead of DOACs in our study was excessively high. In fact, the prescription of DOACs in elderly patients with AF in Spain is lower than in other similar countries.[38] This may be related, at least in part, with the administrative restrictions for the prescription of DOACs in Spain.

This study has some limitations that should be commented. Since this study has an observational and cross-sectional design, not all potential confounders can be controlled. However, this design may represent clinical practice more accurately. In addition, as in this study only anticoagulated elderly patients (≥ 75 years old) were included, the information provided by our data cannot be extended to the whole population with AF, or with different healthcare system. Moreover, despite meticulousness in recording data, because dependency, frailty, cognitive impairment and comorbidity were defined using only one scale in defining them, this could translate into an underdiagnosis of some patients.

In conclusion, the elderly patients with nonvalvular AF treated with oral anticoagulants in Internal Medicine departments in Spain have a high thromboembolic risk, a high number of comorbidities, but less than 30% of patients have a high risk of bleeding. Approximately 44% of patients have some degree of dependency, 43% are fragile, and 32% have cognitive impairment. Patients with any of these conditions are older and have a worse clinical profile. All these conditions are related, since the presence of one of them promotes the presence of the others. The majority of patients are anticoagulated with VKA, regardless dependency or frailty, but with a trend to a higher prescription of VKA in those patients with cognitive impairment. According with previous trials, our study showed that in patients with dependency, frailty or cognitive impairment, the thromboembolic risk was very high, but only one third of these patients had a high risk of bleeding. So unless strong and real contraindication, every elderly patient with AF should receive long-term oral anticoagulation.