Journal of the Neurological Sciences 379 (2017) 16

Contents lists available at ScienceDirect

Journal of the Neurological Sciences

journal homepage: www.elsevier.com/locate/jns

1.4 times increase in atrial brillation-related ischemic stroke and TIA

over 12 years in a stroke center
Qiong Yang a,b, Leonid Churilov c, Dongsheng Fan b, Stephen Davis a, Bernard Yan a,
a
Department of Medicine and Neurology, Melbourne Brain Centre at the Royal Melbourne Hospital, University of Melbourne, Parkville, Australia
b
Department of Neurology, Peking University Third Hospital, Beijing, China
c
Florey Institute of Neuroscience and Mental Health, Melbourne, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Background and Purpose: Prevalence of atrial brillation (AF) has quadrupled in the past 50 years in the general
Received 26 February 2017 population. However, there is uncertainty regarding prevalence of AF over time in ischemic stroke patients given
Received in revised form 24 April 2017 the aging population and enhanced surveillance of AF. We aimed to explore the changing prevalence of AF as well
Accepted 10 May 2017 as other risk factors, stroke subtypes, investigations and pre-stroke medications among ischemic stroke and tran-
Available online 11 May 2017
sient ischemic attack (TIA) patients.
Methods: We performed a retrospective analysis of data from a prospective database of consecutive patients with
Keywords:
Trends
acute ischemic stroke and TIA from 2004 to 2015. Trends in risk factors and other variables year by year were an-
Risk factor alyzed using logistic regression or median regression.
Ischemic stroke Results: Among 6275 patients (median age [interquartile range] 74 [6282] years, 56% males), the prevalence of
Atrial brillation AF increased 1.4 times over 12 years (from 23.3% to 32.7%, P b 0.001). The increase in the prevalence of AF
remained signicant after adjustment for age and the use of Holter monitoring. There was also a signicant in-
crease in prevalence of hypertension (67.4% to 77.3%), structural heart disease (9.8% to 10.5%), and previous
TIA (10.9% to 13.7%) and a signicant decrease in prevalence of dyslipidemia (71.8% to 49.4%).
Conclusions: There was a 1.4 times increase in the prevalence of AF among consecutive ischemic stroke and TIA
patients in the past 12 years in a hospital-based registry. More active screening of the general population for
AF may be warranted in order to decrease the overall stroke burden.
2017 Elsevier B.V. All rights reserved.

1. Introduction there was an increase in the detection of AF in ischemic stroke patients

The overall prevalence of atrial brillation (AF) is 1%2% of the gen-
eral population and increases to 6%15% at 80 years old [1]. The preva-
lence of AF quadrupled from 1958 to 67 to 19972007, which increased
in men from 20.4 to 96.2 per 1000 person-years, and in women from
13.7 to 49.4 per 1000 person-years, respectively [2]. Risk factors of AF
include aging, hypertension, heart failure, coronary artery disease, atrial
dilatation, myocardial stretch, and genetic factors [3]. There is current
interest in the concept of atrial myopathy as the harbinger of AF [4].
AF increases the risk of ischemic stroke nearly 5-fold [5]. The preva-
lence of AF among patients with ischemic stroke varies between 15 and
38% [68]. However, the changing prevalence of AF among ischemic
stroke patients over long time periods remains uncertain, given the
aging of population, heightened awareness and surveillance of AF. The
increased burden of AF may impact upon post- stroke strategies of in-
vestigations, and the proportional use of anticoagulation. Of note,
Corresponding author at: Royal Melbourne Hospital, 300 Grattan Street, Parkville,
Victoria 3050, Australia.
E-mail address: Bernard.yan@mh.org.au (B. Yan).
from 24.6% to 31.7% between 1995 and 1999 and 20102013 in Poland
[9] and from 13.2% to 20% between 2003 and 2007 in Greece [10]. How-
ever, cardioembolic stroke inclusive of all cardiac pathology increased
from 26% to 56% in patients with minor stroke or transient ischemic at-
tack (TIA) without a signicant increase in strokes secondary to AF from
2002 to 2005 to 20092012 in Canada [11]. Prevalence of cardioembolic
stroke increased from 2002 to 2010 in Korea, however, the prevalence
of AF was not reported [12].
Stroke secondary to AF is associated with more severe neurological
impairment, increased disability and mortality, and higher stroke recur-
rence [5,13,14]. Moreover, hemorrhagic transformation, leading to early
neurological deterioration and adverse outcome, develops more fre-
quently in patients with cardioembolic stroke [15]. Increased AF preva-
lence and awareness of its clinical consequences will lead to improved
screening not restricted to routine 12-lead ECG, but also to extended
electrocardiographic monitor and implantable devices such as loop re-
corders [2,16].
To date, the reported changing prevalences of AF over long periods
are inconsistent and there is a lack of data from Australia. On this
basis, we aimed to explore the effect of time on the prevalence of AF

http://dx.doi.org/10.1016/j.jns.2017.05.022
0022-510X/ 2017 Elsevier B.V. All rights reserved.
2 Q. Yang et al. / Journal of the Neurological Sciences 379 (2017) 16
and other risk factors among acute ischemic stroke and TIA patients
over time in an Australian comprehensive stroke center, and to investi-
gate the trends in ischemic stroke subtypes, diagnostic investigations,
and pre-stroke medications. We hypothesized that the proportion of
AF among stroke patients would increase over time periods in an
Australian stroke center.
2. Methods
2.1. Study setting
We performed a retrospective analysis of data from a prospectively
collected database of consecutive patients with acute ischemic stroke
or transient ischemic stroke (TIA) admitted to the Royal Melbourne
Hospital stroke unit from 2004 to 2015.
According to Melbourne Health Annual Report 20142015, the
Royal Melbourne Hospital serves a population base of over 1 million.
Its catchment includes North West of Melbourne, and the rural area
named Hume. The Stroke Care Unit of Royal Melbourne Hospital treats
approximately 800 inpatients a year from these areas. According to
2011 Census report on North West of Melbourne, 36.1% of population
were born in overseas. The regions of birth were Europe (including
southern and Eastern Europe and North-West Europe) 11. 8%,
North Africa and the Middle East 7.8%, Asia (including South-East
Asia, North-East Asia, Southern and Central Asia) 7.8%, Oceania (exclud-
ing Australia) 1.9%, Americas 0.8%, and Sub-Saharan Africa 0.6%.
We included patients with acute ischemic stroke or TIA admitted to
the Royal Melbourne Hospital from January 2004 to December 2015. Pa-
tients with intracerebral hemorrhage or stroke mimics were excluded.
This study was approved by the Ethics Committee of the Royal
Melbourne Hospital.
2.2. Data collection
The demographic data including gender, age and continent of birth
were collected. The stroke types including ischemic stroke and TIA
were collected. Vascular risk factors were collected including current
cigarette smoking, hypertension, diabetes mellitus, dyslipidemia, ische-
mic heart disease, AF, structural heart disease (SHD), peripheral vascu-
lar disease, previous history of TIA, and previous history of stroke. We
collected levels of serum lipids including total cholesterol, triglycerides,
low-density lipoprotein (LDL), and high-density lipoprotein (HDL).
Cardiac investigations including Holter monitoring, transthoracic
echocardiography (TTE), transoesophageal echocardiography (TOE),
vascular investigations including cerebral digital subtraction angiogra-
phy (DSA), computed tomography angiography (CTA), and carotid du-
plex ultrasound, and brain imaging including MRI and CT were
collected.
We also collected the use of pre-stroke medications including anti-
platelet therapies, anticoagulants including warfarin and the novel
oral anticoagulants (NOACs), lipid lowering medications, antihyperten-
sive agents, and hypoglycemic drugs.
2.3. Denition of variables
AF was dened as either a documented past history of AF, or an
in-hospital electrocardiographic evidence of AF. Dyslipidemia was
dened as either having a past history, or abnormalities in the levels
of lipids when admitted (total cholesterol level 4.0 mmol/L, low-
density lipoprotein [LDL] 2.5 mmol/L, high-density lipoprotein
[HDL] 1.0 mmol/L, or triglycerides level 1.5 mmol/L). Hyperten-
sion was dened as either having a past history or documented treat-
ment with antihypertensive agents. SHD was dened as non-
coronary cardiac diseases including abnormality or defect of the
heart muscle, the heart valves, or congenital diseases of vessels.
The subtype of ischemic stroke was classied according to the Trial
of Org 10172 in the Acute Stroke Treatment (TOAST) classication sys-
tem by neurologists: large-artery atherosclerosis, cardioembolism,
small-vessel occlusion, stroke of other determined etiology, and stroke
of undetermined etiology [17].
2.4. Statistical methods
Continuous variables were presented as median with interquartile
range (1st quartile to 3rd quartile, IQR) due to the nature of the under-
lying distribution. Categorical variables were presented as a number of
valid observations with corresponding proportions and 95% condence
intervals (CI).
The associations between time and various sociodemographic and
clinical characteristics (gender, continent of birth, stroke risk factors, in-
vestigations performed, pre-stroke medications) were analyzed using
binary logistic regression. Associations between time and ischemic sub-
types according to TOAST classication were analyzed by comparing pa-
tients with or without each subtype using binary logistic regression. The
association between time and patients' age were analyzed using median
regression due to the nature of the underlying distributions. Statistical
signicance was dened as P b 0.05 for all the tests. The statistical
analyses were performed using Stata version 14.0 (StataCorp LP, College
Station, TX, USA).
3. Results
In the time period 2004 to 2015, there were 6275 consecutive pa-
tients with ischemic stroke (4926 [78.5%]) and TIA (1349 [21.5%]).
There were 3512 (56%) men with median age (IQR) 71 (6180) years
and 2763 (44%) were women with median age (IQR) 77 (6584)
years. The clinical characteristics are shown in Table 1.
3.1. Time trends in demographics data
No signicant change in gender or age of the patients over time (P =
0.419, P N 0.999, respectively) was observed. As for the continent of
birth, there was an increase in proportion of patients who were born
in Oceania (including Australia, New Zealand, etc.) from 46.8% to
60.7% (odds ratio [OR] per year, 1.052; 95% CI, 1.0371.068; P b 0.001)
and a decrease in proportion of patients who were born in Europe
(from 38.5% to 27.2%; OR per year, 0.956; 95% CI, 0.9420.971; P b
0.001). No signicant association between time and the proportions of
patients born in North America, Asia, South America and Africa were
observed.
3.2. Time trends in prevalence of AF
For all patients with either ischemic stroke or TIA, the prevalence of
AF increased signicantly over time (from 23.3% to 32.7%; OR of AF per
year, 1.048; 95% CI, 1.0311.065; P b 0.001) (Fig. 1). The 40% relative in-
crease remained signicant after adjusting for age (OR of AF per year,
1.049; 95% CI, 1.0311.067; P b 0.001), also after adjusting for age and
proportion of use of Holter monitoring over the last 5 years (OR of AF
per year, 1.115; 95% CI, 1.0471.186; P = 0.001).
For patients with ischemic stroke only, the prevalence of AF in-
creased signicantly over time (from 23.5% to 36.9%; OR of AF per
year, 1.052; 95% CI, 1.0341.071; P b 0.001) (Fig. 1). This association
remained signicant after adjusting for age (OR of AF per year, 1.056;
95% CI, 1.0371.077; P b 0.001), also after adjusting for age and propor-
tion of use of Holter monitoring (data available for 20112015) (OR of
AF per year, 1.148; 95% CI 1.0711.230; P b 0.001). However, for patients
with TIA only, no signicant association between the presence of AF and
time was observed (from 22.1% to 15.5%; OR of AF per year, 1.003; 95%
CI, 0.9631.045; P = 0.871) (Fig. 1).
 Q. Yang et al. / Journal of the Neurological Sciences 379 (2017) 16 3

Table 1 decrease in dyslipidemia from 71.8% to 49.4% (OR per year, 0.916; 95%
Clinical characteristics of total acute ischemic stroke and TIA patients admitted in 2004 CI, 0.9010.930; P b 0.001).
2015.
No signicant changes over time in prevalence of other stroke risk
Total patients (n = 6275) factors were observed (diabetes mellitus [OR per year, 1.000; 95% CI,
Demographics data 0.9841.017; P = 0.991], ischemic heart disease [OR per year, 1.006;
Male, no (%) 3512 (56.0) 95% CI, 0.9891.023; P = 0.510], peripheral vascular disease [OR per
Age (years), median (IQR) 74 (6282) year, 1.015; 95% CI, 0.9811.049; P = 0.390] and smoking [OR per
Continent of birth, no (%)
year, 0.988; 95% CI, 0.9701.007; P = 0.212]).
Oceania (including Australia, New Zealand, etc.) 3259 (51.9)
Europe 2153 (34.3) In addition, the proportion of patients with prior history of TIA in-
Asia 549 (8.8) creased from 10.9% to 13.7% (OR per year, 1.056, 95% CI, 1.0321.080;
North America 28 (0.5) P b 0.001), while no change in the proportion of patients with prior his-
South America 29 (0.5) tory of stroke was observed (from 20.4% to 21.4%; OR per year, 1.016;
Africa 96 (1.5)
Unknown 162 (2.6)
95% CI, 0.9981.034; P = 0.083).

Stroke type, no (%) 3.4. Time trends in subtypes according to TOAST classication
TIA 1349 (21.5)
Ischemic stroke 4926 (78.5)
Among ischemic stroke patients, there was an increase in prevalence
Subtypes in ischemic stroke patients, no (%), n = 4926 of cardioembolic stroke (from 24.8% to 35.5%, OR per year, 1.044; 95% CI,
large-artery atherosclerosis 539 (10.9)
1.0261.062; P b 0.001) and stroke of other determined etiology (from
cardioembolism 1534 (31.1)
small-vessel occlusion 183 (3.7) 3.2% to 3.6%, OR per year, 1.044; 95% CI, 1.0041.085; P = 0.029), and
Other determined etiology 234 (4.8) a decrease in stroke of undetermined etiology from 51.0% to 42.6%
Undetermined etiology 2419 (49.1) (OR per year, 0.952; 95% CI, 0.9360.968; P b 0.001). However, there
Unknown 17 (0.4)
was no trend in large artery strokes (OR per year, 1.009; 95% CI,
Vascular risk factors 0.9831.036; P = 0.506) and small vessel strokes (OR per year, 1.014;
Diabetes mellitus, no (%) 1610 (25.7) 95% CI, 0.9711.058; P = 0.536).
Hypertension, no (%) 4651 (74.1)
Dyslipidemia, no (%) 4172 (66.5)
Ischemic heart disease, no (%) 1579 (25.2)
3.5. Time trends in use of investigations
Structural heart disease, no (%) 546 (8.7)
Peripheral vascular disease, no (%) 310 (4.9) The proportion of patients who undergone Holter monitoring during
Smoking, no (%) 1138 (18.1) hospitalization increased from 17.4% to 20.6% (OR per year 1.092; 95%CI
Atrial brillation, no (%) 1753 (27.9)
1.0211.168; P = 0.01) from 2011 to 2015 (Table 2). The use of TOE de-
Previous history of stroke, no (%) 1375 (21.9)
Previous history of TIA, no (%) 712 (11.4) creased from 10.6% to 6.9% from 2004 to 2015 while no signicant
change over time in the use of TTE was observed (P b 0.001 and P =
Investigations
0.714, respectively) (Table 2).
TOE, no (%) 454 (7.2)
TTE, no (%) 275 (4.4) In addition, patients scheduled to have Holter monitoring in the out-
Holter monitoring, no (%)a 521 (20.2) patient setting increased from 16.3% to 30.6% from 2011 to 2015 and pa-
DSA, no (%) 736 (11.7) tients scheduled to have TTE or TOE increased from 20.7% to 37.8% from
CTA, no (%) 2177 (34.7)
2004 to 2015.
Carotid ultrasound, no (%) 3079 (49.1)
CT, no (%) 6189 (98.6)
The use of CTA and DSA increased (from 0% to 69.7%, P b 0.001 and
MRI, no (%) 2496 (39.8) from 14.5% to 17.1%, P = 0.031, respectively), whereas the use of carotid
duplex ultrasound decreased over time (from 66.7% to 22.1%, P b 0.001)
Pre-stroke medications
Antiplatelet therapies, no (%) 2733 (43.6)
(Table 2).
Anticoagulants, no (%) 558 (8.9)
Types of Anticoagulant 3.6. Time trends of pre-stroke medications
Warfarin, no (% in patients taking anticoagulants) 490 (87.8)
NOACs, no (% in patients taking anticoagulants) 58 (10.4)
Pre-stroke use of medications including anticoagulants, lipid lower-
Heparin, no (% in patients taking anticoagulants) 14 (2.5)
Anticoagulants in patients with AFb, no (%) 441 (25.2) ing medications, antihypertensive agents, and hypoglycemic drugs in-
Lipid lowering medications, no (%) 2408 (38.4) creased in total patients whereas no signicant changes over time in
Statins, no (%) 2334 (37.2) pre-stroke use of antiplatelet therapies in total patients and anticoagu-
Antihypertensive agents, no (%) 4064 (64.8) lants in AF patients were observed (Table 3).
Hypoglycemic drugs, no (%) 1098 (17.5)

TIA, transient ischemic attack, TTE, transthoracic echocardiography; TOE, 4. Discussion

transoesophageal echocardiography; DSA, cerebral digital subtraction angiography; CTA,
computed tomography angiography; NOACs, novel oral anticoagulants.
a
Data of whether patients performed Holter monitoring or not were available from
2011 to 2015. And the data of 2585 patients were available and the data of 25 patients
were not available during this time period.
b
The number of patients with AF was 1753.
3.3. Time trends in other risk factors
Trends of risk factors other than AF were shown in Figs. 2 and 3.
There was an increase over time in prevalence of some risk factors: hy-
pertension from 67.4% to 77.3% (OR per year, 1.042; 95% CI, 1.025
1.060; P b 0.001), and SHD from 9.8% to 10.5% (OR per year, 1.042;
95% CI, 1.0161.069; P = 0.002). On the other hand, there was a
4.1. Trends in AF
In this study, we showed that there was a 1.4 times increase in pro-
portion of AF from 23.3% to 32.7% among total patients with acute ische-
mic stroke and TIA in the past 12 years. The increase in the prevalence of
AF remained signicant after adjustment for age and the use of Holter
monitoring. As for the risk factors of AF, there was an increase of hyper-
tension and SHD, whereas we did not show changing pattern in age,
gender, and ischemic heart disease over time.
The prevalence of AF among ischemic stroke patients varied be-
tween 15 and 38% in different studies [68]. However, there have
been few studies focusing on the trend in AF among ischemic stroke pa-
tients. In line with ndings of our study, a study in Poland examined
4 Q. Yang et al. / Journal of the Neurological Sciences 379 (2017) 16

Fig. 1. Trends in proportion of AF year by year from 2004 to 2015. AF increased in total patients and in patients with ischemic stroke signicantly (P b 0.001, and P b 0.001, respectively), but
not in patients with TIA (P = 0.871). AF, atrial brillation; TIA, transient ischemic attack.

3973 patients and showed an increase in AF among ischemic stroke pa-
tients from 24.6% to 31.7% with an increase in cardioembolic strokes
from 24.8% to 28.5% between 1995 and 1999 and 20102013 [9]. How-
ever, both the age of patients and the use of Holter monitoring increased
during this time period in this study [9]. Furthermore, this study did not
adjust for age and the use of Holter monitoring when analyzing the time
trend in AF. On this basis, no conclusion could be drawn regarding the
mechanism of AF prevalence.
The increase in the prevalence of AF among total patients was not
explained by aging or increasing use of Holter monitoring. Possible rea-
sons included the increase in risk factors such as hypertension and SHD
and the decrease in ischemic stroke patients without AF caused by other
etiologies in this population.
The increase in AF over time remained signicant in patients with is-
chemic stroke but not in patients with TIA after stratication by stroke
type. Our study ndings pertaining to TIA were consistent with a
study in Canada which included 3950 patients with minor stroke or
TIA [11]. The study from Canada reported that AF did not change signif-
icantly (from 8% to 11%) between 2002 and 2005 and 20092012 [11].
The time trends in AF between ischemic stroke patients and TIA patients
may be different. Compared with 1538% patients with ischemic stroke
who have AF, only 10.4% of the patients with TIA and minor stroke had
AF [18]. The difference in prevalence and time trend of AF between is-
chemic stroke and TIA patients may be that AF-related stroke are
associated with greater clinical impairment compared with other
strokes [5,13,14]. A different study showed that the proportion of AF
in ischemic stroke patients increased with the increase of stroke severi-
ty (NIHSS 04, 29.7%; NIHSS 15, 53.3%) [19].
No signicant increase in the proportion of use of pre-stroke antico-
agulants in stroke patients with AF was observed in our study, in con-
trast to the study in Poland which reported an increase in use of pre-
stroke anticoagulants [9]. In all the patients with pre-stroke anticoagu-
lants, there were 87.8% patients using warfarin and 10.4% using novel
oral anticoagulants. The percentage of time in therapeutic range for pa-
tients treated with warfarin was only 62% to 68% in clinical trials, and
48% in US outpatient practices [5]. The benet of use anticoagulants
may be limited by suboptimal anticoagulation including poor adherence
to guidelines [5]. Adherence to anticoagulation medications and the use
of novel oral anticoagulants to improve the uptake of anticoagulants is
critical in the management of AF.
4.2. Trends in other risk factors
With increased use of pre-stroke lipid lowering medications, the
proportion of dyslipidemia decreased. On the other hand, the propor-
tions of large artery strokes did not change over time in our study. The
study in Poland reported an increase of large artery atherosclerosis re-
lated stroke and a decreased of hyperlipidemia over time [9].

Fig. 2. Trends in risk factors including dyslipidemia, hypertension, diabetes mellitus and smoking year by year from 2004 to 2015. There was a signicant increase over time in
hypertension (P b 0.001) and a signicant decrease in dyslipidemia (P b 0.001).
 Q. Yang et al. / Journal of the Neurological Sciences 379 (2017) 16 5

Fig. 3. Trends in risk factors including ischemic heart disease, structural heart disease, peripheral vascular disease and previous TIA year by year from 2004 to 2015. There was a signicant
increase over time in structural heart disease (P = 0.002), and previous TIA (P b 0.001). TIA, transient ischemic stroke.

Conversely, the study in Canada reported a decrease of large artery ath-
erosclerosis related stroke, a decrease in the levels of TC, TG and LDL and
an increase in the level of HDL [11].
Our study showed the proportion of hypertension increased from
67.4% to 77.3% with an increase in use of pre-stroke antihypertensive
agents from 57.6% to 63.8% from 2004 to 2015, which was consistent
with the study in Poland [9].
In addition, our study showed a slight increase in SHD from 9.8% to
10.5% in the past 12 years. It was reported that there was an increase
in PFO from 8% to 19% in the study in Canada [11]. However, there
was no time trend in the potential source of cardioembolism including
different heart diseases in the study in Korea [12].
4.3. Limitations
First, this is a single-center hospital-based study, not an epidemio-
logical study, which may not be fully representative for the whole coun-
try. There could be referral bias on the basis of patients with more
severe stroke being referred to our stroke center. Second, this was a ret-
rospective study and some data were missing. NIHSS score was not al-
ways available so that the change in stroke severity over time was not
known. In addition, the proportion of patients who underwent cardiac
investigations in our study was not high. We noted that some patients
were scheduled to have Holter monitoring, TTE or TOE in an outpatient
setting. The results of patients in this group were not available.

Table 2
Changes in use of investigations during hospitalization from 2004 to 2015.

Investigations 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 ORa (95%CI) P value
n = 387 n = 464 n = 550 n = 632 n = 541 n = 587 n = 504 n = 578 n = 261 n = 574 n = 576 n = 621

Cardiac investigations, %
TOE 10.6 11.2 9.1 7.1 9.6 7.2 6.0 4.7 2.7 6.3 5.0 6.9 0.931 b0.001
(0.9040.957)
TTE 5.7 4.1 5.5 4.0 4.4 2.4 3.4 5.0 4.2 2.8 5.2 6.1 1.007 0.714
(0.9721.043)
Holter N/A N/A N/A N/A N/A N/A N/A 17.4 10.5 24.1 22.8 20.6 1.092 0.01
monitoringb (1.0211.168)

Vascular investigations, %
DSA 14.5 11.6 9.1 8.5 13.3 11.6 10.3 11.1 10.3 11.7 11.5 17.1 1.025 0.031
(1.0021.048)
CTA 0 0.9 6.2 7.0 1.3 44.3 49.8 47.4 54 61.1 65.6 69.7 1.516 b0.001
(1.4831.549)
Carotid ultrasound 66.7 69.2 73.1 69.2 49.4 61.8 41.1 45.3 32.6 31.7 27.4 22.1 0.810 b0.001
(0.7980.829)

Brain imaging, %
CT 97.9 99.1 98 99.7 98.5 98.8 98.8 99.5 95.4 98.1 99 98.7 0.987 0.684
(0.9281.050)
MRI 33.9 28 30 25.2 32.7 32.5 33.1 46.7 44.4 59.6 57.3 51.2 1.136 b0.001
(1.1181.153)

TOE, transoesophageal echocardiography; TTE, transthoracic echocardiography; DSA, cerebral Digital Subtraction Angiography; CTA, computed tomography angiography.
a
ORs reported per year.
b
Data of whether patients performed Holter monitoring or not were available from 2011 to 2015. And the data of 2585 patients were available and the data of 25 patients were not
available during this time period.
6 Q. Yang et al. / Journal of the Neurological Sciences 379 (2017) 16

Table 3
Changes in use of pre-stroke medications from 2004 to 2015.

Pre-stroke medications 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 ORa (95%CI) P
n= n= n= n= n= n= n= n= n= n= n= n= value
387 464 550 632 541 587 504 578 261 574 576 621

Antiplatelet therapies, % 35.9 40.7 46.9 41.6 44.6 42.8 47.4 43.9 44.4 44.4 46.2 42.2 1.013 0.075
(0.9991.028)
Anticoagulants, % 5.9 10.3 6.7 9.0 7.0 9.9 8.3 7.4 6.9 8.4 12.3 12.1 1.041(1.0161.069) 0.002
In patients with AFb, % 16.7 29.9 24.4 37.1 21.8 28.4 22.8 19.1 17.7 23.5 28.4 32.0 1.020 0.210
(0.9891.052)
Lipid lowering 29.5 32.1 34.2 35.1 37.7 38.7 41.7 39.5 42.5 46.3 44.8 37.2 1.046 b0.001
medications, % (1.0311.062)
Statins, % 25.6 31.5 33.5 33.7 37.7 37.5 40.3 38.8 41.8 44.8 43.2 36.4 1.049 b0.001
(1.0331.065)
Antihypertensive agents, % 57.6 63.4 62.2 61.6 66.2 67.3 69.6 65.9 60.2 69.2 66.2 63.8 1.021 0.008
(1.0051.036)
Hypoglycemic drugs, % 14.0 16.6 15.3 17.7 16.8 16.5 16.1 16.4 15.7 22.0 22.1 18.2 1.033 0.001
(1.0141.053)
a
ORs reported per year.
b
The number of patients with AF was 1753.

Therefore, we concede that the proportion of AF may be higher. Multi-
center studies are needed to trace the trends in risk factors, especially
the trends in AF and AF related strokes.
5. Conclusions
We showed a 1.4 times increase in the prevalence of AF among
ischemic stroke and TIA patients in the past 12 years in our stroke
center. The increase was not explained by aging or the use of Holter
monitoring. More active screening of the general population for AF
may be warranted in order to decrease the overall stroke burden.
References
[1] J. Andrade, P. Khairy, D. Dobrev, S. Nattel, The clinical prole and pathophysiology of
atrial brillation: relationships among clinical features, epidemiology, and mecha-
nisms, Circ. Res. 114 (2014) 14531468.
[2] R.B. Schnabel, X. Yin, P. Gona, M.G. Larson, A.S. Beiser, D.D. McManus, et al., 50 year
trends in atrial brillation prevalence, incidence, risk factors, and mortality in the
framingham heart study: a cohort study, Lancet 386 (2015) 154162.
[3] J.M. Anumonwo, J. Kalifa, Risk factors and genetics of atrial brillation, Cardiol. Clin.
32 (2014) 485494.
[4] B.W. Calenda, V. Fuster, J.L. Halperin, C.B. Granger, Stroke risk assessment in atrial -
brillation: Risk factors and markers of atrial myopathy, Nat. Rev. Cardiol.
(2016)http://dx.doi.org/10.1038/nrcardio.2016.106 (Published online 07 July
2016).
[5] C.A. Sanoski, Current approaches to anticoagulation for reducing risk of atrial bril-
lation-related stroke, J. Pharm. Pract. 26 (2013) 204213.
[6] K. Bjorn-Mortensen, F. Lynggaard, Pedersen M. Lynge, High prevalence of atrial
brillation among greenlanders with ischemic stroke - atrial brillation found
in more than 30% of cases, Int. J. Circumpolar Health 72 (2013) 22628.
[7] N.E. Andrew, A.G. Thrift, D.A. Cadilhac, The prevalence, impact and economic impli-
cations of atrial brillation in stroke: what progress has been made?
Neuroepidemiology 40 (2013) 227239.
[8] L.A. Sposato, L.E. Cipriano, G. Saposnik, E.R. Vargas, P.M. Riccio, V. Hachinski,
Diagnosis of atrial brillation after stroke and transient ischaemic attack: a
systematic review and meta-analysis, The Lancet Neurology. 14 (2015)
377387.
[9] J.P. Bembenek, M. Karlinski, T.A. Mendel, M. Niewada, I. Sarzynska-Dlugosz, A.
Kobayashi, et al., Temporal trends in vascular risk factors and etiology of
urban polish stroke patients from 1995 to 2013, J. Neurol. Sci. 357 (2015)
126130.
[10] P. Kotsaftis, G. Ntaios, C. Savopoulos, R. Kiparoglou, D. Agapakis, M. Baltatzi, et al.,
Trend in incidence of cardiovascular risk factors in elderly and over-aged stroke pa-
tients between 2003 and 2007 in greece, Arch. Gerontol. Geriatr. 50 (2010)
e31e35.
[11] C. Bogiatzi, D.G. Hackam, A.I. McLeod, J.D. Spence, Secular trends in ischemic stroke
subtypes and stroke risk factors, Stroke 45 (2014) 32083213.
[12] K.H. Jung, S.H. Lee, B.J. Kim, K.H. Yu, K.S. Hong, B.C. Lee, et al., Secular trends in ischemic
stroke characteristics in a rapidly developed country: results from the korean stroke
registry study (secular trends in korean stroke), Circ. Cardiovasc. Qual. Outcomes
5 (2012) 327334.
[13] V. Thijs, K. Butcher, Challenges and misconceptions in the aetiology and manage-
ment of atrial brillation-related strokes, Eur. J. Int. Med. 26 (2015) 461467.
[14] C. Marini, F. De Santis, S. Sacco, T. Russo, L. Olivieri, R. Totaro, et al., Contribution of
atrial brillation to incidence and outcome of ischemic stroke: results from a popu-
lation-based study, Stroke 36 (2005) 11151119.
[15] M. Paciaroni, G. Agnelli, F. Corea, W. Ageno, A. Alberti, A. Lanari, et al., Early hemor-
rhagic transformation of brain infarction: rate, predictive factors, and inuence on
clinical outcome: results of a prospective multicenter study, Stroke 39 (2008)
22492256.
[16] F. Bridge, V. Thijs, How and when to screen for atrial brillation after stroke: insights
from insertable cardiac monitoring devices, J. Stroke 18 (2016) 121128.
[17] H.P. Adams, B.H. Bendixen, L.J. Kappelle, J. Biller, B.B. Love, D.L. Gordon, et al., Classi-
cation of subtype of acute ischemic stroke. Denitions for use in a multicenter clin-
ical trial. Toast. Trial of org 10172 in acute stroke treatment, Stroke 24 (1993) 3541.
[18] P. Amarenco, P.C. Lavallee, J. Labreuche, G.W. Albers, N.M. Bornstein, P. Canhao, et al.,
One-year risk of stroke after transient ischemic attack or minor stroke, N. Engl. J.
Med. 374 (2016) 15331542.
[19] N. Hannon, O. Sheehan, L. Kelly, M. Marnane, A. Merwick, A. Moore, et al., Stroke as-
sociated with atrial brillationincidence and early outcomes in the north dublin
population stroke study, Cerebrovasc. Dis. 29 (2010) 4349.