New Appraisal of Atrial Fibrillation Burden
and Stroke Prevention
Rod Passman, MD, MSCE; Richard A. Bernstein, MD, PhD
A trial fibrillation (AF) quintuples the risk of stroke, and
patients with stroke and AF are the most likely to be
left dead or permanently disabled.1–4 More than 5 million
people in the United States and >8 million people in the
European Union were estimated to have AF in 2010.5–7 The
lifetime risk of AF is high, with a quarter of 40-year-old
patients expected to develop AF during the course of their
lifetime.8 Because of aging populations, the prevalence of
AF in the United Sates is expected to exceed 12 million by
2030, whereas in the European Union estimates 17.9 mil-
lion cases by 2060.6,7 The risk of stroke can be stratified by
various scores, including CHADS2 (congestive heart failure,
hypertension, age ≥75 years, diabetes mellitus [1 point], and
previous stroke or transient ischemic attack [TIA; 2 points])
and the CHA2DS2-VASc score (congestive heart failure/left
ventricular dysfunction, hypertension [1 point], age ≥75
years [2 points], diabetes mellitus [1 point], previous stroke
or TIA [2 points], vascular disease, age 65–75 years, and
Downloaded from http://ahajournals.org by on January 22, 2019
female sex [1 point]).9,10 Multiple trials have demonstrated
the efficacy of anticoagulation in reducing the risk of stroke
by 65% to 80% in patients with clinically detected AF and
other risk factors.10–13
AF burden can be defined as the time spent in AF per unit
of time (day, week, month, etc). It is noteworthy that the bur-
den of AF is not part of any risk stratification scoring system.
In fact, current practice views AF as a dichotomous variable;
one that is either present—whether paroxysmal, persistent,
or permanent—or absent. Implicit in this perspective is an
assumption that the burden AF is not relevant to stroke risk.
This assumption is based on results from older cohort stud-
ies and anticoagulation trials that showed that paroxysmal AF
carries the same stroke risk, and responds as well to antico-
agulation, as persistent or permanent AF. This perspective is
reinforced by major guidelines that do not distinguish stroke
risk based on whether AF is paroxysmal (AF that terminates
spontaneously or with intervention within 7 days of onset),
persistent (continuous AF that is sustained >7 days), or perma-
nent (ie, when the patient and clinician make a joint decision
to stop further attempts to restore or maintain sinus rhythm),
Received July 24, 2015; final revision received November 29, 2015; accepted December 1, 2015.
From the Electrophysiology Section, Cardiology Division, Bluhm Cardiovascular Institute, Feinberg School of Medicine of Northwestern University,
Chicago, IL (R.P.); and Stroke Program, Davee Department of Neurology, Feinberg School of Medicine of Northwestern University, Chicago, IL (R.A.B.).
Correspondence to Richard A. Bernstein, MD, PhD, Northwestern Stroke Program, Davee Department of Neurology, 710 North Lake Shore Dr, Chicago,
IL 60611. E-mail Richard.bernstein4@gmail.com
(Stroke. 2016;47:570-576. DOI: 10.1161/STROKEAHA.115.009930.)
© 2016 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org
570
or whether AF burden has been decreased or eliminated with
antiarrhythmic drugs, ablation, or surgery.10
Stroke neurologists are increasingly asked to make anti-
coagulation recommendations for patients with exceedingly
brief and isolated episodes of AF. This makes it crucial for
neurologists to understand the gaps in knowledge about the
significance of these brief episodes of AF, the risk of stroke
associated with these episodes, and the implications for stroke
prevention. Rather than thinking of AF as a dichotomous vari-
able (present or absent), neurologists may have to view AF
as a continuous variable, with different amounts (burdens)
of AF carrying different stroke risks and treatment implica-
tions. This article seeks to update stroke neurologists on this
new way of thinking about AF. Ongoing clinical trials should
clarify the use of this view of AF, and this review prepares
neurologists to critically appraise these trials.
The current practice of treating AF as a dichotomous
variable—of not stratifying stroke risk or choosing antithrom-
botic therapy based on burden—is an outgrowth of data col-
lected at a time when only high-burden AF was detectable.
For example, the AF described in the Framingham study was
detected by routine ECG, physical examination, or at the time
of symptoms.4 In contrast, the advent of implantable cardiac
rhythm management devices (ie, dual-chamber pacemakers,
implantable cardiac defibrillators [ICDs], and implantable
cardiac monitors [ICMs]) with automatic AF detection, have
now made it possible to detect AF that occurs only a few times
per year and as briefly as a minute or two in duration. Patients
with this de minimus AF may be biologically different than
those with the high-burden AF studied in older landmark tri-
als, but this possibility has not been systematically studied.
We think it is the time to reexamine the assumptions
about AF burden that underlie current practice and guidelines
because ongoing trials may overturn them.
Brief History of AF and Stroke
In the first half of the 20th century, autopsy series of patients
with rheumatic mitral valve disease and chronic AF by Harvey
and Levine14 determined that auricular fibrillation definitely
increases the incidence of auricular thrombosis. A second
DOI: 10.1161/STROKEAHA.115.009930
 Passman and Bernstein   AF Burden and Stroke Risk    571
autopsy series performed by Hay and Levine15 found that the
presence of AF in patients with mitral stenosis significantly
increased the risk of cardiac thrombus formation compared with
those without known AF. In the 1970s, the Framingham study
was the first to demonstrate that the risk of stroke extended to
those with AF and without mitral stenosis, although the risk
was substantially lower in the former (5- versus 17-fold).4
The 1980s brought early attempts to delineate the association
between paroxysmal AF and stroke, but were limited by AF
detection by electrocardiogram and symptoms. Takahashi et al16
found embolic complications in 8 of 94 patients with paroxys-
mal atrial fibrillation, and a Danish study reviewing data accu-
mulated between 1940 and 1957 showed a yearly incidence of
emboli in 2.0% of patients with paroxysmal atrial fibrillation
and 5.1% of those with chronic AF.17 The clear connection
between stroke and nonrheumatic AF (so called nonvalvular
AF) paved the way for trials evaluating the role of anticoagula-
tion for stroke prevention in nonvalvular AF given its success in
patients with valvular AF. These trials also provided the oppor-
tunity to correlate stroke risk with AF duration. In a substudy of
the Stroke Prevention in Atrial Fibrillation (SPAF) I–III trials
performed between 1987 and 1997, stroke risk was compared in
aspirin-treated patients with paroxysmal and persistent or per-
manent AF and found near-identical strokes risks between the 2
groups (3.2% versus 3.3% per year).18
These data formed the basis of current recommendations
that ignore AF burden when risk-stratifying patients for stroke.
However, these data may be outdated because of secular trends
including better control of hypertension and other stroke risk
Downloaded from http://ahajournals.org by on January 22, 2019
factors, which may lower the stroke risk in patients with AF.
For example, an analysis of the stroke rates in 6563 aspirin-
treated patients with AF from the contemporary ACTIVE-A/
AVERROES databases showed that the risk of stroke, after
adjusting for known risk factors, was 2.1, 3.0, and 4.2% for par-
oxysmal, persistent, and permanent AF, respectively (P<0.05
for all comparisons).19 In fact, AF pattern (a gross reflection of
burden) was the second strongest predictor of stroke after pre-
vious stroke or TIA. This finding shows that the risk of stroke
correlates with burden, and suggests that efforts at stroke pre-
vention may have to take burden into account.
If AF burden is important, then assessing burden becomes
critical. Unfortunately, neither symptoms nor short-term
monitoring are reliable in determining who has AF, or how
much AF they have. Approximately 40% of patients with AF
are completely asymptomatic and those who do experience
symptoms are aware of only 5% to 20% of episodes.20,21 The
relevance of this for stroke neurologists is that one quarter
of all AF does not become symptomatic except by causing a
stroke. Long-term monitoring with implantable devices pro-
vide a unique opportunity to detect and quantitate AF burden,
correlate AF duration with stroke occurrence, and potentially
manage decision making surrounding anticoagulation in a
select subset of patients with AF.
Device-Detected Atrial High Rate Episodes
Correlate With True Atrial Tachyarrhythmias
Pacemakers and ICDs with atrial leads can detect local atrial
depolarizations at the endocardial surface.22 Atrial rates above
a certain cut-off (typically 170–220 bpm) can be categorized
as atrial high rate episodes (AHRE). AHRE detection can be
triggered by any of the atrial tachyarrhythmias, including AF,
atrial flutter, or atrial tachycardia, with AF and atrial flutter
representing the most common arrhythmias. Current pace-
makers and ICDs have been shown to detect AF with a high
degree of accuracy. AHRE detection by mode switching algo-
rithms designed to prevent ventricular tracking of prolonged
atrial tachyarrhythmias have a sensitivity and specificity of
98% and 100%, respectively.22,23 In contrast to pacemakers
and defibrillators with leads in the right atrium, leadless ICMs
detect AF based on incoherence of the R–R interval. The over-
all accuracy of the ICM for AF duration is 98.5%.24
Association Between Device-Detected AHRE and
Stroke
AHRE are definitely associated with stroke. However, the bur-
den of AHRE sufficient to raise stroke risk is unclear, with
various studies suggesting AF duration anywhere from 5
minutes to >24 hours is associated with increased stroke risk.
Furthermore, no study has yet determined if anticoagulation of
AHRE reduces the risk of stroke to a similar degree as it does
in clinically manifest AF. Complicating this analysis, most
studies linking AHRE to stroke did not make any attempt to
determine the mechanism of AHRE-associated stroke. These
studies are summarized in Table.
In an ancillary study of the Mode Selection Trial (MOST),
312 patients with a median age of 74 years who had pace-
makers placed for sinus node dysfunction were prospectively
studied for a median of 27 months.25 Their pacemakers were
programmed to log AHRE when the atrial rate was >220 bpm
for >5 consecutive minutes. The incidence of AHRE during
the study was 51.3%. Multivariate analysis showed that the
presence of any AHRE >5 minutes in duration was an inde-
pendent predictor of death or nonfatal stroke, with a hazard
ratio of 2.8 (95% confidence interval [CI], 1.5–5.2) compared
with those without any AHRE. The Asymptomatic Atrial
Fibrillation and Stroke Evaluation in Pacemaker Patients and
the Atrial Fibrillation Reduction Atrial Pacing Trial (ASSERT)
study showed that subclinical AHRE were not only common
in patients with implantable cardiac devices but also associ-
ated with an increased risk of stroke, similar to that seen in the
MOST trial.26 In this study, 2580 patients with no history of
AF who were ≥65 years of age and had a history of hyperten-
sion, in whom a dual-chamber pacemaker or ICD had been
recently placed, were monitored for 3 months for the pres-
ence of AHRE (defined as an atrial rate of >190 bpm for >6
minutes). They were then followed for a mean of 2.5 years for
the primary outcome of ischemic stroke or systemic embo-
lism. At 3 months, 10.1% of patients demonstrated subclinical
AHRE. Having any subclinical AHRE >6 minutes was signifi-
cantly associated with the primary outcome of ischemic stroke
or systemic embolism with a hazard ratio of 2.49 (95% CI,
1.28–4.85), although the annual rates of stroke and systemic
embolization did not become statistically significant until an
AF duration of ≈18 hours was reached when the patients with
AHRE were stratified according to duration quartiles of the
longest AHRE episode.
 572  Stroke  February 2016

Table.  Selected Studies Evaluating the Correlation of AF Burden and Stroke or Systemic Embolism
Author Pts (n) Study Type/Inclusion Criteria Monitoring Method/Duration Outcome
Glotzer et al14
312 Ancillary analysis of multicenter Dual-chamber PPM ● 10 patient (3.2%) developed stroke
RCT (MOST) for a median of 27 mo ● Atrial arrhythmia ≥5 min: HR, 2.8; P=0.0011 for death
or nonfatal stroke
Capucci et al8 725 Prospective, registry study Dual-chamber PPM ● 14 patients (1.9%) had an arterial thromboembolic
for a median of 22 mo event
● AF episode lasting >24 h: HR , 3.1; P=0.044 for
adj
embolic events
● No AF episodes or AF 5 min to 24 h: no difference in

embolic events
Glotzer et al9 2486 Prospective, observational study Dual-chamber PPM ● Daily AT/AF burden <5.5 h: no difference in rate of
(TRENDS) or ICD for a mean of 1.4 y thromboembolism compared with no AT/AF (both with
1.1% annual rate)
● Daily AT/AF burden ≥ 5.5 h: trend toward higher annual

rate of thromboembolism (2.2%) compared with no AT/

AF, but not statistically significant (P=0.06)
● 30-d cumulative AT/AF burden <10.8 h: no difference

in rate of thromboembolism compared with no AT/AF

group (P=0.96)
● 30-d cumulative AT/AF burden ≥10.8 h: trend toward

increased risk of thromboembolism (HRadj, 2.2; P=0.06)

compared with no AT/AF
Healey et al10 2580 Primary analysis of a multicenter Dual-chamber PPM or ● Atrial tachyarrhythmia >6 min: HR, 1.76; P=0.05 for
RCT (ASSERT). ICD for a mean of 2.5 y stroke or systemic embolism compared with patients
with no arrhythmia
● Atrial tachyarrhythmia <17.7 h: annual rate of stroke or

systemic embolism 1.2%

● Atrial tachyarrhythmia >17.7 h: annual rate of stroke or

systemic embolism 4.9%

Shanmugam et al13 560 Ancillary analysis from 2 CRT device for a mean of 1 y ● 11 patients (2%) had a thromboembolic events
Downloaded from http://ahajournals.org by on January 22, 2019

prospective multicenter ● Atrial tacharrhythmia ≥3.8 h a day: HR, 9.4; P=0.006

observational studies of CHF for stroke or systemic embolism compared with
patients with CRT patients with no arrhythmia
● No significant increase risk of thromboembolic events

in patients with ≥3.8 h a day vs <3.8 h a day: HR, 2.4;

P=0.23
AF indicates atrial fibrillation; ASSERT, Asymptomatic Atrial Fibrillation and Stroke Evaluation in Pacemaker Patients and the Atrial Fibrillation Reduction Atrial
Pacing Trial; AT, atrial tachycardia; CHF, congestive heart failure; CRT, cardiac resynchronization therapy; HRadj, adjusted hazard ratio; ICD, implantable cardioverter-
defibrillator; MOST, Mode Selection Trial; PPM, permanent pacemaker; and RCT, randomized control trial.

A higher burden of AF was found to significantly increase
the risk of stroke in the TRENDS study (A Prospective Study
of the Clinical Significance of Atrial Arrhythmias Detected
by Implanted Device Diagnostics).27 This trial was a prospec-
tive cohort study of patients with ≥1 stroke risk factor (heart
failure, hypertension, age ≥65 years, diabetes mellitus, or
previous thromboembolic event) who were scheduled for a
dual-chamber pacemaker or ICD with atrial monitoring capa-
bilities. The primary outcome of the study was the incidence
of stroke, TIA, or thromboembolic event. AHRE were defined
as atrial rates >175 bpm lasting ≥20 s. The AHRE burden
was defined as the maximum number of hours of AHRE on
any 1 day during a 30-day period, and may have included
>1 episode of AHRE per day. The median burden of AHRE
among all 30-day rolling windows that contained AHRE was
5.5 hours. Patients were divided into 3 groups for analysis:
those with zero AHRE, those with <5.5 hours of AHRE, and
those with ≥5.5 hours of AHRE during any 30-day period.
The annualized risk of stroke, TIA, or thromboembolism was
1.1%, 1.1%, and 2.4% for patients with zero, <5.5 hours, and
≥5.5 hours of AHRE, respectively. Compared with patients
who had a zero AHRE burden, patients with <5.5 hours of
AHRE had a hazard ratio of 0.98 (95% CI, 0.34–2.82) for
stroke, TIA, or thromboembolism. Those with ≥5.5 hours of
AHRE had a hazard ratio of 2.20 (95% CI, 0.96–5.05). Thus,
in this study, patients with ≥5.5 hours of AHRE were roughly
twice as likely to have a stroke, TIA, or thromboembolism as
those without AHRE, whereas those with <5.5 hours of AHRE
were at no increased risk.
Finally, a longer duration of device-detected AF was
required to significantly increase stroke risk in a study by
Capucci et al28 that assessed the incidence of arterial embolism
in 725 patients with bradycardia, a history of symptomatic AF
(either paroxysmal or persistent), and an indication for a dual-
chamber pacemaker. The incidence of arterial embolism dur-
ing the median follow-up time of 22 months was 1.9%. The
risk of embolism, adjusted for other known risk factors, was
found to be 3.1× higher (95% CI, 1.1–10.5) in patients with
device-detected AF lasting >24 hours compared with those
with <24 hours of AF or no AF. In fact, the occurrence of
 Passman and Bernstein   AF Burden and Stroke Risk    573
device-detected AF lasting >5 minutes but <24 hours was not
associated with a significantly higher risk of embolic events
compared with those with no AF in this study. These studies
are summarized in the Table.
In sum, these studies demonstrate an association between
the burden of AHRE and stroke risk, but offer no consensus
on a threshold effect between the 2. Furthermore, these stud-
ies did not explicitly investigate an interaction between tra-
ditional stroke risk factors, AF burden, and stroke risk. This
potential interaction was studied by Botto et al,29 who evalu-
ated the combination of duration of device-detected AHRE
and CHADS2 score on stroke risk. The study evaluated 568
patients with pacemakers and a history of AF during a 1-year
period. There was a 2.5% stroke rate during the 1-year fol-
low-up. Patients were divided into groups based on duration
of device-detected AHRE (none, between 5 minutes and 24
hours, and >24 hours) and CHADS2 score (0, 1, 2, or ≥3). Two
subpopulations of patients were identified with significantly
different stroke risks. A low stroke risk of 0.8% per year was
identified in patients who had zero AHRE with CHADS2 ≤2,
had <24 hours of AHRE with CHADS2 ≤1, or had >24 hours
of AHRE with CHADS2=0. A high stroke risk of 5% (P=0.035
compared with the first patient population) was identified in
patients who had zero AHRE but with CHADS2 >2, had <24
hours of AHRE with CHADS2 >1, or had >24 hours of AHRE
with CHADS2 ≥1. This study suggests that AHRE burden may
be most helpful in differentiating stroke risk in patients with
an intermediate CHADS2 score (1 or 2). Although patients
with a low CHADS2 score of 0 are at low risk regardless of
Downloaded from http://ahajournals.org by on January 22, 2019
AF duration, and patients with a high CHADS2 score of ≥3 are
at high risk regardless of AF duration. Of note, those patients
with a CHADS2 score of 1 or 2 and a history of AF but no
documented recurrences had a stroke risk similar to that of a
CHADS2 0 patient. This concept is illustrated schematically
in the Figure.
Temporal Association of AHRE and Stroke
Substudies of the 2 trials showing an association between
AHRE and stroke, TRENDS and ASSERT, have paradoxi-
cally shown that AHRE are not necessarily temporally related
to stroke occurrence. In a substudy of the TRENDS trial, the
40 patients enrolled in TRENDS who experienced stroke or
TIA and who had at least 30 days of monitoring before the
occurrence of an ischemic event were evaluated.30 AHRE of at
least 5 minutes were detected before stroke in 20 (50%) of the
Figure.  Theoretical interaction between stroke risk factors and
atrial fibrillation (AF) burden.
40 patients. In contrast, 9 (45%) of the 20 patients with AHRE
detected before stroke did not have any AHRE detected in the
30 days preceding stroke. Thus, 29 (73%) of the 40 patients
with stroke had no AHRE within the 30 days before their
event. Of the 20 patients who had AHRE before stroke, only 6
(30%) had AHRE at the time of stroke. In the 14 patients who
had AHRE before stroke, the mean time from the last recorded
AHRE and stroke was relatively long at 168±199 days (range,
3–642 days). Still, those patients who had ≥5.5 hours of AF
were more than twice as likely to have a thromboembolism
event within 30 days because those with shorter AF durations
and no recorded AF, and the 30-day stroke risk was higher
even within the same individual after a longer AF episode.
This suggests that there may be a temporal association only
between longer AHRE episodes with stroke.
A substudy of the ASSERT trial showed similar results.31
Of the 51 patients in ASSERT with stroke, subclinical AF was
detected in 26 (51%). Of these patients, only 18 (35%) had
subclinical AF before stroke. Only 4 (8%) had subclinical AF
within 30 days before stroke, and only 1 patient (2%) had sub-
clinical AF at the time of stroke. In patients with AF >30 days
before stroke, AF occurred a median of 339 days (25th–75th
percentile, 211–619 days) before stroke. AF was detected only
after stroke in 8 (16%) patients.
These analyses suggest that AHRE burden correlates with
stroke risk, but may not always be directly causative of stroke.
There are several potential explanations for this latter finding.
First, the link between AHRE burden and stroke could be an
epiphenomenon: patients with more AHRE may have more
severe vascular risk factors.32 Second, it may be that many
patients had strokes because of causes other than AF. Finally,
the event rate may have been insufficient to demonstrate a
temporal association. Consistent with this last point, the larg-
est study to date which enrolled 9850 patients with implanted
devices showed that the odds ratio for the risk of stroke within
30 days of an AHRE lasting ≥5.5 hours was 5.2 compared
with a control period 90 to 120 days before.33
On the basis of the available data, subclinical AHRE that
would likely go otherwise undetected are clearly associated
with stroke. However, it remains unclear if this association is
causative or simply a marker of overall risk. If in fact AF is an
epiphenomenon, then maintaining sinus rhythm will not com-
pletely protect against strokes, particularly in patients with
multiple other risk factors.
Treatment Based on Burden: Anticoagulation Based
on Remote Monitoring
Most implantable cardiac rhythm management devices now
have the ability to send transmissions remotely either manu-
ally or automatically. These transmissions, which can occur as
often as once a day, include detailed information about ambi-
ent arrhythmias and device function, allowing for continu-
ous monitoring of the cardiac rhythm. Several studies have
shown the ability of remote monitoring to decrease the time
from AHRE occurrence to detection and treatment when com-
pared with conventional monitoring with periodic remote or
in-office device interrogations at fixed intervals in an outpa-
tient clinic.34–39 These studies lead to the possibility that early
 574  Stroke  February 2016
recognition and treatment of subclinical AF could decrease
stroke while minimizing exposure to anticoagulation.
In an analysis of the Comparative Follow Up Schedule
With Home Monitoring (COMPAS) trial, patients showed a
trend toward lower stroke rates when followed by continuous
remote monitoring compared with standard in-office follow-
up.38 However, this study was designed primarily to determine
the safety of remote monitoring and was neither powered to
show a difference in stroke rates between groups nor designed
to specify an intervention protocol once AF was discov-
ered. Similarly, the Lumos-T Safely Reduces Routine Office
Device Follow-Up (TRUST) study showed a nonsignificant
trend toward a lower stroke rate in the group of patients with
continuous remote monitoring compared with those moni-
tored with standard in-office follow-up, although again the
study was not designed to look at specific interventions for
device-detected AHRE.36
The IMPACT (Randomized Trial of Anticoagulation
Guided by Remote Rhythm monitoring in Patients with ICD
and Resynchronization Devices) trial attempted to address the
question of management of device-detected AF.39 This trial
was designed to investigate the use of remote monitoring com-
bined with a predefined anticoagulation strategy compared with
conventional device evaluation and physician-directed antico-
agulation in patients with dual-chamber ICDs or cardiac resyn-
chronization therapy devices. Warfarin was the anticoagulant
used in the vast majority of patients. Patients with a CHADS2
≥1 with or without a history of paroxysmal AF were random-
ized in a 1:1 fashion to 1 of 2 arms. In the standard-of-care arm,
Downloaded from http://ahajournals.org by on January 22, 2019
patients received in-office follow-up with physician-directed
anticoagulation if AF was detected. Patients in the treatment
arm were monitored remotely for device-detected AHRE (≥200
bpm for 36 of 48 beats) and anticoagulation was started and
stopped based on a combination of CHADS2 score and duration
of AHRE. Patients with a CHADS2 score <2 were started on
anticoagulation with warfarin if they had >48 continuous hours
of AHRE, and warfarin was stopped if AHRE were not detected
for 30 consecutive days. Patients with a CHADS2 score of 3 to
4 started anticoagulation with warfarin if they had >24 hours
of AHRE within a 48-hour window, and warfarin was stopped
if AHRE were not detected for 90 consecutive days. Finally,
patients with a CHADS2 score of 5 to 6 were started on antico-
agulation with warfarin if they had any duration of AHRE. Once
started, warfarin was not stopped in this group. The primary end
point of the study was freedom from the composite of stroke,
systemic embolism, or major bleeding. A total of 2718 patients
were randomized and followed for ≤5 years. However, there
was no significant difference between groups in the primary out-
come, and the trial was terminated early because of futility. At 5
years, the Kaplan–Meier estimate of the primary end point was
86.8% in the remote monitored group and 87.9% in the stan-
dard of care group (P=0.73). One major limitation of this study
is that warfarin was used for anticoagulation in the majority of
patients. Because warfarin can take 5 to 7 days to become thera-
peutic, patients in the treatment arm who had device-detected
AF may not have been anticoagulated during the critical window
immediately after the detection of AF. In addition, international
normalized ratio levels were subtherapeutic about a third of the
time, and many patients were not treated according to the study
protocol. Event rates during the trial were lower than expected
and the combined end point of bleeding and stroke may have
obscured any benefit for stroke prevention in this study.
Several ongoing and planned studies test the link between
device-detected episodes of subclinical AF, stroke, and the
impact of anticoagulation. The Apixaban for the Reduction
of Thrombo-Embolism in patients with device-detected Sub-
Clinical Atrial fibrillation (ARTESiA) study has been designed
to look specifically at the treatment of patients with subclinical
device-detected AF, with no history of clinical AF (clinicaltri-
als.gov NCT01938248). The trial has begun enrollment, with
a goal of enrolling 4000 patients. Patients with device-detected
subclinical AF will be randomized to treatment with either
aspirin or apixaban. Patients will be followed for an estimated
3 years for the primary end point of stroke or systemic embo-
lism. This study will be one of the first randomized trials to
evaluate treatment for patient with subclinical AF and may elu-
cidate whether anticoagulation for device-detected AF reduces
stroke risk in the same way it does for clinical AF.
Although much recent attention has been paid to the asso-
ciation between device-detected subclinical AF episodes and
stroke, the same technology used to identify AF could poten-
tially also be used to limit oral anticoagulation exposure in
patients with a history of AF but with low or absent AF bur-
dens either spontaneously or as the result of a rhythm con-
trol intervention, such as ablation or antiarrhythmic therapies.
The REACT.COM study used daily home monitoring from
an ICM (Reveal XT, Medtronic Inc) and used an AF duration
of >1 hour as the threshold to initiate oral anticoagulation for
30 days. The pilot study showed a 94% reduction in time on
oral anticoagulation compared with expected.40 In contrast to
the IMPACT trial, this pilot study focused on lower risk (ie,
CHADS2 score 1 or 2) patients and used rapid onset novel oral
anticoagulations instead of warfarin. The planned REACT-AF
trial will randomize CHA2DS2-VASc 1 to 4 patients with non-
permanent AF to either chronic novel oral anticoagulation
therapy or targeted therapy using an ICM.
All these studies implicitly test the link between AF (or
AHRE) burden and stroke by only anticoagulating patients
with AF when they have the arrythmia documented by an
implantable device. If this strategy is successful, it would sug-
gest that AF is not a dichotomous variable, that anticoagula-
tion is only needed when the arrhythmia is present, and that
maintenance of sinus rhythm might indeed lower stroke risk
enough to obviate the need for anticoagulation.
Areas of Uncertainty
Although it seems clear that AHRE detected by ICMs and
other implanted are associated with stroke, several major areas
of uncertainty remain. First, whether there is a particular bur-
den of AHRE needed to raise the risk of stroke is unclear. The
studies presented above have demonstrated a wide range of
durations (>5 minutes to >24 hours) needed to significantly
increase the risk of stroke. Interestingly, all the studies show-
ing an association between AHRE and stroke have shown an
association with durations of <48 hours, which has tradition-
ally been the cut-off used before cardioversion to determine if
 Passman and Bernstein   AF Burden and Stroke Risk    575
the stroke risk with cardioversion is high enough to warrant
either precardioversion anticoagulation or a transesophageal
echocardiogram to rule out left atrial appendage thrombus.10 It
may also be that the duration of AF in combination with other
clinical risk factors, such as the CHADS2 or the CHA2DS2-
VASc score, may be a better predictor of stroke risk than the
presence and duration of AF alone. Second, the mechanism
of stroke in patients with device-detected AF remains to be
determined. The temporal dissociation of most AHRE from
stroke observed in some, but not all, studies raises the pos-
sibility that AHRE may not be the direct cause of stroke in
all patients and that for some patients the possibility exists
that short episodes of AF are simply a marker for other stroke
mechanisms. Further research into this area is needed. Finally,
whether treatment with anticoagulation for subclinical device-
detected AHRE reduces stroke risk as it does for clinical AF
is currently unknown. Planned and ongoing studies can be
expected to answer most if not all of these questions in the
not too distant future. It is important to stress, however, that at
present, there is no clear clinical trial evidence that AF burden
should influence anticoagulation decisions.
Disclosures
Dr Bernstein reports fees for research support, consulting, and speak-
ing, from Medtronic, Inc, Boehringer Ingelheim, and Pfizer/Bristol
Myers Squibb. Dr Passman reports research support from National
Institutes of Health, and Consultant or Advisory Board, and speak-
ers fees, from Medtronic and Janssen; and fees for speaking and
Advisory Boards for Pfizer/Bristol Myers Squibb.
Downloaded from http://ahajournals.org by on January 22, 2019
References
1. Indredavik B, Rohweder G, Lydersen S. Frequency and effect of
optimal anticoagulation before onset of ischaemic stroke in patients
with known atrial fibrillation. J Intern Med. 2005;258:133–144. doi:
10.1111/j.1365-2796.2005.01512.x.
2. Saxena R, Lewis S, Berge E, Sandercock PA, Koudstaal PJ. Risk of early
death and recurrent stroke and effect of heparin in 3169 patients with
acute ischemic stroke and atrial fibrillation in the International Stroke
Trial. Stroke. 2001;32:2333–2337.
3. Kimura K, Minematsu K, Yamaguchi T; Japan Multicenter Stroke
Investigators’ Collaboration (J-MUSIC). Atrial fibrillation as a predic-
tive factor for severe stroke and early death in 15,831 patients with acute
ischaemic stroke. J Neurol Neurosurg Psychiatry. 2005;76:679–683. doi:
10.1136/jnnp.2004.048827.
4. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent
risk factor for stroke: the Framingham Study. Stroke. 1991;22:983–988.
5. Go AS, Hylek EM, Phillips KA, Chang Y, Henault LE, Selby JV, et al.
Prevalence of diagnosed atrial fibrillation in adults: national implications
for rhythm management and stroke prevention: the AnTicoagulation
and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA.
2001;285:2370–2375.
6. Krijthe BP, Kunst A, Benjamin EJ, Lip GY, Franco OH, Hofman A, et
al. Projections on the number of individuals with atrial fibrillation in the
European Union, from 2000 to 2060. Eur Heart J. 2013;34:2746–2751.
doi: 10.1093/eurheartj/eht280.
7. Colilla S, Crow A, Petkun W, Singer DE, Simon T, Liu X. Estimates
of current and future incidence and prevalence of atrial fibrillation in
the U.S. adult population. Am J Cardiol. 2013;112:1142–1147. doi:
10.1016/j.amjcard.2013.05.063.
8. Lloyd-Jones DM, Wang TJ, Leip EP, Larson MG, Levy D, Vasan RS, et
al. Lifetime risk for development of atrial fibrillation: the Framingham
Heart Study. Circulation. 2004;110:1042–1046. doi: 10.1161/01.
CIR.0000140263.20897.42.
9. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford
MJ. Validation of clinical classification schemes for predicting stroke:
results from the National Registry of Atrial Fibrillation. JAMA.
2001;285:2864–2870.
10. January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC
Jr, et al; American College of Cardiology/American Heart Association
Task Force on Practice Guidelines. 2014 AHA/ACC/HRS guideline
for the management of patients with atrial fibrillation: a report of the
American College of Cardiology/American Heart Association Task
Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll
Cardiol. 2014;64:e1–e76. doi: 10.1016/j.jacc.2014.03.022.
11. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al;
ROCKET AF Investigators. Rivaroxaban versus warfarin in nonvalvu-
lar atrial fibrillation. N Engl J Med. 2011;365:883–891. doi: 10.1056/
NEJMoa1009638.
12. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna
M, et al; ARISTOTLE Committees and Investigators. Apixaban versus
warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–
992. doi: 10.1056/NEJMoa1107039.
13. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh
A, et al; RE-LY Steering Committee and Investigators. Dabigatran
versus warfarin in patients with atrial fibrillation. N Engl J Med.
2009;361:1139–1151. doi: 10.1056/NEJMoa0905561.
14. Harvey EA, Levine SA. A study of uninfected mural thrombi of the
heart. Am J Med Sci. 1930;180:365.
15. Hay WE, Levine SA. Age and auricular fibrillation as independent fac-
tors in auricular mural thrombus formation. Am H J. 1942;24:1–3.
16. Takahashi N, Seki A, Imataka K, Fujii J. Clinical features of parox-
ysmal atrial fibrillation. An observation of 94 patients. Jpn Heart J.
1981;22:143–149.
17. Petersen P, Godtfredsen J. Embolic complications in paroxysmal atrial
fibrillation. Stroke. 1986;17:622–626.
18. Hart RG, Pearce LA, Rothbart RM, McAnulty JH, Asinger RW, Halperin
JL. Stroke with intermittent atrial fibrillation: incidence and predictors
during aspirin therapy. JACC. 2000;35:183–187.
19. Vanassche T, Lauw MN, Eikelboom JW, Healey JS, Hart RG, Alings M,
et al. Risk of ischaemic stroke according to pattern of atrial fibrillation:
analysis of 6563 aspirin-treated patients in ACTIVE-A and AVERROES.
Eur Heart J. 2015;36:281–288, 27a. doi: 10.1093/eurheartj/ehu307.
20. Page RL, Wilkinson WE, Clair WK, McCarthy EA, Pritchett EL.
Asymptomatic arrhythmias in patients with symptomatic paroxys-
mal atrial fibrillation and paroxysmal supraventricular tachycardia.
Circulation. 1994;89:224–227.
21. Strickberger SA, Ip J, Saksena S, Curry K, Bahnson TD, Ziegler PD.
Relationship between atrial tachyarrhythmias and symptoms. Heart
Rhythm. 2005;2:125–131. doi: 10.1016/j.hrthm.2004.10.042.
22. Passman RS, Weinberg KM, Freher M, Denes P, Schaechter A,
Goldberger JJ, et al. Accuracy of mode switch algorithms for detection of
atrial tachyarrhythmias. J Cardiovasc Electrophysiol. 2004;15:773–777.
doi: 10.1046/j.1540-8167.2004.03537.x.
23. Swerdlow CD, Schsis W, Dijkman B, Jung W, Sheth NV, Olson WH, et
al. Detection of atrial fibrillation and flutter by a dual-chamber implant-
able cardioverter-defibrillator. For the Worldwide Jewel AF Investigators.
Circulation. 2000;101:878–885.
24. Pürerfellner H, Pokushalov E, Sarkar S, Koehler J, Zhou R, Urban L, et
al. P-wave evidence as a method for improving algorithm to detect atrial
fibrillation in insertable cardiac monitors. Heart Rhythm. 2014;11:1575–
1583. doi: 10.1016/j.hrthm.2014.06.006.
25. Glotzer TV, Hellkamp AS, Zimmerman J, Sweeney MO, Yee R,
Marinchak R, et al; MOST Investigators. Atrial high rate episodes
detected by pacemaker diagnostics predict death and stroke: report
of the Atrial Diagnostics Ancillary Study of the MOde Selection
Trial (MOST). Circulation. 2003;107:1614–1619. doi: 10.1161/01.
CIR.0000057981.70380.45.
26. Healey JS, Connolly SJ, Gold MR, Israel CW, Van Gelder IC, Capucci
A, et al; ASSERT Investigators. Subclinical atrial fibrillation and
the risk of stroke. N Engl J Med. 2012;366:120–129. doi: 10.1056/
NEJMoa1105575.
27. Glotzer TV, Daoud EG, Wyse DG, Singer DE, Ezekowitz MD, Hilker
C, et al. The relationship between daily atrial tachyarrhythmia burden
from implantable device diagnostics and stroke risk: the TRENDS
study. Circ Arrhythm Electrophysiol. 2009;2:474–480. doi: 10.1161/
CIRCEP.109.849638.
28. Capucci A, Santini M, Padeletti L, Gulizia M, Botto G, Boriani G, et al;
Italian AT500 Registry Investigators. Monitored atrial fibrillation dura-
tion predicts arterial embolic events in patients suffering from bradycar-
dia and atrial fibrillation implanted with antitachycardia pacemakers. J
Am Coll Cardiol. 2005;46:1913–1920. doi: 10.1016/j.jacc.2005.07.044.
 576  Stroke  February 2016
29. Botto GL, Padeletti L, Santini M, Capucci A, Gulizia M, Zolezzi
F, et al. Presence and duration of atrial fibrillation detected by con-
tinuous monitoring: crucial implications for the risk of thromboem-
bolic events. J Cardiovasc Electrophysiol. 2009;20:241–248. doi:
10.1111/j.1540-8167.2008.01320.x.
30. Daoud EG, Glotzer TV, Wyse DG, Ezekowitz MD, Hilker C, Koehler
J, et al; TRENDS Investigators. Temporal relationship of atrial tachyar-
rhythmias, cerebrovascular events, and systemic emboli based on
stored device data: a subgroup analysis of TRENDS. Heart Rhythm.
2011;8:1416–1423. doi: 10.1016/j.hrthm.2011.04.022.
31. Brambatti M, Connolly SJ, Gold MR, Morillo CA, Capucci A, Muto C,
et al; ASSERT Investigators. Temporal relationship between subclinical
atrial fibrillation and embolic events. Circulation. 2014;129:2094–2099.
doi: 10.1161/CIRCULATIONAHA.113.007825.
32. Inoue H, Nozawa T, Okumura K, Jong-Dae L, Shimizu A, Yano K.
Prothrombotic activity is increased in patients with nonvalvular atrial
fibrillation and risk factors for embolism. Chest. 2004;126:687–692. doi:
10.1378/chest.126.3.687.
33. Turakhia MP, Ziegler PD, Schmitt SK, Chang Y, Fan J, Than CT, et al.
Atrial Fibrillation burden and short-term risk of stroke: case-crossover
analysis of continuously recorded heart rhythm from cardiac electronic
implanted devices. Circ Arrhythm Electrophysiol. 2015;8:1040–1047.
doi: 10.1161/CIRCEP.114.003057.
34. Lazarus A. Remote, wireless, ambulatory monitoring of implantable pace-
makers, cardioverter defibrillators, and cardiac resynchronization therapy
systems: analysis of a worldwide database. Pacing Clin Electrophysiol.
2007;(30 suppl 1):S2–S12. doi: 10.1111/j.1540-8159.2007.00595.x.
35. Ricci RP, Morichelli L, Santini M. Remote control of implanted
devices through Home Monitoring technology improves detection and
clinical management of atrial fibrillation. Europace. 2009;11:54–61.
doi: 10.1093/europace/eun303.
Downloaded from http://ahajournals.org by on January 22, 2019
36. Varma N, Epstein AE, Irimpen A, Schweikert R, Love C; TRUST
Investigators. Efficacy and safety of automatic remote monitoring
for implantable cardioverter-defibrillator follow-up: the Lumos-T
Safely Reduces Routine Office Device Follow-up (TRUST) trial.
Circulation. 2010;122:325–332. doi: 10.1161/CIRCULATIONAHA.
110.937409.
37. Crossley GH, Boyle A, Vitense H, Chang Y, Mead RH; CONNECT
Investigators. The CONNECT (Clinical Evaluation of Remote
Notification to Reduce Time to Clinical Decision) trial: the value of
wireless remote monitoring with automatic clinician alerts. J Am Coll
Cardiol. 2011;57:1181–1189. doi: 10.1016/j.jacc.2010.12.012.
38. Mabo P, Victor F, Bazin P, Ahres S, Babuty D, Da Costa A, et al;
COMPAS Trial Investigators. A randomized trial of long-term remote
monitoring of pacemaker recipients (the COMPAS trial). Eur Heart J.
2012;33:1105–1111. doi: 10.1093/eurheartj/ehr419.
39. Martin DT, Bersohn MM, Waldo AL, Wathen MS, Choucair WK, Lip
GY, et al; IMPACT Investigators. Randomized trial of atrial arrhythmia
monitoring to guide anticoagulation in patients with implanted defibril-
lator and cardiac resynchronization devices. Eur Heart J. 2015;36:1660–
1668. doi: 10.1093/eurheartj/ehv115.
40. Passman R, Leong-Sit P, Andrei A-C, Huskin A, Tomson TT, Bernstein
R, et al. Targeted Anticoagulation for Atrial Fibrillation Guided by
Continuous Rhythm Assessment with an Insertable Cardiac Monitor:
The Rhythm Evaluation for Anticoagulation with Continuous
Monitoring (REACT.COM) Pilot Study [published online ahead
of print November 23, 2015]. J Cardiovasc Electrophysiol. http://
onlinelibrary.wiley.com/enhanced/doi/10.1111/jce.12864. Accessed
November 30, 2015.
Key Words: atrial fibrillation ◼ cohort studies ◼ diabetes mellitus ◼ heart
failure ◼ stroke