Increased risk of ischemic stroke associated with new onset atrial fibrillation complicating acute coronary syndrome- a systematic review and meta-analysis

1. Increased risk of ischemic stroke associated with new-onset atrial
fibrillation complicating acute coronary syndrome: A systematic review
and meta-analysis
Jiachen Luo, Hongqiang Li, Xiaoming Qin, Baoxin Liu, Jinlong Zhao, Guli Maihe, Zhiqiang Li, Yidong Wei ⁎,1
Department of Cardiology, Shanghai Tenth People's Hospital, Tongji University School of Medicine, 301 Middle Yanchang Road, Jingan District, Shanghai, People's Republic of China
a b s t r a c ta r t i c l e i n f o
Article history:
Received 21 December 2017
Received in revised form 28 March 2018
Accepted 20 April 2018
Available online xxxx
Background: Atrial fibrillation has been established as a major risk factor of ischemic stroke, however, the influ-
ence of new-onset atrial fibrillation (NOAF) complicating acute coronary syndrome (ACS) on ischemic stroke re-
mains controversial. This meta-analysis aimed to validate the association between NOAF complicating ACS and
ischemic stroke.
Methods: We identified randomized controlled trials and cohort studies comparing the ischemic stroke risk be-
tween patients with NOAF and sinus rhythm after ACS by searching MEDLINE, EMBASE and Cochrane Central
Register of Controlled Trials databases. We included studies reporting the number of ischemic stroke events or
their risk estimates at the longest follow-up. We pooled risk ratios (RRs) using a random-effects model. This
meta-analysis is registered in PROSPERO (CRD42017079858).
Results: In the 14 included studies (n = 292,774, 5 randomized controlled trials and 9 cohort studies), NOAF was
associated with an increased risk of ischemic stroke (RR: 2.84, 95% confidence interval [CI]: 1.91–4.23; 6 studies),
especially for patients with ST-segment elevation myocardial infarction (RR: 4.01, 95% CI: 2.61–6.18; 3 studies).
In addition, the detrimental impact persisted in patients with transient NOAF (RR: 3.05, 95% CI: 1.63–5.70; 3 stud-
ies). The pooled result from a sensitivity analysis in which all individual components in the CHA2DS2-VASc score
(heart failure, hypertension, age, diabetes, previous stroke, vascular disease and female sex) had been adjusted
further validated the association between NOAF and ischemic stroke (RR: 2.32, 95% CI: 1.53–3.52; 4 studies).
Conclusions: NOAF is significantly associated with ischemic stroke events in patients with ACS, even after adjust-
ment for several important ischemic stroke risk factors.
© 2017 Elsevier B.V. All rights reserved.
Keywords:
Acute coronary syndrome
Atrial fibrillation
Ischemic stroke
Meta-analysis
1. Introduction
Stroke is a rare but devastating clinical event after acute coronary
syndrome (ACS) and has been validated as a strong predictor for subse-
quent mortality [1–3]. During the past decades, the incidence of stroke
after ACS is gradually falling, which may be attributable to the rigorous
secondary prevention of atherosclerotic diseases [4,5], as well as the de-
creased incidence of hemorrhagic stroke due to the shift in reperfusion
strategies from fibrinolysis to percutaneous coronary intervention [6].
Until now, ischemic stroke has increasingly been perceived as the lead-
ing cause of stroke after ACS [2].
Atrial fibrillation (AF) has been established as a major risk factor for
ischemic stroke [7]. In fact, AF is a common finding in patients with ACS
as the result of their similar risk factors and can be categorized as pre-
existing and new-onset AF (NOAF) based on their temporal relation-
ships with ACS. In a previous meta-analysis, Jabre et al. showed both
AF types were significantly associated with increased mortality and sug-
gested more attention should be paid to their management [8]. Unlike
the confirmed ischemic stroke susceptibility of pre-existing AF [9], it is
still controversial whether the NOAF complicating ACS is also associated
with an increased risk of ischemic stroke [10–15]. Hence, we aimed to
perform a systematic review and meta-analysis to quantify the risk of is-
chemic stroke associated with post-ACS NOAF.
2. Methods
2.1. Literature search and selection criteria
The present meta-analysis was performed following the Preferred Reporting Items for
Systematic reviews and Meta-Analyses (PRISMA) guideline [16] and was registered in the In-
ternational Prospective Register for Systematic Reviews (PROSPERO: CRD42017079858). We
searched Ovid MEDLINE, Ovid EMBASE and Cochrane Central Register of Controlled Trials da-
tabases from their inception to October 2017, using the combination of medical subjects
heading terms and keywords: “acute coronary syndrome”, “myocardial infarction”, “coronary
thrombosis”, “atrial fibrillation”, “cerebrovascular disorders”, “cerebrovascular accident”,
International Journal of Cardiology xxx (2017) xxx–xxx
⁎ Corresponding author.
E-mail address: TJYidong_wei@163.com. (Y. Wei).
1
This author takes responsibility for all aspects of the reliability and freedom from bias
of the data presented and their discussed interpretation.
IJCA-26369; No of Pages 7
https://doi.org/10.1016/j.ijcard.2018.04.096
0167-5273/© 2017 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
International Journal of Cardiology
journal homepage: www.elsevier.com/locate/ijcard
Please cite this article as: J. Luo, et al., Increased risk of ischemic stroke associated with new-onset atrial fibrillation complicating acute coronary
syndrome: A systematic..., Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2018.04.096

2. “ischemic stroke”, “brain infarction”, without any language limitation. The reference lists of
retrieved studies and prior reviews were also screened for other eligible studies.
Studies considered in our study were randomized controlled trials (RCTs) and cohort
studies comparing the ischemic stroke risk between patients with NOAF and sinus rhythm
(SR) after index ACS. We excluded studies that did not report the number of stroke events
or their risk estimates and those in which no distinction could be made between hemor-
rhagic and ischemic stroke. In addition, “reviews”, “editorials”, “letters”, “case reports”,
“conference abstracts”, and “case-control studies” were also excluded.
We classified NOAF as transient NOAF, persisting NOAF or any NOAF. NOAF was
defined as AF occurring for the first time after the ACS with no history of AF in medical re-
cords. Transient NOAF was defined as NOAF occurring during hospitalization with SR at
discharge. Persisting NOAF was defined as NOAF occurring during hospitalization with
AF at discharge. If no distinction about the status of NOAF at discharge was made, NOAF
was classified as any NOAF.
2.2. Data extraction
Four reviewers working independently and using a standardized form extracted data
from all eligible studies, including baseline characteristics of studies and patients and the
number of ischemic stroke events or their risk estimates. If several risk estimates were
available in the same study, the most fully adjusted result corresponding to the longest
follow-up duration was extracted. We further tried to contact corresponding authors of
studies for missing data through E-mail. Discrepancies were resolved by consensus.
2.3. Quality evaluation
For the purpose of our study, we dealt with all eligible RCTs as cohort studies, with the
population being treated as a whole without considering the randomization process. The
Newcastle-Ottawa Scale (NOS) was used to evaluate the quality of studies. A quality score
was calculated according to three major components: selection (0–4 points), comparabil-
ity (0–2 points) and outcome (0–3 points) [17]. Notably, whether the individual compo-
nents in the CHA2DS2-VASc score had been adjusted was used for comparability
assessment. The age was chosen as the major risk factor, and other risk factors were
heart failure, hypertension, diabetes, previous stroke/transient ischemic attack (TIA), vas-
cular disease and female sex (Table S1 [supplements]). Good comparability was consid-
ered if one or two stars were obtained. A total score of seven or more was considered as
a high-quality study.
2.4. Outcomes and subgroup analyses
The primary study endpoint was the ischemic stroke. TIA would be an alternative
when ischemic stroke was not reported. Subgroup analyses were performed to compare
the outcomes according to study type, sample size, geographic location, comparability,
and publication date. In addition, we explored the risk of ischemic stroke in patients
with either ST-segment elevation myocardial infarction (STEMI) or transient NOAF.
2.5. Sensitivity analyses
To confirm the robustness of our analyses, several sensitivity analyses were per-
formed including: 1) statistical models (fixed- and random-effects); 2) limited to studies
with large sample (≥10,000), with all components in the CHA2DS2-VASc score being ad-
justed, conducted in multiple centers, or in which ischemic stroke events were measured
after discharge; 3) exclusion of studies with the largest sample or the most outlier result,
with atrial flutter being included, or in which coronary artery bypass grafting surgery was
performed.
2.6. Statistical analysis
Descriptive analyses were demonstrated as frequencies for categorical variables and
standardized means (standard deviations) or median (interquartile) for continuous vari-
ables. We used random-effects model described by DerSimonian and Laird to calculate
pooled risk ratios (RR) and 95% confidence intervals (CI) [18]. Heterogeneity was evalu-
ated with the χ2
based-Q-statistic test, and I2
was used to quantify the inconsistency. I2
b 25%, 25%–50% and N50% suggested low, moderate and high heterogeneity, respectively.
Univariate meta-regression models were used to determine the interactions between sub-
groups. Publication bias was evaluated using Egger's test [19]. A value of p b 0.05 (2 sided)
was considered statistically significant. All analyses were performed using Stata software
version 14 (StataCorp, College Station, Texas).
3. Results
3.1. Characteristics of the included studies
As demonstrated in Fig. 1, our initial literature search identified 1198
studies. After title and abstract screening, 1153 studies were excluded
and full-text review retrieved 45 studies; 31 studies were further ex-
cluded according to exclusion criteria and 14 studies including 5 retro-
spectives from RCTs [20–24] and 9 cohort studies [12–15,25–29] were
available for the final analysis. Atrial flutter and fibrillation were treated
as a whole in 4 studies [15,23,24,29] and 6 studies only included STEMI
patients [12,20,21,23,24,27]. Most of NOAF events were evaluated dur-
ing hospitalization except for 3 studies in which on-admission NOAF
were included [15,25,26]. All studies had reported the ischemic stroke
except for one in which only TIA was available [24]. Table 1 showed
the details of included studies.
3.2. Characteristics of the included patients
The incidence of NOAF was 7.4% (95% CI: 5.8%–9.0%). Patients with
NOAF were older (70.1 ± 3.4 years vs 61.9 ± 2.8 years), more likely to
be women (31.8 ± 4.2% vs 24.5 ± 4.5%) and had more baseline co-
existing conditions (e.g., hypertension, diabetes, myocardial infarction,
etc.) than those with SR. In addition, the CHA2DS2-VASc score was sig-
nificantly higher in patients with NOAF (4.2 ± 0.1 vs 3.1 ± 0.2). Fur-
thermore, patients with NOAF were more likely to receive oral
anticoagulants (17.0% vs 4.0%) and less likely to receive aspirin (84.8%
vs 87.3%) or P2Y12 inhibitors (41.6% vs 49.0%) at discharge. Details of pa-
tients' characteristics were demonstrated in Table 2.
3.3. Quality evaluation
Quality evaluation by NOS revealed a median score of 7 (range, 4–9).
Furthermore, six studies with good comparability demonstrated an ex-
cellent quality (median 8, range 7–9), whereas the other 8 studies only
had a median score of 6 (range, 4–6) (Table S2 [supplements]).
Accounting for the high heterogeneity from the pooled result of all el-
igible studies, and the origins of which could not be determined by
performing subgroup analyses and meta-regression analyses (Table S3
[supplements]), we decided to report only stroke risk estimates from 6
studies that with good comparability and high quality [13,21,23,27–29].
3.4. Ischemic stroke associated with NOAF complicating ACS
The incidence of ischemic stroke after ACS was 1.6% (95% CI: 0.5%–
2.8%), and ischemic stroke rates at three periods: in-hospital, 1 month
to 1 year and ≥1 year were 0.9%, 1.2%, and 3.7%, respectively. Post-ACS
NOAF was associated with an increased risk of ischemic stroke com-
pared with those in SR (RR: 2.84; 95% CI: 1.91–4.23; p b 0.01) (Fig.
2A). After removing the GRACE registry [29], only a low heterogeneity
was observed and the significance of the pooled result remained (RR:
3.21; 95% CI: 2.36–4.37; p b 0.01) (Fig. 2B). Of note, in the GRACE regis-
try, only in-hospital ischemic stroke events were evaluated. No risk of
publication bias was showed by the Egger's test (p = 0.15).
3.5. Subgroup and sensitivity analyses
In a subgroup analysis of patients with STEMI [21,23,27], NOAF was
significantly associated with an increased risk of ischemic stroke (RR:
4.01; 95% CI: 2.61–6.18; p b 0.01) (Fig. S1A [supplements]). When sub-
group analysis was performed with respect to transient NOAF
[13,27,28], the RR of ischemic stroke was 3.05 (95% CI: 1.63–5.70; p b
0.01) (Fig. S1B [supplements]).
We conducted a sensitivity analysis pooling studies in which all
components in the CHA2DS2-VASc score had been adjusted
[13,21,28,29], the detrimental impact of post-ACS NOAF was still of
great significance (RR: 2.32, 95% CI: 1.53–3.52; p b 0.01). Details of sen-
sitivity analyses were demonstrated in Fig. S2 (supplements).
4. Discussion
4.1. Main findings
The current meta-analysis demonstrates the mean incidence of is-
chemic stroke after ACS is 1.6%. NOAF complicating ACS is significantly
2 J. Luo et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Please cite this article as: J. Luo, et al., Increased risk of ischemic stroke associated with new-onset atrial fibrillation complicating acute coronary
syndrome: A systematic..., Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2018.04.096

3. associated with an increased risk of ischemic stroke, especially for pa-
tients with STEMI, after adjustment for several important ischemic
stroke risk factors. Moreover, transient NOAF is even associated with is-
chemic stroke events.
4.2. Incidence of ischemic stroke after ACS
Ischemic stroke is an infrequent clinical event after ACS. In a previ-
ous meta-analysis performed by Witt et al., the ischemic stroke rates
at hospital stay (1.1%) and 1 month (1.2%) were similar to our study
(0.9% and 1.2%, respectively) [30]. However, as we analyzed the mean
cumulative rate of stroke over 1 year (mean follow-up: 45 months) as
a whole, it was not surprising to observe such a higher ischemic stroke
rate (3.7%) in our study compared with that in Witt et al. (2.1% at 1
year).
4.3. NOAF complicating ACS and ischemic stroke
NOAF is one of the most common arrhythmias after ACS with a re-
ported incidence ranging from 4% to 19% [8]. Although the increased
mortality associated with the post-ACS NOAF has been validated by nu-
merous studies [8,29], it is still unknown whether the post-ACS NOAF
has a similar influence on ischemic stroke. In a previous study, Zusman
et al. showed that the NOAF following myocardial infarction was associ-
ated with a nearly 35-fold increased risk of stroke during follow-up
(mean: 41 months; hazard ratio [HR]: 34.6, 95% CI: 4.0–296.8) [10].
However, the limited number of patients and events (14 events out of
300 patients) made their results seem to be less precise, as evidenced
by such a wide 95% CI. In contrast, with the use of data from Danish Na-
tional Patient Registry, Bang et al. conducted a retrospective analysis
with a total of 89,703 patients with MI being analyzed and at the end
of 5-year follow-up, NOAF complicating myocardial infarction was dem-
onstrated as an independent predictor for fatal or non-fatal stroke (HR:
2.34, 95% CI: 2.12–2.57 and HR: 2.47, 95% CI: 2.24–2.73, respectively)
[31]. Nevertheless, the lack of data on stroke etiology made a more com-
prehensive understanding of the prognostic implication of NOAF un-
available. Differently, in the present meta-analysis with a relatively
large population, we can validate that the NOAF was significantly asso-
ciated with an increased risk of ischemic stroke given all stroke events
could be clearly categorized as ischemic origins. To our best knowledge,
this is the first meta-analysis of clinical studies on the ischemic stroke
risk of NOAF after ACS
Despite the strong association between post-ACS NOAF and ische-
mic stroke events, it remains unclear whether the NOAF is a “causal
risk factor” or rather a “risk indicator” for ischemic stroke after ACS. As
exposure always precedes the outcome, the temporal relationship is a
pivotal factor in causality establishment [32]. For example, in the
ASSERT (Asymptomatic Atrial Fibrillation and Stroke Evaluation in Pace-
maker Patients and the Atrial Fibrillation Reduction Atrial Pacing Trial)
study, Brambatti et al. demonstrated that only 8% AF events were de-
tected within 30 days before index stroke with the use of implanted de-
vices, 16% of stroke victims had their first AF event after strokes.
Fig. 1. PRISMA flow diagram of included studies.
3J. Luo et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Please cite this article as: J. Luo, et al., Increased risk of ischemic stroke associated with new-onset atrial fibrillation complicating acute coronary
syndrome: A systematic..., Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2018.04.096

4. Table 1
Characteristics of included studies.
Study Year Country Single/multicenter Years of
study
Total/NOAF
population, n
NOAF types1
NOAF evaluation ACS
types
Endpoints Timing of endpoint
measurement
In-hospital
fibrinolysis, %
(NOAF/SR)
In-hospital PCI, %
(NOAF/SR)
Retrospective from RCTs
HORIZON AMI Trial [23] 2014 United States Multicenter 2005–2007 3281/147 Any2
In-hospital STEMI IS 3 years 0/0 100/100
APEX-AMI Trial [21] 2009 Multination Multicenter 2004–2006 5742/342 Any In-hospital STEMI IS 3 months 0/0 100/100
GISSI-3 Trial [22] 2001 Italy Multicenter 1991–1993 17,749/1386 Any In-hospital MI IS In-hospital 65/73 NR
GUSTO-III Trial [24] 2000 Multination Multicenter 1995–1997 13,858/906 Any2
In-hospital STEMI TIA In-hospital 100/100 0/0
GUSTO-I Trial [20] 1997 Multination Multicenter 1990–1993 40,891/3254 Any In-hospital STEMI IS In-hospital 100/100 0/0
Cohort studies
SWEDEHEART Registry [28] 2016 Sweden Multicenter 2000–2009 155,071/11742 Any/transient/persistent In-hospital MI IS 3 months NR 24/48
Braga et al. [25] 2015 Portugal Single 2009–2012 1373/142 Any On-admission/in-hospital ACS IS 3 months NR 55/69
González et al. [26] 2015 Mexico Single 2006–2013 6705/220 Any On-admission/in-hospital ACS IS In-hospital 5/5 18/164
ARIAM Registry [15] 2014 Spain Multicenter 2001–2011 39,237/1568 Any2
On-admission/in-hospital ACS IS In-hospital 60/58 66/69
Viliani et al. [12] 2012 Spain Single 2004–2008 913/92 Any/transient/persistent In-hospital STEMI IS In-hospital 0/0 100/100
Bishara et al. [13] 2011 Israel Single 2000–2009 2402/174 Transient In-hospital MI IS and TIA 1 year NR 43/52
Asanin et al. [14] 2009 Serbia Single 1996–1998 3210/320 Transient In-hospital MI IS 7 years3
25/24 NR
Siu et al. [27] 2007 China Single 1997–2005 431/59 Transient In-hospital STEMI IS 38.5 months3
39/32 17/134
GRACE Registry [29] 2003 Multination Multicenter 1999–2001 21,785/1221 Any2
In-hospital ACS IS In-hospital 50/55 25/32
AMI = acute myocardial infarction; ACS = acute coronary syndrome; AF = atrial fibrillation; AMI = acute myocardial infarction; IS = ischemic stroke; NOAF = new-onset atrial fibrillation; NR = not report; PCI = percutaneous coronary interven-
tion; STEMI=ST-segment elevation myocardial infarction; RCT = randomized controlled trial; SR = sinus rhythm; TIA = transient ischemic attack.
1
Transient AF means NOAF only presents during hospital stay with sinus rhythm at discharge; Persistent AF means NOAF presents both during hospital stay and at discharge; Any AF means NOAF cannot be categorized as transient or persistent
derives;
2
Studies include atrial flutter;
3
Mean follow-up durations;
4
Data represent primary PCI.
4J.Luoetal./InternationalJournalofCardiologyxxx(2017)xxx–xxx
Pleasecitethisarticleas:J.Luo,etal.,Increasedriskofischemicstrokeassociatedwithnew-onsetatrialfibrillationcomplicatingacutecoronary
syndrome:Asystematic...,IntJCardiol(2017),https://doi.org/10.1016/j.ijcard.2018.04.096

5. Therefore, they drew a conclusion that subclinical AF may simply be a
risk marker of stroke due to the lack of confirmed temporal relationship
[33]. Differently, as reported in the CRYSTAL AF (The Cryptogenic Stroke
and Underlying AF) trial, although cryptogenic stroke occurred before
the AF detection, Sanna et al. still recognized the AF as the underlying
cause of observed cryptogenic stroke given the high HR of 6.4–8.8 for
AF detection [34], thus also indicating the importance of association
strength. In the present study, we performed a sensitivity analysis in
which all ischemic stroke events were measured after discharge to en-
sure the NOAF occurred before the ischemic stroke, the pooled result
demonstrated that the NOAF was associated with almost a 3.2-fold in-
creased risk of subsequent ischemic stroke (Fig. S2 [supplements]),
thus also revealing a strong association between them.
Several underlying mechanisms have been proposed to explain the
links between NOAF and ischemic stroke, including cardiac emboliza-
tion, worsening heart failure [14], recurrent AF susceptibility
[13,14,27] and coexisting conditions (e.g., hypertension, diabetes [4,5],
etc.). Notably, the recurrent AF may be a major pathophysiological
mechanism by which NOAF significantly increases the risk of stroke,
as shown in the Asanin et al., it was the recurrence of AF during
follow-up (adjusted RR: 5.08, 95% CI: 1.92–13.42, p = 0.001) rather
than initial NOAF that was independently associated with long-term
stroke events after adjustment for confounding factors [14].
However, there are still some pitfalls in establishing unequivocal
causality. First, the scarcity of clinical data evaluating the correlation be-
tween different NOAF burdens (e.g., paroxysmal, persistent or perma-
nent) and ischemic stroke risk made the dose-response relationship
cannot be validated. Nevertheless, we still could speculate that such a
dose-response relationship might exist, as was reported in the Batra
et al., the risk estimate of stroke for persistent NOAF (HR: 2.77, 95% CI:
2.03–3.77) was higher than that for transient NOAF (HR: 1.87, 95% CI:
Table 2
Baseline characteristics of participants.
Baseline variables Total population
(n)
NOAF SR p
value
Age, years 275,025 70.1 ± 3.4 61.9 ± 2.8 b0.001
Female sex, % 88,875 31.8 ± 4.2 24.5 ± 4.5 b0.001
BMI, kg/m2
11,160 27.8 ± 0.5 27.2 ± 0.1 0.154
Hypertension, % 139,137 56.5 ± 11.0 48.8 ± 9.6 b0.001
Diabetes, % 63,838 27.1 ± 7.9 22.7 ± 7.5 b0.001
Hyperlipidemia, % 19,267 40.7 ± 7.9 44.6 ± 6.5 b0.001
Current smoker, % 46,618 28.0 ± 12.6 36.8 ± 11.0 b0.001
Previous MI, % 62,735 20.6 ± 6.2 18.1 ± 6.8 b0.001
Previous
revascularization, %
8688 9.5 ± 3.5 10.4 ± 3.3 b0.001
Previous stroke/TIA, % 13,660 8.1 ± 3.2 4.9 ± 2.0 b0.001
Previous HF, % 18,639 13.0
(7.5–29.5)
9.0
(3.5–14.5)
b0.001
HF at admission1
, % 22,421 34.3 ± 13.7 15.6 ± 5.2 b0.001
LVEF, % 88,646 46.2 ± 2.9 50.0 ± 1.4 0.016
STEMI, % 167,705 78.5 ± 24.4 76.4 ± 26.6 b0.001
CHA2DS2-VASc score 3718 4.2 ± 0.1 3.1 ± 0.2 0.024
Fibrinolysis, % 82,941 55.5 ± 33.5 55.9 ± 34.6 b0.001
PCI treatment, % 100,081 54.8 ± 35.0 59.8 ± 33.5 b0.001
Aspirin, % 207,592 84.8 ± 17.7 87.3 ± 21.2 b0.001
P2Y12 inhibitors, % 121,304 41.6 ± 27.6 49.0 ± 32.5 b0.001
Oral-anticoagulant, % 9414 17.0
(6.5–19.0)
4.0
(2.5–4.5)
b0.001
Statins, % 14,974 61.6 ± 20.9 65.9 ± 21.7 b0.001
ACEI/ARB, % 42,546 57.9 ± 21.9 55.8 ± 26.3 b0.001
Values are demonstrated as n, mean ± SD or median (interquartile).
ACEI = angiotensin converting enzyme inhibitors; ARB = angiotensin-II receptor
blockers; BMI = body mass index; CHA2DS2-VASc = congestive heart failure, hyperten-
sion, age ≥ 75 yrs, diabetes, previous stroke and/or TIA, vascular diseases, age 65–74 yrs,
female gender; HF = heart failure; LVEF = left ventricular ejection fraction; MI = myo-
cardial infarction; Other abbreviations refer to Table 1.
1
HF at admission refers to Killip class N I.
Fig. 2. Summary forest plot of ischemic stroke risk associated with NOAF complicating ACS. (A) Ischemic stroke risk and NOAF after ACS. (B) Ischemic stroke risk and NOAF after MI. The size
of each square is proportional to the study's weight. The solid line across the square represents the 95% CI. The dotted line in the forest plot shows random-effects pooled risk estimate. ACS
= acute coronary syndrome; CI = confidence interval; MI = myocardial infarction; NOAF = new-onset atrial fibrillation; RR = risk ratio.
5J. Luo et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Please cite this article as: J. Luo, et al., Increased risk of ischemic stroke associated with new-onset atrial fibrillation complicating acute coronary
syndrome: A systematic..., Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2018.04.096

6. 1.33–2.63) when compared with SR, respectively [28]. Second, neither
has RCT been conducted to evaluate the benefits of post-ACS NOAF
treatment on subsequent ischemic stroke nor are animal models or lab-
oratory findings available to validate the association between post-ACS
NOAF and ischemic stroke. Taken together, the hypothesis that post-
ACS NOAF is a risk factor for ischemic stroke is very much alive and vi-
able, but a great deal of work involving both basic research and specifi-
cally designed RCT still needs to be done.
4.4. NOAF complicating ACS and antithrombotic strategy
Given the fact that NOAF complicating ACS is significantly associated
with an increased risk of ischemic stroke, anticoagulation therapy may
be effective in reducing stroke risk and improving mortality [14,27]. In
an observational study with respect to transient NOAF after myocardial
infarction, Bishara et al. reported the risk estimate for stroke/TIA in pa-
tients receiving antiplatelet agents (HR: 3.28,95% CI: 1.82–5.93) was
higher than that in those receiving oral-anticoagulants (HR: 1.97, 95%
CI: 0.48–8.12) [13]. By contrast, Tangelder et al. demonstrated that com-
pared with aspirin plus placebo, a dual-antithrombotic therapy com-
posed of aspirin and ximelagatran was not associated with a
decreased risk of ischemic stroke (HR: 0.24; 95% CI: 0.02–2.30) [35].
However, as the present meta-analysis was not made to evaluate the
benefits of anticoagulation therapy on stroke prevention with respect
to the post-ACS NOAF, further studies are warranted to explore the op-
timal antithrombotic strategy in this setting.
4.5. Limitations
The major limitation was including observational data from ran-
domized trials and cohort studies for the purpose of our work, which
could subject this analysis to potential bias. Second, due to the lack of
patient-level data, we failed to test for interactions at the patient-level
covariates. Third, although all included studies had stated that patients
with a medical history of AF were excluded, the possibility of asymp-
tomatic AF episodes before ACS should be noted, which might result
in an overestimation of the ischemic stroke risk associated with post-
ACS NOAF given the confirmed ischemic stroke susceptibility of silent
AF [36]. In fact, this is an inherent limitation for all studies on the topic
of NOAF. However, accounting for the better detection of silent AF in
the contemporary clinical reality [37], as well as the relatively lower
prevalence and incidence of subclinical AF in patients without a medical
history of AF compared with those unselected patients [38], the poten-
tial misclassification of NOAF might has little influence on the interpre-
tation of our pooled results. Fourth, since the management of post-ACS
NOAF regarding either anticoagulation or cardioversion therapy during
the follow-up period had not been reported, we could not estimate their
effects on the risk of subsequent stroke. Finally, we also could not eval-
uate the influence of different NOAF burdens on ischemic stroke risk be-
cause no data were available in our study.
5. Conclusions
NOAF complicating ACS is significantly associated with an increased
risk of ischemic stroke, especially for patients with STEMI, even after ad-
justment for several important ischemic stroke risk factors. Therefore,
closer attention with respect to stroke prevention should be paid to pa-
tients with NOAF after index ACS.
Funding
This work was supported by the National Natural Science Founda-
tion of China [grant numbers 81270193, 30800466] to Dr. Yidong Wei.
Conflict of interest
The authors report no relationship that could be construed as a con-
flict of interest.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.ijcard.2018.04.096.
References
[1] N. Naderi, H. Masoomi, T. Mozaffar, S. Malik, Patient characteristics and comorbidi-
ties associated with cerebrovascular accident following acute myocardial infarction
in the United States, Int. J. Cardiol. 175 (2014) 323–327.
[2] J.S. Saczynski, F.A. Spencer, J.M. Gore, et al., Twenty-year trends in the incidence of
stroke complicating acute myocardial infarction: Worcester Heart Attack Study,
Arch. Intern. Med. 168 (2008) 2104–2110.
[3] A. Budaj, K. Flasinska, J.M. Gore, et al., Magnitude of and risk factors for in-hospital
and postdischarge stroke in patients with acute coronary syndromes: findings
from a global registry of acute coronary events, Circulation 111 (2005) 3242–3247.
[4] U. Kajermo, A. Ulvenstam, A. Modica, T. Jernberg, T. Mooe, Incidence, trends, and
predictors of ischemic stroke 30 days after an acute myocardial infarction, Stroke
45 (2014) 1324–1330.
[5] A. Ulvenstam, U. Kajermo, A. Modica, T. Jernberg, L. Soderstrom, T. Mooe, Incidence,
trends, and predictors of ischemic stroke 1 year after an acute myocardial infarction,
Stroke 45 (2014) 3263–3268.
[6] E. Binsell-Gerdin, A. Graipe, J. Ogren, T. Jernberg, T. Mooe, Hemorrhagic stroke the
first 30 days after an acute myocardial infarction: incidence, time trends and predic-
tors of risk, Int. J. Cardiol. 176 (2014) 133–138.
[7] C.T. January, L.S. Wann, J.S. Alpert, et al., AHA/ACC/HRS guideline for the manage-
ment of patients with atrial fibrillation: a report of the American College of Cardiol-
ogy/American Heart Association Task Force on Practice Guidelines and the Heart
Rhythm Society, J. Am. Coll. Cardiol. 64 (2014) (2014) e1–76.
[8] P. Jabre, V.L. Roger, M.H. Murad, et al., Mortality associated with atrial fibrillation in
patients with myocardial infarction: a systematic review and meta-analysis, Circula-
tion 123 (2011) 1587–1593.
[9] D. Tanne, U. Goldbourt, M. Zion, H. Reicher-Reiss, E. Kaplinsky, S. Behar, Frequency
and prognosis of stroke/TIA among 4808 survivors of acute myocardial infarction.
The SPRINT study group, Stroke 24 (1993) 1490–1495.
[10] O. Zusman, G. Amit, H. Gilutz, D. Zahger, The significance of new onset atrial fibril-
lation complicating acute myocardial infarction, Clin. Res. Cardiol. 101 (2012)
17–22.
[11] T. Mooe, P. Eriksson, B. Stegmayr, Ischemic stroke after acute myocardial infarction.
A population-based study, Stroke 28 (1997) 762–767.
[12] D. Viliani, D. Vivas, M. Chung, et al., Prognosis of different types of atrial fibrillation in
the primary angioplasty era, Coron. Artery Dis. 23 (2012) 511–516.
[13] R. Bishara, G. Telman, F. Bahouth, J. Lessick, D. Aronson, Transient atrial fibrillation
and risk of stroke after acute myocardial infarction, Thromb. Haemost. 106 (2011)
877–884.
[14] M.R. Asanin, Z.M. Vasiljevic, M.D. Matic, et al., The long-term risk of stroke in pa-
tients with acute myocardial infarction complicated with new-onset atrial fibrilla-
tion, Clin. Cardiol. 32 (2009) 467–470.
[15] M. Almendro-Delia, M.J. Valle-Caballero, J.C. Garcia-Rubira, et al., Prognostic impact
of atrial fibrillation in acute coronary syndromes: results from the ARIAM registry,
Eur. Heart J. Acute Cardiovasc. Care 3 (2014) 141–148.
[16] D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman, Preferred reporting items for system-
atic reviews and meta-analyses: the PRISMA statement, BMJ 339 (2009), b2535. .
[17] G.S.B. Wells, D. O'Connell, J. Peterson, V. Welch, M.T.P. Losos, The Newcastle-Ottawa
Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-
AnalysesAvailable at http://www.ohri.ca/programs/clinical_epidemiology/oxford.
asp 2013, Accessed date: 25 October 2017.
[18] R.L.N. DerSimonian, Meta-analysis in clinical trials, Control. Clin. Trials 7 (1986)
177–188.
[19] M.D.S.G. Egger, M. Schneider, C. Minder, Bias in meta-analysis detected by a simple,
graphical test, BMJ 315 (1997) 629–634.
[20] B.S. Crenshaw, S.R. Ward, C.B. Granger, A.L. Stebbins, E.J. Topol, R.M. Califf, Atrial fi-
brillation in the setting of acute myocardial infarction: the GUSTO-I experience.
Global utilization of streptokinase and TPA for occluded coronary arteries, J. Am.
Coll. Cardiol. 30 (1997) 406–413.
[21] R.D. Lopes, L.E. Elliott, H.D. White, et al., Antithrombotic therapy and outcomes of
patients with atrial fibrillation following primary percutaneous coronary interven-
tion: results from the APEX-AMI trial, Eur. Heart J. 30 (2009) 2019–2028.
[22] F. Pizzetti, F.M. Turazza, M.G. Franzosi, et al., Incidence and prognostic significance of
atrial fibrillation in acute myocardial infarction: the GISSI-3 data, Heart 86 (2001)
527–532.
[23] A.G. Rene, P. Genereux, M. Ezekowitz, et al., Impact of atrial fibrillation in patients
with ST-elevation myocardial infarction treated with percutaneous coronary inter-
vention (from the HORIZONS-AMI [Harmonizing Outcomes With Revascularization
and Stents in Acute Myocardial Infarction] trial), Am. J. Cardiol. 113 (2014) 236–242.
[24] C.K. Wong, H.D. White, R.G. Wilcox, et al., New atrial fibrillation after acute myocar-
dial infarction independently predicts death: the GUSTO-III experience, Am. Heart J.
140 (2000) 878–885.
6 J. Luo et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Please cite this article as: J. Luo, et al., Increased risk of ischemic stroke associated with new-onset atrial fibrillation complicating acute coronary
syndrome: A systematic..., Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2018.04.096

7. [25] C.G. Braga, V. Ramos, J. Martins, et al., Impact of atrial fibrillation type during acute
coronary syndromes: clinical features and prognosis, Rev. Port. Cardiol. 34 (2015)
403–410.
[26] H. Gonzalez-Pacheco, M.F. Marquez, A. Arias-Mendoza, et al., Clinical features and
in-hospital mortality associated with different types of atrial fibrillation in patients
with acute coronary syndrome with and without ST elevation, J. Cardiol. 66
(2015) 148–154.
[27] C.W. Siu, M.H. Jim, H.H. Ho, et al., Transient atrial fibrillation complicating acute in-
ferior myocardial infarction: implications for future risk of ischemic stroke, Chest
132 (2007) 44–49.
[28] G. Batra, B. Svennblad, C. Held, et al., All types of atrial fibrillation in the setting of
myocardial infarction are associated with impaired outcome, Heart 102 (2016)
926–933.
[29] R.H. Mehta, O.H. Dabbous, C.B. Granger, et al., Comparison of outcomes of patients
with acute coronary syndromes with and without atrial fibrillation, Am. J. Cardiol.
92 (2003) 1031–1036.
[30] B.J. Witt, K.V. Ballman, R.D. Brown Jr., R.A. Meverden, S.J. Jacobsen, V.L. Roger, The in-
cidence of stroke after myocardial infarction: a meta-analysis, Am. J. Med. 119
(2006) (354.e351-359).
[31] C.N. Bang, G.H. Gislason, A.M. Greve, et al., New-onset atrial fibrillation is associated
with cardiovascular events leading to death in a first time myocardial infarction
population of 89,703 patients with long-term follow-up: a nationwide study, J.
Am. Heart Assoc. 3 (2014), e000382. .
[32] A.B. Hill, The environment and disease: association or causation? Proc. R. Soc. Med.
58 (1965) 295–300.
[33] M. Brambatti, S.J. Connolly, M.R. Gold, et al., Temporal relationship between subclin-
ical atrial fibrillation and embolic events, Circulation 129 (2014) 2094–2099.
[34] T. Sanna, H.C. Diener, R.S. Passman, et al., Cryptogenic stroke and underlying atrial
fibrillation, N. Engl. J. Med. 370 (2014) 2478–2486.
[35] M.J. Tangelder, L. Frison, D. Weaver, et al., Effect of ximelagatran on ischemic events
and death in patients with atrial fibrillation after acute myocardial infarction in the
efficacy and safety of the oral direct thrombin inhibitor ximelagatran in patients
with recent myocardial damage (ESTEEM) trial, Am. Heart J. 155 (2008) 382–387.
[36] T.V. Glotzer, P.D. Ziegler, Silent atrial fibrillation as a stroke risk factor and
anticoagulation indication, Can. J. Cardiol. 29 (2013) S14–23.
[37] P. Kirchhof, S. Benussi, D. Kotecha, et al., ESC guidelines for the management of atrial
fibrillation developed in collaboration with EACTS, Eur. Heart J. 37 (2016)
2893–2962.
[38] R. Mahajan, T. Perera, A.D. Elliott, et al., Subclinical device-detected atrial fibrillation
and stroke risk: a systematic review and meta-analysis, Eur. Heart J. 39 (2018)
1407–1415.
7J. Luo et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Please cite this article as: J. Luo, et al., Increased risk of ischemic stroke associated with new-onset atrial fibrillation complicating acute coronary
syndrome: A systematic..., Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2018.04.096