原Early 著

or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

Early or Delayed Anticoagulation in Patients with Atrial

Fibrillation Upon Acute Ischemic Stroke: A Systematic
Review and Meta-Analysis
Yen-Ling Lo1, Hsin-Hsi Tsai1, 2, Li-Kai Tsai1, 3
1
Departments of Neurology, National Taiwan University Hospital, Taipei, Taiwan.
2
Departments of Neurology, National Taiwan University Hospital Bei-Hu Branch, Taipei, Taiwan.
3
Departments of Neurology, National Taiwan University Hospital Hsinchu Branch, Hsinchu, Taiwan.

ABSTRACT
Background: The optimal timing to initiate oral anticoagulation in acute ischemic stroke for patients
with atrial fibrillation (AF) remains undetermined. This meta-analysis aims to investigate the effectiveness
and safety between early and delayed use of oral anticoagulation after acute ischemic stroke in patients with
AF.
Method: Relevant literature was collected from the PubMed and Scopus electronic database platforms.
4 randomized controlled trials (RCTs) and 7 observational studies from 204 articles were included for meta-
analysis. We compared the incidence of recurrent ischemic stroke, intracranial hemorrhage and composite
outcomes between those receiving early or delayed anticoagulation.
Results: In RCTs and observational studies of 8,874 patients with AF upon acute ischemic stroke, early
initiation of NOACs or warfarin showed significant reduction in composite outcome using the random-
effects model (relative risk [RR]: 0.75; 95% confident interval [CI]: 0.58-0.96) as compared to delayed
administration of anticoagulants. In 10,743 patients, early anticoagulation led to significant reduction in
recurrent ischemic stroke (RR: 0.72; 95% CI: 0.51-0.91) than delayed anticoagulation. On the other hand, in
10,002 patients with AF, there was no significant difference in the risk of intracranial hemorrhage between
two groups (RR: 0.91; 95% CI: 0.67-1.22). To restrict the analysis on two large high-quality RCTs with
total 2,901 patients, early anticoagulation with NOACs showed a trend toward lower risk to develop events
of composite outcome than late administration of NOACs (RR: 0.75; 95% CI: 0.54-1.04).
Conclusion: In patients with AF-related acute ischemic stroke, our meta-analysis demonstrated that
early initiation of anticoagulation is superior than delayed administration in reducing composite outcome
and recurrent ischemic strokes without increasing on the risk of intracranial hemorrhage.

Keywords: Atrial fibrillation, early, ischemic stroke, intracranial hemorrhage, anticoagulants.

Introduction 0.51%, respectively. 1, 2 AF increases the risk of

stroke by five times, and approximately 20% of
The prevalence of atrial fibrillation (AF) all strokes are linked to AF. 3 Survivors of AF-
in Taiwan and all over the world are 1.1% and related stroke tend to experience more significant

Corresponding author: Hsin-Hsi Tsai, MD, PhD, Departments of Neurology and Stroke Center, National Taiwan University
Hospital, Taipei, Taiwan.
E-mail: [email protected]
DOI: 10.6318/FJS.202403_6(1).0004

72
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

disability and face an increased risk of recurrent incorporates the latest studies available.
strokes compared to patients with other stroke
causes.4, 5 For patients with AF, warfarin, a vitamin
Methods
K antagonist, has been found to reduce the risk of
stroke by about 60%. Non-vitamin K antagonist
oral anticoagulants (NOACs) selectively inhibit Search methodology
thrombin or activated factor X, presenting potential We searched PubMed and Scopus database
6, 7
benefits compared to warfarin. Therefore, between June 1985 to May 2023 using a
current guidelines from the Taiwan Stroke Society combination of search terms: (atrial fibrillation)
(TSS) and the American Heart Association/ AND ((early) OR (delay)) AND (anticoagulant)
American Stroke Association (AHA/ASA) suggest AND ((ischemic stroke) OR (ischaemic stroke))
to use NOACs or warfarin for secondary stroke AND ((intracranial hemorrhage) OR (intracerebral
prevention for AF patients with elevated risks of hemorrhage)).
8, 9
stroke. Studies were eligible for (1) published in
Anticoagulation therapy decreases the chance English; (2) were randomized controlled trials
of recurrent stroke and systemic embolism, but (RCTs) or observation studies that evaluated the
it also elevates the risk of cerebral hemorrhage timing of NOAC or warfarin initiation and their
especially during the acute phase following effects on composite outcomes (recurrent ischemic
an ischemic stroke in AF patients. 8 In clinical stroke, transient ischemic attack [TIA], systemic
practice, the timing for initiating anticoagulants embolism, major bleeding, intracranial hemorrhage
upon acute ischemic stroke in patients with AF [ICH], and/or vascular death) in patients with AF
is an important issue. However, the optimal and acute ischemic stroke. We excluded review
timing to start anticoagulants for AF patients articles and studies in which title and abstract are
with acute ischemic stroke is still unclear. The incompatible with keywords.
Guideline of the TSS suggests the “1-3-6-12”
rule for starting anticoagulant treatment primarily Data extraction and risk of bias
based on expert opinions. 9 The 2021 guideline assessment
from the AHA/ASA suggests that patients with Studies were screened focusing on titles and
large cerebral infarction might have a greater risk abstracts to recognize suitable articles for full-
of hemorrhagic transformation; the initiation of text assessment by two reviewers (author Y-L.L.
oral anticoagulation after stroke onset at 14-day and L-K.T.). We used a spreadsheet to collected
8
delay is only moderately recommendated. There the following information: authors; country
have been increasing observation studies and of origin; study design; median age; type and
randomized controlled trials addressing the optimal number of participants; initiation time of NOACs/
timing of initiating anticoagulation after stroke. warfarin; follow-up time, composite outcomes and
In the absence of clear and uniform guideline secondary outcome including recurrent ischemic
recommendations, we endeavored to address stroke and ICH. We collected data from RCTs
this clinically unsettled issue by undertaking a and observational studies in accordance with the
systematic literature review and meta-analysis that Preferred Reporting Items for Systematic Reviews

73
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

and Meta-Analysis (PRISMA) statement. follow-up period using the random-effects model.
The Cochrane risk of bias tool for RCT To assess the overall heterogeneity, I 2 test was
studies was evaluated using the danger of bias tool utilized with I2 statistic. I2 25%, 50%, and 75%
that encompasses six domains of bias: selection indicates low, moderate and high heterogeneity,
bias, performance bias, detection bias, attrition respectively.10 P-value < 0.05 was considered as
bias, reporting bias, and publication bias. The bias statically significant.
was appraised by funnel plots of study results
plotted against participant size using Review
Results
Manager5.

The process of data synthesis and Included studies

statistical analysis A search conducted on the PubMed and
Data were examined by R version 4.3.1 Scopus databases resulted in a collection of 204
software with meta package. We pooled risk studies. A PRISMA flow diagram is displayed
ratio (RR) and 95% confidence intervals (CI) of in Figure 1. Consequently, four RCTs and seven
composite and secondary outcomes over 30-day observational studies were included for further

Figure 1. PRISMA 2020 flow diagram for systematic reviews.

74
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

meta-analysis.11-21 experienced acute ischemic stroke in eligible 7

The summary of four RCT studies is presented of 11 studies. In the random effects model, early
in Table 1. In a total of 3,172 participants (female, anticoagulation showed significantly lower risk of
11-14
40.9% to 63.4%; mean age, 70 to 78 years), composite outcome than delayed anticoagulation
the definition of early and delayed administration (RR: 0.75; 95% CI: 0.58-0.96) (Figure 2). The
of anticoagulants was different across RCTs. Two heterogeneity was low with heterogeneity I2 of
of them sub-grouped the administration time 33% (P = 0.18).
12,
according to clinical stroke severity or infarct size When focusing on four RCTs with 3,172
14
. In two studies comparing the outcomes between participants, early administration of anticoagulants
anticoagulation using NOACs and warfarin, we showed an insignificant trend toward lower risk
considered the NOAC arm to be early treatment than delayed administration of composite outcome
group and the warfarin arm to be delayed treatment (RR:0.83; 95% CI: 0.67-1.01) (Figure 2). The
group since the peak plasma levels are achieved heterogeneity was low with heterogeneity I 2 of
within 4 hours after NOAC while after 5 days after 0% (P = 0.77). We also conducted an analysis on
warfarin administration. In quality assessment, the only two large high-quality RCTs assessing early
category of “blinding participants and personnel” vs. delayed initiation of NOAC.13,14 The early
showed a high risk of bias in which two studies treatment group (n = 1,456) showed non-significant
were designed to be open-labeled and the other two decreases in the composite outcome (RR: 0.75;
were single masking. Other categories presented 95% CI: 0.54-1.04) than delayed treatment group (n
as low bias risk including random sequence = 1,445) and there was low heterogeneity with I2
generation, allocation concealment, blinding of of 0% (P = 0.73).
outcome assessment, incomplete outcome data, Three of the 7 observational studies were
and selective reporting. eligible for analyses of composite outcome.
The characteristics of seven observational Patients receiving early anticoagulants treatment
studies were presented in Table 2.15-21 In a total of (n = 2,505) showed significantly lower risk of
7,756 participants (female, 30.6% to 62%; mean composite outcome than those receiving delayed
age, 64.5 to 80 years), the primary outcomes were treatment (n = 3,197) (RR: 0.63; 95% CI: 0.32-
composite events including recurrent ischemic 1.24) (Figure 2). Meaningful heterogeneities were
stroke, TIA, systemic embolism, major bleeding, observed across these observational studies (I2 =
ICH, and/or cerebrovascular death. These studies 70%, P = 0.03).
also had varying definitions for the early and
delayed administration of anticoagulants. Risk of recurrent ischemic
stroke between early vs. delayed
Composite outcome comparisons initiation of OAC
between early vs. delayed Ten of the 11 studies (n = 10,744) were
initiation of OAC eligible to compare the risks for recurrent ischemic
The result of early treatment group (n = 4,097) stroke between early (n = 5,292) and delayed
was compared to that from delayed treatment group (n = 5,452) treatment groups (Figure 3). The
(n = 4,777) in 8,874 participants with AF who had heterogeneity was low with I2 of 15% (P = 0.31).

75
76
Table 1. Summary of four randomized controlled trials
Author Country Trial Study Type of Sample size(% Initiation time of Primary Conclusion
(year) allocation participants female) and medication(Stroke severity) composite
ratio and mean age outcome (time,
masking poststroke)
Hong et South Triple 1:1, open Mild AIS 95 (42.1) vs. Early: rivaroxaban ICH/ recurrent No differences in the
al., 2017 Korea AXEL label within 5 days 88 (40.9); < 5 days after AISD IS (30 days) primary composite
with AF 70.4 years elayed: warfarin, outcome (49.5% vs
< 5 days after AIS 54.5%; P = 0.49).
Labovitz et USA AREST 1:1, open AIS within 3–5 41 (63.4) vs. Early: apixaban Recurrent IS, Similar outcome
al., 2021 label days or TIA 47 (48.9); Day 0-3 after TIA TIA, and fatal between two groups
within 3 days 73.5 years Day 3-5 after small-sized stroke (180 with the risk of
with AF AIS days) primary composite
Day 7-9 after medium- outcome of 17.1% and
sized AIS 25.5% (P=0.44).
Delayed: warfarin
> 7 days
Oldgren et Sweden TIMING 1:1, single AIS within 72 450 (46) vs. Early: NOAC, ≦4 days Recurrent IS, Early initiation was
al., 2022 (outcomes hours with AF 438 (46.3); Delayed: NOAC, 5-10 days symptomatic noninferior to delayed
assessor) 78.3 years intracerebral start of NOAC with
hemorrhage, nonsignificant lower
or all-cause rates of composite
mortality (90 outcome (OR:
days) 0.78 [0.48-1.25]);
no symptomatic
intracerebral
hemorrhage in each
group.
Fischer et Switzerland ELAN 1:1, single AIS lasting 1006 (45.6) vs. Early: NOAC Major A trend toward lower
al., 2023 (outcomes more than 24 1007 (45.3); < 2 days after minor or bleeding, risk of composite
assessor) hours with AF 77 years moderate AIS recurrent outcome in early than
Day 6 or 7 after major AIS IS, systemic later anticoagulation
Delayed: NOAC embolism and/ groups (OR: 0.7
Day 3 or 4 after minor AIS or vascular [0.44-1.14]); only 2
Day 6 or 7 after moderate death (30 days) symptomatic ICH
AIS occurred in each group
Day 12-14 after major AIS (0.2%)
Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

AF indicates atrial fibrillation; AIS, acute ischemic stroke; ICH, intracranial hemorrhage; IS, ischemic stroke; NOAC, non-vitamin K antagonist oral anticoagulant;
OR, odds ratio; TIA, transient ischemic attack.
 Table 2. Summary of seven observational studies

Author Country Study design Type of Sample size Initiation time of medication Outcome (time, post- Conclusion
(year) participants (% female; (Stroke severity) stroke)
mean age)
Cappellari Italy Prospective Stroke 97 (61.8) Early: NOACs 1-3 days ICH and recurrent IS (7 No link between starting
et al., 2016 observational with non- 50 (62); Delayed: NOACs 4-7 days days) NOACs early and ICH
valvular AF 79 years
Wilson et UK Prospective AF with 358 (43) Early: NOACs ≤4 days Recurrent IS, TIA, ICH, Composite outcome 2% vs
al., 2019 observational acute 997 (41); Delayed: NOACs ≥ 5 days or death due to any cause 5% (no difference)
ischemic 76 years (90 days)
stroke/TIA
Yaghi et USA Retrospective AF with 862 (50.3) Early: NOACs median 5 Recurrent ischemic event Early NOAC treatment
al., 2020 observational acute 389 (49.4); days and symptomatic ICH associated with less infarct
ischemic 76-78 years Delayed: warfarin median 2 (90 days) and similar ICH risk
stroke days
D'Anna et Germany Retrospective AF with 300 (46) Early: NOACs 7 ± 9.4 days Major (intracranial and Using NOACs within 7 days
al., 2020 observational acute 259 (44.4); Delayed: warfarin 11.9 extracranial) bleeding, are safe in both extracranial
ischemic 80 years ±19.7 days Recurrent IS, TIA and and intracranial bleeding
stroke mortality (16 days) complications.
Bakr et al., Saudi Arabia Retrospective AF with 55 (50) Early: NOACs or warfarin Recurrent IS and ICH (90 Higher risk of intracerebral
2020 observational acute 42 (30.6); 1~6 days to 180 days) hemorrhage in early
ischemic 64.5 years Delayed: NOACs or treatment group
stroke/TIA warfarin 7~14 days
Kimura et Japan Prospective AF with 785 (40) Early: NOACs,≦1 (TIA), Recurrent IS, systemic Early initiation of
al., 2022 observational acute 1012 (41); ≦2 (minor), ≦3 (mod), embolism, severe NOACs reduces the risk
ischemic 77 years ≦4 (severe) bleeding, and death (90 of recurrence stroke or
stroke/TIA Delayed: NOACs, >1, >2, days) systemic embolism without
>3, >4 days increasing major bleeding.
Marchis et Switzerland Prospective AF with 1362 (46.8) Early: NOACs ≦5 days Recurrent IS, ICH and Composite outcome 1.8%
al., 2022 observational acute 1188 (47.7); Delayed: NOACs > 5 days any-cause of mortality vs 1.5% (no difference)
ischemic 77 years (30 days)
stroke
Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

AF indicates atrial fibrillation; AIS, acute ischemic stroke; ICH, intracranial hemorrhage; IS, ischemic stroke; NOAC, non-vitamin K antagonist oral anticoagulant;

77
TIA, transient ischemic attack.
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

Figure 2. Forest plot of risk in primary composite outcome.

Figure 3. Forest plot of risk in recurrent ischemic stroke.

Forest plot demonstrated significantly lower risk 3), the forest plot demonstrated significantly lower
of recurrent ischemic stroke for early than delayed risk of recurrent ischemic stroke for early initiation
initiation of anticoagulants (RR: 0.72; 95% CI: of anticoagulants than delayed administration
0.57-0.91). (RR: 0.73; 95% CI: 0.55-0.99). The heterogeneity
In 4 RCT studies with 3,134 patients (Figure was low with I2 of 0% (P = 0.79). When focusing

78
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

on the two large high-quality RCTs,13, 14 the early of anticoagulants (RR: 0.91; 95% CI: 0.67-1.22).
treatment group (n = 1,434) showed significant The heterogeneity was low (I2 = 0%, P = 0.91).
decreases in the risk of recurrent ischemic stroke In our respective subgroup analyses within
(RR: 0.62; 95% CI: 0.39-0.98) than delayed 3 RCT studies with 2,246 patients or within 7
treatment group (n = 1,429) and there was low observational studies with 7,756 patients, there
2
heterogeneity with I of 0% (P = 0.69). were no significant differences in ICH risk between
In 6 eligible observational studies with 7,609 early and delayed administration of anticoagulants
patients, the analysis showed non-significant (RR: 1.08; 95% CI: 0.71-1.67 for RCTs; RR: 0.77;
difference between two groups (RR: 0.78; 95% CI: 95% CI: 0.51-1.16 for observational studies). The
0.51-1.20) with moderate heterogeneity between heterogeneity was low between studies.
studies (I2 =47%; P = 0.09) (Figure 3).

Discussion
Risk of intracranial hemorrhage
between early vs. delayed This meta-analysis aims to assess whether
initiation of OAC the early administration of NOACs or warfarin
The comparison between early vs. delayed in patients with AF during acute ischemic stroke
treatment of OAC for risk of ICH was shown yields superior outcomes compared to the delayed
in Figure 4. The forest plot in 3 RCTs and 7 initiation of anticoagulants. Four RCTs and seven
observational studies (n = 10,002) revealed no observational studies were eligible for analyses,
significant difference in risk of ICH between early and we demonstrated that the early treatment group
(n = 4,939) and delayed (n = 5,063) administration showed lower risk of composite outcome and less

Figure 4. Forest plot of risk in intracranial hemorrhage.

79
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

recurrent ischemic stroke compared to the delayed Until now, there are two large-scale high-
treatment group without increasing the risk of quality RCTs investigating early vs. delayed use
ICH. The results showed similar trend in subgroup of NOAC after AF-related stroke. The TIMING
analyses within RCTs or within observational (Timing of oral anticoagulant therapy in acute
studies, respectively. Overall, early administration ischemic stroke with atrial fibrillation) clinical trial
of oral anticoagulants in AF patients upon acute is a noninferiority open-labeled RCT with subject
ischemic stroke should be preferred over delayed number of 450 in early anticoagulation group (≤4
initiation of anticoagulation in general. days) and 438 in delayed anticoagulation group
Early anticoagulation therapy decreases the (5-10 days). The study demonstrated that early
risk of recurrent embolism, but it also elevates the initiation of anticoagulants was noninferior to
risk of hemorrhagic transformation. Patients with delayed start of NOAC after acute ischemic stroke
acute ischemic stroke associated with AF were with nonsignificant lower rates of ischemic stroke
not included in pivotal RCTs comparing NOACs and death. Notably, there was no ICH event in both
and warfarin on effectiveness and safety, likely groups.13 The ELAN (Early versus Late initiation
due to concerns regarding the potential risk of of DOAC in post-ischemic stroke patients with
3
hemorrhagic transformation. Further observational Atrial fibrillation) trial is a randomized controlled
studies discussing about timing of anticoagulation open-label study with subject number of 1006 in
did not demonstrate consistent results. For early anticoagulation group (≤2 days for minor and
example, initiation of NOACs within 2 days after moderate stroke and 6-7 days for major stroke)
acute ischemic stroke is associated with relative and 1007 in delayed anticoagulation group (3-4
high rate (5%) of hemorrhagic transformation. 22 days for minor stroke; 6-7 days for moderate
On the other hand, the SAMURAI-NVAF study stroke; 12-14 days for major stroke). The study
showed that after initiation of NOACs, no ICH was reported a lower rate of composite poor outcome
23
observed within 4 days after stroke. In addition, in early than later anticoagulation groups (1.18%
the recommendation of various guidelines was difference) with only 2 symptomatic ICH occurred
primarily based on expert consensus to initiate in each group (0.2%). 14 Our meta-analysis
anticoagulation after acute ischemic stroke. The focusing on these two RCTs demonstrated that
Canadian Heart and Stroke Foundation, European early anticoagulation with NOACs showed a trend
Heart Rhythm Association, and TSS suggest to toward lower risk to develop events of composite
follow the ‘1-3-6-12’ rule, according to different outcome than late administration of NOACs, and
9, 24, 25
stroke severity. The AHA/ASA recommended therefore, supported a future direction of early
to administer NOACs to patients with AF who initiation of anticoagulation for these patients. Two
have experienced an acute ischemic stroke between additional RCTs are still undergoing including
8
4 and 14 days after the event. The results of OPTIMAS (Optimal Timing of Anticoagulation
our meta-analysis support that early initiation of after Ischemic Stroke) and START (Optimal delay
anticoagulation prevents more composite outcome time to initiate anticoagulation after ischemic
and recurrent ischemic stroke without increasing stroke in atrial fibrillation) trials. The final results
risk of ICH, although the exact timing and strategy of these two important clinical trials are highly
differed significantly cross studies. anticipated.

80
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

Our study has several limitations: (1) The References

variability in study designs and the timing of
anticoagulation across different studies poses a 1. Lippi G, Sanchis-Gomar F, Cervellin G.
challenge to the direct clinical application of the Global epidemiology of atrial fibrillation:
concept advocating early anticoagulant use; (2) An increasing epidemic and public health
In both RCTs and observational studies, certain challenge. Int J Stroke 2021;16:217-221.
investigations compared outcomes between 2. Chiang CE, Wu TJ, Ueng KC, et al. 2016
NOACs and warfarin. It is crucial to recognize Guidelines of the Taiwan Heart Rhythm
that the results may be influenced by the choice Society and the Taiwan Society of Cardiology
of anticoagulant type rather than being solely for the management of atrial fibrillation. J
attributable to the timing of anticoagulant Formos Med Assoc 2016;115:893-952.
administration; (3) All of the 4 RCTs were not 3. Seiffge DJ, Werring DJ, Paciaroni M, et
double blinded with two of them being open- al. Timing of anticoagulation after recent
label trials and two of them being single masking ischaemic stroke in patients with atrial
studies, leading to a potential risk of biased fibrillation. Lancet Neurol 2019;18:117-126.
reporting for clinical events; (4) There is racial 4. Camm AJ, Kirchhof P, Lip GYH, et al.
deviation for the RCTs that the trial population Guidelines for the management of atrial
consisted primarily of participants from European fibrillation: the task force for the management
centers, where there was a high proportion of of atrial fibrillation of the European Society of
White individuals; (5) The majority of RCTs and Cardiology (ESC). Eur Heart J 2010;31:2369-
observational studies focused on patients with 2429.
mild to moderate stroke severity, a group generally 5. Wolf PA, Abbott RD, Kannel WB. Atrial
perceived to carry a lower risk of ICH when early fibrillation as an independent risk factor for
oral anticoagulation is initiated. Therefore, the stroke: the Framingham Study. Stroke 1991;22:
findings may not apply to individuals with severe 983-988.
stroke or with large infarct size. Future large RCTs 6. Atrial Fibrillation Investigators. Risk factors
are needed with a special target in this subgroup. for stroke and efficacy of antithrombotic
In conclusion, for patients with AF- therapy in atrial fibrillation: analysis of pooled
related acute ischemic stroke, our meta-analysis data from five randomized controlled trials.
demonstrated that early initiation of anticoagulation Arch Intern Med 1994;154:1449-1457.
is superior than delayed administration for reducing 7. Ruff CT, Giugliano RP, Braunwald E, et al.
composite outcome and recurrent ischemic stroke Comparison of the efficacy and safety of new
without posing additional risk on ICH. Further oral anticoagulants with warfarin in patients
large RCTs are still mandatory not only to confirm with atrial fibrillation: a meta-analysis of
the benefits of early anticoagulation in these randomised trials. Lancet 2014;383:955-962.
patients, but to delineate the exact timing strategy 8. Powers WJ, Rabinstein AA, Ackerson T, et
of drug initiation with or without stratifying stroke al. Guidelines for the early management of
severity for practical clinical application. patients with acute ischemic stroke: 2019
update to the 2018 Guidelines for the early

81
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

management of acute ischemic stroke: a versus late anticoagulation for ischaemic stroke
Guideline for healthcare professionals from the associated with atrial fibrillation: multicentre
American Heart Association/American Stroke cohort study. J Neurol Neurosurg Psychiatry
Association. Stroke 2019;50:e344-e418. 2019;90:320-325.
9. Lin YJ, Huang YT, Tsui PC, et al. 2019 17. Bakr AIA, AlOmar RS, Nada MAF, et al.
Taiwan Stroke Society Guideline on the use of Timing to start anticoagulants after acute
non-vitamin K antagonist oral anticoagulants ischemic stroke with non-valvular atrial
for prevention of stroke in patients with atrial fibrillation. J Neurol Sci 2020;409:116582.
fibrillation. Formosan Journal of Stroke 2019; 18. D'Anna L, Filippidis FT, Antony S, et al. Early
1:143-184. initiation of direct anticoagulation after stroke
10. Higgins JPT, Thompson SG, Deeks JJ, et al. in patients with atrial fibrillation. Eur J Neurol
Measuring inconsistency in meta-analyses. 2020;27:2168-2175.
BMJ 2003;327:557-560. 19. Yaghi S, Mistry E, Liberman, AL, et al.
11. H o n g K S , K w o n S U , L e e S H , e t a l . Anticoagulation type and early recurrence in
Rivaroxaban vs warfarin sodium in the ultra- cardioembolic stroke: the IAC study. Stroke
early period after atrial fibrillation-related mild 2020;51:2724-2732.
ischemic stroke: A randomized clinical trial. 20. Marchis GMD, Seiffge DJ, Schaedelin S,
JAMA Neurology 2017;74:1206-1215. et al. Early versus late start of direct oral
12. Labovitz AJ, Rose DZ, Fradley MG, et al. anticoagulants after acute ischaemic stroke
Early apixaban use following stroke in patients linked to atrial fibrillation: an observational
with atrial fibrillation: results of the AREST study and individual patient data pooled
trial. Stroke 2021;52:1164-1171. analysis. J Neurol Neurosurg Psychiatry 2022;
13. Oldgren J, Åsberg S, Hijazi Z, et al. Early 93:119-125.
versus delayed non-vitamin K antagonist oral 21. Kimura S, Toyoda K, Yoshimura S, et al.
anticoagulant therapy after acute ischemic Practical “1-2-3-4-Day” rule for starting direct
stroke in atrial fibrillation (TIMING): oral anticoagulants after ischemic stroke with
a registry-based randomized controlled atrial fibrillation: combined hospital-based
noninferiority study. Circulation 2022;146: cohort study. Stroke 2022;53:1540-1549.
1056-1066. 22. Smythe MA, Parker D, Garwood CL, et al.
14. Fischer U, Koga M, Strbian D, et al. Early Timing of initiation of oral anticoagulation
versus later anticoagulation for stroke with after acute ischemic stroke in patients with
atrial fibrillation. N Engl J Med, 2023;388: atrial fibrillation. Pharmacotherapy 2020;40:
2411-2421. 55-71.
15. Cappellar M, Carletti M, Danese A, et al. 23. Toyoda K, Arihiro S, Todoet K, et al. Trends
Early introduction of direct oral anticoagulants in oral anticoagulant choice for acute stroke
in cardioembolic stroke patients with patients with nonvalvular atrial fibrillation
non-valvular atrial fibrillation. J Thromb in Japan: the SAMURAI-NVAF study. Int J
Thrombolysis 2016;42:393-398. Stroke 2015;10:836-842.
16. Wilson D, Ambler G, Banerjee G, et al. Early 24. Wein T, Lindsay MP, Côté R, et al. Canadian

82
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

stroke best practice recommendations: 2018 European Heart Rhythm Association

secondary prevention of stroke, sixth edition Practical Guide on the use of non-vitamin
practice guidelines, update 2017. Int J Stroke K antagonist oral anticoagulants in patients
2018;13:420-443. with atrial fibrillation: executive summary.
25. Steffel J, Verhamme P, Potpara TS, et al. The Europace 2018;20:1231-1242.

83
 Early or Delayed Anticoagulation in Patients with Atrial Fibrillation Upon Acute Ischemic Stroke: A Systematic Review and Meta-Analysis

心房顫動患者急性缺血性中風後的早期或延遲抗凝
治療：系統性回顧與統合分析

羅彥伶 1、 蔡欣熹 1, 2、蔡力凱 1, 3

1
台大醫院 神經部暨腦中風中心
2
台大醫院北護分院 神經科
3
新竹台大分院 神經部

摘 要
背景：針對心房顫動患者，何時在急性缺血性中風後開始使用warfarin或新型口服抗凝血劑
(NOAC)，至今最佳策略仍然未知。過去臨床試驗或觀察性研究探討了早期使用NOAC的利弊，但
這些研究的方法學和結論不盡相同，需要統合分析以提供整體的療效比較。
研究方法：由Pubmed及Scopus電子資料庫平台收尋相關醫學文獻，共可得到204篇文章。依照
設定的納入及排除條件後，共獲得4篇隨機的臨床試驗與7篇觀察性研究進行統合分析。我們嘗試比
較早期或延遲抗凝治療(warfarin或NOAC)之預後，包括複合性預後指標、再次復發的缺血性中風或
顱內出血的相關風險。
研究結果：於臨床試驗與觀察性研究之8,874名心房顫動患者，隨機效應模型的統合分析顯
示，早期給藥較晚期給藥顯著減少複合式預後指標(相對風險：0.75；95%信賴區間：0.58-0.96)。而
針對次級指標，於10,743名心房顫動患者，早期給藥較晚期給藥顯著減少復發性缺血性腦中風 (相
對風險：0.72；95%信賴區間：0.51-0.91)；而於10,002名心房顫動患者，早期給藥和晚期給藥之顱
內出血風險相似(相對風險：0.91；95%信賴區間：0.67-1.22)。若統合分析僅侷限於2個大型高品質
之NOAC隨機對照試驗，於2,901名心房顫動患者中，早期給藥較晚期給藥呈現複合式預後指標減少
的趨勢，但統計並不顯著(相對風險：0.75；95%信賴區間：0.54-1.04)。
研究結論：在心房顫動的急性缺血性腦中風患者，於中風後早期給予口服抗凝血劑比延後給予
藥物治療，可能可以減少整體的不良預後。其中早期給予抗凝血藥物似乎可降低復發性缺血性中風
的發生率，且無增加中風後出血發生的機會。

關鍵詞：心房顫動，早期使用，缺血性中風，顱內出血，抗凝血劑

通訊作者：蔡欣熹醫師，台大醫院神經部暨腦中風中心
E-mail: [email protected]

84