Aspirin resistance is associated with

increased stroke severity and infarct

volume
Mi Sun Oh, MD
Kyung-Ho Yu, MD, PhD
Ju-Hun Lee, MD, PhD
San Jung, MD
Chulho Kim, MD
Min Uk Jang, MD
Juneyoung Lee, PhD
Byung-Chul Lee, MD,
PhD

ABSTRACT

Objectives: To investigate whether aspirin resistance is associated with initial stroke severity and
infarct volume, using diffusion-weighted imaging (DWI) in patients with acute ischemic stroke that
occurred while taking aspirin.

Methods: We studied a total of 310 patients who were admitted within 48 hours of acute ischemic
stroke onset. All patients had been taking aspirin for at least 7 days before stroke onset. Aspirin
resistance, defined as high residual platelet reactivity (HRPR) on aspirin treatment, was measured
using the VerifyNow assay and defined as an aspirin reaction unit $550. Initial stroke severity
was assessed using the NIH Stroke Scale (NIHSS) score. Infarct volume was measured using DWI.
Results: HRPR occurred in 86 patients (27.7%). The initial NIHSS score (median [interquartile

Correspondence to
Dr. B.-C. Lee:
ssbrain@hallym.ac.kr
or Dr. Oh:
iyyar@hallym.ac.kr

range]) was higher in patients with HRPR than in the non-HRPR group (6 [315] vs 3 [18], p ,
0.001). DWI infarct volumes were also larger in the HRPR group compared to the non-HRPR
group (5.4 [0.843.2] vs 1.7 [0.410.3], p 5 0.002). A multivariable median regression analysis
showed that HRPR was significantly associated with an increase of 2.1 points on the NIHSS
(95% confidence interval 0.84.0, p , 0.001) and an increase of 2.3 cm3 in DWI infarct volume
(95% confidence interval 0.43.9, p , 0.001).

Conclusions: Aspirin resistance is associated with an increased risk of severe stroke and large
infarct volume in patients taking aspirin before stroke onset. Neurology 2016;86:18081817
GLOSSARY
ARU 5 aspirin reaction unit; ASPECTS 5 Alberta Stroke Program Early CT Score; CE 5 cardioembolic; CI 5 confidence
interval; DWI 5 diffusion-weighted imaging; HRPR 5 high residual platelet reactivity; IQR 5 interquartile range; LAA 5 large
artery atherosclerosis; NIHSS 5 NIH Stroke Scale; OE 5 other determined etiology; SVO 5 small vessel occlusion; TOAST 5
Trial of Org 10172 in Acute Stroke Treatment; UE 5 undetermined etiology.

Supplemental data
at Neurology.org

Aspirin is the most frequently used antiplatelet agent for primary and secondary prevention of
ischemic stroke or TIA.1,2 Numerous studies have suggested that prestroke use of antiplatelet
agents reduces initial stroke severity and infarct volume,37 by inhibiting platelet aggregation,
thrombus formation, and thrombus propagation.8,9 Aspirin also has actions including neuroprotective and anti-inflammatory effects,8,10 which reduce stroke severity by limiting the infarct
size. However, some patients with aspirin resistance, defined as high residual platelet reactivity
(HRPR) on aspirin treatment, exhibit an increased risk of recurrent stroke and poor clinical
outcomes.1114 Some studies have reported that patients with HRPR have more severe
strokes,13,1517 whereas other studies have not found a relationship between HRPR and stroke
severity.18,19 These disparate results of previous studies may be attributable to insufficient number of study participants, methodologic differences, diverse outcome measurement time points,
heterogeneous stroke subtypes, or clinically imprecise neurologic scales.13,15,18 Infarct volume is
strongly correlated with clinical stroke severity and is therefore an important surrogate of stroke
outcome.20 The relationship between HRPR and actual volumetric measurements of infarct size
by diffusion-weighted imaging (DWI) has not been assessed previously. In the present study, we
investigated the effects of HRPR on clinical stroke severity and infarct volume using DWI in
patients who were taking aspirin before stroke onset.
From the Department of Neurology (M.S.O., K.-H.Y., J.-H.L., S.J., C.K., M.U.J., B.-C.L.), Hallym University College of Medicine, Hallym
Neurological Institute; and Department of Biostatistics (J.L.), Korea University College of Medicine, South Korea.
Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

1808

2016 American Academy of Neurology

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

METHODS Patients. We screened patients with acute ischemic stroke or TIA who were admitted to Hallym University
Sacred Heart Hospital between October 2008 and May 2014. Patients were eligible for the study if they were admitted within 48
hours of stroke onset, exhibited evidence of acute ischemic stroke
on DWI, had taken aspirin for at least 7 days before the index
stroke, and were classified as having HRPR within 24 hours of
admission. We excluded patients who were taking an additional
antiplatelet or anticoagulant before admission or who had a prestroke modified Rankin Scale score .2 (figure e-1 on the
Neurology Web site at Neurology.org).

Standard protocol approvals, registrations, and patient

consents. The Hallym University Sacred Heart Hospital institutional review board approved the study design and protocol.
Informed consent was obtained from each patient or their authorized relative.

Data collection. Demographic and clinical data were prospectively collected on admission, including information on patients
demographics, hospitalizations, prior treatments, IV tissue plasminogen activator, and health conditions (table 1). Stroke subtypes were classified as large artery atherosclerosis (LAA), small
vessel occlusion (SVO), cardioembolic (CE), undetermined etiology (UE), or other determined etiology (OE), according to the
Trial of Org 10172 in Acute Stroke Treatment (TOAST) classifications.21 Laboratory data were collected, including values for
white blood cells, hemoglobin, high-sensitivity C-reactive
protein, fasting blood glucose, total cholesterol, and low-density
lipoprotein cholesterol. Prior medications were determined by
interviews with patients or their relatives, or preadmission
prescriptions. In addition, patients or their relatives were
questioned regarding aspirin compliance.
Platelet reactivity assessment. Blood samples to test for platelet reactivity were obtained within 12 hours of admission and
before replacing or adding another antiplatelet or anticoagulant
agent. Residual platelet reactivity on aspirin treatment was determined using the VerifyNow Aspirin Assay (Accumetrics, San
Diego, CA), which is a rapid point-of-care system using
turbidimetric optical detection of platelet aggregation to
fibrinogen and arachidonic acid in whole blood. The assay
measures the change in light transmission produced by agonistinduced platelet aggregation, yielding a value in aspirin reaction
units (ARUs). HRPR was defined as an ARU $550.22 Patients
were categorized into either the HRPR or non-HRPR group.
Outcomes. Initial stroke severity was assessed by the NIH
Stroke Scale (NIHSS) score at the time of admission.23 All patients underwent DWI (3.0-tesla MRI) within 24 hours of
hospitalization. The imaging parameters for DWI were as
follows: repetition time of 3,000 milliseconds, echo time of 56
milliseconds, matrix number 120 3 124, voxel size 1.9 3
1.85 mm3, interslice gap 2 mm, and 2 b values of 0 and 1,000
s/mm2. Infarct volumes indicated by DWI were measured with
MIPAV software (Medical Image Processing, Analysis, and
Visualization, version 3.0; NIH, Bethesda, MD).24 Acute
diffusion lesions were identified on a slice-by-slice basis using a
semiautomatic segmentation method, consulting apparent
diffusion coefficients to distinguish acute from nonacute
diffusion signals. DWI infarct volumes were calculated by
multiplying slice thickness by total areas of lesions. The images
were independently analyzed by 2 blinded and experienced stroke
neurologists (M.U.J. and C.K.).
Statistical analyses. To compare baseline characteristics
between the HRPR and non-HRPR groups, the Student t test

or MannWhitney U test for continuous variables, and Pearson

x2 test, Fisher exact test, or x2 trend test for categorical variables
were used, as appropriate. Associations between raw ARU and
initial NIHSS scores or DWI infarct volumes were examined
using Spearman rank correlation analysis. Interrater reliability
for assessing DWI infarct volume was calculated using the
interclass correlation coefficient, with .0.80 as a threshold for
satisfactory agreement.
A quantile regression analysis was used to evaluate the effects
of HRPR on the initial NIHSS scores and DWI infarct volumes.
Quantile regression model allows for analyses of outcomes that
are nonnormally distributed and that have nonlinear relationships
with predictor variables. In addition, the quantile regression
model has advantages over traditional regression models by allowing an estimation of the effects of predictor variables at different
quantiles of conditional distribution for each of the outcome variables, rather than at a single measure of the central tendency of its
distribution, such as mean value.25,26 First, the effects of HRPR
on median NIHSS scores and DWI infarct volumes were tested
using bivariate and multivariable median regression analyses.
After that, we performed bivariate and multivariable quantile
regression analyses to determine whether the effects of HRPR
differed across quantiles of the conditional distributions for initial
NIHSS scores and DWI infarct volumes. The quantiles examined
were 10th, 25th, 50th, 75th, and 90th quantiles, which we denoted as Q-10, Q-25, Q-50, Q-75, and Q-90, respectively. Age,
sex, and variables with p , 0.05 in bivariate analyses were
included in multivariable models. Further multivariable median
regression (Q-50) was performed to investigate whether the effects of HRPR on initial NIHSS scores and DWI infarct volumes
were different according to subtypes of ischemic stroke based on
the TOAST classification. For these stratified analyses, we combined UE and OE subtypes because of insufficient number of
patients with OE. All p values were 2-tailed and those less than
0.05 were considered statistically significant. Statistical analyses
were performed using SPSS version 22.0 and the quantreg package27 of the R version 3.0.1 (the R Foundation for Statistical
Computing).
RESULTS The study included 310 patients (mean
age 71.5 6 10.8 years, 43.5% women), and 309
patients were taking 100 mg/d of aspirin before the
index ischemic stroke. Only one patient was taking
200 mg/d of aspirin before the stroke. The median
NIHSS score at presentation was 4 points (interquartile range [IQR] 29) and the median DWI infarct
volume was 2.3 cm3 (IQR 0.516.3). There was good
interrater reliability in the assessment of DWI infarct
volumes (interclass correlation coefficient 5 0.99).
Raw ARUs ranged from 217 to 668, as measured
by the VerifyNow assay (median 486.5, IQR
423.8556.3). Spearman rank correlation analysis
showed that the raw ARUs were significantly positively correlated with both initial NIHSS scores
(r 5 0.23; p , 0.001) and DWI infarct volumes
(r 5 0.15; p 5 0.01) (figure 1). Figure 1 also indicates
the associations between raw ARUs and both initial
NIHSS scores and DWI infarct volumes at different
percentiles.
Of 310 patients, 86 (27.7%) had HRPR. Comparisons of demographic, clinical, and laboratory
Neurology 86

May 10, 2016

1809

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Table 1

Clinical characteristics of the study participants

Total patients (N 5 310)

Characteristic

HRPR group
(n 5 86, 27.7%)

Non-HRPR group
(n 5 224, 72.3%)

Age, y

72.2 (10.6)

71.2 (10.9)

0.43

Sex, female

36 (41.9)

99 (44.2)

0.71

p Value

Vascular risk factors

History of TIA or stroke

35 (40.7)

82 (36.6)

0.51

History of CAD

7 (8.1)

37 (16.5)

0.06

Hypertension

77 (89.5)

206 (92.0)

0.49

Diabetes mellitus

33 (38.4)

92 (41.1)

0.66

Dyslipidemia

36 (41.9)

100 (44.6)

0.66

Atrial fibrillation

25 (29.1)

57 (25.4)

0.52

Current smoking

18 (20.9)

42 (18.8)

0.66

Antihypertensive drug

60 (69.8)

192 (85.7)

Statin

24 (27.9)

73 (32.6)

NSAID

8 (9.3)

22 (9.8)

Prestroke medication

Prestroke mRS score

,0.001
0.43
0.89
0.39

56 (65.1)

153 (68.3)

16 (18.6)

45 (20.1)

14 (16.3)

26 (11.6)

TOAST classification

0.91

Large artery atherosclerosis

29 (33.7)

71 (31.7)

Small vessel occlusion

19 (22.1)

58 (25.9)

Cardioembolism

22 (25.6)

58 (25.9)

Other determined

2 (2.3)

1 (0.4)

Undetermined

14 (16.3)

36 (16.1)

10 (11.6)

13 (5.8)

Administration of IV tPA

0.80

Median onset to arrival time, h

5.8 (1.319.2)

12.8 (3.425.8)

0.02

Median onset to ARU, h

15.3 (5.730.0)

19.4 (8.435.2)

0.14

Median onset to DWI, h

20.6 (11.534.7)

23.4 (13.740.1)

0.18

Mean admission NIHSS score

9.1 (8.1)

5.3 (5.5)

,0.001

Median admission NIHSS score

6 (315)

3 (18)

,0.001

Mean DWI volume, cm

49.6 (99.8)

19.9 (55.5)

0.01

5.4 (0.843.2)

1.7 (0.410.3)

0.002

ARU

620 (571.5657.0)

442 (415.3495.0)

SBP, mm Hg

142 (130160)

140 (130155)

0.07

DBP, mm Hg

80 (7190)

80 (7090)

0.54

WBCs, 3103/mL

8.7 (6.710.5)

7.6 (6.29.7)

0.03

Median DWI volume, cm

Laboratory data, median (IQR)

,0.001

Hemoglobin, g/dL

13.6 (11.714.5)

13.3 (12.114.5)

0.79

hsCRP, mg/L

5.6 (1.915.1)

2.7 (0.97.6)

0.003

Glucose, mg/dL

109 (93147)

105 (90139)

0.20

Total cholesterol, mg/dL

169 (148207)

170 (148198)

0.79

Continued

1810

Neurology 86

May 10, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Table 1

Continued
Total patients (N 5 310)

Characteristic
LDL cholesterol, mg/dL

HRPR group
(n 5 86, 27.7%)

Non-HRPR group
(n 5 224, 72.3%)

106 (84135)

104 (85124)

p Value
0.73

Abbreviations: ARU 5 aspirin reaction unit; CAD 5 coronary artery disease; DBP 5 diastolic blood pressure; DWI 5
diffusion-weighted imaging; HRPR 5 high residual platelet reactivity; hsCRP 5 high-sensitivity C-reactive protein;
IQR 5 interquartile range; LDL 5 low-density lipoprotein; mRS 5 modified Rankin Scale; NIHSS 5 NIH Stroke Scale;
NSAID 5 nonsteroidal anti-inflammatory drug; SBP 5 systolic blood pressure; TOAST 5 Trial of Org 10172 in Acute
Stroke Treatment; tPA 5 tissue plasminogen activator; WBC 5 white blood cell.
Values given are n (%), mean (SD), or median (IQR), as appropriate. The p values are for Student t test, MannWhitney
U test, Pearson x2 test, and Fisher exact test, or x2 trend test with linear by linear association, as appropriate.

characteristics between the HRPR and non-HRPR

groups are presented in table 1. HRPR patients were
more likely to arrive at the hospital earlier and to have
higher levels of white blood cells and high-sensitivity
C-reactive protein, whereas non-HRPR patients were
more likely to have been taking antihypertensive
medication before the index stroke. There were no
differences between groups in age, sex, risk factors,
distribution of the TOAST classification, administration of tissue plasminogen activator, or median time
between symptom onset and imaging. The median
initial NIHSS score and DWI infarct volume in the
HRPR group were significantly higher than in the
non-HRPR group.
Using bivariate median regression analysis, the
median initial NIHSS scores were higher in the
HRPR group (6 [IQR 315]) compared to
the non-HRPR group (3 [IQR 18]), resulting in
a statistically significant median difference of 3
points (95% confidence interval [CI] 0.58.5;
table e-1). After adjustment for age, sex, and variables with p , 0.05 in the bivariate analyses (tables 1
and e-1), the median adjusted initial NIHSS score
was 2 points higher for the HRPR group compared
to the non-HRPR group (median difference, 2.1;
95% CI 0.84.0) (table 2).
The HRPR group showed higher median DWI
infarct volume than did the non-HRPR group (5.4
[IQR 0.843.2] vs 1.7 [IQR 0.410.3], respectively),
in which the difference was statistically significant
(2.5 [95% CI 0.510.2]) with bivariate median
regression analysis (table e-1). HRPR was also a significant predictor of a large DWI infarct volume
(median volume difference between groups, 2.3
[95% CI 0.43.9]) from multivariable median regression analysis (table 2).
To examine whether the effects of HRPR differed
across the quantiles of the conditional distributions of
the initial NIHSS scores and DWI infarct volumes,
quantile regression analysis was performed. In a bivariate quantile regression analysis, the Q-50, Q-75, and
Q-90 of NIHSS score distributions in the HRPR

group were significantly higher compared to the

non-HRPR group, by 3 points (Q-50), 7 points
(Q-75), and 6 points (Q-90), respectively. However,
the Q-10 and Q-25 values were not significantly different between the 2 groups (table e-2). After adjusting for age, sex, and other significant covariates from
bivariate quantile regression analysis, the effect of
HRPR on the initial NIHSS score remained significant at the upper quantiles (Q-50, Q-75, and
Q-90) (table 3).
Bivariate quantile regression indicated that the
DWI infarct volumes were larger in the HRPR group
compared to those in the non-HRPR group by 0.5
cm3 (Q-25), 2.5 cm3 (Q-50), 32.8 cm3 (Q-75), and
101.9 cm3 (Q-90), respectively (table e-3), but there
were no differences between groups at Q-10. Significant associations between HRPR and DWI infarct
volume remained only at the upper quantiles (Q-50,
Q-75, and Q-90) by multivariable quantile regression
analysis (table 3).
These significant effects of HRPR on the upper
quantiles (Q-50, Q-75, and Q-90) of the distributions of the initial NIHSS scores as well as DWI
infarct volumes, but not on the lower quantiles
(Q-10 and Q-25), can be seen in figure 2, which
represented quantile estimates of the HRPR effect
on outcome variables along with its 95% CIs. We
also added standard ordinal least square estimate of
the HRPR effect (red solid line) and its 95% CI (red
dashed lines) in the figure, in which we can see its
underestimation of the effect of HRPR in severe
stroke patients and, conversely, an overestimation of
the effect in mild stroke patients.
When patients were stratified using the TOAST
classification, the effects of HRPR and initial NIHSS
scores or DWI infarct volumes were statistically significant in LAA (p , 0.001 and p , 0.001, respectively) and UE or OE (p 5 0.03 and p 5 0.01,
respectively), but not in SVO (p 5 0.52 and p 5
0.52, respectively) or CE (p 5 0.77 and p 5 0.26,
respectively) groups, in multivariable median regression analyses (tables e-4 to e-6).
Neurology 86

May 10, 2016

1811

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Figure 1

Scatterplot at different quantiles of (A) initial NIHSS score and (B) DWI infarct volumes

Lines from bottom to top indicate 0.1, 0.25, 0.50, 0.75, and 0.9 quantiles, respectively. The red dashed line indicates Spearman rank
correlation coefficient (rho) in overall. ARU 5 aspirin reaction unit; DWI 5 diffusion-weighted imaging; NIHSS 5 NIH Stroke Scale.

DISCUSSION Our study found that aspirin resistance is

associated with more severe stroke and larger infarct volume in patients who were taking aspirin at the time of
1812

Neurology 86

their index stroke. Specifically, these associations were

more apparent in the upper quantiles of initial NIHSS
score and DWI infarct volume distributions than in the

May 10, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Table 2

Multivariable median regression analyses for the initial NIHSS scores and for DWI infarct volumes
Median initial NIHSS score

Median DWI infarct volume

Estimate (95% CI)

Adjusted p valuea

Estimate (95% CI)

Age, y

0.10 (0.03 to 0.16)

,0.001

0.03 (20.02 to 0.05)

Sex, female

0.73 (20.08 to 1.86)

HRPR (compared to non-HRPR)

2.06 (0.81 to 3.97)

Hypertension

21.24 (24.53 to 2.55)

0.26

21.48 (24.19 to 1.50)

Atrial fibrillation

1.41 (20.33 to 4.74)

0.16

4.46 (1.06 to 12.35)

,0.001

Antihypertensive drug

20.18 (23.68 to 2.12)

0.82

0.31 (21.47 to 1.42)

0.76

Statin

0.49 (20.94 to 1.80)

0.50

Reference

Reference

2.02 (0.73 to 3.51)

0.003

20.63 (21.88 to 0.53)

0.45

1.35 (20.87 to 3.30)

0.08

1.00 (20.67 to 4.86)

0.31

20.02 (20.04 to 0.01)

0.29

0.01 (20.02 to 0.04)

0.61

0.18
,0.001

1.09 (20.04 to 2.93)

2.33 (0.42 to 3.91)

Adjusted p valueb
0.43
0.10
,0.001
0.27

Prestroke mRS score

2
Median onset to arrival time
TOAST classification
Small vessel occlusion

Reference

Large artery atherosclerosis

0.66 (20.03 to 1.77)

0.33

Reference
2.60 (0.65 to 3.83)

Cardioembolism

2.65 (0.68 to 4.80)

0.02

15.05 (7.08 to 23.24)

Other determined or
undetermined

20.63 (21.34 to 0.78)

0.46

2.12 (0.61 to 4.92)

0.002
,0.001
0.04

Administration of IV tPA

WBCs, 310 /mL

0.36 (0.20 to 0.60)

Hemoglobin, g/dL

20.19 (20.54 to 0.08)

0.16

hsCRP, mg/L

0.01 (0.02 to 0.05)

0.03

20.01 (20.03 to 0.14)

0.37

Total cholesterol, mg/dL

0.01 (20.02 to 0.05)

0.71

LDL cholesterol, mg/dL

20.01 (20.07 to 0.01)

0.65

,0.001

12.04 (1.62 to 35.98)

,0.001

0.21 (20.02 to 0.58)

0.07

Abbreviations: CI 5 confidence interval; DWI 5 diffusion-weighted imaging; HRPR 5 high residual platelet reactivity; hsCRP 5
high-sensitivity C-reactive protein; LDL 5 low-density lipoprotein; mRS 5 modified Rankin Scale; NIHSS 5 NIH Stroke Scale;
TOAST 5 Trial of Org 10172 in Acute Stroke Treatment; tPA 5 tissue plasminogen activator; WBC 5 white blood cell.
The p values are from the multivariable quantile regression analysis.
a
Variables adjusted are age, sex, hypertension, atrial fibrillation, antihypertensive drug use, prestroke mRS score, delay
time from symptom onset to hospital arrival, TOAST classification, WBCs, hemoglobin, and hsCRP, which showed p , 0.05
from bivariate analyses between the HRPR and initial NIHSS scores.
b
Variables adjusted are age, sex, hypertension, atrial fibrillation, antihypertensive drug use, statin use, prestroke mRS
score, delay time from symptom onset to hospital arrival, TOAST classification, IV tPA, WBCs, hsCRP, total cholesterol, and
LDL cholesterol, which showed p , 0.05 from bivariate analyses between the HRPR and the DWI infarct volumes.

lower quantiles. We also observed that the associations of

aspirin resistance with stroke severity and infarct volume
may differ by stroke mechanism.
NIHSS scores at the time of hospital admission
were higher in patients with HRPR than in patients
in the non-HRPR group, after adjusting for age,
sex, and the observed increase of 2 points in the
adjusted median NIHSS scores. These findings are
consistent with previous reports that HRPR is significantly associated with higher NIHSS scores at hospital admission.13,1517 However, these previous studies
were small; the total number of study participants was
less than 150. Except for 2 studies,16,17 less than half
of the patients were taking long-term aspirin therapy

before the index stroke. In addition, 2 other small

studies found no relationship between HRPR and
stroke severity.18,19 In our study, the effect of HRPR
on initial stroke severity was more convincingly demonstrated because of a larger sample size. We also
evaluated the effect of HRPR on stroke severity using
clinical assessment of neurologic deficits and the
tissue-based imaging parameter, DWI. Only 2 previous studies have investigated the association between
HRPR and infarct size, using initial Alberta Stroke
Program Early CT Score (ASPECTS).16,19 One of
these studies reported that HRPR was associated with
more extensive tissue injury on the initial CTASPECTS, but the other study found no significant
Neurology 86

May 10, 2016

1813

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Table 3

Multivariable quantile regression analyses for the effects of high residual platelet reactivity on the
initial NIHSS score and on DWI infarct volume across the distribution
Initial NIHSS score

Percentile

Estimate

95% CI

10th

0.49

20.24 to 0.78

25th

1.03

20.08 to 2.10

DWI infarct volume

Adjusted p valuea

Adjusted p valueb

Estimate

95% CI

0.24

0.09

20.07 to 0.41

0.35

0.06

0.19

20.04 to 0.55

0.28

50th

2.06

0.95 to 4.12

,0.001

1.54

0.13 to 4.18

0.02

75th

4.39

1.99 to 6.29

,0.001

5.80

2.26 to 32.75

0.01

90th

4.36

2.60 to 7.74

0.001

36.71

1.12 to 17.015

,0.001

Abbreviations: CI 5 confidence interval; DWI 5 diffusion-weighted imaging; HRPR 5 high residual platelet reactivity; hsCRP 5
high-sensitivity C-reactive protein; LDL 5 low-density lipoprotein; mRS 5 modified Rankin Scale; NIHSS 5 NIH Stroke Scale;
TOAST 5 Trial of Org 10172 in Acute Stroke Treatment; tPA 5 tissue plasminogen activator; WBC 5 white blood cell.
The p values are from the multivariable quantile regression analysis.
a
Variables adjusted are age, female sex, hypertension, atrial fibrillation, prior antihypertensive drug use, prestroke mRS score,
delay time from symptom onset to hospital arrival, TOAST classification, WBCs, hemoglobin, hsCRP, fasting total cholesterol,
and fasting LDL cholesterol, which showed p , 0.05 from bivariate analyses between the HRPR and initial NIHSS scores.
b
Variables adjusted are age, sex, hypertension, prior antihypertensive drug use, prior statin use, atrial fibrillation, prestroke
mRS score, delay time from symptom onset to hospital arrival, TOAST classification, IV tPA, WBCs, hsCRP, fasting glucose,
fasting total cholesterol, and fasting LDL cholesterol, which showed p , 0.05 from bivariate analyses between the HRPR
and DWI infarct volumes.

association between HRPR and infarct size. In addition, initial CT is not a logical imaging technique to
assess early ischemic change, and the anatomical regions coved by ASPECTS are limited to those with
anterior circulation. However, our study demonstrated a distinct association between HRPR and
infarct volume, as measured by DWI, which indicates
brain tissue with failing cellular energy-dependent
processes during early ischemic stroke. Thus, DWI
is a more refined method for assessing changes in early
ischemic tissue, compared to initial CT-ASPECTS.
In addition, early DWI infarct volumes strongly correlate with final infarct volumes on T2-weighted
imaging.28
The question of whether HRPR is a cause or a
marker of severe stroke remains unclear. Insufficient
inhibition of platelet aggregation in patients with
HRPR may lead to a larger thrombus and accumulating microvascular thrombi, resulting in more severe
stroke and larger infarct volumes.9,29 Conversely,
HRPR may represent the burden of atherosclerotic
disease. In some cases, larger areas of ischemic tissue
may produce more profound prothrombotic inflammatory reactions and may thereby contribute to the
development of HRPR-platelet aggregation during
the acute phase of atherosclerotic disease.15,3033 We
hypothesized that the effects of HRPR may differ
across the conditional percentiles of initial NIHSS
scores and DWI infarct volumes, based on the severity of thrombosis. To explore this hypothesis, we
performed quantile regression analyses for associations between HRPR and both initial stroke severity
and infarct volume, finding that the relationships
were more pronounced at the upper quantiles for
1814

Neurology 86

both variables. Upward sloping curves plotted using

quantile regression analyses indicated that HRPR had
a greater negative association with stroke severity in
patients with greater thrombosis. Although our study
design does not allow us to make conclusions about
cause and effect relationships between HRPR and
stroke severity, the findings extend knowledge of
the relationship between HRPR and stroke severity.
In our study, the distribution of stroke subtypes
using the TOAST classification was not different
between the HRPR and non-HRPR groups. In contrast, a previous study reported that HRPR occurred
more frequently in lacunar stroke compared with CE
stroke, as defined by the Oxfordshire Community
Stroke Project,18 although other recent studies
showed no association between HRPR and stroke
subtypes using the TOAST classification.16,17 However, stroke subtypes were not thoroughly investigated in previous studies because these studies only
explored the relationship between HRPR and stroke
subtypes.1618 No previous study had investigated
whether the effects of HRPR on stroke severity differ
by stroke mechanism. Meanwhile, our findings demonstrate that the differential effects of HRPR on
stroke severity and infarct volume depend on the
stroke mechanism. The significant associations of
HRPR with higher NIHSS scores and larger DWI
infarct volumes were found in the LAA and UE or
OE classification groups based on adjusted analyses,
whereas the relationship did not occur for stroke due
to CE and SVO. These results suggest that the preventive effects of aspirin might be related to the
underlying stroke mechanisms, specifically that
HRPR has a larger influence on stroke severity in

May 10, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Figure 2

Effect of HRPR on quantiles of (A) initial NIHSS score and (B) DWI infarct volume by multivariable
quantile regression analysis

Lines and gray bands represent connected slope coefficient estimates obtained at 0.1, 0.25, 0.5, 0.75, and 0.9 quantiles and their
CIs, respectively. Red horizontal lines indicate ordinal least squares estimates (solid line) with their 95% CIs (dashed lines). CI 5
confidence interval; DWI 5 diffusion-weighted imaging; HRPR 5 high residual platelet reactivity; NIHSS 5 NIH Stroke Scale.

patients with atherosclerotic stroke than in patients

with nonatherosclerotic stroke. They are consistent
with a previous study demonstrating that the effect
of prestroke antiplatelet use on the initial NIHSS
score was statistically significant in LAA but not in
SVO and CE.34
Our study has several limitations. First, this was a
single hospital-based observational study, although patients were consecutively screened and prospectively

enrolled. Thus, we could not control for unmeasured

confounding factors and hence our findings should
be interpreted with caution when generalizing to other
patients with ischemic stroke. Second, there is a possibility of the prevalence of HRPR being overestimated
in our study because compliance with aspirin therapy
was ascertained by interviews with the patients or their
relatives. The most reliable assessment of aspirin compliance is to measure plasma concentrations of aspirin
Neurology 86

May 10, 2016

1815

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

and salicylate. However, it is not possible to detect the

plasma concentrations of aspirin and salicylate at hospital admission because plasma concentrations of aspirin exhibit a half-life of 15 to 20 minutes. Finally, we
used only one platelet function test to assess HRPR.
Additional laboratory tests that detect platelet
cyclooxygenase-1 function, as measured by serum
thromboxane B2 levels, should be performed to verify
the aspirin compliance and aspirin responsiveness.35
However, the VerifyNow assay is a simple and rapid
point-of-care test for assessing HRPR, which is readily
accessible in clinical practice.
Nevertheless, our study also has several strengths
compared to previous studies. First, the sample size
is larger than previous studies. Second, this is the first
published study to report on the relationship between
HRPR and DWI infarct volume. Third, by using a
quantile regression model, a more complete understanding of the effect of HRPR on the distribution
of stroke severity and infarct volume can be obtained.
In conclusion, we found that patients with HRPR
were more likely to have severe strokes and large infarcts compared to patients in the non-HRPR group.
This negative effect of aspirin resistance would be further significantly worsening to patients with larger
atherosclerotic burden. Therefore, laboratory tests
for aspirin resistance should be considered before
modifying an antiplatelet regimen in patients who
experience new or recurrent ischemic stroke while
taking aspirin.
AUTHOR CONTRIBUTIONS
Study concept and design: Drs. M.S. Oh and B.-C. Lee. Acquisition of
data: Drs. M.S. Oh, K.-H. Yu, and B.-C. Lee. Analyses and interpretation of data: Drs. M.S. Oh, C. Kim, M.U. Jang, K.-H. Yu, J.-H. Lee,
S. Jung, and B.-C. Lee. Drafting of the manuscript: Dr. M.S. Oh. Critical revision of the manuscript for important intellectual content: Drs. M.
S. Oh, K.-H. Yu, J.-H. Lee, S. Jung, C. Kim, M.U. Jang, B.-C. Lee, and
J. Lee. Statistical analysis: Drs. M.S. Oh and J. Lee.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

STUDY FUNDING
This study was funded by grants from the Korea Healthcare Technology
R&D project, Ministry of Health and Family Welfare, Republic of Korea
(HI10C2020).

15.

DISCLOSURE

16.

The authors report no disclosures relevant to the manuscript. Go to

Neurology.org for full disclosures.

Received October 7, 2015. Accepted in final form January 29, 2016.

REFERENCES
1. Meschia JF, Bushnell C, Boden-Albala B, et al. Guidelines
for the primary prevention of stroke: a statement for
healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2014;45:
37543832.
2. Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for
the prevention of stroke in patients with stroke and transient
ischemic attack: a guideline for healthcare professionals from
1816

Neurology 86

17.

18.

19.

the American Heart Association/American Stroke Association. Stroke 2014;45:21602236.

Grotta JC, Lemak NA, Gary H, Fields WS, Vital D. Does
platelet antiaggregant therapy lessen the severity of stroke?
Neurology 1985;35:632636.
Wilterdink JL, Bendixen B, Adams HP Jr, Woolson RF,
Clarke WR, Hansen MD. Effect of prior aspirin use
on stroke severity in the Trial of Org 10172 in
Acute Stroke Treatment (TOAST). Stroke 2001;32:
28362840.
Sanossian N, Saver JL, Rajajee V, et al. Premorbid antiplatelet use and ischemic stroke outcomes. Neurology
2006;66:319323.
Dowlatshahi D, Hakim A, Fang J, Sharma M; Investigators of the Registry of the Canadian Stroke Network. Pre
admission antithrombotics are associated with improved
outcomes following ischaemic stroke: a cohort from the
Registry of the Canadian Stroke Network. Int J Stroke
2009;4:328334.
Ovbiagele B, Buck BH, Liebeskind DS, et al. Prior antiplatelet use and infarct volume in ischemic stroke.
J Neurol Sci 2008;264:140144.
Joseph R, Han E, Tsering C, Grunfeld S, Welch KM.
Platelet activity and stroke severity. J Neurol Sci 1992;
108:16.
Huang Z, Teng C, Lee T, Shun C, Wang C. Combined
use of aspirin and heparin inhibits in vivo acute carotid
thrombosis. Stroke 1993;24:829836.
Grilli M, Pizzi M, Memo M, Spano P. Neuroprotection
by aspirin and sodium salicylate through blockade of NFkappaB activation. Science 1996;274:13831385.
Krasopoulos G, Brister SJ, Beattie WS, Buchanan MR.
Aspirin resistance and risk of cardiovascular morbidity:
systematic review and meta-analysis. BMJ 2008;336:
195198.
Eikelboom JW, Hankey GJ, Thom J, et al. Incomplete
inhibition of thromboxane biosynthesis by acetylsalicylic
acid: determinants and effect on cardiovascular risk. Circulation 2008;118:17051712.
Ozben S, Ozben B, Tanrikulu AM, Ozer F, Ozben T.
Aspirin resistance in patients with acute ischemic stroke.
J Neurol 2011;258:19791986.
Yi X, Zhou Q, Lin J, Chi L. Aspirin resistance in Chinese
stroke patients increased the rate of recurrent stroke and
other vascular events. Int J Stroke 2013;8:535539.
Schwammenthal Y, Tsabari R, Shenkman B, et al. Aspirin
responsiveness in acute brain ischaemia: association with
stroke severity and clinical outcome. Cerebrovasc Dis
2008;25:355361.
Zheng AS, Churilov L, Colley RE, Goh C, Davis SM,
Yan B. Association of aspirin resistance with increased
stroke severity and infarct size. JAMA Neurol 2013;70:
208213.
Coignion C, Poli M, Sagnier S, et al. Interest of antiplatelet drug testing after an acute ischemic stroke. Eur Neurol
2015;74:135139.
Englyst NA, Horsfield G, Kwan J, Byrne CD. Aspirin
resistance is more common in lacunar strokes than embolic
strokes and is related to stroke severity. J Cereb Blood
Flow Metab 2008;28:11961203.
El-Mitwalli A, Azzam H, Abu-Hegazy M, Gomaa M,
Wasel Y. Clinical and biochemical aspirin resistance in
patients with recurrent cerebral ischemia. Clin Neurol
Neurosurg 2013;115:944947.

May 10, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

20.

21.

22.

23.

24.

25.

26.

27.

Saunders DE, Clifton AG, Brown MM. Measurement of

infarct size using MRI predicts prognosis in middle cerebral artery infarction. Stroke 1995;26:22722276.
Adams HP Jr, Bendixen BH, Kappelle LJ, et al. Classification of subtype of acute ischemic stroke: definitions for
use in a multicenter clinical trial. TOAST. Trial of Org
10172 in Acute Stroke Treatment. Stroke 1993;24:3541.
Mirkhel A, Peyster E, Sundeen J, et al. Frequency of aspirin resistance in a community hospital. Am J Cardiol
2006;98:577579.
Brott T, Adams HP Jr, Olinger CP, et al. Measurements
of acute cerebral infarction: a clinical examination scale.
Stroke 1989;20:864870.
Buck BH, Liebeskind DS, Saver JL, et al. Association of
higher serum calcium levels with smaller infarct volumes in
acute ischemic stroke. Arch Neurol 2007;64:12871291.
Wei Y, Pere A, Koenker R, He X. Quantile regression
methods for reference growth charts. Stat Med 2006;25:
13691382.
Wing JJ, Baek J, Sanchez BN, et al. Differences in initial
stroke severity between Mexican Americans and nonHispanic whites vary by age: the Brain Attack Surveillance
in Corpus Christi (BASIC) project. Cerebrovasc Dis 2014;
38:362369.
Koenker R. Quantreg: Quantile Regression. R Package
Version 5.05. Vienna: R Foundation for Statistical Computing; 2013.

28.
29.

30.

31.

32.

33.

34.

35.

Albers GW. Diffusion-weighted MRI for evaluation of

acute stroke. Neurology 1998;51:S47S49.
Michelson AD, Cattaneo M, Eikelboom JW, et al. Aspirin
resistance: position paper of the Working Group on Aspirin Resistance. J Thromb Haemost 2005;3:13091311.
Roden-Jullig A, Britton M, Malmkvist K, Leijd B. Aspirin
in the prevention of progressing stroke: a randomized controlled study. J Intern Med 2003;254:584590.
Choudhri TF, Hoh BL, Zerwes HG, et al. Reduced
microvascular thrombosis and improved outcome in acute
murine stroke by inhibiting GP IIb/IIIa receptor-mediated
platelet aggregation. J Clin Invest 1998;102:13011310.
Borna C, Lazarowski E, van Heusden C, Ohlin H,
Erlinge D. Resistance to aspirin is increased by STelevation myocardial infarction and correlates with adenosine diphosphate levels. Thromb J 2005;3:10.
Valles J, Santos MT, Fuset MP, et al. Partial inhibition of
platelet thromboxane A2 synthesis by aspirin is associated
with myonecrosis in patients with ST-segment elevation
myocardial infarction. Am J Cardiol 2007;99:1925.
Kim WJ, Ko Y, Yang MH, et al. Differential effect of
previous antiplatelet use on stroke severity according to
stroke mechanism. Stroke 2010;41:12001204.
Cox D, Maree AO, Dooley M, Conroy R, Byrne MF,
Fitzgerald DJ. Effect of enteric coating on antiplatelet
activity of low-dose aspirin in healthy volunteers. Stroke
2006;37:21532158.

NEW!
Without Borders A curated collection featuring
advances in global neurology
This Neurology special interest Web site is the go-to source for tracking science and politics of
neurology beyond the United States, featuring up-to-the-minute blogs, scholarly perspectives, and
academic review of developments and research from Neurology journals and other sources.
Curated by Gretchen L. Birbeck, MD, MPH.
Expand your world view at Neurology.org/woborders.

Neurology 86

May 10, 2016

1817

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Aspirin resistance is associated with increased stroke severity and infarct volume
Mi Sun Oh, Kyung-Ho Yu, Ju-Hun Lee, et al.
Neurology 2016;86;1808-1817 Published Online before print April 8, 2016
DOI 10.1212/WNL.0000000000002657
This information is current as of April 8, 2016
Updated Information &
Services

including high resolution figures, can be found at:

http://www.neurology.org/content/86/19/1808.full.html

Supplementary Material

Supplementary material can be found at:

http://www.neurology.org/content/suppl/2016/04/08/WNL.0000000000
002657.DC1.html

References

This article cites 34 articles, 18 of which you can access for free at:
http://www.neurology.org/content/86/19/1808.full.html##ref-list-1

Subspecialty Collections

This article, along with others on similar topics, appears in the

following collection(s):
Infarction
http://www.neurology.org//cgi/collection/infarction

Permissions & Licensing

Information about reproducing this article in parts (figures,tables) or in

its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions

Reprints

Information about ordering reprints can be found online:

http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright 2016 American Academy of Neurology. All
rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.