close

Вход

Забыли?

вход по аккаунту

?

Apolipoprotein E phenotype and the efficacy of intravenous tissue plasminogen activator in acute ischemic stroke.

код для вставкиСкачать
Apolipoprotein E Phenotype and the
Efficacy of Intravenous Tissue Plasminogen
Activator in Acute Ischemic Stroke
Joseph Broderick, MD,1 Mei Lu, PhD,2 Christy Jackson, MD,3 Arthur Pancioli, MD,4 Barbara C. Tilley, PhD,5
Susan C. Fagan, PharmD,6 Rashmi Kothari, MD,7 Steven R. Levine, MD,8 John R. Marler, MD,9
Patrick D. Lyden, MD,3 E. Clark Haley, Jr., MD,10 Thomas Brott, MD,11 James C. Grotta, MD,12 and the
NINDS t-PA Stroke Study Group
We used stored plasma samples from 409 patients in the National Institute of Neurological Diseases and Stroke
(NINDS) tissue plasminogen activator (t-PA) Stroke Trial to examine the relationship between an apolipoprotein (Apo)
E2 or an Apo E4 phenotype and a favorable outcome 3 months after stroke, the risk of intracerebral hemorrhage, and
the response to intravenous t-PA therapy. For the 27 patients with an Apo E2 phenotype who were treated with t-PA,
the odds ratio (OR) of a favorable outcome at 3 months was 6.4 [95% confidence interval (CI) 2.7–15.3%] compared to
the 161 patients without an Apo E2 phenotype who were treated with placebo. The 190 patients treated with t-PA who
did not have an Apo E2 phenotype also had a greater, though less pronounced, likelihood of a favorable outcome (OR
2.0, 95% CI 1.2–3.2%) than patients without an Apo E2 phenotype treated with placebo. For the 31 patients with an
Apo E2 phenotype treated with placebo, the OR of a favorable 3 month outcome was 0.8 (95% CI 0.4 –1.7%) compared
to the 161 patients without an Apo E2 phenotype treated with placebo. This interaction between treatment and Apo E2
status persisted after adjustment for baseline variables previously associated with 3 month outcome, for differences in the
baseline variables in the two treatment groups and in the Apo E2-positive and -negative groups, and for a previously
reported time-to-treatment ⴛ treatment interaction ( p ⴝ 0.03). Apo E4 phenotype, present in 111 (27%) of the 409
patients, was not related to a favorable 3 month outcome, response to t-PA, 3 month mortality, or risk of intracerebral
hemorrhage. We conclude that the efficacy of intravenous t-PA in patients with acute ischemic stroke may be enhanced
in patients who have an Apo E2 phenotype, whereas the Apo E2 phenotype alone is not associated with a detectable
benefit on stroke outcome at 3 months in patients not given t-PA. In contrast to prior studies of head injury and stroke,
we could not detect a relationship between Apo E4 phenotype and clinical outcome.
Ann Neurol 2001;49:736 –744
Apolipoprotein E (Apo E) is a member of a family of
lipid-associated proteins whose isoforms have been implicated as an important modifier of several neurologic
conditions, including Alzheimer’s disease, intracerebral
hemorrhage (ICH), head trauma, and ischemic
stroke.1–20 Apo E has three common isoforms E2, E3,
and E4, which are coded by the alleles e2, e3, and e4,
respectively, at a single locus on chromosome 19.1–3
These alleles combine to make six Apo E genotypes:
E2/2, E2/3, and E2/4 (referred to as the E2 category);
E3/3 (E3 category); and E4/3 and E4/4 (E4 category).
Apo E has been linked to outcome and survival fol-
lowing acute injury of the central nervous system. The
presence of an E4 allele has been associated with a poor
outcome following severe head trauma10 –12 and with
poorer survival in patients with an ICH.17 McCarron
and colleagues13 found that patients with an ischemic
stroke who had an Apo E4 allele had a small but statistically greater likelihood of survival, but Nicoll and
colleagues14 found no relationship between Apo E4 genotype and outcome in patients with acute cerebral infarction. Both the Apo E418,19 and the Apo E218 –20
alleles have been associated with an increased risk of
lobar ICH in the elderly, although data regarding Apo
From the 1Department of Neurology, University of Cincinnati,
Cincinnati, OH; 2Department of Biostatistics and Research Epidemiology, Henry Ford Health System, Detroit, MI; 3University of
California Stroke Center, San Diego, CA; 4Department of Emergency Medicine, University of Cincinnati, Cincinnati, OH; 5Department of Biometry and Epidemiology, Medical University of
South Carolina, Charleston, SC; 6University of Georgia, Augusta,
GA; 7Borgess Research Institute, Kalamazoo, MI; 8Wayne State
University Stroke Program, Detroit, MI; 9Division of Stroke and
Trauma, National Institute of Neurological Diseases and Stroke, Bethesda, MD; 10Department of Neurology, University of Virginia
Health System, Charlottesville, VA;
Mayo Clinic, Jacksonville, FL; and
University of Texas, Houston, TX.
736
© 2001 Wiley-Liss, Inc.
11
12
Department of Neurology,
Department of Neurology,
Received Mar 30, 2000, and in revised form Dec 15. Accepted for
publication Dec 27, 2000.
Published online 25 May 2001.
Address correspondence to Dr Broderick, Department of Neurology, University of Cincinnati, 231 Bethesda Ave., ML 0525, Cincinnati, OH 45267-0525. E-mail: joseph.broderick@uc.edu
E4 are contradictory.18 –20 We report a study of patients from the National Institute of Neurological Diseases and Stroke (NINDS) Tissue Plasminogen Activator (t-PA) Stroke Trial that was designed to investigate
the relationship of Apo E phenotype, specifically the
presence of either an Apo E4 or an Apo E2 phenotype,
to stroke outcome, risk of ICH, and response to intravenous t-PA.
Materials and Methods
The NINDS t-PA Stroke Study consisted of two sequential placebo-controlled, randomized, double-blind,
multicenter clinical trials. Details have been reported in
several previous publications.21–24 Patients had to have
an acute ischemic stroke with a measurable neurologic
deficit on the National Institute of Health Stroke Scale
(NIHSS)21–25 and be treated with sudy medication
within 3 hours of symptom onset. Detailed clinical,
historical, and laboratory data were collected on each
patient at baseline as previously reported.21 Patient
outcome data were systematically collected at 2 hours,
24 hours, 7–10 days, and 3 months following stroke.
Records of adverse events such as death and symptomatic/asymptomatic ICH were collected as the event
occurred. The primary outcome measure, a favorable
outcome at 3 months, was defined as a complete or
near-complete recovery on each of the four clinical outcome scales: a Barthel Index score of 95 or 100,26 a
Modified Rankin Scale of 0 or 1,27 a Glasgow Outcome Scale of 1,28 and an NIHSS score of 0 or 1. The
study was approved by the Institutional Review Board
at each participating hospital, and informed consent
was obtained for each study patient.
Computerized tomography (CT) images of the brain
were obtained and processed 24 hours, 7–10 days, and
3 months after stroke onset. The methodology of the
processing of CT images and the measurement of lesion volume have been previously reported.24
Blood samples were obtained from patients at baseline, 2 hours after start of therapy, and 24 hours poststroke and collected in citrated tubes. Plasma from
these samples was used to determine fibrinogen levels
at these three time points by a central laboratory. The
remaining plasma was frozen and stored at –70°C for
future blood studies that examined factors potentially
related to the risk of bleeding.
We combined, categorized, or ranked data before
linking the patient data with the Apo E phenotypes.
The coordinating center staff then merged the laboratory data and clinical information from the original database files by patient ID and then created a merged
file with all complete matches and IDs. After this
merger was completed, the identifying information,
such as the original patient ID and other patient information that could be linked to the patient’s medical
record number, was removed from the merged file.
This process was documented and signed by at least
two witnesses. The coordinating staff then destroyed
other evidence that could link patient laboratory data
to identifying clinical information.
The stored 24 hour plasma samples were used to determine the specific Apo E phenotypes using simplified
isoelectric focusing/immunoblotting techniques to
identify Apo E isoforms as reported by Kataoka and
colleagues.29 The serum had been stored for a median
time of 38 months. Because patient serum (and not
white blood cells) was available for analysis, patient
DNA was not used for this laboratory test. This
method has been demonstrated to have a 98% concordance rate with Apo E genotypes in 431 samples of
patients who also underwent genotyping.29 These laboratory tests were performed by Penn Medical Laboratory at the Medlantic Research Institute.
Statistical Methods
We were primarily interested in investigating the association between Apo E phenotype (specifically the presence of an Apo E4 or an Apo E2 phenotype) and 3
month favorable outcome, symptomatic or asymptomatic ICH within 36 hours of onset of treatment, and
death within 90 days with respect to treatment with
t-PA/placebo. Our prospective primary hypotheses,
prior to determination of Apo E phenotypes, were as
follows. 1) Apo E phenotype is significantly related to
the likelihood of a favorable outcome at 3 months as
measured by the Global Outcome odds ratio (OR). 2)
Apo E phenotype interacts significantly with treatment
assignment with respect to 1) the occurrence of any
ICH during the first 36 hours, 2) a favorable 3 month
outcome, and 3) the lesion volume as measured by CT
of the head at 3 months.
We tested whether the 409 patients who had blood
samples available for analysis were comparable to the
215 patients without a blood sample in the NINDS
t-PA Trial cohort. Thirty-five variables, including time
from stroke onset to treatment, were selected (see
Appendix).
To determine whether there were significant differences between patients with and without a blood sample, we used a ␹2 test for categorical variables (or Fisher’s exact test when a cell proportion was too small)
and a t-test or Wilcoxon test for continuous variables,
depending on whether the variable was normally distributed. The remainder of the analyses involve only
patients who had blood samples available for analysis.
We identified baseline variables that were unbalanced
between t-PA and placebo-treated patients and compared the distribution of baseline variables in patients
with and without an Apo E2 or Apo E4 phenotype
using the previously described analysis approach.
To investigate the relationship of Apo E2 and Apo
E4 phenotypes with clinical outcome measures, we per-
Broderick et al: Apo E and t-PA Efficacy
737
formed analyses for each individual clinical outcome
for all placebo- and t-PA-treated patients and included
treatment assignment as a covariate. The analyses that
examined the relationship of Apo E2 and Apo E4
phenotypes and the risk of ICH within 36 hours of
treatment onset were carried out initially using the
t-PA-treated group only and subsequently using both
t-PA- and placebo-treated patients. A logistic regression
model was carried out for a single binary outcome, eg,
Barthel Index 95 or 100 (yes, no). In addition, a global
statistical test was used to assess the effects of Apo E2
and Apo E4 on a binary measure of favorable outcome
based on all four clinical outcome measures.30 A global
statistical test (GST) summarizes treatment efficacy
across a number of different outcome measures taken
on the same patient and is useful when the outcomes
measured represent the same underlying construct. In
the case of stroke, the underlying construct is “favorable outcome” as measured by the following binary
outcomes: NIH stroke scale of 0 or 1, a Barthel Index
of 95 or 100, a Modified Rankin Scale of 0 or 1, and
a Glasgow Outcome Scale of 1. These outcomes measuring “a favorable outcome” are assessed on the same
patient, and the outcomes are correlated. Generalized
estimating equations are used to take the inpatient correlations among outcomes into account in estimating
the treatment effect, its variance, and the test statistic.
An interaction between Apo E2/Apo E4 and treatment was considered significant when the p value for
interaction was ⱕ0.10.31 (In all analyses of interactions, the main effects that are combined to form the
interaction are included in the model.) If no interaction was detected, we assumed that the effect of Apo
E2/Apo E4 was the same for t-PA- or placebo-treated
patients. Detection of an interaction indicated that the
effect of treatment (the efficacy of t-PA compared to
placebo) depended on the presence or absence of an
Apo E2 or Apo E4 phenotype. The log-rank test was
used to test the effect of an Apo E2 or Apo E4 phenotype with respect to the deaths within 90 days after
adjusting for the treatment variable.
In analyzing the effect of Apo E2 or Apo E4 phenotype on CT lesion volumes at 3 months, a Poisson
regression model was used on the cube roottransformed CT lesion volume to reduce the skewness
and variability.24 A negative coefficient indicated a reduction of CT lesion volume in the presence of an Apo
E2 or an Apo E4 phenotype compared to the absence
of an Apo E2 or an Apo E4 phenotype or a reduction
in CT lesion volume in t-PA-treated patients compared
to the placebo-treated patients. The median and interquartiles of CT lesion volume were reported as descriptive analyses.
All the above-mentioned analyses were then rerun
adjusting for 1) the baseline variables unbalanced between t-PA and placebo as well as the baseline variables
738
Annals of Neurology
Vol 49
No 6
June 2001
associated with the presence of Apo E2/E4 phenotypes
and 2) the baseline variables known to be significantly
related to 3 month outcome in our previously reported
models of outcome using the entire NINDS t-PA
Stroke Trial cohort.22–24 A time-from-stroke-onset-totreatment ⫻ treatment interaction was also added into
the model for 3 month favorable outcome based on
previously demonstrated interaction of time and efficacy of t-PA therapy.32
For descriptive purposes, the adjusted log odds and
95% confidence intervals (CI) are presented comparing
each Apo E2/treatment category [ie, “Apo E2(⫹)/treatment(⫹),” “Apo E2(⫹)/treatment(–),” “Apo E2(–)/
treatment(⫹)” to the referent group “Apo E2(–)/treatment(–)].” The same analyses are presented for Apo
E4/treatment categories.
Results
Among the 624 patients in the NINDS rt-PA Stroke
Trial, 312 were treated with t-PA and 312 were treated
with placebo. For our analysis, we used only the 409
patients (217 treated with t-PA and 192 treated with
placebo) with an available plasma sample. The phenotypes for the 409 patients were Apo E2–2 (1%), Apo
E2–3 (12%), Apo E2– 4 (1%), Apo E3–3 (60%), Apo
E3– 4 (24%), and Apo E4 – 4 (2%).
Only three baseline variables were significantly different between the 409 patients with a blood sample
and the 215 patients without a blood sample ( p ⬍
0.05). Patients with a blood sample were more likely to
be male (61% vs 51%) and had a greater frequency of
cardiac disease (51% vs 42%) or atherosclerosis (34%
vs 25%) compared to those without a sample. We
could not detect differences between those with and
without blood samples in any of the four 3 month outcomes (all p values ⬎0.12). Among the 409 patients
with a blood sample, six baseline variables (age, ranked
weight, ranked total dose delivered, baseline pulse pressure, aspirin, and the baseline NIHSS) were unbalanced ( p ⬍ 0.05) between the t-PA and the placebo
treatment groups.
When compared to those persons who did not have
an Apo E2 phenotype, patients with an Apo E2 phenotype were more likely to be black (47% vs 26%,
p⫽0.002), to have a lower median baseline NIHSS (eg,
22% vs 8% of patients with an NIHSS ⱕ 5,
p⫽0.02), to be hypertensive (81% vs 66%,p ⫽ 0.02),
and to have a history of cardiac disease (64% vs 49%,
p ⫽ 0.04). Patients with an Apo E2 phenotype also
were more likely to be taking a calcium channel
blocker prior to the stroke (36% vs 25%, p ⫽ 0.06).
Compared to those persons without an E4 phenotype,
persons with an Apo E4 phenotype were less likely to
have a systolic blood pressure ⬎190 mm Hg at admission (8% vs 13%, p ⫽ 0.05) and less likely to have a
prior hematologic disorder (1% vs 5%, p ⫽ 0.05).
Relationship of Apo E2 and Apo E4 Phenotype to 3
Month Clinical Outcome and Response to t-PA
Table 1a shows the analysis of 3 month outcomes by
treatment and Apo E2 phenotype (yes/no) categories
for the individual and global tests without adjusting for
any covariates. Apo E2 itself was not associated with
any of the 3 month favorable outcomes ( p ⬎ 0.42),
but there was an interaction between treatment and
Apo E2 phenotype ( p ⫽ 0.03). Patients treated with
t-PA who had an Apo E2 phenotype had a much
greater likelihood of a favorable outcome by any of the
four 3 month measures (range of OR from 3.6 to 6.3)
compared to those patients treated with placebo and
without an E2 phenotype. Patients treated with t-PA
who did not have an Apo E2 phenotype also had a
higher likelihood of a favorable outcome by any of the
four 3 month measures (range of odds ratios from 1.6
to 2.0) compared to those treated by placebo and without an E2 phenotype.
The interaction between treatment and Apo E2 phenotype persisted after adjusting for imbalances in baseline variables between treatment groups and between
Table 1. Three-Month Outcome by Apo E2 Status Without (a) or With (b) Adjusting for Baseline Covariates
t-PA
Outcome
NIHSSa
0 or 1%
OR
95% CI
Barthel
95 or 100%
OR
95% CI
Rankin
0 or 1%
OR
95% CI
Glasgow
1%
OR
95% CI
Global
OR
95% CI
NIHSSb
0 or 1%
OR
95% CI
Barthel
95 or 100%
OR
95% CI
Rankin
0 or 1%
OR
95% CI
Glasgow
1%
OR
95% CI
Global
OR
95% CI
Placebo
Apo E2(⫹)
(n ⫽ 27)
Apo E2(⫺)
(n ⫽ 190)
Apo E2(⫹)
(n ⫽ 31)
Apo E2(⫺)c
(n ⫽ 161)
63
6.3
2.7, 15.1
33
1.8
1.1, 2.9
19
0.9
0.3, 2.4
21
1.0
NA
0.04
70
3.6
1.5, 8.7
52
1.6
1.1, 2.5
39
1.0
0.4, 2.1
40
1.0
NA
0.16
67
5.5
2.3, 13.1
43
2.0
1.3, 3.2
23
0.8
0.3, 2.0
27
1.0
NA
0.06
67
4.6
1.9, 10.9
46
1.9
1.2, 3.0
35
1.3
0.6, 2.8
30
1.0
NA
0.29
5.4
2.4, 12.1
1.8
1.2, 2.6
1.0
0.5, 2.0
1.0
NA
63
7.4
2.1, 25.8
33
2.0
1.1, 3.6
19
0.9
0.3, 2.8
21
1.0
NA
0.08
70
4.8
1.4, 16.8
52
2.1
1.2, 3.7
39
1.0
0.4, 2.7
40
1.0
NA
0.31
67
8.1
2.1, 31.4
43
2.2
1.2, 4.1
23
0.6
0.2, 2.0
27
1.0
NA
0.05
67
4.0
1.2, 13.6
46
2.2
1.3, 4.0
35
1.0
0.4, 2.9
30
1.0
NA
0.49
6.4
2.7, 15.3
2.0
1.2, 3.2
0.8
0.4, 1.7
1.0
NA
p Values for Interaction
0.03
0.01
a
Without adjusting for baseline covariates.
Baseline covariates: age (categorized), rank of the actual weight, rank of total dose delivered, baseline pulse pressure, prior aspirin use, NIHSS
(5 groups), hypertension, race, cardiac history, age ⫻ NIHSS. Admission mean blood pressure (MBP), age ⫻ admission MBP, diabetes, and
the variable early CT finding.
c
Reference group for calculation of odds ratio.
b
Broderick et al: Apo E and t-PA Efficacy
739
persons with and without an Apo E2 phenotype as well
as adjusting for baseline variables previously associated
with a favorable outcome at 3 months (Table 1b;
p ⫽ 0.01). The likelihood of a favorable outcome by
any of the four 3 month measures for patients who
were t-PA treated and with an Apo E2 phenotype became even higher (range of OR from 4.0 to 8.1) compared to those without an E2 phenotype treated by
placebo. When we adjusted for a time-from-stroke-onset-to-treatment ⫻ treatment interaction that was previously identified,32 the interaction between Apo E2
and treatment also remained ( p ⫽ 0.03). No treatment ⫻ Apo E4 interaction was detected ( p ⬎ 0.49),
even after adjusting for the variables mentioned above
(Table 2a,b).
Relationship of Apo E2 and Apo E4 Phenotype to
Intracerebral Hemorrhage
No treatment and Apo E2/E4 interactions were detected with respect to symptomatic ICH or all ICH
within 36 hours of treatment onset ( p ⬎ 0.40). Neither Apo E4 nor Apo E2 was significantly associated
Table 2. Three-Month Outcome by Apo E4 Status Without (a) or With (b) Adjusting for Baseline Covariates
t-PA
Outcome
NIHSSa
0 or 1%
OR
95% CI
Barthel
95 or 100%
OR
95% CI
Rankin
0 or 1%
OR
95% CI
Glasgow
1%
OR
95% CI
Global
OR
95% CI
NIHSSb
0 or 1%
OR
95% CI
Barthel
95 or 100%
OR
95% CI
Rankin
0 or 1%
OR
95% CI
Glasgow
1%
OR
95% CI
Global
OR
95% CI
Placebo
Apo E4(⫹)
(n ⫽ 53)
Apo E4(⫺)
(n ⫽ 164)
Apo E4(⫹)
(n ⫽ 58)
Apo E4(⫺)c
(n ⫽ 134)
40
2.7
1.4, 5.5
35
2.3
1.3, 3.9
24
1.3
0.6, 2.8
19
1.0
NA
0.85
53
1.7
0.9, 3.2
54
1.8
1.1, 2.9
40
1.0
0.5, 1.9
40
1.0
NA
0.89
43
2.5
1.3, 5.0
46
2.9
1.7, 4.8
33
1.6
0.8, 3.2
23
1.0
NA
0.20
47
2.2
1.1, 4.2
49
2.3
1.4, 3.8
36
1.4
0.7, 2.7
29
1.0
NA
0.40
2.1
1.2, 3.9
2.1
1.4, 3.2
1.2
0.6, 2.1
1.0
NA
40
2.4
1.0, 5.6
35
2.4
1.2, 4.7
24
1.2
0.5, 2.7
19
1.0
NA
0.80
53
1.6
0.7, 3.7
54
2.5
1.4, 4.5
40
0.8
0.4, 1.8
40
1.0
NA
0.72
43
2.2
0.9, 5.3
46
3.9
2.0, 7.8
33
1.4
0.6, 3.4
23
1.0
NA
0.12
47
2.0
0.9, 4.6
49
2.9
1.6, 5.5
36
1.3
0.6, 2.8
29
1.0
NA
0.29
2.0
1.0, 4.1
2.6
1.6, 4.3
1.1
0.5, 2.3
1.0
NA
p Value for Interaction
0.72
0.49
a
Without adjusting for baseline covariates.
Baseline covariates: age (categorized), rank of the actual weight, rank of total dose delivered, baseline pulse pressure, prior aspirin use, NIHSS
(5 groups), hypertension, race, cardiac history, age ⫻ NIHSS. Admission mean blood pressure (MBP), age ⫻ admission MBP, diabetes, and
the variable early CT finding.
c
Reference group for calculation of odds ratio.
b
740
Annals of Neurology
Vol 49
No 6
June 2001
with either symptomatic ICH or with all ICH, even
after adjustments for the other covariates in the primary hemorrhage model22 or the covariates unbalanced
between groups with and without an Apo E2 or Apo
E4 phenotype ( p ⬎ 0.30).
Relationship of Apo E2 and Apo E4 Phenotype and
Mortality at 3 Months
No treatment ⫻ Apo E2 phenotype or treatment ⫻
Apo E4 interactions were detected with respect to mortality at 3 months. The mortality rate at 90 days was
17% in patients with an Apo E2 phenotype vs 19% in
patients without an Apo E2 phenotype ( p ⫽ 0.66) and
19% in patients with Apo E4 phenotype vs 19% in
patients without an Apo E4 phenotype ( p ⫽ 0.95),
adjusting for treatment status. After adjusting for the
variables unbalanced between the t-PA and placebo
treatment groups and variables associated with outcome, the results remained the same.
Relationship Between Apo E2 and Apo E4 Phenotype
and CT Lesion Volumes at 24 Hours and at
3 Months
Among the 409 patients with a blood sample, 404 patients had a lesion volume for determination of 3
month outcomes. Patients with an Apo E2 phenotype
had a smaller median CT lesion volume (8.5 cm3, interquartile range of 1.0 –36.0 cm3) compared to patients without an Apo E2 phenotype (20.0 cm3, interquartile range of 2.0 –99.0 cm3). However, no
interaction between Apo E2 status and treatment on 3
month CT lesion volume was detected after adjusting
for imbalances in baseline variables and for the other
covariates associated with CT lesion volume in a previously reported model (age, age ⫻ treatment interactions, baseline NIHSSS, early CT findings of ischemia,
old CT lesion volume at baseline, NIHSS ⫻ early CT
findings ischemia interaction, NIHSS ⫻ old CT lesion
volume interaction, and presumptive stroke subtype).
There was no relationship between Apo E4 phenotype
and volume of cerebral infarction.
Discussion
The clinical efficacy of intravenous t-PA in patients
with acute ischemic stroke appeared to be enhanced in
those patients in the NINDS t-PA Stroke Trial who
had an Apo E2 phenotype. No such benefit for an Apo
E2 phenotype was demonstrated in placebo-treated patients, as measured either by clinical outcome or by CT
lesion volume at 3 months. The three most likely explanations of our data are: 1) An Apo E2 phenotype or
another factor associated with an Apo E2 phenotype
enhances thrombolysis, and/or the brain’s response to
early reperfusion in the presence of t-PA; 2) an Apo E2
phenotype is associated with decreased reperfusion injury; or 3) our findings occurred by chance or repre-
sent an artifact resulting from imbalances in other variables (known or unknown) that may affect outcome
following ischemic stroke.
We carefully considered the last explanation first. In
a previously reported multivariable model of outcome,
we identified variables that could be plausibly related
to a favorable 3 month outcome.23 When we adjusted
for imbalances in these variables, the interaction between the Apo E2 status and treatment with t-PA persisted. However, the small number of patients with an
Apo E2 phenotype in our study (n ⫽ 58) raises the
unanswered question of whether the relationship between Apo E2 and response to t-PA in the NINDS
t-PA Stroke Trial represents an artifact from imbalances in other variables or a chance finding rather than
a true biological relationship.
The only variable that previously had been reported
to be associated significantly with the response to t-PA
in the NINDS t-PA Stroke Trial is the time from onset
of symptoms to start of treatment.32 The relationship
of time to the efficacy of t-PA therapy is consistent
with what is known about the pathophysiology of focal
brain ischemia. The biological mechanism of our observation remains to be demonstrated. Because chance
cannot be ruled out as an explanation, our finding
must be confirmed in a separate group of patients
treated with t-PA within 3 hours of onset in whom the
Apo E phenotype or genotype is identified.
We prospectively chose to look at the relationship of
Apo E4 and Apo E2 phenotypes to 3 month outcome,
risk of ICH, mortality, and response to t-PA because of
previously published data concerning the modifying effect of Apo E in diseases of the central nervous system
as well as in cardiovascular disease.1–17, 33 In contrast
to prior studies of ICH, cerebral infarction, and head
trauma, our study found no relationship between the
presence of an Apo E4 phenotype and a favorable 3
month outcome, risk of ICH, 3 month mortality, or
response to t-PA. Unlike the present study, prior studies of ICH and ischemic stroke did not adjust for imbalances in important variables that are known to be
associated with outcome following stroke,13,17 such as
the baseline neurologic deficit and the volume of brain
hemorrhage or cerebral infarction. This may also explain the conflicting results regarding Apo E4 and outcome in two studies of patients with cerebral infarction13,14 and a third study of patients with ICH.17
Additionally, we performed a simulation using the
Rankin Scale as the outcome measure and assuming
the observed prevalence of Apo E4 and an interaction
of the magnitude found for Apo E2. The power of our
study to detect an interaction of that magnitude for
Apo E4 was 64%. Whereas the power of the global test
would be greater than or equal to the power of the test
using only the Rankin Scale, the power would likely be
⬍80%. If the Apo E4 effects on the outcomes listed
Broderick et al: Apo E and t-PA Efficacy
741
above were smaller than the effect for Apo E2, we
would have even lower power to detect these differences. Thus the negative results for Apo E4 require
confirmation in a larger study.
In both studies of the relationship between outcome
following head trauma and Apo E4 genotype, the Glasgow Coma Scale (GCS) scores at baseline, the most
important predictor of outcome in patients with head
trauma, were substantially lower in patients with an
Apo E4 allele.11,12 After adjustment for baseline variables such as the GCS and exclusion of 6 patients who
were not contacted directly to assess outcome, the relationship between Apo E4 genotype and severe disability or death was only of borderline significance in
the first study. In the second study, the authors found
no difference in the likelihood of an unfavorable outcome (severe disability or death) in persons with and
without an E4 allele. Patients with an Apo E4 allele
were more likely to have a fair or unfavorable outcome
after adjustment for age and time of unconsciousness.
The authors of the two studies of head trauma that
relate Apo E4 status to outcome hypothesize that the
presence of Apo E4 allele results in impaired neuronal
sprouting and reorganization following brain injury.11,12,34 –38 If this hypothesis concerning brain injury
and Apo E4 status is true, we would expect poorer outcome in the placebo-treated patients in our study who
had an Apo E4 phenotype. We were unable to demonstrate such a relationship in our patients. Certainly,
the pathophysiologies of traumatic brain injury and
ischemic stroke are different, and we cannot rule out
the possibility that Apo E4 plays a different role in the
modification or recovery from injury following brain
trauma.
If the Apo E2 isoform directly or indirectly modifies
the effect of t-PA in acute ischemic stroke or the
brain’s response to reperfusion, this likely occurs because of its unique binding properties compared to the
Apo E4 and E3 isoforms.39 – 43 The Apo E2 dimer
forms a complex with amyloid beta peptides more efficiently than the Apo E3 dimer and much more efficiently than the Apo E4 isoform.44,45 This is one proposed mechanism for the protective effect of Apo E2
and the increased risk of Apo E4 with regard to Alzheimer’s disease. Amyloid beta peptides markedly stimulate plasminogen activation by t-PA in vitro,46,47
which may provide a possible clue for increased activity
of t-PA in stroke patients with an Apo E2 allele. Apo E
also has been shown to stimulate endothelial production of heparin sulfate,48and heparin may enhance the
lytic effect of t-PA in stroke patients.49 We did not
find an increased risk of symptomatic ICH in patients
with an Apo E2 phenotype, which one might expect in
the setting of increased t-PA activity. However, the
number of ICHs in our study was small and provide
limited power to detect such a relationship.
742
Annals of Neurology
Vol 49
No 6
June 2001
A relationship between other apolipoproteins and fibrinolysis has been reported. Apo A isoforms are
known to modify fibrinolysis in vitro and in vivo by
binding to fibrin and competing with plasminogen for
fibrin binding sites.50 –54 In an in vivo model of clot
lysis, Biemond and colleagues50 found that high concentrations of Apo (a) incorporated into clots significantly reduced tPA-induced thrombolysis. Sangrar and
colleagues54 have also demonstrated in vitro that Apo
(a) attenuates t-PA-mediated glu-plasminogen activation. We hypothesize that the Apo E2 isoform, like the
Apo (a) isoforms, may directly or indirectly modify the
thrombolytic activity of t-PA because of its binding
properties, which are different from the Apo E3 and
E4 isoforms.
Testing our hypothesis requires first that our findings are replicated in another group of stroke patients
who are treated with t-PA within 3 hours of onset.
Large thrombolytic studies in patients with myocardial
infarction may have serum samples that could be evaluated for Apo E phenotype to see whether increased
thrombolysis at initial angiography following intravenous t-PA is seen in patients with an Apo E2 phenotype. Some of the in vitro and in vivo studies of Apo
(a) could be replicated with Apo E isoforms. Further
exploration of differential binding of Apo E isoforms,
particularly Apo E2, to endothelial receptors or substances that are part of the thrombolytic and coagulation cascade may provide additional insights. Replication of our findings and identification of an underlying
mechanism could lead to a more effective treatment for
patients with acute ischemic stroke.
This work was supported by NIH contracts NO1-NS-02382, NO1NS-02374, NO1-NS-02377, NO1-NS-02379, NO1-NS-02373,
NO1-NS-02378, NO1-NS-02376, and NO1-NS-02380.
References
1. Corder E, Sanders A, Strittmatter W, et al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in
late onset familites. Science 1993;261:921–923.
2. Payami I, Montee K, Kaye J, et al. Alzheimer’s disease, apolipoprotein E4, and gender. JAMA 1994;271:1316 –1317.
3. Polvikoski T, Sulkava R, Haltia M, et al. Apolipoprotein E,
dementia, and cortical deposition of ␤-amyloid protein. N Engl
J Med 1995;333:1242–1247.
4. Corder E, Saunders A, Risch N, et al. Protective effect of apolipoprotein E type 2 allele for late onset Alzheimer’s disease.
Nature Genet 1994;7:180 –184.
5. Talbot C, Lendon C, Craddock N, et al. Protection against
Alzheimer’s disease with apoE e2. Lancet 1994;343:1432–1433.
6. Bickeboller H, Campion D, Brice A, et al. Apolipoprotein E
and Alzheimer’s disease: genotype-specific risks by age and sex.
Am J Hum Genet 1997;60:439 – 446.
7. Mayeux R, Ottman R, Maestre G, et al. Synergistic effects of
traumatic head injury and apolipoprotein-E4 in patients with
Alzheimer’s disease. Neurology 1995;45:555–557.
8. Henderson A, Easteal S, Jorm A, et al. Apolipoprotein E allele
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
E4, dementia, and cognitive decline in a population sample.
Lancet 1995;346:1387–1390.
Haan M, Shemanski L, Jagust W, et al. The role of APOE e4
in modulating effects of other risk factors for cognitive decline
in elderly persons. JAMA 1999;282:40 – 46.
Sorbi S, Nacmias B, Piacentini S, et al. ApoE as a prognostic
factor for post-traumatic coma. Nature Med 1995;1:852.
Teasdale G, Nicoll J, Murray G, Fiddles M. Association of apolipoprotein E polyporphism with outcome after head injury.
Lancet 1997;350:1069 –1071.
Friedman G, Froom P, Sazbon L, et al. Apolipoprotein E-e4
genotype predicts a poor outcome in survivors of truamatic
brain injury. Neurology 1999;52:244 –248.
McCarron M, Muir K, Weir C, et al. The apolipoprotein E e4
allele and outcome in cerebrovascular disease. Stroke 1998;29:
1882–1887.
Muir K, McCarron M, Curry Y, et al. Apolipoprotein E genotype and functional outcome in ischemic stroke (abstract). Neurology 1999;52:A149.
Sheng H, Laskowitz D, Mackensen G, et al. Apolipoprotein E
deficiency worsens outcome from global cerebral ischemia in
the mouse. Stroke 1999;30:1118 –1124.
Laskowitz D, Sheng H, Bart R, et al. Apolipoprotein E deficient mice have increased susceptibility to focal cerebral ischemia. J Cereb Blood Flow Metab 1997;17:753–758.
Alberts M, Graffafninno C, McClenny C, et al. Effect of apoE
genotype and age on survival after intracerebral hemorrhage
(abstract). Stroke 1996;27:183.
Greenberg S, Briggs M, Hyman B, et al. Apolipoprotein E4 is
associated with the presence and earlier onset of hemorrhage in
cerebral amyloid angiopathy. Stroke 1996;27:1333–1337.
Alberts M. Intracerebral hemorrhage and vascular malformations. In: Alberts M, editor. Genetics of cerebrovascular disease.
Armonk, NY: Futura Publishing Company; 1999:209 –236.
McCarron M, Nicoll J. High frequency of apolipoprotein E e2
allele is specific for patients with cerebral amyloid angiopathyrelated haemorrhage. Neurosci Lett 1998;247:45– 48.
NINDS rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med 1995;333:1581–
1587.
NINDS t-PA Stroke Study Group. Intracerebral hemorrhage
after intravenous t-PA therapy for ischemic stroke. Stroke 1997;
28:2109 –2118.
NINDS t-PA Stroke Study Group. Generalized efficacy of t-PA
for acute stroke: subgroup analysis of the NINDS t-PA Stroke
Trial. Stroke 1997;28:2119 –2125.
NINDS rt-PA Stroke Study Group. Effect of rt-PA on ischemic
stroke lesion size by computed tomography: preliminary results
from the NINDS rt-PA Stroke Trial (abstract). Stroke 1998;29:
287.
Brott T, Adams HJ, Olinger C, et al. Measurements of acute
cerebral infarction: a clinical examination scale. Stroke 1989;20:
864 – 870.
Mahoney F, Barthel D. Functional evaluation: the Barthel index. MD State Med J 1965;14:61– 65.
Van Swieten JC, Koudstaal PJ, Visser MC, et al. Interobserver
agreement for the assessment of handicap in stroke patients.
Stroke 1988;19:604 – 607.
Jennett B, Bond M. Assessment of outcome after severe brain
injury: a practical scale. Lancet 1975;1:480 – 484.
Kataoka S, Paidi M, Howard BV. Simplified isoelectric
focusing/immunoblotting determiniation of apoprotein E phenotype. Clin Chem 1994;40:11–13.
Tilley B, Marler J, Geller N, et al. Use of global test for multiple outcomes in stroke trials with application to the National
Institute of Neurological Disorders and Stroke t-PA Stroke
Trial. Stroke 1996;27:2136 –2142.
31. Fleiss JL. Analysis of data from multiclinic trials. Control Clin
Trials 1986;7:267–275.
32. Marler J, Tilley B, Lu M, et al. Earlier treatment associated
with better outcome: the NINDS t-PA Stroke Study. Neurology (submitted).
33. Bohnet K, Regis-Bailly A, Vincent-Viry M, et al. Apolipoprotein E genotype E4/E2 in the STANISLAS cohort study—
dominance of the E2 allele? Ann Hum Genet 1996;60:509 –
516.
34. Nathan B, Bellosta S, Sanan D, et al. Differential effects of
apolipoproteins E3 and E4 on neuronal growth in vitro. Science 1994;264:850 – 852.
35. Handelmann G, Boyles J, Weisgraber K, et al. Effects of apolipoproteins E, B-very low density lipoproteins and cholesterol
on the extension of neurites by rabbit dorsal root ganglion neurons in vitro. J Lipid Res 1992;33:1677–1688.
36. Holtzmann D, Pitas R, Kilbridege J, et al. Low density lipoprotein receptor-related protein mediates apolipoprotein
E-dependent neurite outgrowth in a central nervous systemderived neuronal cell line. Proc Natl Acad Sci USA 1995;92:
9480 –9484.
37. Strittmatter W, Saunders A, Goedert M, et al. Isoform-specific
interactions of apolipoprotein E with microtubule-associated
protein tau: implications for Alzheimer disease. Proc Natl Acad
Sci USA 1994;91:11183–11186.
38. DeMattos R, Curtiss L, Williams D. A minimally lipidated
form of cell-derived apolipoprotein E exhibits isoform-specific
stimulation of neurite outgrowth in the absence of exogenous
lipids or lipoproteins. J Biol Chem 1998;273:4206 – 4212.
39. Havel R, Yamada N, Shames D. Role of apolipoprotein E in
lipoprotein metabolism. Am Heart J 1987;113:470 – 474.
40. Davignon J, Gregg R, Sing C. Apolipoprotein E polymorphism
and atherosclerosis. Arteriosclerosis 1988;8:1–21.
41. Dong L, Innerarity T, Arnold K, et al. The carboxyl terminus
in apolipoprotein E2 and the seven amino acid repeat in apolipoprotein E-Leiden: role in receptor-binding activity. J Lipid
Res 1998;39:1173–1180.
42. Feussner G, Albanese M, Valencia A. Three-dimensional structure of the LDL receptor-binding domain of the human apolipoprotein E2 variant. Atherosclerosis 1996;126:177–184.
43. Schmidt R, Schmidt H, Fazekas F, et al. Apolipoprotein E
polymorphism and silent microangiopathy-related cerebral
damage. Results of the Austrailian Stroke Prevention Study.
Stroke 1997;28:951–956.
44. Aleshkov S, Abraham C, Zannis V. Interaction of nascent of
ApoE2, ApoE3, and ApoE4 isoforms expressed in mammalian
cells with amyloid peptide beta. Relevance to Alzheimer’s disease. Biochemistry 1997;36:10571–10580.
45. Yang D, Smith J, Zhou Z, et al. Characterization of the binding of amyloid-beta peptide to cell culture-derived native apolipoprotein E2, E3, and E4 isoforms and to isoforms from human plasma. J Neurochem 1997;68:721–725.
46. Kingston I, Castro M, Anderson S. In vitro stimulation of
tissue-type plasminogen activator by Alzheimer amyloid betapeptide analogues. Nature Med 1995;1:138 –142.
47. Wendt S, Wetzels I, Gunzler W. Amyloid beta peptides stimulate tissue-type plasminogen activator but not recombinant
prourokinase. Thromb Res 1997;85:217–224.
48. Paka L, Kako Y, Obunike J, Pillarisetti S. Apolipoprotein E
containing high density lipoprotein stimulates endothelial production of heparan sulfate rich in biologically active heparinlike domains. a potential mechanism for the anti-atherogenic
actions of vascular apolipoprotein e. J Biol Chem 1999;274:
4816 – 4823.
49. del Zoppo G, Higashida R, Furlan A, et al. for the PROCT
Investigators. PROACT: a phase II randomized trial of recom-
Broderick et al: Apo E and t-PA Efficacy
743
50.
51.
52.
53.
54.
binant pro-urokinase by direct arterial delivery in acute middle
cerebral artery stroke. Stroke 1998;29:4 –11.
Biemond B, Friedrich P, Koschinsky M, et al. Apolipoprotein(a) attenuates endogenous fibrinolysis in the rabbit jugular
vein thrombosis model in vivo. Circulation 1997;96:1612–
1615.
Angles-Cano E, Hervio L, Rouy D, et al. Effects of lipoprotein(a) on the binding of plaminogen to fibrin and its activation
by fibrin-bound tissue-type plasminogen activator. Chem Phys
Lipid 1994;67/68:369 –380.
Hervio L, Chapman M, Thillet J, et al. Does the apolipoprotein(a) heterogeneity influence lipoprotein(a) effects on fibrinolysis? Blood 1993;82:392–397.
Palabrica T, Liu A, Aronovitz M, et al. Antifibrinolytic activity
of apolipoprotein(a) in vivo: human apolipoprotein(a) transgenic mice are resistant to tissue plasminogen activatormediated thrombolysis. Nature Med 1995;1:256 –259.
Sangrar W, Bajzar L, Nesheim M, Koschinsky M. Antifibrinolytic effect of recombinant apolipoprotein(a) in vitro is primarily due to attenuation of t-PA-mediated Glu-plasminogen activation. Biochemistry 1995;34:5151–5157.
Appendix 1. List of baseline variables of
interest
1. Age category (years): ⬍50 (50.54), (55.59),
(60.64) (65.69), (70.74), (75.79), (80.84) or
⬎84
2. Race: black, white, and Hispanic
3. Gender
4. Smoking in previous year
5. Drinking problems
6. Drinking in past 24 hours
7. History of diabetes
744
Annals of Neurology
Vol 49
No 6
June 2001
8.
9.
10.
11.
12.
13.
14.
History of hypertension
History of atherosclerosis
History of atrial fibrillation
History of other cardiac disease
Prior stroke
Baseline stroke subtype (TDX1, TDX2, TDX3)
Baseline stroke scale category (NIHSS): ⬍5,
5–10, 11–15, 16 –20, or ⬎20
15. Early CT findings (w/o thrombus)
16. Weight (ranked)
17. Percent of correct dose (ranked)
18. Total dose delivered (ranked)
19,20. Admission blood pressure (MAP) as continuous variable or ⬎130 mmHg
21,22. Baseline blood pressure (MAP) as continuous variable or ⬎130 mmHg
23. Admission systolic BP ⬎190 mmHg
24. Admission diastolic BP ⬎100 mmHg
25. Baseline systolic BP ⬎190 mmHg
26. Baseline diastolic BP ⬎190 mmHg
27. Admission glucose ⬎300 mmHg
28. Hepatic disease
29. Malignancy
30. Hyperlipidemia
31. Prosthetic heart valve
32. Aspirin (NSAID)
33. Heparin
34. Calcium Channel Blocker
35. Time from stroke onset to the treatment
(ranked)
Документ
Категория
Без категории
Просмотров
1
Размер файла
168 Кб
Теги
stroki, efficacy, ischemia, intravenous, phenotypic, plasminogen, apolipoprotein, activator, tissue, acute
1/--страниц
Пожаловаться на содержимое документа