Thrombosis Research (2007) 119, 265 — 274
intl.elsevierhealth.com/journals/thre
REVIEW ARTICLE
What is the optimal pharmacological prophylaxis
for the prevention of deep-vein thrombosis and
pulmonary embolism in patients with acute
ischemic stroke?
Pieter W. Kamphuisen a,b,*, Giancarlo Agnelli a
a
Stroke Unit and Division of Internal and Cardiovascular Medicine, University of Perugia, Perugia, Italy
Division of Vascular Medicine, Department of Internal Medicine, Radboud University Medical Center
Nijmegen, Nijmegen, The Netherlands
b
Received 30 August 2005; received in revised form 22 March 2006; accepted 22 March 2006
Available online 3 May 2006
Abstract
Background: Pulmonary embolism after acute ischemic stroke (AIS) is associated
with a high in-hospital mortality. The benefit from pharmacological prophylaxis for
venous thromboembolism (VTE) is uncertain probably due to doubts about the
optimal agent and dose. We evaluated the benefit/risk ratio of different
anticoagulant regimens in the prevention of VTE in patients with AIS.
Methods: The MEDLINE, EMBASE, and Cochrane Library databases were searched up
to January 2005. Randomized controlled trials (RCT) comparing early administration
of either low-molecular-weight heparin (LMWH) or unfractionated heparin (UFH) with
control were included. Endpoints were objectively diagnosed deep-vein thrombosis
(DVT), pulmonary embolism, intracranial hemorrhage (ICH), and extracranial
hemorrhage (ECH). Low-dose UFH was arbitrarily defined as V 15,000 IU/day, lowdose LMWH as V6000 IU/day or weight-adjusted dose of V 86 IU/kg/day.
Results: Sixteen trials involving 23,043 patients with AIS met the inclusion criteria.
The number of events was small and different doses of anticoagulant treatment
were used. Compared to control, high-dose UFH was associated with a reduction in
pulmonary embolism (OR = 0.49, 95% confidence interval (CI) = 0.29—0.83), but also
with an increased risk of ICH (OR = 3.86, 95% CI = 2.41—6.19) and ECH (OR = 4.74, 95%
CI = 2.88—7.78). Low-dose UFH decreased the thrombosis risk (OR = 0.17, 95%
* Corresponding author. Stroke Unit and Medicina Interna e Cardiovascolare, Universita
` di Perugia, Via Enrico Dal Pozzo, 06126
Perugia, Italy. Tel.: +39 075 5783395; fax: +39 075 5733642.
E-mail address: [email protected] (P.W. Kamphuisen).
0049-3848/$ - see front matter D 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.thromres.2006.03.010
266
P.W. Kamphuisen, G. Agnelli
CI = 0.11—0.26), but had no influence on pulmonary embolism (OR = 0.83, 95%
CI = 0.53—1.31); the risk of ICH or ECH was not statistically significant increased
(OR = 1.67, 95% CI = 0.97—2.87 for ICH; and OR = 1.58, 95% CI = 0.89—2.81 for ECH,
respectively). High-dose LMWH decreased both DVT (OR = 0.07, 95% CI = 0.02—0.29)
and pulmonary embolism (0.44, 95% CI = 0.18—1.11), but this benefit was offset by an
increased risk for ICH (OR = 2.01, 95% CI = 1.02—3.96) and ECH (OR = 1.78, 95%
CI = 0.99—3.17). Low-dose LMWH reduced the incidence of both DVT (OR = 0.34, 95%
CI = 0.19—0.59) and pulmonary embolism (OR = 0.36, 95% CI = 0.15—0.87), without an
increased risk of ICH (OR = 1.39, 95% CI = 0.53—3.67) or ECH (OR = 1.44, 95%
CI = 0.13—16). For low—dose LMWH, the numbers needed to treat were 7 and 38
for DVT and pulmonary embolism, respectively.
Conclusions: Indirect comparison of low and high doses of UFH and LMWH suggests
that low-dose LMWH have the best benefit/risk ratio in patients with acute ischemic
stroke by decreasing the risk of both DVT and pulmonary embolism, without a clear
increase in ICH or ECH.
D 2006 Elsevier Ltd. All rights reserved.
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . .
Methods. . . . . . . . . . . . . . . . . . . . . . . . . . .
Search strategy . . . . . . . . . . . . . . . . . . . . .
Criteria for study selection . . . . . . . . . . . . . .
Outcomes . . . . . . . . . . . . . . . . . . . . . . . .
Data extraction and statistical analysis . . . . . . .
Results . . . . . . . . . . . . . . . . . . . . . . . . . . .
Incidence of DVT . . . . . . . . . . . . . . . . . . . .
Incidence of symptomatic pulmonary embolism. . .
Incidence of symptomatic intracranial hemorrhage.
Incidence of major extracranial hemorrhage . . . .
Discussion. . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgments . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . .
Introduction
Stroke is the third highest cause of death in the
Western world. Pulmonary embolism is a major
contributor to in-hospital death after stroke.
Although the rate of clinically overt pulmonary
embolism after stroke has been estimated to be
less than 1% [1], pulmonary emboli account for up
to 50% of early deaths after stroke [2,3].
Anticoagulant prophylaxis is effective in preventing pulmonary embolism in hospitalized
patients, and reduces mortality related to this
disease after surgery [4]. Current guidelines recommend the use of prophylaxis in stroke patients
with risk factors for venous thromboembolism (VTE)
[5,6]. However, intracranial bleeding seems to be
associated with early anticoagulation in acute
stroke and may outweigh the benefit of prevention
of VTE [7].
The risk of intracranial bleeding after acute
stroke seems to be related to the dose of the used
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
266
267
267
267
267
267
267
268
270
270
272
272
273
273
anticoagulant agent. In the International Stroke
Trial (IST), the risk of major intracranial hemorrhage associated with 25,000 IU/day of unfractionated heparin (UFH) was more than doubled
compared to 10,000 IU/day [1]. Also for lowmolecular-weight heparin (LMWH) there seems to
be a dose-dependent bleeding risk [8]. Although
these two agents have been shown to be effective
in preventing VTE after stroke, the optimal benefit/risk ratio between the risk of VTE and the risk of
bleeding in patients with acute ischemic stroke has
not been established. This ratio could be influenced
by the type and the dose of the prophylactic agent.
There are no clinical trials with adequate sample
sizes that directly compared the clinical benefit
from UFH and LMWH or their doses. Given this
absence, we performed a systematic review to
assess the effectiveness of LMWH or UFH in
reducing the rate of objectively diagnosed deepvein thrombosis (DVT), pulmonary embolism, without an excessive risk of intracranial hemorrhage in
VTE prophylaxis in acute ischemic stroke
patients with acute ischemic stroke. We tried to
identify regimens with a favorable benefit/risk
profile.
Methods
Search strategy
We performed a search for randomized controlled
trials in which early treatment with either UFH or
LMWH was compared with control in patients with
acute ischemic stroke. We looked specifically for
studies reporting on the incidence of DVT or
pulmonary embolism and intracranial and extracranial hemorrhage We searched electronic databases
(MEDLINE and EMBASE) and the Cochrane library for
studies published up to January 2005. Keywords
were cerebrovascular disorders, stroke, venous
thrombosis, thromboembolism, pulmonary embolism, intracranial bleeding, prophylaxis, heparin,
low-molecular-weight heparin, and randomized
controlled trial. Meeting abstracts were scanned,
and the reference lists of the identified articles
were manually checked for additional publications.
Criteria for study selection
We included studies that met the following criteria:
(1) Randomized controlled trials that included at
least one arm of the study with either LMWH or UFH
among patients with acute ischemic stroke. Control
patients were defined as patients that either
received placebo or no treatment. Antiplatelet
treatment was allowed as long as the distribution
of patients who used this treatment was not
different between the comparison groups. (2)
Randomization within 14 days after the acute
ischemic event. (3) Objective diagnosis of venous
thrombosis by 125I-fibrinogen scan, contrast venography, or venous ultrasound of the leg. (4) Pulmonary embolism reported either as primary or
secondary outcome. (5) Intracranial hemorrhage
confirmed by either CT or MRI scanning after
clinical deterioration or at autopsy. The search
was not restricted to studies that performed a
systematic screening for either DVT or pulmonary
embolism.
Outcomes
The primary outcome of this review was the
incidence of adverse clinical outcome events
(DVT, pulmonary embolism, or major intracranial
or extracranial hemorrhage) which occurred dur-
267
ing treatment and follow-up. Major intracranial
bleeding was defined as acute or subacute events
of neurological impairment, with hemorrhage
proven by CT scan or autopsy. Major extracranial
hemorrhage was defined as fatal bleeding or a
bleeding requiring transfusion or hospitalization.
DVT or pulmonary embolism had to be objectively
diagnosed.
Data extraction and statistical analysis
The aim of our review was to compare the
incidence of outcome events in patients receiving
UFH or LMWH compared to control. In addition, we
analyzed the effect of high dose versus low dose of
these agents, again compared to control. From
each study we extracted information on the odds
ratios (OR) and its 95% confidence interval (CI) of
the separate adverse outcome events (DVT, pulmonary embolism, intracranial and extracranial hemorrhage). If the odds ratios and 95% CI were not
provided, they were calculated from the number of
events occurring in the intervention or control
groups. For all analyses, summary estimates were
obtained by taking an inverse variance weighted
average of the log odds ratio from individual
studies. Heterogeneity was assessed using standard
v 2 tests. We compared the occurrence of outcome
events in the treatment group and in the control
group using Cox proportional hazard models. The
number needed to treat and the number needed to
harm were calculated.
We performed subgroup analyses of high- and
low-dose LMWH including heparinoids or UFH. To
facilitate a comparison between different doses,
UFH high dose was arbitrarily defined as a dose
higher than 15,000 IU daily, UFH low dose as a dose
lower than 15,000 IU daily. Low dose of LMWH was
arbitrarily defined as a fixed dose of less than
6000 IU daily. Higher doses of LMWH were considered LMWH high dose. Concerning weight-adjusted
doses of LMWH, 86 IU/kg/day was considered as
LMWH low dose while 86 IU/kg twice a day was
considered LMWH high dose.
Results
We identified 681 potentially eligible articles. After
scanning the abstracts and titles, we selected 26
articles for more detailed evaluation, of which 16
trials fulfilled the predefined inclusion criteria
(Table 1) [1,9—23]. Two of these trials were
published only as an abstract [18,22]. The total
number of patients included in the analysis was
268
Table 1
P.W. Kamphuisen, G. Agnelli
Characteristics of the studies included in the overview
Prophylaxis regimen
Unfractionated heparin
5000 IU or 12,500 IU bd s.c.
5000 IU td s.c.
5000 IU td s.c.
5000 IU td s.c.
5000 IU td s.c.
Dose-adjusted continuous i.v., APTT 50—70 s
Dose-adjusted continuous i.v., APTT 50—70 s
LMWH
Dalteparin 2500 IU bd s.c.
Dalteparin 3000—5500 IU od s.c.
Dalteparin 15,000 IU od s.c.
Nadroparin 4100 IU od or bd s.c.
Nadroparin 15,000 IU bd s.c. for 1 week,
then 7500 IU bd s.c.
Nadroparin 86 IU/kg od or td s.c.
Mesoglycan 50 mg td i.m. (5 days),
then 144 mg od p.o.
Danaparoid 750 IU bd s.c.
Danaparoid 2500 IU i.v. bolus, then
dose-adjusted continuous i.v., antiXa 0.6—0.8
No. of
patients
Time from stroke
to treatment
Treatment
(days)
Follow-up
(months)
Reference
19435
32
65
305
131
45
225
b48
b48
b48
b48
b14
b48
b48
h
h
h
h
days
h
h
14
14
7
14
7
10
7
6
1
12
3
1
0.5
12
IST [1]
McCarthy et al. [9]
Duke et al. [10]
McCarthy et al. [11]
Pambianco et al. [12]
Hakim et al. [13]
Duke et al. [14]
60
103
30
312
120
b72
b72
b48
b48
b48
h
h
h
h
h
14
14
14
14
14
0.5
1
0.5
6
3
Prins et al. [15]
Sandset et al. [16]
Elias et al. [21]
Kay et al. [17]
Kwiecinski et al. [22]
767
57
b24 h
b48 h
14
30
6
1
Hommel et al. [18]
Cazzato et al. [19]
75
1281
b168 h
b24 h
14
7
3
3
Turpie et al. [20]
TOAST [23]
od = once daily; bd = twice daily; td = thrice daily; s.c. = subcutaneous; i.m. = intramuscular; i.v. = intravenous; p.o. = per os.
23,043. Most trials excluded patients with a high
risk of bleeding. In most of the trials prophylaxis
was started within 48 h after stroke onset. 20,238
patients received UFH and 2805 patients LMWH.
Five studies assessed the effect of low-dose UFH
[1,9—12], and 3 studies the effect of high-dose UFH
[1,13,14] (Table 1). In two studies patients were
treated with intravenous UFH [13,14] and in the
other six studies with subcutaneous UFH [1,9—12].
Six studies assessed the effect of low-dose LMWH
[15—20], and five of high-dose LMWH [17,18,21—
23]. Three trials used dalteparin, three nadroparin,
two danaparoid and one a mesoglycan. In one study
LMWH was given intravenously [23]. One study used
bodyweight-adjusted doses of LMWH [18]. Three
studies, FISS [17], FISS-bis [18], and IST [1],
investigated two different treatment doses. In the
IST [1] we selected patients randomised to either
heparin or control, including the patients that were
allocated to aspirin, since aspirin use was the same
in both arms.
The method of randomization was adequate in 10
studies [1,10,14—20,23], while in 5 studies sealed
envelopes were used [9,11—13,21]. In one study, the
method of randomization was unclear [22]. Also not
all studies used a double-blind design [1,13,21,22,
12], which could have influenced the final results.
CT scan of the head to rule out intracranial
hemorrhage was performed in most studies. In two
studies, a CT scan was seldom performed [9,11].
The treatment period in the different trials varied
considerably, from 7 to 30 days (Table 1). In only
one study, treatment duration was longer than 14
days [19]. Also follow-up was generally relatively
short (Table 1), ranging from 14 days to 12 months.
The IST, the by far largest included study had a
duration of treatment of 14 days and a follow-up of
6 months. Overall, only 4 patients were excluded
from the trials after randomization. 91 of the
23,043 patients (0.7%) were lost to follow-up, 83
patients allocated to anticoagulation, and 89 of the
controls. In three patients lost to follow-up the
result of allocation was unknown. In one study by
Kwiecinski et al., no data on follow-up or exclusion
were provided.
Data on the concomitant use of aspirin were
generally insufficiently provided. In most studies,
like the IST, aspirin use was permitted, but exact
numbers of patients using antiplatelet agents were
lacking.
Incidence of DVT
Eleven studies were included in the analysis of the
efficacy of the different prophylactic regimens,
involving 2484 patients. Eight studies systematically sought for asymptomatic and symptomatic DVT.
Six studies investigated DVT by 125I-fibrinogen
scanning [9—11,15,20,21], one by contrast venography [16], and one by Doppler-ultrasound [12].
Three studies defined DVT as a secondary endpoint
and diagnostic tests were only performed at clinical
VTE prophylaxis in acute ischemic stroke
269
suspicion [17,22,23]. Four studies analyzed the
efficacy of low-dose UFH, four studies investigated
low-dose LMWH or high-dose LMWH. There were no
trials that investigated the incidence of DVT with
high-dose UFH. In none of the studies, systematic
screening for DVT was performed after the end of
the treatment or follow-up period.
The total incidence of thrombosis was 4.8% in
patients receiving any type of prophylaxis and 17% in
the control patients (OR = 0.19, 95% CI = 0.13—0.27).
A. Deep Vein Thrombosis
Study
or sub-category
Treatment
n/N
01 High dose anticoagulant prophylaxis
0/15
Elias 1990
0/101
FISS 1995
0/62
Kwiecinski 1995
2/638
TOAST 1998
816
Subtotal (95% CI)
Total events: 2 (Treatment), 26 (Control)
2
Test for heterogeneity: Chi = 3.22, df = 3 (P = 0.36), I2 = 6.9%
Test for overall effect: Z = 4.06 (P < 0.0001)
Control
n/N
OR (fixed)
95% CI
Weight
%
12/15
1/53
3/58
10/628
754
02 Low dose anticoagulant prophylaxis
2/16
12/16
McCarthy 1977
0/35
3/30
Duke 1980
32/144
117/161
McCarthy 1986
2/50
7/25
Turpie 1987
6/30
15/30
Prins 1989
15/42
17/50
Sanset 1990
0/101
1/53
FISS 1995
3/64
3/67
Pambianco 1995
482
432
Subtotal (95% CI)
Total events: 60 (Treatment), 175 (Control)
2
2
Test for heterogeneity: Chi = 27.19, df = 7 (P = 0.0003), I = 74.3%
Test for overall effect: Z = 8.55 (P < 0.00001)
1298
Total (95% CI)
Total events: 62 (Treatment), 201 (Control)
2
Test for heterogeneity: Chi = 31.17, df = 11 (P = 0.001), I2 = 64.7%
Test for overall effect: Z = 9.56 (P < 0.00001)
7.41
1.19
2.19
6.14
16.94
0.01
0.17
0.13
0.19
0.10
[0.00,
[0.01,
[0.01,
[0.04,
[0.03,
0.19]
4.31]
2.51]
0.89]
0.31]
6.42
2.27
52.55
5.48
7.34
6.10
1.19
1.71
83.06
0.05
0.11
0.11
0.11
0.25
1.08
0.17
1.05
0.21
[0.01,
[0.01,
[0.06,
[0.02,
[0.08,
[0.46,
[0.01,
[0.20,
[0.14,
0.31]
2.23]
0.18]
0.56]
0.79]
2.55]
4.31]
5.40]
0.30]
100.00
1186
0.01
0.1
1
Favours treatment
10
OR (fixed)
95% CI
0.19 [0.13, 0.27]
100
Favours control
B. Pulmonary Embolism
Study
or sub-category
Treatment
n/N
01 High dose anticoagulant prophylaxis
0/24
Hakim 1983
1/5
Elias 1990
0/112
FISS 1995
0/62
Kwiecinski 1995
20/4860
IST 1997
5/245
FISS-bis 1998
2/638
TOAST 1998
5946
Subtotal (95% CI)
Total events: 28 (Treatment), 55 (Control)
2
Test for heterogeneity: Chi = 2.37, df = 4 (P = 0.67), I2 = 0%
Test for overall effect: Z = 3.23 (P = 0.001)
02 Low dose anticoagulant prophylaxis
0/50
Turpie 1987
1/28
Cazzato 1989
1/30
Prins 1989
2/42
Sanset 1990
0/101
FISS 1995
2/64
Pambianco 1995
33/4856
IST 1997
4/272
FISS-bis 1998
5443
Subtotal (95% CI)
Total events: 43 (Treatment), 56 (Control)
Test for heterogeneity: Chi2 = 7.34, df = 7 (P = 0.39), I2 = 4.6%
Test for overall effect: Z = 1.79 (P = 0.07)
11389
Total (95% CI)
Total events: 71 (Treatment), 111 (Control)
2
Test for heterogeneity: Chi = 11.36, df = 12 (P = 0.50), I2 = 0%
Test for overall effect: Z = 3.51 (P = 0.0004)
Control
n/N
OR (fixed)
95% CI
Weight
%
OR (fixed)
95% CI
0/21
1/15
0/53
2/58
41/4859
7/125
4/628
5759
2.18
34.79
7.74
3.42
48.48
Not estimable
3.50 [0.18, 69.34]
Not estimable
0.18 [0.01, 3.85]
0.49 [0.28, 0.83]
0.35 [0.11, 1.13]
0.49 [0.09, 2.69]
0.47 [0.30, 0.74]
2/25
0/29
2/30
2/50
1/53
1/67
41/4859
7/125
5238
2.79
0.40
1.65
1.48
1.66
0.81
34.68
8.05
51.52
0.09
3.22
0.48
1.20
0.17
2.13
0.80
0.25
0.70
0.34
10997
100.00
0.01
0.1
Favours treatment
1
10
[0.00,
[0.13,
[0.04,
[0.16,
[0.01,
[0.19,
[0.51,
[0.07,
[0.47,
2.02]
82.38]
5.63]
8.91]
4.31]
24.07]
1.27]
0.88]
1.03]
0.59 [0.44, 0.79]
100
Favours control
Figure 1 Odds ratio’s for DVT (A) and pulmonary embolism (B) in controls and patients receiving low-dose and highdose anticoagulation after acute ischemic stroke.
270
P.W. Kamphuisen, G. Agnelli
Fig. 1A shows the odds ratio’s of DVT for low dose and
high dose of anticoagulant prophylaxis. There was
significant heterogeneity between the different
studies (v 2 = 31.2, p = 0.001), probably caused by the
different diagnostic methods. Both high- and lowdose prophylaxis reduced DVT after ischemic stroke.
In patients assigned to low-dose UFH, the
incidence of DVT was 14%, whereas this was 49%
in patients assigned to placebo, resulting in an odds
ratio of 0.17 (95% CI = 0.11—0.26, number needed to
treat = 3) (Table 2). The incidence of DVT was 10% in
the patients receiving low-dose LMWH and 25% in
patients receiving placebo (OR = 0.34, 95%
CI = 0.19—0.59, number needed to treat = 7). No
studies on high-dose UFH reported data on DVT.
The incidence of DVT was 0.2% in the patients
receiving high-dose LMWH and 3.4% in patients
receiving placebo (OR = 0.07, 95% CI = 0.02—0.29,
number needed to treat = 31).
Furthermore, we restricted our analysis to the
incidence of symptomatic or proximal DVT. Trials
with low-dose heparin did not report these incidences separately. In the two trials with low-dose
LMWH in which the occurrence of symptomatic or
proximal DVT was reported [16,17], this incidence
was 8.4% in patients receiving LMWH and 16% in
patients receiving placebo, which resulted in an OR
of 0.46 (95% CI = 0.21—1.02, number needed to
treat = 13). The incidence of symptomatic or proximal DVT in patients with high-dose LMWH was
reported in three trials [17,22,23], and was 0.2%
among patients receiving LMWH and 1.9% in
patients receiving placebo (OR = 0.13, 95%
CI = 0.03—0.57, number needed to treat = 59).
Incidence of symptomatic pulmonary
embolism
Eleven studies including 22,386 patients provided
data on the incidence of pulmonary embolism. The
diagnostic method for confirmation of pulmonary
embolism was seldom mentioned in the different
trials and, in all trials, it was merely a secondary
outcome. The overall incidence of pulmonary embolism was low, 0.6% in the treatment group and
1.0% in the control patients (OR = 0.59, 95%
CI = 0.44—0.79). Two studies reported the efficacy
of low-dose UFH or high-dose UFH. Six studies
investigated low-dose LMWH and four studies highdose LMWH. There was no statistical heterogeneity
between the different studies.
Fig. 1B shows the results of low- and high-dose
anticoagulant prophylaxis on the incidence of pulmonary embolism. Both high- and low-dose anticoagulants reduced pulmonary embolic events.
When we calculated dose and type of anticoagulant
separately, the incidence of pulmonary embolism
was 0.7% in patients receiving low-dose UFH and
0.9% for controls, resulting in an odds ratio of 0.83
(95% CI = 0.53—1.31) (Table 2). High-dose UFH was
associated with an OR of 0.49 (95% CI = 0.29—0.83,
number needed to treat = 277). Low-dose LMWH
resulted in a decrease of pulmonary embolism, with
an incidence of 1.5% in patients receiving LMWH and
4.1% in patients receiving placebo or no treatment
(OR = 0.36, 95% CI = 0.15—0.87, number needed to
treat = 38). Finally, the incidence of symptomatic
pulmonary embolism was 0.7% in patients receiving
high-dose LMWH and 1.5% in controls (OR = 0.44, 95%
CI = 0.18—1.11, number needed to treat = 120).
Most episodes of pulmonary embolism were
reported as non-fatal. From the IST [1], it was
possible to distinguish fatal and non-fatal events
within the 14 days of treatment. The incidence of
non-fatal pulmonary embolism was the same for
low- or high-dose heparin (0.16%), and 0.4% for
patients not receiving heparin (OR = 0.38, 95%
CI = 0.17—0.86). It should be noted, however, that
half of these patients, equally distributed, used 300
mg of aspirin.
Incidence of symptomatic intracranial
hemorrhage
From twelve studies data on symptomatic intracranial hemorrhage confirmed by CT scan or autopsy
Table 2 Relative risk of deep-vein thrombosis, pulmonary embolism, intracranial hemorrhage and extracranial
hemorrhage associated with prophylaxis of low-dose or high-dose unfractionated heparin or low-molecular-weight
heparin compared to no prophylaxis after acute ischemic stroke
DVT
OR
Low-dose UFH
High-dose UFH
Low-dose LMWH
High-dose LMWH
PE
95% CI
0.17
0.11—0.26
No studies
0.34
0.19—0.59
0.07
0.02—0.29
ICH
ECH
OR
95% CI
OR
95% CI
OR
95% CI
0.83
0.49
0.36
0.44
0.53—1.31
0.29—0.83
0.15—0.87
0.18—1.11
1.67
3.86
1.39
2.01
0.97—2.87
2.41—6.19
0.53—3.67
1.02—3.96
1.58
4.74
1.44
1.78
0.89—2.81
2.88—7.78
0.13—16
0.99—3.17
DVT = deep-vein thrombosis; PE = pulmonary embolism; ICH = intracranial hemorrhage; ECH = extracranial hemorrhage; OR = odds
ratio; CI = confidence interval.
VTE prophylaxis in acute ischemic stroke
271
were available (22,471 patients). Four of the
twelve studies systematically repeated a CT scan,
usually at the end of the treatment period [13,15—
17]. Two studies reported the safety of low-dose
UFH and three of high-dose UFH. For LMWH, six
studies investigated low dose and three studies high
dose. Sensitivity analysis showed no significant
heterogeneity between the results of the different
studies analyzing the incidence of ICH.
The overall incidence of intracranial bleeding
was 1.4% in the treatment group and 0.5% in the
control patients (OR = 2.39, 95% CI = 1.78—3.20).
Fig. 2A shows that especially high-dose anticoagulation increased the risk of intracranial bleeding.
A. Intracranial Hemorrhage
Study
or sub-category
Treatment
n/N
Control
n/N
OR (fixed)
95% CI
01 High dose anticoagulant prophylaxis
0/24
0/21
Hakim 1983
0/112
0/113
Duke 1986
0/101
1/53
FISS 1995
85/4856
21/4859
IST 1997
15/245
4/125
FISS-bis 1998
14/638
7/628
TOAST 1998
5976
5799
Subtotal (95% CI)
Total events: 114 (Treatment), 33 (Control)
Test for heterogeneity: Chi2 = 5.94, df = 3 (P = 0.11), I2 = 49.5%
Test for overall effect: Z = 5.82 (P < 0.00001)
02 Low dose anticoagulant prophylaxis
1/50
Turpie 1987
0/28
Cazzato 1989
1/30
Prins 1989
2/50
Sanset 1990
0/101
FISS 1995
0/64
Pambianco 1995
35/4860
IST 1997
10/272
FISS-bis 1998
5455
Subtotal (95% CI)
Total events: 49 (Treatment), 27 (Control)
2
Test for heterogeneity: Chi = 2.35, df = 5 (P = 0.80), I2 = 0%
Test for overall effect: Z = 1.77 (P = 0.08)
11431
Total (95% CI)
Total events: 163 (Treatment), 60 (Control)
2
Test for heterogeneity: Chi = 13.52, df = 9 (P = 0.14), I2 = 33.4%
Test for overall effect: Z = 5.82 (P < 0.00001)
Weight
%
OR (fixed)
95% CI
3.02
31.94
7.70
10.68
53.34
Not estimable
Not estimable
0.17 [0.01, 4.31]
4.10 [2.54, 6.63]
1.97 [0.64, 6.07]
1.99 [0.80, 4.96]
3.15 [2.14, 4.64]
1.00
0/25
0/29
0/30
1/53
1/53
0/67
21/4859
4/125
5241
0.74
1.44
3.02
32.28
8.17
46.66
100.00
11040
0.01
0.1
1
Favours treatment
1.55 [0.06, 39.31]
Not estimable
3.10 [0.12, 79.23]
2.17 [0.19, 24.67]
0.17 [0.01, 4.31]
Not estimable
1.67 [0.97, 2.87]
1.15 [0.36, 3.76]
1.52 [0.96, 2.41]
2.39 [1.78, 3.20]
10
100
Favours control
B. Extracranial Hemorrhage
Study
or sub-category
Treatment
n/N
Control
n/N
OR (fixed)
95% CI
01 High dose anticoagulant prophylaxis
0/24
0/21
Hakim 1983
0/102
1/53
FISS 1995
90/4856
19/4859
IST 1997
34/646
17/635
TOAST 1998
5628
5568
Subtotal (95% CI)
Total events: 124 (Treatment), 37 (Control)
2
2
Test for heterogeneity: Chi = 8.12, df = 2 (P = 0.02), I = 75.4%
Test for overall effect: Z = 6.42 (P < 0.00001)
02 Low dose anticoagulant prophylaxis
0/35
Duke 1980
0/50
Turpie 1987
1/28
Cazzato 1989
0/30
Prins 1989
0/52
Sanset 1990
1/101
FISS 1995
0/64
Pambianco 1995
30/4856
IST 1997
5216
Subtotal (95% CI)
Total events: 32 (Treatment), 19 (Control)
Test for heterogeneity: Chi2 = 0.18, df = 2 (P = 0.91), I2 = 0%
Test for overall effect: Z = 1.70 (P = 0.09)
10844
Total (95% CI)
Total events: 156 (Treatment), 56 (Control)
Test for heterogeneity: Chi2 = 12.40, df = 5 (P = 0.03), I2 = 59.7%
Test for overall effect: Z = 6.49 (P < 0.00001)
0/21
0/25
0/29
0/30
0/51
0/53
0/67
19/4859
5135
Weight
%
OR (fixed)
95% CI
3.45
32.80
28.58
64.83
Not estimable
0.17 [0.01, 4.26]
4.81 [2.93, 7.90]
2.02 [1.12, 3.65]
3.33 [2.31, 4.81]
33.22
35.17
Not estimable
Not estimable
3.22 [0.13, 82.38]
Not estimable
Not estimable
1.60 [0.06, 39.88]
Not estimable
1.58 [0.89, 2.82]
1.62 [0.93, 2.83]
100.00
2.73 [2.02, 3.70]
0.82
1.13
10703
0.01
0.1
1
Favours treatment
10
100
Favours control
Figure 2 Odds ratio’s for major intracranial hemorrhage (A) and extracranial hemorrhage (B) in controls and patients
receiving low-dose and high-dose anticoagulation after acute ischemic stroke.
272
High-dose UFH treatment resulted in a nearly
fourfold increased bleeding risk compared to placebo (incidence 1.7% in patients and 0.4% in
controls, OR = 3.86, 95% CI = 2.41—6.19, number
needed to harm = 77) (Table 2). The incidence of
intracranial bleeding was 0.7% in 4924 patients
receiving low-dose UFH and 0.4% in 4926 controls
(OR = 1.67, 95% CI = 0.97—2.87). For LMWH, high
dose doubled the risk of symptomatic intracranial
bleeding (2.6% in patients and 1.5% in controls,
OR = 2.01, 95% CI = 1.02—3.96, number needed to
harm = 91). Low-dose LMWH did not seem to
influence the risk of hemorrhage (2.6% in patients
and 1.9% in controls, OR = 1.39, 95% CI = 0.53—3.67).
Incidence of major extracranial hemorrhage
Ten trials from 21,547 patients reported data on
major extracranial hemorrhage. There was statistical heterogeneity between the studies (v 2 = 31.2,
p = 0.001). High dose seemed more harmful than
low-dose anticoagulation (Fig. 2B). Both low-dose
LMWH and UFH did not increase the risk of major
extracranial bleeding, but incidences were low.
The risk associated with high-dose UFH was nearly
five-fold increased (OR = 4.74, 95% CI = 2.88—7.78,
number needed to harm = 71), and high-dose LMWH
doubled the risk of major extracranial hemorrhage
(OR = 1.78, 95% CI = 0.99—3.17, number needed to
harm = 53).
Discussion
When dealing with the balance between efficacy
and safety of a pharmacological agent for a
specific indication, the treatment dose is crucial;
this is particular to the case for anticoagulant
agents whose efficacy and safety are closely
related to the dose and route of administration.
This review was performed to assess the risk/
benefit ratio of low or medium/high dose of UFH
or LMWH in patients with ischemic stroke and
focused on adverse clinical outcomes like DVT,
pulmonary embolism, and intracranial and extracranial hemorrhage. The main conclusion of our
pooled analysis across different studies is that
low-dose LMWH seems to be associated with the
best benefit/risk ratio. With this regimen, the
numbers needed to treat to prevent a DVT or
pulmonary embolism were 7 and 38, respectively,
while risk of both ICH and ECH was not significantly increased. The other three regimens performed less well as either they increased the
bleeding risk (high-dose UFH and LMWH) or had no
P.W. Kamphuisen, G. Agnelli
obvious effect on pulmonary embolism (low-dose
UFH). Based on these figures, low-dose LMWH
seems to be the most adequate for prophylaxis of
venous thromboembolism in stroke patients.
We have not performed direct comparisons of
UFH and LMWH and we cannot, therefore, conclude with certainty that one anticoagulant regimen is better than the other. The results of our
present study are however supporting the results
of a Cochrane review that evaluated five randomized trials directly comparing LMWH with UFH in
705 stroke patients [24]. LMWH significantly
reduced the incidence of DVT by approximately
50% without increasing the rate of symptomatic
intracranial and extracranial hemorrhage. Very
recently, the PROTECT trial, comparing 3000 U of
the LMWH certoparin once daily and 5000 U UFH
twice daily in 545 patients with acute ischemic
stroke, showed that certoparin was associated
with an incidence of DVT, pulmonary embolism
or death of 7%, compared to 9.7% in the UFH
group ( p = 0.0011), without a difference in bleeding complications [25].
Nearly half of the patients with ischemic stroke
who developed pulmonary embolism die [26]. In
daily practice the clinical burden of pulmonary
embolism in patients with stroke is, however,
underestimated since the clinical symptoms of
stroke may obscure the recognition of this complication. Venous thromboembolism is clinically
recognized in about one quarter of patients with
stroke ultimately found to have DVT or pulmonary
embolism when screened with objective testing
[27]. This is one of the problems we encountered
in the interpretation of the different trials.
Considering the difficulty of diagnosing VTE in
patients with stroke, systematic screening for
these events would have produced more reliable
incidences. However, the outcomes of efficacy
and safety identified for our analysis were not
always (DVT) or even seldom (pulmonary embolism) the primary study outcomes in the individual
trials. This could have caused an underreporting
of these events with consequent underestimation
of their incidence. Furthermore, there is also a
risk of bias due to non-blinding. Another point of
concern is the heterogeneity in the dose of
anticoagulation, conditional on the purpose of
the study. Some studies aimed at preventing
ischemic stroke progression and used therapeutic
doses of UFH or LMWH, while other trials evaluated the prevention of venous thromboembolism
after ischemic stroke with prophylactic doses
[28]. In addition, for the purpose of our analysis
between low and medium/high dose of anticoagulation, we used artificial cut-off levels. The
VTE prophylaxis in acute ischemic stroke
cut-off dose to separate low and high dose of
LMWH (6000 IU/daily) is higher than the prophylactic dose mostly used in clinical practice.
Finally, the duration of antithrombotic prophylaxis
in almost all studies was 14 days or less. Since the
occurrence of pulmonary embolism peaks at 2—4
weeks after stroke onset [29], it is possible that
the duration of prophylaxis was in most of the
cases too short to provide full benefit. Notwithstanding these uncertainties, we think that especially the analysis of the safety of anticoagulation
is solid, since the incidence of ICH was the
primary outcome in all included trials. Since
prevention of DVT with anticoagulant prophylaxis
has been well-established, the favorable safety
profile of low-dose LMWH or UFH makes them
useful as prophylactic agents in patients with
acute ischemic stroke who have risk factors for
VTE, as propagated by the American College of
Chest Physicians (ACCP) [5] and the Stroke Council
of the American Stroke Association [6]. A recently
published Cochrane review on anticoagulants for
acute ischemic stroke did not support the routine
use of pharmacological prophylaxis in acute
ischemic stroke, but the included trials were not
analyzed for different doses of the anticoagulant
agent [7]. Our results thereby exceed these
observations, since we encountered for several
treatment doses and we tried to identify an
optimal benefit/risk ratio.
A considerable amount of patients included in
our analysis did not receive an antiplatelet agent in
association with LMWH or UFH. Since aspirin is
currently recommended in patients with acute
stroke, the question is whether LMWH or UFH are
safe when given in association with aspirin. On the
other hand, the combination of aspirin and lowdose LMWH might be more beneficial in the
prevention of venous thromboembolism than aspirin alone, as it was suggested in a review by the
Cochrane collaboration [30].
Although the results of our analysis advocate the
use of low-dose LMWH in patients with acute stroke
to prevent venous thromboembolism, it may be
argued that compression stockings are sufficient.
However, while waiting for clinical trials showing a
definitive benefit from elastic stockings, these
should be recommended as prophylaxis only in
patients who have a high bleeding risk associated
with LMWH [5,31—33].
In conclusion, a low dose of LMWH or UFH seems
effective and relatively safe as a prophylactic
agent to prevent venous thromboembolism in
patients with ischemic stroke. In these patients,
the benefit—risk profile seems best for low-dose
LMWH.
273
Acknowledgments
Pieter W. Kamphuisen was financially supported by
an unrestricted grant from the Niels Stensen
Foundation, The Netherlands.
References
[1] The International Stroke Trial (IST): a randomised trial of
aspirin, subcutaneous heparin, both, or neither among
19 435 patients with acute ischaemic stroke. International Stroke Trial Collaborative Group. Lancet 1997;349:
1569 – 81.
[2] Bounds JV, Wiebers DO, Whisnant JP, Okazaki H. Mechanisms and timing of deaths from cerebral infarction. Stroke
1981;12:474 – 7.
[3] Wijdicks E, Scott J. Pulmonary embolism associated with
acute stroke. Mayo Clin Proc 1997;72:297 – 300.
[4] Collins R, Scrimgeour A, Yusuf S, Peto R. Reduction in fatal
pulmonary embolism and venous thrombosis by perioperative administration of subcutaneous heparin. N Engl J Med
1988;318:1162 – 73.
[5] Geerts WH, Pineo GF, Heit JA, Bergqvist D, Lassen MR,
Colwell CW, et al. Prevention of venous thromboembolism:
the Seventh ACCP Conference on Antithrombotic and
Thrombolytic Therapy. Chest 2004;126:338S – 400S.
[6] Adams HP, Adams RJ, Brott T, del Zoppo GJ, Furlan A,
Goldstein LB, et al. Stroke council of the American Stroke
Association. Guidelines for the early management of
patients with ischemic stroke. A scientific statement from
the Stroke Council of the American Stroke Association.
Stroke 2003;34:1056 – 83.
[7] Gubitz G, Sandercock P, Counsell C. Anticoagulants for
acute ischaemic stroke (Cochrane Review). The Cochrane
Library, Issue 4. Chichester, UK7 John Wiley and Sons, Ltd.;
2005.
[8] Diener HC, Ringelstein EB, von Kummer R, Langohr HD,
Bewermeyer H, Landgraf H, et al. Treatment of acute
ischemic stroke with the low-molecular-weight heparin
certoparin: results of the TOPAS trial. Therapy of Patients
With Acute Stroke (TOPAS) Investigators. Stroke 2001;32:
22 – 9.
[9] McCarthy ST, Turner JJ, Robertson D, Hawkey CJ, Macey DJ.
Low dose heparin as a prophylaxis against deep-vein
thrombosis after acute stroke. Lancet 1977;ii:800 – 1.
[10] Duke RJ, Turpie AGG, Bloch RF, Trebilcock RG, Reivich M,
Hurtig HI. Clinical trial of low-dose subcutaneous heparin
for the prevention of stroke progression: natural history of
acute partial stroke and stroke-in-evolution. Cerebrovasc
Dis. New York7 Raven Press;, 1983. p. 399 – 405.
[11] McCarthy ST, Turner J. Low-dose subcutaneous heparin in
the prevention of deep-vein thrombosis and pulmonary
emboli following acute stroke. Age Ageing 1986;15:84 – 8.
[12] Pambianco G, Orchard T, Landau P. Deep vein thrombosis:
prevention in stroke patients during rehabilitation. Arch
Phys Med Rehabil 1995;76:324 – 30.
[13] Hakim AM, Ryder-Cooke A, Melanson D. Sequential computerized tomographic appearances of strokes. Stroke
1983;14:893 – 7.
[14] Duke RJ, Bloch RF, Turpie AGG, Trebilcock RG, Bayer N.
Intravenous heparin for the prevention of stroke progression in acute partial stable stroke. Ann Intern Med 1986;
105:825 – 8.
274
[15] Prins MH, Gelsema R, Sing AK, van Heerde LR, den
Ottolander GJ. Prophylaxis of deep venous thrombosis
with a low-molecular-weight heparin (Kabi 2165/Fragmin)
in stroke patients. Haemostasis 1989;19:245 – 50.
[16] Sandset PM, Dahl T, Stiris M, Rostad B, Scheel B, Abildgaard
U. A double-blind and randomized placebo-controlled trial
of low molecular weight heparin once daily to prevent
deep-vein thrombosis in acute ischemic stroke. Semin
Thromb Hemost 1990;16:25 – 33 [suppl].
[17] Kay R, Wong KS, Yu YL, Chan YW, Tsoi TH, Ahuja AT, et al.
Low-molecular-weight heparin for the treatment of acute
ischemic stroke. N Engl J Med 1995;333:1588 – 93.
[18] Hommel Mfor the FISS bis Investigators Group. Fraxiparine
in ischaemic stroke study (FISS bis). Cerebrovasc Dis
1998;8:19 [Abstract].
[19] Cazzato G, Zorzon M, Mase G, Antonutti L, Iona LG. Il
mesoglicano nelle ischemie cerebrali acute a focolaio
[Mesoglycan in acute focal cerebral ischemia]. Rev Neurol
1989;59:121 – 6.
[20] Turpie AG, Levine MN, Hirsh J, Carter CJ, Jay RM, Powers
PJ, et al. Double-blind randomised trial of ORG 10172
low-molecular-weight heparinoid in prevention of deepvein thrombosis in thrombotic stroke. Lancet 1987;1:
523 – 6.
[21] Elias A, Milandre L, Lagrange G, Aillaud MF, Alonzo B,
Toulemonde F, et al. Prevention of deep venous thrombosis
of the leg by a very low molecular weight heparin fraction
(CY 222) in patients with hemiplegia following cerebral
infarction: a randomized pilot study (30 patients). Rev Med
Interne 1990;11:95 – 8.
[22] Kwiecinski H, Pniewski J, Kaminska A, Szyluk B. A randomized trial of fraxiparine in acute ischemic stroke. Cerebrovasc Dis 1995;5:234 [Abstract].
[23] The Trial of ORG 10172 in Acute Stroke Treatment (TOAST)
Investigators B. Low molecular weight heparinoid, ORG
10172 (danaparoid), and outcome after acute ischemic
stroke. JAMA 1998;279:1265 – 72.
P.W. Kamphuisen, G. Agnelli
[24] Counsell C, Sandercock P. Low-molecular-weight heparins
or heparinoids versus standard unfractionated heparin for
acute ischaemic stroke (Cochrane Review). The Cochrane
Library, Issue 4. Chichester, UK7 John Wiley and Sons, Ltd.;
2005.
[25] Diener HC, Ringelstein EB, von Kummer R, Landgraf H,
Koppenhagen K, Harenberg J, et al. Prophylaxis of thrombotic and embolic events in acute ischemic stroke with the
low-molecular-weight heparin certoparin: results of the
PROTECT trial. Stroke 2006;37:139 – 44.
[26] Heuschmann PU, Kolominsky-Rabas PL, Misselwitz B, Hermanek P, Leffmann C, Janzen RW, et alGerman Stroke
Registers Study Group. Predictors of in-hospital mortality
and attributable risks of death after ischemic stroke: the
German Stroke Registers Study Group. Arch Intern Med
2004;164:1761 – 8.
[27] Kelly J, Rudd A, Lewis RR, Coshall C, Moody A, Hunt BJ.
Venous thromboembolism after acute ischemic stroke: a
prospective study using magnetic resonance direct thrombus imaging. Stroke 2004;35:2320 – 5.
[28] Kamphuisen PW, Agnelli G, Sebastianelli M. Prevention
of venous thromboembolism after acute ischemic stroke.
J Thromb Haemost 2005;3:1187 – 94.
[29] Viitanen M, Winblad B, Asplund K. Autopsy-verified causes
of death after stroke. Acta Med Scand 1987;222:401 – 8.
[30] Berge E, Sandercock P. Anticoagulants versus antiplatelet
agents for acute ischaemic stroke (Cochrane Review). The
Cochrane Library, Issue 4. Chichester, UK7 John Wiley and
Sons, Ltd.; 2005.
[31] Kelly J, Rudd A, Lewis R, Hunt BJ. Venous thromboembolism
after acute stroke. Stroke 2001;32:262 – 7.
[32] Adams Jr HP. Effective prophylaxis for deep vein thrombosis
after stroke: low-dose anticoagulation rather than stockings alone: for. Stroke 2004;35:2911 – 2.
[33] Dennis MS. Effective prophylaxis for deep vein thrombosis
after stroke: low-dose anticoagulation rather than stockings alone: against. Stroke 2004;35:2912 – 3.