Hemorrhagic Transformation in Acute Ischemic Stroke

The MAST-E Study

Assia Jaillard, MD; Catherine Cornu, MD; Anne Durieux, MD; Thierry Moulin, MD;
Florent Boutitie, PhD; Kennedy R. Lees, MD; Marc Hommel, MD; on behalf of the MAST-E Group

Background and PurposeHemorrhagic transformation (HT) is the most critical complication of thrombolytics in clinical
trials in acute stroke. The aim of this study was to determine the rates and the predictors of HT in the Multicenter Acute
Stroke TrialEurope (MAST-E) study.
MethodsWe performed a post hoc analysis of MAST-E data designed to assess the safety and efficacy of streptokinase
administered intravenously within 6 hours of stroke onset. HT included all intracerebral hemorrhages and symptomatic
hemorrhages (SHT) associated with clinical worsening. The predictors of HT and SHT were determined using
multivariate modeling.
ResultsAmong the 310 patients included, 159 patients had HT and 37 SHT (97 and 33 in the streptokinase group and
62 and 4 in the placebo group, respectively). Patients with SHT had significantly more atrial fibrillation, diabetes
Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017

mellitus, no heparin use, streptokinase treatment, and early CT signs. In the multivariate analysis, HT was predicted by
early CT signs and streptokinase treatment. SHT was predicted by diabetes mellitus, early CT signs, streptokinase
treatment, and the interaction between streptokinase treatment and decreased level of consciousness. Among the
streptokinase-treated patients, the same predictors remained.
ConclusionsThe relative risks of HT after streptokinase were in the same range in MAST-E as in other streptokinase and
tPA trials. Early CT signs were strong predictors of both HT and SHT, stressing that these patients are at high risk of
bleeding. In our study, the predictors of HT and SHT were similar to those of tPA trials in acute stroke. (Stroke.
1999;30:1326-1332.)
Key Words: cerebral hemorrhage n clinical trials n streptokinase n stroke, acute

U ntil recently, the incidence of intracranial bleeding after

acute ischemic stroke was not well known. It could vary
between 18% and 71%,1,2 4 according to whether the infor-
mation was from autopsy or CT scan studies. Hemorrhagic
transformation (HT), either in control groups or after
thrombolytic treatment, has been assessed in recent large
clinical trials in acute stroke, with systematic admission and
control CT scan assessment. HT includes symptomatic hem-
orrhages (SHT) that are associated with clinical worsening
and those that are asymptomatic.59 The risk of bleeding is
present for both streptokinase and tPA, because cerebral HT
accounts for a major cause of early hazard in acute-stroke
patients, ranging from 26 to 154 extra deaths for each 1000
patients treated in the recent studies.5,6,8 10 Predictors of HT
have been evidenced in 2 tPA clinical trials11,12 but have not
yet been reported in a streptokinase trial. Streptokinase is
widely used in the treatment of myocardial infarction (MI),
and cerebral infarction is one of the major complications after
MI. Because the major risk carried by a thrombolytic agent,
either in stroke or MI, is cerebral bleeding, to delineate the
profile of patients having a high risk of bleeding and to
identify predictors of HT remains a major issue. We per-
formed a post hoc analysis of the Multicenter Acute Stroke
TrialEurope (MAST-E) data to study the occurrence and the
predictors of cerebral HT in both the streptokinase and
placebo groups.
Subjects and Methods
MAST-E was a multicenter, double-blind, controlled clinical trial
designed to assess the safety and the efficacy of streptokinase in
acute cerebral ischemia. The design and the results of the study have
been reported elsewhere.7,13 Briefly, patients were eligible if they
presented with clinical signs attributed to ischemia in the middle
cerebral artery (MCA) territory. Either 1 500 000 U of streptokinase
or placebo were administered intravenously after a CT scan within a
maximum of 6 hours after the onset of symptoms. Patients with signs
of hemorrhagic stroke on prerandomization CT were excluded.
Efficacy was evaluated on a binary criterion combining mortality and

Received February 16, 1999; accepted March 30, 1999.

From the Department of Clinical and Biological Neurosciences, Stroke Unit, INSERM U 438, University Hospital, Grenoble, France (A.J., M.H.);
Service de Pharmacologie, Clinique EA 643, Claude Bernard University, Lyon, France (C.C., F.B.); Service de Neurologie, University Hospital,
Clermont-Ferrand, France (A.D.); Service de Neurologie, University Hospital, Besancon, France (T.M.); and the Acute Stroke Unit, University
Department of Medicine and Therapeutics, Western Infirmary, Glasgow, UK (K.R.L.).
Correspondence to Assia Jaillard, Service de NeurologieUnite dUrgences Cerebrovasculaires, Centre Hospitalier Universitaire de Grenoble, BP
217-38043 Grenoble Cedex, France. E-mail Assia.Jaillard@ujf-grenoble.fr
1999 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org

1326
 Jaillard et al Hemorrhagic Transformation 1327

severe disability at 6 months. Safety was evaluated on the mortality

at 10 days, and symptomatic intracerebral hemorrhage and clinically
silent intracerebral hemorrhage were assessed by control CT on day
5, or earlier in the event of clinical deterioration. These events were
validated and adjudicated by the Critical Events Reviewing Com-
mittee, which was blinded to the treatment assignment. All CT scans
were reviewed independently by 3 trained neurologists who were
unaware of patient treatment assignment. CT scan imaging was
scored with the use of a standard questionnaire, including previous
lesions of stroke, cerebral atrophy, leukoaraiosis, and early signs of
stroke.14 CT findings considered to be early CT scan signs of brain
ischemia included loss of the density contrast of the lentiform
nucleus,15 loss of the density contrast of the insular ribbon,16 and
hemispheric sulcus effacement, either alone or in association with a
hyperdense MCA sign.17
Cerebral HT, assessed either on CT scans by the Neuroradiologic
Reviewing Committee or by autopsy, was considered to be both the
symptomatic and asymptomatic cerebral HT. SHT were defined as
clinical deterioration temporally related to HT documented by CT
scan or autopsy, as adjudicated independently by the Critical Events
Reviewing Committee.
Demographic characteristics, medical history, vascular risk fac-
Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017

tors, baseline clinical characteristics and neurological state

(MAST-E score, the items and the sum) were collected. Adminis-
tration of anticoagulants or antiplatelets agents during hospitalization
was allowed to give the placebo group patients access to the best
treatment available in the investigators opinion. Use of heparin or
antiplatelets agents during hospitalization was recorded.

Statistical Analysis
We analyzed HT and SHT using the x 2 test, Student t test, or the
Fishers exact test, as appropriate. Subgroup analyses were per-
formed according to the following variables: age; sex; body weight;
atrial fibrillation; history of hypertension, diabetes mellitus, previous
stroke, or transient ischemic attack of the brain; systolic blood
pressure and diastolic blood pressure; MAST-E score18; hand, arm,
and leg paresias; level of consciousness at admission; side of the
ischemia; cerebral atrophy; early CT signs of lentiform nucleus
attenuation, insular ribbon contrast attenuation, hemispheric sulcus
effacement, and hyperdense MCA; streptokinase treatment alloca-
tion; antithrombotic-associated treatment (heparin, antiplatelet
agents); and delay from stroke onset to treatment.
A logistic forward stepwise regression model was used to define
predictors of HT after adjustment for the effect of confounding
variables.19 We chose to include in the initial model all variables
associated with hemorrhage in univariate analysis with a value of
P,0.2. We also included all variables reported to be related to
cerebral hemorrhages in other studies: age, sex, weight, hyperten-
sion, previous stroke or transient ischemic attack, atrial fibrillation,
systolic and diastolic blood pressures, and severity at admis-
sion.12,20,21 Because of the small number of patients, logistic regres-
sion was performed on symptomatic cerebral hemorrhages both in Figure 1. A, Seventy-nine-year-old patient with 4 hours 30 min-
the total population and in the streptokinase group, which included utes of new left dense hemiparesis. Early CT signs are observed
most SHT. We tested the interactions among the selected variables in the right MCA. Note lentiform nucleus contrast attenuation
and refit the model. We chose a value of P50.05 as a level of insular zone contrast attenuation and hemispheric sulcus atten-
statistical significance and did not correct for multiple comparisons. uation. B, The CT scan 5 days later showed hypodensity in the
We assessed the sensivity, specificity, positive and negative predic- deep and superficial territories of the right MCA with HT.
tive values, and efficiency of each mutivariate model for HT
and SHT. second CT scan was performed (1 in each group), poor
quality of the CT scan (3 in the streptokinase group and 8 in
Results the placebo group), or the second CT scan being performed
Among the 310 patients included (156 in the streptokinase .10 days after stroke (1 in the streptokinase group and 3 in
group and 154 in the placebo group), 157 patients had an HT, the placebo group). Among the 16 patients who died without
96 (61%) in the streptokinase group and 61 (31%) in the CT scan control, 3 had autopsy. Thus, the analysis included
placebo group. The mean delay of CT from the onset of 280 patients for HT and 310 for SHT.
stroke was 150 minutes (first and third quartiles, 123 and 223
minutes). Control CT scan data were missing for 33 patients, Hemorrhagic Transformation
because of either early death (12 patients in the streptokinase Comparisons of baseline characteristics, treatments received,
group and 4 in the placebo group), discharge before the and CT scan signs between patients with and without HT are
 1328 Stroke July 1999

TABLE 1. Clinical Characteristics and Treatments of Patients With HT

Characteristic HT (n5157) No HT (n5123) P SHT (n537) No SHT (n5273) P
Clinical status at admission
Male, % 92 (57.9) 71 (56.3) 0.79 22 (59.5) 153 (56.0) 0.69
Mean age (range), y 69.1 (22 to 92) 69.9 (32 to 94) 0.63 70.6 (38 to 92) 69.1 (22 to 94) 0.50
Weight (SD), kg 71.2 (12.8) 70.8 (13.0) 0.38 71.4 (12.9) 72.9 (12.5) 0.54
Mean systolic blood pressure (SD), mm Hg 152.7 (23.1) 153.4 (22.2) 0.80 156.9 (23.1) 153.2 (22.8) 0.34
Mean diastolic blood pressure (SD), mm Hg 84.5 (13.2) 85.9 (12.3) 0.36 84.6 (14.8) 85.1 (12.8) 0.83
Atrial fibrillation, % 54 (34.0) 32 (25.4) 0.12 16 (43.2) 75 (27.5) 0.05
Right brain ischemia, % 76 (47.8) 51 (40.5) 0.22 19 (51.4) 122 (44.7) 0.45
Mean MAST clinical score (SD) 24.1 (11.4) 25.8 (12.5) 0.22 23.2 (11.7) 25.2 (11.9) 0.34
MAST score ,20, % 52 (32.7) 61 (48.4) 0.45 16 (43.2) 77 (28.2) 0.06
Decreased level of consciousness, % 84 (52.9) 60 (47.0) 0.64 23 (66.2) 135 (49.5) 0.10
Delay from stroke to treatment (SD), min 269 (57.6) 267.9 (63.4) 0.87 254.1 (62.0) 269.9 (60.0) 0.13
Medical history
History of hypertension, % 76 (47.8) 61 (48.4) 0.92 21 (56.8) 127 (46.5) 0.24
Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017

History of diabetes mellitus, % 20 (12.6) 10 (7.9) 0.20 9 (24.3) 24 (8.8) 0.004

Previous stroke, % 9 (5.7) 13 (10.3) 0.14 3 (8.1) 19 (7.0) 0.79
Previous transient ischemic attack, % 13 (8.2) 19 (15.1) 0.07 2 (5.4) 32 (11.7) 0.25
Medication
Streptokinase treatment 97 (61.0) 46 (36.5) ,0.001 33 (89.2) 123 (45.1) ,0.001
Heparin use within 10 d 111 (69.8) 91 (72.2) 0.66 15 (40.5) 201 (73.6) ,0.001
Antiplatelet agents use within 10 d 42 (26.4) 36 (26.8) 0.69 6 (16.2) 81 (29.8) 0.09

reported in Tables 1 and 2. Streptokinase treatment and early
CT scan signs (Figure 1) differed significantly between the 2
groups and remained significantly linked with HT in the
multivariate stepwise logistic regression (Table 3). Age, body
weight, atrial fibrillation, systolic and diastolic blood pres-
sures, the delay from stroke onset to treatment, and heparin
and antiplatelet agent use were not kept in the multivariate
model. The specificity of the model was 44.9% (95% CI, 39%
to 51%), the sensitivity 80% (95% CI, 75% to 85%), the
positive predictive value 66% (95% CI, 60% to 72%), the
negative predictive value 63% (95% CI, 57% to 69%), and
the global efficiency 65% (95% CI, 59% to 71%). When we
selected the streptokinase group, early CT signs remained the
sole predictors of HT (Table 3).
Symptomatic Hemorrhages
Among the 157 patients with HT, 37 had SHT (33 in the
streptokinase group and 4 in the placebo group). Comparisons of
the baseline characteristics, treatments, and baseline CT scan
signs in patients with and without SHT, are reported in Tables 1
and 2. SHT was significantly more frequent in patients with
atrial fibrillation, diabetes mellitus, no heparin use, streptokinase
treatment, and early CT scan signs (hemispheric sulcus attenu-
ation and insular zone contrast attenuation; Figure 2). The
multivariate model included streptokinase treatment, diabetes
mellitus, hemispheric sulcus attenuation, and the interaction
between a decreased level of consciousness and streptokinase
treatment (ie, the patients who had a decreased level of con-
sciousness and were treated with streptokinase). Within the
streptokinase group, diabetes mellitus, hemispheric sulcus atten-
uation, and decreased level of consciousness were kept in the
final model (Table 4). The predictive model for SHT had 19%
sensitivity (95% CI, 15% to 24%), 99% specificity (95% CI,
98% to 100%), 78% positive predictive value (95% CI, 73% to
83%), 89% negative predictive value (95% CI, 85% to 93%),
and 89% efficiency (95% CI, 85% to 93%).

TABLE 2. Baseline CT Scan Characteristics of Patients With HT and SHT

Baseline CT Scan Characteristics HT No HT P SHT No SHT P
Delay from stroke onset to CT (SE), min 151.3 (15.4) 155.9 (15.8) 0.84 120.7 (39.5) 154.6 (11.1) 0.3

Early CT scan signs, n (%) 114 (73.5) 59 (50.0) ,0.001 28 (75.7) 153 (61.0) 0.04
Hemispheric sulcus effacement 89 (57.4) 46 (39.0) 0.003 25 (67.6) 117 (46.6) 0.02
Lentiform nucleus contrast attenuation 88 (56.8) 41 (34.7) ,0.001 22 (59.5) 114 (45.4) 0.11
Insular zone contrast attenuation 86 (55.5) 42 (35.6) 0.001 23 (62.2) 113 (45.0) 0.05
Dense middle cerebral artery 39 (25.2) 28 (23.7) 0.78 9 (24.3) 60 (23.9) 0.95

Leukoaraiosis, n (%) 18 (11.5) 17 (14.3) 0.49 6 (16.2) 35 (13.8) 0.69

Previous stroke, n (%) 38 (24.2) 31 (26.1) 0.73 7 (18.9) 66 (26.0) 0.35

Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017
Figure 2. A, Eighty-nine-year-old patient with 2 hours 20 min-
utes of new left dense hemiparesis and drowsiness. Early CT
signs are observed in the right MCA. Note lentiform nucleus
contrast attenuation insular zone contrast attenuation and hemi-
spheric sulcus attenuation. B, The CT scan 1 day later showed
symptomatic HT in the total territory of the right MCA with mid-
line shift and contralateral ventricular mass effect.
Discussion
Incidence of HT and SHT
The rate of spontaneous HT ranged from 38% to 71% in
autopsy studies13 and from 13% to 43% in the recent CT
studies.14,2227 In controlled clinical trials28,29 of thrombolysis
for acute MI, rates of 14 and 30% of HT have been reported
in patients with ischemic stroke. Recent series30 32 of patients
Jaillard et al Hemorrhagic Transformation 1329
TABLE 3. Results of Logistic Regression Model for HT
Streptokinase
Total Population Population
(n5273) (n5135)
Variable OR (95% CI) P OR (95% CI) P
Streptokinase treatment 2.6 (1.64.4) 0.001 zzz zzz
Early CT signs 2.6 (1.54.4) 0.002 3.0 (1.46.5) 0.004
treated with tPA for acute ischemic stroke reported rates of
HT from 10% to 30%. Recent controlled clinical trials in
acute ischemic stroke reported rates of total HT over the first
5 days from 3.2% to 37% in the placebo group and from
10.6% to 44% in the thrombolyzed group (Table 5).
The rate of SHT has not been reported for all studies. After
thrombolysis for MI, 5 of 27 patients (19%) suffering from an
ischemic stroke developed SHT.28 In recent series of patients
treated with t-PA for acute ischemic stroke, SHT occurred in
3.6% of 85 consecutive patients,27 9.6% of 104 patients,30 5%
of 100 patients32 and 7% of 30 patients.31 In controlled
clinical trials of thrombolysis for acute stroke, SHT ranged
from 0.6% to 7% in the placebo group and from 6.4% to 20%
in the thrombolyzed group (Table 5). In our study, the rates of
total HT and SHT were 43.7% and 2.6%, respectively, in the
placebo group compared with 67.8% and 21.2% in the
streptokinase group. The rates in the placebo group are within
the range of published rates of HT and SHT. In the streptoki-
nase group the rates of HT and SHT are higher than those in
the treatment groups of other acute stroke trials (Table 5).
However, the rates of HT and SHT differed significantly from
one study to another. Such heterogeneity between trials in the
incidence of HT and SHT may result in part from sample
fluctuations. In each trial, the proportion of HT and SHT is
small, and thus populations of HT, and particularly SHT, are
small. By chance, sample fluctuations may result in major
differences in HT and SHT rates. Potential biases in the HT
rate estimation also need to be addressed. First, a selection
bias related to the baseline characteristics of the patients, such
as geographic or ethnic factors, severity, mechanism (cardiac
embolism versus in situ thrombosis), topography of stroke
(ICA versus MCA), delay to inclusion, and time period over
which the HT were evaluated could have selected groups of
patients with very different bleeding risks. Second, biases in
the care given to patients could have occurred; for example,
associated treatments were different between studies. Third,
classification biases such as the assessment of HT and SHT
may have occurred, because there is no shared criteria for the
assessment and classification of HT on CT. This is particu-
larly true for petechial HT.33 The classification of hemor-
rhages as symptomatic or not may also be controversial,
because neurological deterioration may or may not be due to
HT. Using crude rates rather than relative risks to compare
the incidence of HT and SHT between these trials may be
misleading. After thrombolysis, the relative risks are ranged
from 1.165 to 3.110 for HT and from 2.75 to 10.010 for SHT
(Table 5). No significant difference between streptokinase
and tPA in hemorrhagic risk can be evidenced. In controlled
thrombolytic trials for acute MI, an excessive number of
 1330 Stroke July 1999

TABLE 4. Results of Multivariate Analyses in the Total Population and

Restricted to Streptokinase-Treated Patients
Streptokinase
Total Population Population
(n5288) (n5139)

Variable OR (95% CI) P OR (95% CI) P

Hemispheric sulcus attenuation 2.6 (1.25.7) 0.02 3.1 (1.37.4) 0.01
Diabetes mellitus 3.0 (1.17.9) 0.03 3.7 (1.310.6) 0.01
Decreased level of consciousness zzz 0.19 2.7 (1.16.4) 0.03
Streptokinase treatment 5.9 (1.918.7) 0.004 zzz zzz
Consciousness-streptokinase interaction 2.6 (1.16.2) 0.03 zzz zzz

hemorrhagic strokes has been reported for tPA.34 Therefore,
only a direct comparison of streptokinase and tPA in an acute
ischemic stroke trial could permit to assess the difference in
risk of bleeding between the 2 drugs.
Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017
diabetes mellitus, early CT signs, streptokinase treatment and
the interaction between streptokinase treatment and a de-
creased level of consciousness. When we ran the model of
SHT among the streptokinase-treated patients, the same

predictors remained. This suggests that streptokinase in-

Predictors of HT and SHT
In controlled thrombolysis trials for MI, risk factors for HT
have been reported to be low body weight,29,35 elderly
age, 20,21,29,35 hypertension at admission, 20,29,35 and
thrombolytic treatment assignment.29,35 Data for acute ische-
mic stroke are available from trials and from clinical and
autopsy series. In acute ischemic stroke trials, body weight
and hypertension have not been evidenced as risk factors, but
advanced age and tPA treatment were associated with in-
creased risk of parenchymal hemorrhage in ECASS11 and of
both SHT and total HT in the NINDS.12 Other risk factors for
SHT in acute stroke thrombolytic studies include a cardio-
embolic mechanism of stroke,11,28,31 prior MI,32 stroke sever-
ity,12 size of infarct,22 and early CT signs.12,31 In autopsy
studies1,3 and recent clinical series without thrombolysis, HT
was associated with severity, cardioembolic mechanism of
the stroke27 and early CT signs.26
In our study, HT was predicted by both early CT signs and
streptokinase treatment, whereas SHT was predicted by
creases the risk of SHT among patients who would have
presented with asymptomatic hemorrhage had the treatment
been the placebo. This is consistent with data from ECASS I
and ECASS II, in which tPA treatment was associated with an
increased risk of parenchymal hemorrhage but not of hemor-
rhagic infarction.11 However, SHT has not been assessed in
ECASS I.
The presence of early CT signs predicted both HT and SHT
in the multivariate analysis, which is in agreement with other
thrombolytic studies.11,12,31,32 In MAST-E, early CT signs
were present in 63% of the patients, the same order of
magnitude as the Australian Streptokinase Trial (57.7%) but
higher than in MAST-I (4.7%), ECASS (31%),11 and the
NINDS study (5%).12 Early CT signs are reported from 31%
to 92% in series focused on early CT signs among patients
with MCA ischemia, according to the delay from onset to CT
scan (4 to 8 hours).14 16,36 39 These discrepancies may be
related to the patient selection and to CT sign measurement

TABLE 5. Incidence and Risk of Cerebral HT According to Thrombolytic Therapy in Recent Controlled
Clinical Trials
Total HTs SHTs

Enrolled Thrombolytic Thrombolytic

Trial Patients Treatment Placebo RR (95% CI) Treatment Placebo RR (95% CI)
MAST-E* 310 96 (67.6) 61 (43.3) 1.6 (1.22.0) 33 (21.2) 4 (2.6) 8.1 (3.022.0)
MAST-I 622 81 (26.0) 38 (12.3) 2.9 (1.94.3) 25 (8.0) 4 (1.3) 6.2 (2.217.6)
ASK 340 56 (32.0) 28 (13.2) 1.9 (1.32.8) 23 (16.0) 5 (3.0) 5.3 (1.915)
NINDS 624 33 (10.6) 11 (3.2) 3.1 (1.66.0) 20 (6.4) 2 (0.6) 10.0 (2.442.2)
ECASS I\ 620 134 (44.0) 113 (37.0) 1.16 (0.951.4) 62 (20) 20 (7) 2.7 (136.2)
ECASS II 800 197 (48.4) 155 (40.2) 1.2 (1.041.4) 36 (8.9) 13 (3.4) 2.6 (1.44.8)
Values are number (percent).
*Data were available for 142 patients in the streptokinase group and 141 in the placebo group for HT and for 156 patients in the
streptokinase group and 154 in the placebo group for SHT.
Data were available for 311 patients in the streptokinase group and 309 in the placebo group.
Data were available for 174 patients in the streptokinase group and 166 in the placebo group.
Data were available for 311 patients in the tPA group and 312 in the placebo group.
\Data were available for 307 patients in the tPA group and 302 in the placebo group. Parenchymal hemorrhage data are reported
instead of SHT, which were not available.
Data were available for 407 patients in the tPA group and 386 in the placebo group.
 Jaillard et al
biases. There were differences in the severity of stroke, in the
delay from stroke onset to CT scan, in the criteria used for
early CT signs assessment, and also in the quality of CT
scans. Actually, the absence of shared definition of early CT
signs and the subtle brain appearance changes they charac-
terize may be important contributors to the discrepancies
between the rates reported. This is supported by the low
intraobserver and interobserver concordance14,40 for early CT
sign assessment. Although important metrological and teach-
ing efforts could be done to reach a high level of reliability,41
this emphasizes that a treatment based on such subtle signs
might be very difficult to introduce into routine practice.
We found an association between stroke severity (MAST
score of ,20, or low level of consciousness) and SHT in
univariate analyses, although it did not reach a level of
significance. In the multivariate model, decreased level of
consciousness was a predictor of SHT only in patients
assigned to streptokinase. Stroke severity has been demon-
Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017
strated to be a predictor for SHT in NINDS12 and for HI in
ECASS.11 This confirms that patients with such conditions
should not receive thrombolytic treatment.
Our data indicate that diabetes mellitus is associated with
SHT. Hyperglycemia at baseline was associated with an
increased risk of SHT in the NINDS trial,10 and history of
diabetes mellitus was associated with parenchymal hematoma
in a stroke series.26 Both hyperglycemia and diabetes were
predictors of HT and SHT in tPA-treated patients.42 However,
although diabetes mellitus is a well-known risk factor for
ischemic stroke, whether it is a risk factor for hemorrhage has
yet to be determined.2
Our univariate analysis showed a possible association
between atrial fibrillation and both HT (P50.12) and SHT
(P50.05). A relation between cardioembolic strokes and HT
has been reported in autopsy studies1,3 in thrombolysis
series,31 in ECASS,11 and in trials in MI.28 The high rate of
cardioembolic stroke in the MAST-E population (30%) may
have promoted a high rate of HT.
The high crude rates of HT and SHT in our streptokinase
group could be attributed to the heavy use of heparin or
aspirin as associated treatment within the first 48 hours.
However, both heparin and aspirin were associated with a low
incidence of SHT (Table 3). Moreover, heparin administra-
tion appeared to be a predictor of low risk of SHT in our
multivariate analysis, suggesting a paradoxical protective
effect. Heparin appears to be an established risk factor for HT
in acute stroke.43 45 However, some studies have not reported
an increased risk of bleeding during heparin treatment, either
in association with thrombolytics in a meta-analysis of trials
for MI35 or alone.46 Because heparin was not used randomly
in MAST-E, it is likely that early occurrence of HT prevented
the investigator from using antithrombotic drugs within the
first hours. Therefore, because we thought that including this
variable in the model could be misleading, we did not keep
heparin treatment in our multivariate models.
It has been suggested30 that a long delay from stroke onset
to thrombolytic treatment is related to a high rate of SHT. We
explored this hypothesis, but neither univariate nor multivar-
iate analysis demonstrated any association between HT or
SHT and a longer delay. These results are supported by
Hemorrhagic Transformation 1331
findings from ECASS I and ECASS II, in which patients
treated between 3 and 6 hours after stroke did not have a
higher risk of parenchymal hemorrhage than those treated
within the first 3 hours,11 and by the NINDS study,12 in which
patients treated between 90 and 180 minutes did not have a
higher risk of SHT than those treated within 90 minutes of
onset.
To explore the dose effect in HT occurrence, because a
fixed dose of streptokinase was used, we studied the relation-
ship between HT and body weight. We found no relationship,
but in a recent study fibrinolytic blood parameters were
strongly correlated with the body mass index.47 These con-
flicting results suggest that the dose issue should be addressed
further.
Conclusion
We report the post hoc analysis of a controlled clinical trial
using streptokinase in acute ischemic stroke (MAST-E). In
both the streptokinase and placebo groups, higher absolute
rates of HT and SHT than in other studies were found, but the
relative risks of streptokinase were in the same range as in
other streptokinase and tPA trials. Only a direct comparison
of streptokinase and tPA in acute ischemic stroke would
allow assessment of the differential effects of the 2 drugs. In
addition, the issue of the dose of streptokinase should be
addressed before planning other trials with streptokinase in
acute stroke.
Through use of logistic regression analysis, early CT signs
were identified as a strong predictor of both HT and SHT,
stressing that this group of patients has a peculiarly high risk
of bleeding. The precise role of diabetes mellitus in the
occurrence of SHT must be assessed. Our results indicate that
a decreased level of consciousness was a predictor of SHT in
streptokinase-treated patients. This is consistent with the
lower rate of SHT in studies that excluded patients with coma
or stupor.5,9,10
Acknowledgment
This study was sponsored financially with a grant from the French
Ministry of Health (PHRC 1993).
References
1. Fisher CM, Adams RD. Observations on brain embolism with special
reference to the mechanism of hemorrhagic infarction. J Neuropathol Exp
Neurol. 1951;10:9293.
2. Jorgensen L, Torvik A. Ischaemic cerebrovascular disease in an autopsy
series, part 2: prevalence, location, pathogenesis and clinical course of
cerebral infarcts. J Neurol Sci. 1969;9:285320.
3. Lodder J, Krijne-Kubat B, Broekman J. Cerebral hemorrhagic infarction
at autopsy: cardiac embolic cause and the relationship to the cause of
death. Stroke. 1986;17:626 629.
4. Moulin T, Crepin-Leblond T, Chopard JL, Bogousslavsky J. Hemor-
rhagic infarct. Eur Neurol. 1994;34:64 77.
5. Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R,
Boysen G, Bluhmki E, Hoxter G, Mahagne MH, et al. Intravenous
thrombolysis with recombinant tissue plasminogen activator for acute
hemispheric stroke: the European Cooperative Stroke Study (ECASS).
JAMA. 1995;274:10171025.
6. Multicentre Acute Stroke Trial-Italy (MAST-I) Group. Randomised con-
trolled trial of streptokinase, aspirin, and combination of both in treatment
of acute ischaemic stroke. Lancet. 1995;346:1509 1514.
7. The Multicenter Acute Stroke Trial-Europe Study Group. Thrombolytic
therapy with streptokinase in acute ischemic stroke. N Engl J Med.
1996;335:145150.
 1332 Stroke July 1999
8. Donnan GA, Davis SM, Chambers BR, Gates PC, Hankey GJ, McNeil JJ,
Rosen D, Stewart-Wynne EG, Tuck RR. Streptokinase for acute ischemic
stroke with relationship to time of administration. JAMA. 1996;276:
961966.
9. Hacke W, Kaste M, Fieschi C, von Kummer R, Davalos A, Meier D,
Larrue V, Bluhmki E, Davis S, Donnan G, Schneider D, Diez-Tejedor E,
Trouillas P. Randomised double-blind placebo-controlled trial of
thrombolytic therapy with intravenous alteplase in acute ischaemic stroke
(ECASS II). Lancet. 1998;352:12451251.
10. The National Institute of Neurological Disorders and Stroke rt-PA Stroke
Study Group. Tissue plasminogen activator for acute ischemic stroke.
N Engl J Med. 1995;333:15811587.
11. Larrue V, von Kummer R, del Zoppo G, Bluhmki E. Hemorrhagic
transformation in acute ischemic stroke: potential contributing factors in
the European Cooperative Acute Stroke Study. Stroke. 1997;28:957960.
12. The NINDS rt-PA Stroke Study Group. Intracerebral hemorrhage after
intravenous rt-PA therapy for ischemic stroke. Stroke. 1997;28:
2109 2118.
13. The MAST Group. Protocol for the Multicenter Acute Stroke Trial-
Thrombolysis Study. Clin Trials Metaanal. 1993;28:329 344.
14. Moulin T, Cattin F, Crepin-Leblond T, Tatu L, Chavot D, Piotin M, Viel
JF, Rumbach L, Bonneville JF. Early CT signs in acute middle cerebral
artery infarction: predictive value for subsequent infarct locations and
Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017
outcome. Neurology. 1996;47:366 375.
15. Tomura N, Uemura K, Inugami, Fujita HHigano S, Shishido F. Early CT
finding in cerebral infarction: obscuration of the lenticular nucleus.
Radiology. 1988;168:463 467.
16. Truwit CL, Barkovich AJ, Gean-Marton A, Hibri N, Norman D. Loss of
the insular ribbon:another early CT sign of acute middle cerebral artery
infarction. Radiology. 1990;176:801 806.
17. Tomsick TA, Brott TG, Chambers AA. Hyperdense middle cerebral
artery sign on CT: efficacy in detecting middle cerebral artery thrombosis.
AJNR Am J Neuroradiol. 1990;11:473 477.
18. Orgogozo JM, Asplund K, Boysen G. A unified form for neurological
scoring of hemispheric stroke with motor impairment. Stroke. 1992;23:
1678 1689.
19. SPSS for Windows. Chicago, Ill: SPSS Inc; 1990.
20. Sloan MA, Price TR, Petito CK, Randall AM, Solomon RE, Terrin ML,
Gore J, Collen D, Kleiman N, Feit F, et al. Clinical features and patho-
genesis of intracerebral hemorrhage after rt-PA and heparin therapy for
acute myocardial infarction: the Thrombolysis in Myocardial Infarction
(TIMI) II pilot and randomized clinical trial combined experience. Neu-
rology. 1995;45:649 658.
21. Gebel JM, Sila CA, Sloan MA, Granger CB, Mahaffey KW, Weisen-
berger J, Green CL, White HD, Gore JM, Weaver WD, Califf RM, Topol
EJ. Thromboysis-related intracranial hemorrhage: a radiographic analysis
of 244 from the GUSTO-1 Trial with clinical correlation. Stroke. 1998;
29:563569.
22. Lodder J. CT-detected hemorrhagic infarction: relation with the size of
the infarct and the presence of midline shift. Acta. Neurol. Scand. 1984;
70:329 335.
23. Ott BR, Zamani A, KLeefield J, Funkelstein HH. The clinical spectrum of
hemorrhagic infarction. Stroke. 1986;17:630 637.
24. Hart RG, Tegeler CH, for the Cerebral Embolism Study Group. Hemor-
rhagic infarction on CT in the absence of anticoagulation therapy. Stroke.
1986;17:558. Letter.
25. Hornig CR, Dorndorf W, Agnoli AL. Hemorrhagic cerebral infarction: a
prospective study. Stroke. 1986;17:179 185.
26. Toni D, Fiorelli M, Bastianello S, Sacchetti ML, Sette G, Argentino C,
Montinaro E, Bozzao L. Hemorrhagic transformation of brain infarct:
predicability in the first 5 hours from stroke onset and influence on
clinical outcome. Neurology. 1996;46:341345.
27. Alexandrov AV, Black SE, Ehrlich LE, Caldwell CB, Norris JW. Pre-
dictors of hemorrhagic transformation occuring spontaneously and on
anticoagulation in patients with acute ischemic stroke. Stroke. 1997;28:
1198 1202.
28. Sloan MA, Price TR, Terrin ML, Forman S, Gore JM, Chaitman BR,
Hodges M, Mueller H, Rogers WJ, Knatterud GL, Braunwald E. Ischemic
cerebral infarction after rt-PA and heparin therapy for acute myocardial
infarction: the TIMI-II pilot randomized clinical trial combined expe-
rience. Stroke. 1997;28:11071114.
29. Gore JM, Granger CB, Simoons ML, Sloan MA, Weaver WD, White HD,
Barbash GI, Van de Werf F, Aylward PE, Topol EJ, et al. Stroke after
thrombolysis: mortality and functional outcomes in the GUSTO-1 Trial.
Circulation. 1995;92:28112818.
30. del Zoppo GJ, Poeck K, Pessin MS, Wolpert SM, Furlan AJ, Ferbert A,
Alberts MJ, Zivin JA, Wechsler L, Busse O, et al. Recombinant tissue
plasminogen activator in acute thrombotic and embolic stroke. Ann
Neurol. 1992;32:78 85.
31. Chiu D, Krieger D, Villar-Cordova C, Kasner SE, Morgenstern LB,
Bratina PL, Yatsu FM, Grotta JC. Intravenous tissue plasminogen acti-
vator for acute ischemic stroke. Feasibility, safety, and efficacy in the first
year of clinical practice. Stroke. 1998;29:18 22.
32. Grond M, Stenzel C, Schmulling S, Rudolf J, Neveling M, Lechleuthner
A, Schneweis S, Heiss WD. Early intravenous thrombolysis for acute
ischemic stroke in a community-based approach. Stroke. 1998;29:
1544 1549.
33. Motto C, Aritzu E, Boccardi E, De Grandi C, Piana A, Candelise L.
Reliability of hemorrhagic transformation diagnosis in acute ischemic
stroke. Stroke. 1997;28:302306.
34. International Study Group. In-hospital mortality and clinical course of
20 891 patients with suspected acute myocardial infarction randomized
between alteplase and strptokinase with or without heparin. Lancet. 1990;
336:7175.
35. Simoons ML, Maggioni AP, Knatterud G, Leimberger JD, de Jaegere P,
van Domburg R, Boersma E, Franzosi MG, Califf R, Schroder R, et al.
Individual risk assessment for intracranial haemorrhage during
thrombolytic therapy. Lancet. 1993;342:15231528.
36. Bozzao L, Bastianello S, Fantozzi LM, Angeloni U, Argentino C, Fieschi
C. Correlation of angiographic and sequential CT findings in patients with
evolving cerebral infarction. AJNR Am J Neuroradiol. 1989;10:
12151222.
37. Bryan RN, Levy LM, Whitlow WD, Killian JM, Preziosi TJ, Rosario JA.
Diagnosis of acute cerebral infarction: comparison of CT and MR
imaging. AJNR Am J Neuroradiol. 1991;12:611 620.
38. Wolpert SM, Bruckmann H, Greenlee R, Wechsler L, Pessin MS, del
Zoppo GJ. Neuroradiologic evaluation of patients with acute stroke
treated with recombinant tissue plasminogen activator. AJNR Am J Neu-
roradiol. 1993;14:313.
39. von Kummer R, Meyding-Lamade U, Forsting M, Rosin L, Rieke K,
Hacke W, Sartor K. Sensivity and pronognostic value of early CT signs
in occlusion of the middle cerebral artery trunk. AJNR Am J Neuroradiol.
1994;15:9 15.
40. Besson G, Moulin, Garnier P, Crepin-Leblond, The MAST Group.
Intraobserver concordance of the Neuroradiologic Reviewing Commitee
in CT scan reviewing in MAST-E. Acta Neurol Scand. 1998;98:292293.
41. von Kummer R. Effect of training in reading CT scans on patient
selection for ECASS II. Neurology. 1998;51:S50 S52.
42. Demchuk AM, Morgenstern LB, Krieger DW, Chi TL, Hu W, Wein TH,
Hardy RJ, Grotta JC, Buchan AM. Serum glucose level, and diabetes
predict tissue plasminogen activatorrelated intracerebral hemorrhage in
acute ischemic stroke. Stroke. 1999;30:34 39.
43. Babikian VL, Kase CS, Pessin MS. Intracerebral hemorrhage in stroke
patients anticoagulated with heparin. Stroke. 1989;20:1500 1503.
44. International Stroke Trial Collaborative Group. The International Stroke
Trial (IST): a randomised trial of aspirin, subcutaneous heparin, both, or
neither among 19435 patients with acute ischaemic stroke. Lancet. 1997;
349:1569 1581.
45. The Publication Committee for the Trial of ORG 10172 in Acute Stroke
Treatment (TOAST) Investigators. Low molecular weight heparinoid,
ORG 10172 (danaparoid), and outcome after acute ischemic stroke: a
randomized controlled trial. JAMA. 1998;279:12651272.
46. Kay R, Wong KS, Yu YL, Chan YW, Tsoi TH, Ahuja AT, Chan FL, Fong
KY, Law CB, Wong A. Low-molecular-weight heparin for the treatment
of acute ischemic stroke. N Engl J Med. 1995;333:1588 1593.
47. Kristensen B, Malm J, Nilsson N, Hultdin J, Carlberg B, Olsson T.
Increased fibrinogen levels and acquired hypofibrinolysis in young adults
with ischemic stroke. Stroke. 1998;29:2261 67.
 Hemorrhagic Transformation in Acute Ischemic Stroke: The MAST-E Study
Assia Jaillard, Catherine Cornu, Anne Durieux, Thierry Moulin, Florent Boutitie, Kennedy R.
Lees and Marc Hommel
on behalf of the MAST-E Group
Downloaded from http://stroke.ahajournals.org/ by guest on August 9, 2017

Stroke. 1999;30:1326-1332
doi: 10.1161/01.STR.30.7.1326
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 1999 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/30/7/1326

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Stroke is online at:

http://stroke.ahajournals.org//subscriptions/