Thrombolytic Therapy for Pulmonary
Embolism*
Frequency of Intracranial Hemorrhage and
Associated Risk Factors
Daniel S. Kanter, MD; Katriina M. Mikkola; Sanjay R. Patel, MD;
J. Anthony Parker, MD, PhD; and Samuel Z. Goldhaber, MD, FCCP
risk factors and frequency of intracranial hemorrhage among
Study objectives: To determine thefor
embolism.
patients undergoing thrombolysis andpulmonary
controlled
A
analysis.
Design: retrospective descriptive
the United States, Canada, and Italy.
Setting: Hospitalized patients at centers inembolism
on perfusion scans or angiography.
Patients: All had evidence of pulmonary
Interventions: None.
Measurements and results: Data were analyzed on 312 patients from five previously reported
studies of pulmonary embolism thrombolysis. The frequency of intracranial hemorrhage up to 14
days after pulmonary embolism thrombolysis was 6 of 312 or 1.9% (95% confidence interval, 0.7
to 4.1%). Two of six intracranial hemorrhages were fatal. Two of the six patients received
of the protocol because they had pre-existing, known intracranial
thrombolysis in violation
BP at the time of hospital admission was significantly elevated in
diastolic
disease. Average
an
who
developed intracranial hemorrhage (90.3 ±15.1 mm Hg) compared with those
patients
who did not (77.6± 10.9 mm Hg; p=0.04). Other baseline characteristics and laboratory data were
similar in both groups. Decreased level of consciousness, hemiparesis, and visual field deficits
were the most common clinical signs of intracranial hemorrhage.
Conclusions: Intracranial hemorrhage after pulmonary embolism thrombolysis is an infrequent
but often grave complication. Meticulous patient screening before administering thrombolysis is
imperative. Diastolic hypertension at the time of hospital admission is a risk factor for intracranial
(CHEST 1997; 111:1241-45)
hemorrhage after pulmonary embolism thrombolysis.
Key words: intracranial hemorrhage; hypertension; pulmonary embolism; thrombolytic therapy
DVT=deep venous thrombosis;
PE=pulmonary embolism; rt-PA=recombinant tissue plasminogen activator
Abbreviations: CI=confidence interval;
T ntracranial hemorrhage
(ICH) is one of the most
serious and feared complications of thrombolysis
for pulmonary embolism (PE), but only limited
information exists on the frequency of its occurrence
-¦"
For related commentary see pages 1246 and 1410
Neurology/Neurosurgery Intensive Care Unit (Dr.
Kanter), Department of Neurology (Dr. Kanter), and Cardio¬
vascular Division, Department of Medicine (Dr. Goldhaber),
Brigham and Women's Hospital; the Division of Nuclear Med¬
icine (Dr. Parker), Beth Israel Hospital, Boston; Lund Univer¬
School of Medicine, Lund, Sweden (Ms. Mikkola); and
sity
Harvard Medical School (Drs. Kanter, Patel, Parker, and Gold¬
haber), Boston.
received August 27, 1996; revision accepted Decem¬
Manuscript
ber 5.
Reprint requests: Samuel Z. Goldhaber, MD, FCCP, Cardiovas¬
*From the
cular Division,
Brigham
Street, Boston, MA 02115
and Women s
Hospital,
75 Francis
ICH=intracranial
hemorrhage;
associated risk factors. Most current knowledge of
ICH comes from studies in
thrombolysis-related
patients with myocardial infarction. However, we do
not know if the same frequencies and risk factors
apply to patients with PE. Therefore, we reviewed
our experience in patients who have undergone PE
to determine the risk factors and fre¬
thrombolysis
of
ICH
up to 14 days after administration of
quency
or
a
thrombolytic agent.
Materials
and
Methods
Data on 312 patients from five previously reported studies17 of
thrombolysis in PE (as an adjunct to heparin anticoagulation)
were analyzed in an overview. The first trial was a dose selection
trial of recombinant tissue plasminogen activator (rt-PA) and all
47 patients received 50 to 90 mg over 2 to 6 h.1-3 The second trial
was a study of 45 patients randomized to rt-PA, 100 mg over 2 h
CHEST/111 /5/MAY, 1997
Downloaded From: http://journal.publications.chestnet.org/ on 09/29/2016
1241
Table
1.Frequency Of ICH in Our Five Trials17
ICH
No.
Fatal
Nonfatal
Total
2/312
4/312
6/312
Table 2.Demographic and Historical Data for
Patients Treated With Thrombolysis
95% CI
0.6
1.3
1.9
0.08-2.3
0.3-3.3
Patient Data
0.7-4.1
Mean age, yr
Age<50yr, %
Male:female
Body weight, kg*
urokinase, 2,000 Units/lb body weight, as a bolus followed by
same dose per hour for up to 24 h.4 In the third trial, 90
patients were randomized to rt-PA, 100 mg over 2 h vs urokinase,
3 million units.5 The fourth trial consisted of 101 patients
randomized to rt-PA, 100 mg over 2 h followed by heparin (46
patients) or heparin alone (55 patients).6 The fifth trial compared
rt-PA, 100 mg over 2 h to 0.6 mg/kg (maximum 50 mg) over 15
min in 90 patients.7 All patients were given heparin infusions at
the conclusion of their thrombolytic regimens. Data from six
patients assigned to thrombolytic therapy who did not receive the
assigned drug and from 55 patients randomized to heparin alone
(in the fourth trial) were excluded from the analysis.
The diagnosis of PE in all patients was based on pulmonary
Prior PE/DVT, %
vs
History of cancer, %
the
angiograms or perfusion scans as described in the original
reports.17 Exclusion criteria were similar in all five trials: signif¬
icant internal bleeding within the previous 6 months, intracranial
disease, operation or biopsy within 10 days, occult blood on stool
significant
examination,
impairment of hepatic
thrombocytopenia,
or
severe
renal function, pregnancy, severe
or any concurrent condition likely to limit survival
hypertension,
to 1
anemia
month.
or
were obtained from case report forms and hospital
discharge summaries, and were entered into a database (SAS
Institute Inc; Cary, NC). ICHs were confirmed by CT. ICH up
to 14 days after thrombolysis w?as considered to be possibly
related to thrombolytic treatment and was counted as an ICH
due to thrombolysis.
Continuous variables were analyzed using the Wilcoxon rank
sum test. Categorical variables were examined with Fisher's
Exact Test. All reported p values are two-tailed.
Data
Results
The overall frequency of ICH was 6 of 312 or 1.9%
(95% confidence interval [CI], 0.7 to 4.1%) (Table
1). Three patients died, two of ICH and one of
recurrent PE or myocardial infarction, but no autop-
Duration of symptoms, d
Patients
Patients
With ICH
Without ICH
68.1±8.1
57.3±17.0
0 35
0.10
1
1.20
1.00
80.2±20.6
83.3±18.8
33 30
1.00
17 18
1.00
2.7±1.9
3.7±3.3
p Valu
0.14
0.76
0.57
*Mean±SD.
performed. The other three were dis¬
from
chargeddeficits.the hospital with minor residual neu¬
rologic
Patients with ICH were older (68.1 ±8.1 years vs
57.3±17.0 years), but this trend did not achieve
statistical significance (p=0.14). No patient younger
than 50 years of age suffered an ICH, and 35% of the
entire cohort was in this age group. There was no
gender
predominance. Body weight, history of prior
PE or deep venous thrombosis (DVT), known can¬
cer, and duration of symptoms were similar in
patients with and without ICH (Table 2).
Baseline characteristics (Table 3) revealed an ele¬
vated diastolic BP in patients who subsequently
suffered an ICH (90.3±15.1 mm Hg vs 77.6±10.9
mm Hg; p^O.04). Hospital admission pulse pres¬
sures, hematocrit levels, platelet counts, and coagu¬
lation values were similar in patients with and with¬
out ICH (Table 3). None of the partial
times drawn within 6 h of ICH
thromboplastin
were
symptoms
markedly elevated (>100 s) in pa¬
tients with ICH. Fibrinogen values were <200
mg/dL in two patients with ICH during the first 24 h
after thrombolysis. However, average fibrinogen lev¬
els did not differ between patients with and without
sies
were
ICH
(Table 3).
Table 4 shows the frequency of ICH by thrombo-
Table 3.Baseline Characteristics
Patients With ICH*
Patients Without ICH*
Characteristics
(n=6)
(n=306)
Hospital admission systolic BP, mm Hg
Hospital admission diastolic BP, mm Hgf
Hospital admission pulse pressure, mm Hg
145.3±33.7
90.3±15.1
55.0±21.2
37.4±6.8
227.0±45.4
34.2±17.7
401 ±304
127.2±18.1
77.6±10.9
49.6±14.5
39.1±5.5
260.3± 105.1
41.4±37.5
364±178
Hematocrit, %
Platelet count, 103/mm3
Activated partial thromboplastin time,
s
Fibrinogen, mg/dL*
Value
0.21
0.04
0.64
0.38
0.47
0.49
0.95
Values are mean±SD.
f p<0.05 on Wilcoxon rank sum test.
'Patients with ICH, n=4; patients without ICH, n=192.
*
1242
Downloaded From: http://journal.publications.chestnet.org/ on 09/29/2016
Clinical
Investigations
Table 4.Intracranial Hemorrhages
Assignment
Percent of nonperfused
47
rt-PA, 50-90 mg 0
4
rt-PA, 100 mg
rt-PA, 0.6 mg/kg bolus
1
(2.9)
(1.7)
Urokinase, 2,000 U/lb/hX24h
Urokinase, 3 million units/2h
0
1 (2.2)
thrombolysis
rapidly
5.4
0.43 6.3
*<3=small size PE; 4-6 moderate size PE; 7-9=massive PE.9
=
45
is an infrequent but at times fatal
thrombolysis The
complication. presence of diastolic hypertension
on hospital admission is associated with a higher risk
of ICH. Younger patients appear to be at very low
risk for thrombolysis-related ICH after PE.
The frequency of ICH after thrombolysis for PE in
our series (1.9%; 95% CI, 0.7 to 4.1%) is slightly
in most studies of thrombolysis for
higher thaninfarction
(0.3 to 1.4%).8 The wide CIs,
myocardialand small number
of hemorrhages render
however,
direct comparisons with larger series imprecise. Two
of the six ICHs occurred late (62.5 and 157 h) after
and were probably due to administra¬
thrombolysis
tion of heparin and not to thrombolysis. Further¬
more, the first and fifth patients (Table 5) received
that actually
thrombolysis in violation of the protocolknown
intra¬
excluded their participation because of
cranial disease (seizures and transient ischemic
Discussion
reverse
lung
Angiogram score* (0-9)
138
59
23
(Table 6).
can
Patients
With ICH Without ICH p Value
0.22 52
41
on scan
No significant differences are appar¬
lyric regimen.
ent given the small number of ICHs. All hemor¬
were primarily in the
rhages were lobar; noneor basal
brainstem, cerebellum,
ganglia (Table 5).
Clinical signs of ICH included confusion or de¬
creased level of consciousness (three), hemiparesis
(two), visual field deficits (two), aphasia (one), head¬
ache (one), and arm paresthesias (one). Four of the
six patients had the onset of ICH within 24 h of
treatment (range for
completion of1.5theto thrombolytic
157 h). The size of PE based on
all patients,
lung scans and angiograms did not differ in patients
with and without ICH
PE
Severity ofPE
Patients
Patients
Without ICH
Patients
With ICH (<
Drug
Table 6.ICH and
by Treatment
right ventric¬
at¬
tacks).
Previous studies of PE thrombolysis report only an
occasional ICH.910 The introduction of CT scanning
ular dysfunction and can reduce the rate of recurrent
PE.6 This overview demonstrates that ICH after PE
Table 5.Patient Data*
Time to
First
History'
Location of ICH
Age/Sex
Medical History
71/F
Prior DVT; seizure;
bilateral TKR
Left parietal
81/M
Bilateral TKR
Left frontal
67/M
Thyroid disease;
Left occipitoparietal RHH
and right parietal
RHH
Left occipital and
small right
of HTN
Neurologic
Symptoms
Admission
Symptoms After BP, mm Fibrinogen/
mg/dL PTT^ s
Thrombolysis/ h Hg
110/75
Drowsiness,
121
45
Outcome
Minor deficits after
rehabilitation
aphasia, and
right-sided
weakness
4-PPD
57/M
Angina pectoris
64/F
Alcoholism; anemia;
occipital
Right frontal
malignant
Right
hemiplegia.>coma
Altered mental
4,5
166/103
231
52
Death
1.5
110/70
358
26
Persistent RHH
62.5
132/88
60
Persistent RHH
157
Death due
194/101
PE
status
or
to recurrent
Ml
melanoma; TIA
65/F
Unknown
Right
temporoparietal
Headache and
arm
22
160/105
72
91
Death
paresthesias
HTN=hypertension; TKR=total knee replacement; RHH=right homonymous hemianopsia; PPD=purified protein derivative; TIA=transient
attack; MI.myocardial infarction.
fTime to symptoms attributed to the ICH after completion of thrombolytic infusion.
* Lowest
fibrinogen value in first 24 h.
§Partial thromboplastin time closest to ICH symptoms (within 6 h).
*
ischemic
CHEST/111 /5/MAY, 1997
Downloaded From: http://journal.publications.chestnet.org/ on 09/29/2016
1243
the mid-1970s allowed earlier and more accurate
diagnoses of ICH which
previously could be proved
or
at
only by craniotomy autopsy. Therefore, studies
done before CT scanning was available may have
underestimated the frequency of ICH.
The largest (to our knowledge) case series of PE
thrombolysisarepatients in(since the introduction of CT
scanning) listed Table 7. Only three ICHs
were described in 455 patients. Combining our
series with those listed in Table 7 yields an ICH
frequency of 9 in 767 patients or 1.2% (95% CI, 0.5
in
to
2.2%).n-17
All ICHs in our series were in lobar locations (ie,
cortical regions or subcortical white matter). This is
consistent with previous series of thrombolysis-re¬
lated and anticoagulant-related ICH.1820 By con¬
trast, ICH due to hypertensive small-vessel disease,
which is presumably the most common cause of ICH
overall, usually occurs in deep structures such as the
basal ganglia, thalamus, pons, or cerebellum.21
Hence, the underlying sites of vascular rupture and
mechanisms for pure hypertensive vs anticoagulation
and thrombolysis-related ICH are probably differ¬
ent. However, both acute and chronic hypertension
may play important contributory roles in these ICHs,
since three of six of our patients had a history of
and higher BPs were found on hospital
hypertension,
admission in those who suffered an ICH. A greater
than expected frequency of systemic arterial hyper¬
tension has also been found in patients with ICH
after myocardial infarction thrombolysis.22
Confusion or altered level of consciousness was
seen in three of six of our patients as an initial sign of
ICH. This is typical for ICH of any kind, which
confusion, focal
commonly produces ordrowsiness,
and cowork¬
nausea.
Sloan
weakness, headache,
in
ers19 noted alteration
consciousness in 65% of
those with ICH in the Thrombolysis in Myocardial
Infarction (TIMI) II trial.
Table 7.Studies ofPE Thrombolysis in the Era of
Head CT Scanning*
Studv
Year
UKEP11
Verstraete et al12
Levine et al13
Diehl et al14
European Cooperative
Study Group15
Meneveau et al16
Sors et al17
Present study
Total 767
?Overall ICH
No. of Patients
Receiving PE
Thrombolysis
1987
1988
34
1990
33
1992
54
129
1992
63
199389
1994
53
1996
No. of Patients
With ICH
In the largest study (to our knowledge) of ICH
after thrombolysis for myocardial infarction, clinical
data and brain imaging on 268 patients with ICH
were analyzed.22 Multivariate regression analysis
showed that age, weight, prior cerebrovascular dis¬
ease, diastolic and systolic BP, history of hyperten¬
sion, and combined thrombolytic therapy or acceler¬
ated thrombolytic therapy were significantly
associated with an increased risk of ICH. There are
similar trends toward increased age, decreased
and increased systolic BP in our patients with
weight,
ICH that might be statistically significant in a larger
series.
The available autopsy evidence on ICH after
myocardial infarction thrombolysis demonstrates
that a high proportion of these patients have cerebral
amyloid angiopathy.19-23 This condition is signifi¬
cantly more common with increasing age and in¬
volves the deposition of protein with amyloid char¬
acteristics in the walls of cortical and leptomeningeal
vessels. This deposition results in structural changes
more brittle and susceptible to
making theAn vessels
association between apoprotein E e4
rupture.
and cerebral amyloid angiopathy has re¬
phenotype
cently been described.24 Patients with amyloid angi¬
and chronic hypertension may be at increased
opathy
risk for thrombolysis-related ICH.
Despite increasing experience with thrombolysis,
the risk of ICH in an individual patient remains
defined. Diastolic hypertension increases the
poorly
risk of ICH after PE thrombolysis, probably in part
changes in the integrity of cerebral vessels.25
through
Patients with intracranial disease or diastolic hyper¬
tension should be offered thrombolysis only if the
benefits of treatment outweigh the risks. Careful
attention must be paid to clinical risk factors for
ICH, because fibrinogen levels do not appear to
provide incremental utility in assessing risk. Risk
factors for ICH identified in patients with myocar¬
dial infarction may be applicable to patients with PE.
Further research is necessary to clarify the underly¬
ing mechanisms for thrombolysis-related ICH. Im¬
detection of patients with vascular injury
proved
from hypertension or cerebral amyloid angiopathy
could permit identification of individuals at high risk
for developing ICH following thrombolysis.
1 Goldhaber
References
SZ, Vaughan DE, Markis JE, et al. Acute pulmo¬
plasminogen activator.
Thrombolytic therapy of acute pulmonary embolism: current
status and future potential. J Am Coll Cardiol 1987; 10:96Bnary
embolism treated with tissue
Lancet 1986; 2:886-89
2
312
frequency (n=767) is 1.2% (95% CI, 0.5 to 2.2%).
1244
Downloaded From: http://journal.publications.chestnet.org/ on 09/29/2016
104B
3 Goldhaber SZ, Markis JE, Kessler CM, et al.
treatment
Perspectives on
of acute pulmonary embolism with tissue plasminClinical
Investigations
ogen activator. Semin Thromb Haemost 1987; 13:171-77
4 Goldhaber SZ, Kessler CM, Heit J, et al. Randomized
controlled trial of recombinant tissue plasminogen activator
versus urokinase in the treatment of acute pulmonary embo¬
lism. Lancet 1988; 2:293-98
5 Goldhaber SZ, Kessler CM, Heit JA, et al. Recombinanttissue-type plasminogen activator versus a novel dosing regi¬
men of urokinase in acute pulmonary embolism: a random¬
ized controlled multi-center trial. J Am Coll Cardiol 1992;
20:24-30
6 Goldhaber SZ, Haire WT>, Feldstein ML, et al. Alteplase
heparin in acute pulmonary embolism: randomised
trial assessing right-ventricular function and pulmonary' per¬
fusion. Lancet 1993; 341:507-11
7 Bolus Alteplase Pulmonary Embolism Group. Reduced dose
bolus alteplase vs conventional alteplase infusion for pulmo¬
nary embolism thrombolysis. Chest 1994; 106:718-24
8 Levine MN, Goldhaber SZ, Gore JM, et al. Hemorrhagic
complications of thrombolytic therapy in the treatment of
myocardial infarction and venous thromboembolism. Chest
versus
1995; 108:291S-301S
9 The Urokinase Pulmonary7 Embolism Trial. A national coop¬
erative study. Circulation 1973; 47(suppl II): 1-108
10 Urokinase Pulmonary Embolism Trial Study Group. Uroki¬
nase streptokinase embolism trial: phase II trials results.
JAMA 1974; 229:1606-13
11 The UKEP Study. Multicentre trial on two local regimens of
urokinase in massive pulmonary embolism. Eur Heart J 1987;
8:2-10
12 Verstraete M, Miller GAN, Bounameaux H, et al. Intravenous
and
intrapulmonary
recombinant
tissue-type plasminogen
pulmonary embolism.
Circulation 1988; 77:353-60
13 Levine M, Hirsh J, Weitz J, et al. A randomized trial of a
single bolus dosage regimen of recombinant tissue plasmin¬
ogen activator in patients with acute pulmonary embolism.
Chest 1990; 98:1473-79
14 Diehl J-L, Meyer G, Igual J, et al. Effectiveness and safety of
bolus administration of alteplase in massive pulmonary em¬
bolism. Am J Cardiol 1992; 70:1477-80
activator in the treatment of massive
European Cooperative Study Group for Pulmonary Embo¬
lism. Effects of intravenous urokinase versus alteplase on total
pulmonary resistance in acute massive pulmonary embolism:
a European multicentre double-blind trial. J Am Coll Cardiol
1992; 19:239-45
16 Meneveau N, Bassand J, Schiele F, et al. Safety of thrombo¬
lytic therapy in elderly patients with massive pulmonary
embolism: a comparison with nonelderly patients. J Am Coll
Cardiol 1993; 22:1075-79
17 Sors H, Pacouret G, Azarian R, et al. Hemodynamic effects of
bolus vs 2-h infusion of alteplase in acute massive pulmonary7
embolism: a randomized controlled multicenter trial. Chest
1994; 106:712-17
18 Uglietta JP, O'Connor CM, Boyko OB, et al. CT patterns of
intracranial hemorrhage complicating thrombolytic ther¬
apy for acute myocardial infarction. Radiology 1991; 181:
15
555-59
19 Sloan MA, Price TR, Petito
CK,
et
al. Clinical features and
pathogenesis of intracerebral hemorrhage after rt-PA and
heparin therapy for acute myocardial infarction: the Throm¬
bolysis in Myocardial Infarction (TIMI) II Pilot and Random¬
ized Clinical Trial combined experience. Neurology 1995;
45:649-58
20 Kase CS, Robinson RK, Stein RW, et al. Anticoagulantrelated intracerebral hemorrhage. Neurology 1985; 35:943-48
21 Wiggins WS, Moody DM, Toole JF, et'al. Clinical and
computerized tomographic study of hypertensive intracere¬
bral hemorrhage. Arch Neurol 1978; 35:832-33
22 Gore JM, Granger CB, Simoons ML, et al. Stroke after
thrombolysis: mortality and functional outcomes in the GUSTO-I trial. Circulation 1995; 92:2811-18
23 Pendlebury WW, Iole ED, Tracy RP, et al. Intracerebral
hemorrhage related to cerebral amyloid angiopathy and t-PA
treatment. Ann Neurol 1991; 29:210-13
24 Greenberg SM, Rebeck GW, Vonsattel JP, et al. Apolipoprotein E epsilon 4 and cerebral hemorrhage associated with
amyloid angiopathy. Ann Neurol 1995; 38:254-59
25 Garcia JH, Ho KL. Pathology of hypertensive arteriopathy.
Clin North Am
Neurosurg
1992; 3:497-507
CHEST/111 /5/MAY, 1997
Downloaded From: http://journal.publications.chestnet.org/ on 09/29/2016
1245