Surgical Neurology 64 (2005) 160 – 164

www.surgicalneurology-online.com
Vascular
Recurrence and extension of lobar hemorrhage related to cerebral amyloid
angiopathy: multivariate analysis of clinical risk factors
Akifumi Izumihara, MDa,*, Michiyasu Suzuki, MDb, Tokuhiro Ishihara, MDc
a
Department of Neurosurgery, Hikari City General Hospital, Hikari, Yamaguchi 743-0022, Japan
b
Department of Neurosurgery, Yamaguchi University School of Medicine, Yamaguchi 755-8505, Japan
c
First Department of Pathology, Yamaguchi University School of Medicine, Yamaguchi 755-8505, Japan
Received 29 June 2004; accepted 2 September 2004

Abstract Background: Many recent studies have analyzed clinical risk factors for the recurrence and
extension of intracerebral hemorrhage. However, they have not been investigated in patients with
lobar hemorrhage related to cerebral amyloid angiopathy (CAA).
Methods: We studied 40 surgically treated patients with lobar hemorrhage diagnosed histologically
as being related to CAA. To determine clinical factors influencing the recurrence and hematoma size
their clinical data (demographics, medical history, and radiographic and laboratory data) were
examined retrospectively and subjected to multivariate analysis.
Results: Twelve patients (30%) had recurrent lobar hemorrhage. Twenty-one patients had a small
hematoma and 19 had a large hematoma. Hypertension was the only significant clinical factor
influencing the recurrence of CAA-related lobar hemorrhage. There was no significant clinical factor
influencing the hematoma size of CAA-related lobar hemorrhage.
Conclusions: The history of hypertension is associated with an increase in the recurrence of CAA-
related lobar hemorrhage.
D 2005 Elsevier Inc. All rights reserved.
Keywords: Amyloid; Lobar hemorrhage; Recurrence; Extension

1. Introduction In many recent studies [1,2,7,8,11,19,21], clinical risk

factors for recurrent hemorrhage and hematoma extension
Recent aggressive diagnosis and treatment of hyperten-
influencing the outcome of patients with ICH have been
sion have reduced the proportion of deep hemorrhage
analyzed. However, they have yet to be investigated in
among all types of intracerebral hemorrhages (ICHs),
patients with CAA-related lobar hemorrhage. The purpose
whereas the proportion of lobar hemorrhage, which is often
of this study is to elucidate clinical factors influencing the
caused by cerebral amyloid angiopathy (CAA), tends to
recurrence and extension (hematoma size) of CAA-related
increase, especially in Europe and North America [5,11,21].
lobar hemorrhage by multivariate analyses.
CAA–related lobar hemorrhage usually affects elderly
normotensive individuals. The hematoma involves the
corticosubcortical region, extends from the cortex to the 2. Patients and methods
subarachnoid space, and is lobular in shape [15]. In
During a period of 11 years (1987-1997), 40 patients
addition, recurrent and multiple hemorrhages are a feature
with lobar hemorrhage were treated surgically and diagnosed
of CAA-related lobar hemorrhage [6,22]. These clinical
histologically as being related to CAA at our institute
characteristics have been reported to be associated with the
and 6 associated hospitals. CAA was diagnosed by tissue
adverse outcome of patients with CAA-related lobar
staining with Congo red and by immunohistochemical
hemorrhage [13].
staining with the use of anti–b-amyloid protein antibody
in surgical specimens of the adjacent brain parenchyma
* Corresponding author. Tel.: +81 833 72 1000; fax: +81 833 72 6018. obtained during hematoma evacuation or brain biopsy. When
E-mail address: mia5anay@siren.ocn.ne.jp (A. Izumihara). a histological examination demonstrated lobar hemorrhage
0090-3019/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.surneu.2004.09.010
 A. Izumihara et al. / Surgical Neurology 64 (2005) 160 –164 161

Table 1 arterial hypertension: systolic blood pressure z 160 mm Hg,

Univariate correlation between 11 independent variables and the recurrence diastolic blood pressure z 95 mm Hg, or both on at least 2
of CAA-related lobar hemorrhage
occasions and/or past or present use of antihypertensive
Variable Without recurrence With recurrence P agents), diabetes (previous diagnosis of diabetes and/or
(n = 28) [n (%)] (n = 12) [n (%)]
past or present use of antidiabetic agents), cigarette
Age (y)
smoking, alcohol drinking, drug habit, head injury, and/or
b 70 (n = 5) 5 (18) 0 (0)
70-79 (n = 24) 15 (54) 9 (75) other brain disease; presence of liver dysfunction (both
z 80 (n = 11) 8 (29) 3 (25) .24 glutamic oxaloacetic transaminase and glutamic pyruvic
Gender transaminase at admission z 50 IU/L), dyscholesterolemia
Men (n = 17) 11 (39) 6 (50) (serum cholesterol at admission z 220 or b 120 mg/dL),
Women (n = 23) 17 (61) 6 (50) .53
and/or any systemic bleeding tendency; use of anticoagulant
Hypertension (n = 17) 9 (32) 8 (67) .09
Diabetes (n = 9) 4 (14) 5 (42) .14 and/or antiplatelet agents.
Cigarette smoking (n = 10) 7 (25) 3 (25) 1 Univariate (2-tailed Fisher exact test or v 2 test) and
Alcohol drinking (n = 13) 10 (36) 3 (25) .78 logistic regression analyses were performed with the use of
Head injury (n = 1) 1 (4) 0 (0) 1 StatView 5.0 software (SAS Institute, Cary, NC). Indepen-
Other brain disease (n = 3) 2 (7) 1 (8) 1
dent variables with a univariable probability value of less than
Liver dysfunction (n = 5) 4 (14) 1 (8) 1
Dyscholesterolemia .5 and those deemed clinically important were selected for
Hypocholesterolemia (n = 4) 4 (14) 0 (0) inclusion in the multivariable model. A probability value of
Hypercholesterolemia (n = 5) 3 (11) 2 (17) .36 less than .05 was considered significant.
Anticoagulant or antiplatelet 2 (7) 0 (0) .97
agents use (n = 2)
3. Results
Twelve patients (30%) had recurrent lobar hemorrhage.
in the presence of severe CAA [12] without other diagnostic Nine of these 12 patients had 1 previous episode of lobar
lesions, the diagnosis of CAA-related lobar hemorrhage hemorrhage and 3 had 2 previous episodes of lobar
was established. hemorrhage. Twenty-one patients had a small hematoma
To determine clinical factors influencing the recurrence and 19 had a large hematoma. Their ages ranged from 61 to
and hematoma size of CAA-related lobar hemorrhage, 91 years, with a mean age of 75.8 years. Five patients were
patients’ clinical records at the time of the latest CAA- aged less that 70 years, 24 were 70 or more, but less than 80
related lobar hemorrhage (the index lobar hemorrhage) were years, and 11 were 80 years or more. Seventeen patients
reviewed retrospectively, including demographics, medical were men and 23 were women. Seventeen patients (43%)
history, previous episode of lobar hemorrhage, and radio- had history of hypertension and 9 (23%) had diabetes. Ten
graphic and laboratory data. During follow-up periods of 3 (59%) of the 17 hypertensive patients were taking antihy-
to 93 months, with a mean of 28.7 months, none had new pertensive agents. Four (44%) of the 9 diabetic patients were
episode of lobar hemorrhage. To reveal clinical risk factors
for the recurrence of ICH, prospective studies are necessary; Table 2
however, in this study, we compared the incidence of the Univariate correlation between 11 independent variables and the hematoma
size of CAA-related lobar hemorrhage
possible clinical factors between patients who have previous
episode of lobar hemorrhage and those who do not. With Variable Small hematoma Large hematoma P
(n = 21) [n (%)] (n = 19) [n (%)]
respect to previous episode of lobar hemorrhage, the
frequency, the hematoma site, the interval between the 2 Age (y)
b 70 (n = 5) 3 (14) 2 (11)
lobar hemorrhage episodes, and whether there was histo- 70-79 (n = 24) 14 (67) 10 (53)
logical diagnosis of CAA were examined. The hematoma z 80 (n = 11) 4 (19) 7 (37) .45
size was evaluated as small ( b50 mL) or large (z 50 mL) by Gender
measuring on preoperative CT scanning according to the Men (n = 17) 11 (52) 6 (32)
formula ABC /2 [9], where A and B represent the largest Women (n = 23) 10 (48) 13 (68) .18
Hypertension (n = 17) 7 (33) 10 (53) .22
perpendicular diameters through the hemorrhage and C Diabetes (n = 9) 5 (24) 4 (21) 1
represents the thickness of the hemorrhage (the number of Cigarette smoking (n = 10) 6 (29) 4 (21) .86
10-mm slices containing hemorrhage). In the cases of Alcohol drinking (n = 13) 7 (33) 6 (32) .91
multiple lobar hemorrhage, which was defined as more Head injury (n = 1) 0 (0) 1 (5) .95
than 2 separate hemorrhages in the multiple lobes, the total Other brain disease (n = 3) 1 (5) 2 (11) .92
Liver dysfunction (n = 5) 4 (19) 1 (5) .41
size of all hematomas was evaluated. In addition, the Dyscholesterolemia
hematoma shape was described as lobular or nonlobular Hypocholesterolemia (n = 4) 3 (14) 1 (5)
(round, oval, or irregular). The independent variables Hypercholesterolemia (n = 5) 3 (14) 2 (11) .57
examined were age (b 70, z 70, and b80, or z 80 years); Anticoagulant or antiplatelet 1 (5) 1 (5) 1
gender; history of hypertension ( previous diagnosis of agents use (n = 2)
162 A. Izumihara et al. / Surgical Neurology 64 (2005) 160 –164

Table 3
Clinical features in 12 patients with recurrent lobar hemorrhage
Patient Age (y) Gender Hypertension Diabetes Cigarette Alcohol Liver Dyscholesterolemia Site of Interval between
smoking drinking dysfunction lobar hemorrhage episodes (mo)
Previous Latest
(index)
1 77 Female + Rt F,a Rt FPa Rt FPa 9, 2
2 70 Female + + Hypercholesterolemia Rt PO,a Rt Fa Lt Fa 70, 12
3 82 Male + Rt Fa Rt OPa 7
4 76 Male + + + Lt O, Rt O Rt OPT a 72, 58
5 74 Male + + Rt F Rt FPT a 12
6 75 Female + + Lt F Lt Fa 5
7 71 Female + Rt P Rt PT a 1
8 80 Male + + + Hypercholesterolemia ? Lt T a 36
9 78 Male + Lt F Lt OPa 56
10 75 Male + + Rt O Rt T a 13
11 80 Female + ? Lt FPa 24
12 77 Female + ? Rt POT a 4
Rt indicates right; Lt, left; F, frontal; P, parietal; T, temporal; O, occipital.
a
Related to CAA.

taking antidiabetic agents. Ten patients (25%) were cigarette
smokers and 13 (33%) were alcohol drinkers. None had
drug habit, 1 (3%) had head injury, and 3 (8%) had other
brain diseases (2 cerebral infarction and 1 chronic subdural
hematoma). Five patients (13%) had liver dysfunction, 4
(10%) had hypocholesterolemia, and 5 (13%) had hyper-
cholesterolemia. None had any systemic bleeding tendency.
Two patients (5%) were taking anticoagulant or antiplatelet
agents.
Univariate analysis demonstrated that 11 independent
variables did not have any significant influence on the
recurrence and hematoma size of CAA-related lobar
hemorrhage (Tables 1 and 2). Five potential clinical factors
(age, hypertension, diabetes, liver dysfunction, and dyscho-
lesterolemia) influencing the recurrence of CAA-related
lobar hemorrhage were selected for inclusion in the
multivariable model. Hypertension was the only significant
clinical factor influencing the recurrence of CAA-related
lobar hemorrhage (odds ratio, 6.55; 95% confidence
interval, 1.04- 41.34). Four potential clinical factors (age,
gender, hypertension, and liver dysfunction) influencing the
hematoma size of CAA-related lobar hemorrhage were
selected for inclusion in the multivariable model. There was
no significant clinical factor influencing the hematoma size
of CAA-related lobar hemorrhage.
In 12 patients with recurrent lobar hemorrhage, 8 patients
(67%) had the history of hypertension. Five (63%) of these
8 hypertensive patients were taking antihypertensive agents.
Hypertension had been diagnosed before previous lobar
hemorrhage in 6 patients and was disclosed during the first
hospitalization in 2 patients.
Twelve patients with previous lobar hemorrhage and 28
patients without previous lobar hemorrhage were followed
for 29.4 and 28.3 months (mean), respectively. Of a total of
15 previous episodes of lobar hemorrhage, 5 were treated
surgically and diagnosed histologically as being related to
CAA (Table 3). The mean interval between the 2 lobar
hemorrhagic episodes ranged from 1 to 72 months, with a
mean of 25.4 months. In 12 patients with recurrent lobar
hemorrhage, 3 patients (patients 1-3) had previous lobar
hemorrhage diagnosed histologically as being related to
CAA and 4 patients (patients 4 -7) had previous lobar
hemorrhage that occurred at the same site as the latest
CAA-related lobar hemorrhage. Five (71%) of these
7 patients with recurrent CAA-related lobar hemorrhage
had history of hypertension and 4 (57%) had diabetes.
One patient (14%) was a cigarette smoker and 2 (29%)
were alcohol drinkers. None had liver dysfunction and
1 (14%) had hypercholesterolemia. Six patients (15%) had
multiple lobar hemorrhage. Four (67%) of these 6 patients
had a large hematoma. Twenty-four patients (60%) had a
lobular-shaped hematoma. Fifteen (63%) of these 24 patients
had a large hematoma.
4. Discussion
ICH has been generally considered a one-time event with
exceptional recurrence [14]. However, many recent studies
have demonstrated that the recurrence of ICH is not rare. In
6 Asian series [1-3,16 -18], 2.7% to 10.8% of patients with
ICH had recurrent hemorrhage. The most common pattern
of recurrence was bdeep-deep,Q which is mainly caused by
hypertension. On the other hand, European and North
American studies have demonstrated that 6.4% to 24% of
patients with ICH had recurrent hemorrhage, and the
recurrence rate of ICH was 2.4% to 3.4% per year
[11,19,21]. In these studies, the lobar location of hemor-
rhage was a risk factor for the recurrence of ICH, whereas
hypertension was not, which are strongly suggestive of
CAA. Recently, O’Donnell et al [20] have reported that the
e4 and e2 alleles of apolipoprotein E, which are associated
with the severity of b-amyloid deposition and vascular
degeneration in CAA, are risk factors for recurrent lobar
hemorrhage, but hypertension is not. However, in their
 A. Izumihara et al. / Surgical Neurology 64 (2005) 160 –164
series, only 10 of 71 patients with lobar hemorrhage were
diagnosed histologically as being related to CAA (definite
CAA). In this study, we examined clinical factors influenc-
ing recurrence in 40 patients with CAA-related lobar
hemorrhage. Our data indicated that hypertension is the
only significant clinical risk factor for the recurrence of
CAA-related lobar hemorrhage. CAA is originally well
known as a common cause of lobar hemorrhage in elderly
nonhypertensive patients. However, several previous studies
have reported that clinical or pathological evidence of
hypertension is common in patients with CAA-related lobar
hemorrhage [4,10]. In this study, 17 (43%) of 40 patients
with CAA-related lobar hemorrhage had a history of
hypertension. On the other hand, among recurrent cases of
ICH, there seem to be more cases that are diagnosed as
being hypertensive due to a transient increase in blood
pressure at the acute and subacute phases of previous ICH,
high blood pressure at the chronic phase of previous ICH,
or advanced age. However, in this study, hypertension
has been diagnosed before previous lobar hemorrhage in
most cases. Therefore, we consider that our result is not
influenced by those biases. In addition, in our previous
immunohistochemical study, 8 (20%) of 41 patients with
CAA had severe CAA with severe cystatin C deposition,
which tended to induce recurrent hemorrhage, but the
difference was not significant [12]. We consider that the
history of hypertension is clinically as important for
the recurrence of CAA-related lobar hemorrhage as severe
CAA and that a strict control of blood pressure may
prevent the recurrence of CAA-related lobar hemorrhage
and improve the outcome of patients with CAA-related
lobar hemorrhage.
Hematoma extension should be investigated from 2
viewpoints, namely, hematoma growth and final hematoma
size. Several recent studies have analyzed clinical risk factors
for hematoma growth and indicated that short time interval
from onset to admission, elevated systolic blood pressure,
liver dysfunction, hemostatic ( platelet and coagulation)
abnormalities, heavy alcohol drinking, consciousness distur-
bance, a large hematoma, and an irregularly shaped
hematoma were associated with hematoma growth [7,8]. In
this study, we were not always able to observe the growth of
the hematoma, because all the patients with lobar hemor-
rhage were treated surgically and diagnosed histologically as
being related to CAA. Therefore, we evaluated hematoma
extension using preoperative hematoma size as a parameter.
Consequently, we were unable to find any significant clinical
factors influencing the hematoma size of CAA-related lobar
hemorrhage. On the other hand, CAA-related lobar hemor-
rhage is characterized by a lobular-shaped hematoma and
multiple hemorrhage [15,22]. Our data indicated that patients
with the lobular-shaped hematoma or multiple hemorrhage
tended to have a large hematoma. However, we exclud-
ed those clinical factors from this multivariate analyses
because they were not clinical factors determined before
hemorrhage. In addition, our previous immunohistochemical
163
study demonstrated that severe CAA with severe cystatin
C deposition was a significant risk factor for hematoma
extension and that loss of vascular smooth muscle was
evident in the amyloid-laden vascular walls in those cases
[12]. These findings suggest that hematoma extension is
associated with simultaneous bleeding from multiple arterio-
les and local hemostatic abnormalities.
References
[1] Arakawa S, Saku Y, Ibayashi S, Nagao T, Fujishima M. Blood
pressure control and recurrence of hypertensive brain hemorrhage.
Stroke 1998;29:1806 - 9.
[2] Bae HG, Jeong DS, Doh JW, Lee KS, Yun IG, Byun BJ. Recurrence
of bleeding in patients with hypertensive intracerebral hemorrhage.
Cerebrovasc Dis 1999;9:102 - 8.
[3] Chen ST, Chiang CY, Hsu CY, Lee TH, Tang LM. Recurrent
hypertensive intracerebral hemorrhage. Acta Neurol Scand 1995;91:
128 - 32.
[4] Cosgrove GR, Leblanc R, Meagher-Villemure K, Ethier R. Cerebral
amyloid angiopathy. Neurology 1985;35:625 - 31.
[5] Feldmann E. Intracerebral hemorrhage. Stroke 1991;22:684 - 91.
[6] Finelli PF, Kessimian N, Bernstein PW. Cerebral amyloid angiopathy
manifesting as recurrent intracerebral hemorrhage. Arch Neurol 1984;
41:330 - 3.
[7] Fujii Y, Takeuchi S, Sasaki O, Minakawa T, Tanaka R. Multivariate
analysis of predictors of hematoma enlargement in spontaneous
intracerebral hemorrhage. Stroke 1998;29:1160 - 6.
[8] Fujii Y, Tanaka R, Takeuchi S, Koike T, Minakawa T, Sasaki O.
Hematoma enlargement in spontaneous intracerebral hemorrhage. J
Neurosurg 1994;80:51 - 7.
[9] Gebel JM, Sila CA, Sloan MA, Granger CB, Weisenberger JP,
Green CL, Topol EJ, Mahaffey KW. Comparison of the ABC/2
estimation technique to computer-assisted volumetric analysis of
intraparenchymal and subdural hematomas complicating the GUSTO-
1 trial. Stroke 1998;29:1799 - 801.
[10] Gilbert JJ, Vinters HV. Cerebral amyloid angiopathy: incidence and
complications in the aging brain. I. Cerebral hemorrhage. Stroke 1983;
14:915 - 23.
[11] Hill MD, Silver FL, Austin PC, Tu JV. Rate of stroke recurrence in
patients with primary intracerebral hemorrhage. Stroke 2000;31:
123 - 7.
[12] Izumihara A, Ishihara T, Hoshii Y, Ito H. Cerebral amyloid angiopathy
associated with hemorrhage: immunohistochemical study of 41 biopsy
cases. Neurol Med Chir (Tokyo) 2001;41:471 - 8.
[13] Izumihara A, Ishihara T, Iwamoto N, Yamashita K, Ito H. Postoper-
ative outcome of 37 patients with lobar intracerebral hemorrhage
related to cerebral amyloid angiopathy. Stroke 1999;30:29 - 33.
[14] Kase CS, Mohr JP. General features of intracerebral hemorrhage. In:
Barnett HJM, Mohr JP, Stein BM, Yatsu FM, editors. Stroke:
pathophysiology, diagnosis, and management. New York7 Churchill
Livingstone; 1986. p. 497 - 523.
[15] Leblanc R, Preul M, Robitaille Y, Villemure J-G, Pokrupa R. Surgical
considerations in cerebral amyloid angiopathy. Neurosurgery 1991;
29:712 - 8.
[16] Lee KS, Bae HG, Yun IG. Recurrent intracerebral hemorrhage due to
hypertension. Neurosurgery 1990;26:586 - 90.
[17] Maruishi M, Shima T, Okada Y, Nishida M, Yamane K, Okita S.
Clinical findings in patients with recurrent intracerebral hemorrhage.
Surg Neurol 1995;44:444 - 9.
[18] Misra UK, Kalita J. Recurrent hypertensive intracerebral hemorrhage.
Am J Med Sci 1995;310:156 - 7.
[19] Neau JP, Ingrand P, Couderq C, Rosier MP, Bailbe M, Dumas P,
Vandermarcq P, Gil R. Recurrent intracerebral hemorrhage. Neurology
1997;49:106 - 13.
164 A. Izumihara et al. / Surgical Neurology 64 (2005) 160 –164

[20] O’Donnell HC, Rosand J, Knudsen KA, Furie KL, Segal AZ, Chiu RI,
Ikeda D, Greenberg SM. Apolipoprotein E genotype and the risk of
recurrent lobar intracerebral hemorrhage. N Engl J Med 2000;342:
240 - 5.
[21] Passero S, Burgalassi L, D’Andrea P, Battistini N. Recurrence of
bleeding in patients with primary intracerebral hemorrhage. Stroke
1995;26:1189 - 92.
[22] Tucker WS, Bilbao JM, Klodawsky H. Cerebral amyloid angiopathy
and multiple intracerebral hematomas. Neurosurgery 1980;7:611 - 4.
hypertension may be a risk factor for hemorrhage
recurrence, even in patients with CAA. The fact that none
of the 40 patients had a recurrence during the 28.7 months
of active surveillance is significant. Control of hyperten-
sion may reduce the risk of hemorrhage, even in patients
with CAA. The optimal level of blood pressure control
remains to be elucidated.

Daniel B. Hier, MD
Department of Neurology
University of Illinois at Chicago
Commentary Chicago, IL 60612-7330, USA

Izumihara et al have studied recurrent hemorrhage in a
group of 40 surgically treated lobar hemorrhages due to
CAA. CAA is distinct from systemic amyloidosis. Amy-
loid protein is deposited in cerebral arteries leading to
spontaneous lobar hemorrhages. Hypertension is generally
not believed to be a risk factor for such hemorrhages.
During the follow-up period of 28.7 months (mean), none
of the patients had a second hemorrhage. This meant that
Izumihara et al could not use survival analysis to examine
risk factors for recurrent hemorrhage. However, the authors
compared 12 patients with prior hemorrhage to 28 patients
without prior hemorrhage. Hypertension was more preva-
lent in the group with prior hemorrhage, suggesting that
I have read with interest the paper of Izumihara et al.
Their small series has been very well documented.
Although it is retrospective, they have taken care to
analyze their data in a statistically valid way. They have
found that hypertension is an independent risk factor for
the recurrence of intracerebral hemorrhage from CAA.
While the series is of insufficient size to prove the
hypothesis, it also appears that diabetes may be an
independent risk factor.
C. David Hunt, MD (Director of Neuroscience)
Maimonides Medical Center
Brooklyn, NY 11219, USA

To be able under all circumstances to practice five things

constitutes perfect virtue; these five things are gravity,
generosity of soul, sincerity, earnestness and kindness.
— Confucius