From the Departments of Neurology (C.A.C.W., V.L.B., V.E.P.),
Ophthalmology (I.C.A.M., B.K.), and Radiology (C.H.), Boston University School
of Medicine, and the Boston University School of Public Health (M.R.W.),
Boston, Mass.
Correspondence to V.L. Babikian, MD, Department of Neurology, Boston University School of Medicine and Boston Veterans Administration Medical Center, 150 S Huntington Ave, Boston, MA 02130.
Methods—Records of 331 consecutive patients examined during a
47-month period at the Neurovascular Laboratory were reviewed. Of the
original 453 intracranial arteries, 186 middle cerebral arteries (MCAs)
satisfied qualifying criteria that excluded patients with cardiac
embolic sources. Forty-five MCAs ipsilateral to the
symptomatic eye constituted the study group. The control
group consisted of 141 asymptomatic MCAs. Microembolus
detection studies were performed on transcranial
Doppler instruments equipped with special software, and the degree
of carotid artery stenosis was measured by cerebral or MR
angiography or by color duplex studies.
Results—Microembolism was detected in 40.0% of study MCAs and
9.2% of controls (P<0.001). In the study group,
microembolic signals were detected in 61.9% of MCAs
tested within a week of symptom onset and 20.8% of those tested
afterward (P<0.001). Severe (
Conclusions—In patients without cardiac embolic sources, cerebral
microembolism is frequently present on the side of retinal
ischemia, particularly during the week after onset of symptoms.
It is often associated with severe stenosis or occlusion of the
ipsilateral carotid artery.
The notion of retinal embolism as an important cause of not only TMB
but also of other forms of retinal ischemia, such as central or
branch retinal artery occlusion, is predominantly based on clinical
observations.10 11 12 13 14 15 In some patients, embolic
material has been observed to course through retinal arterioles
during episodes of transient monocular visual
loss.10 11 However, retinal embolism tends
to occur transiently and intermittently16 and
thus cannot be excluded on clinical grounds alone.
Emboli that originate from ICA or more proximal vascular lesions can
travel not only to the retina but also to cerebral branches of the
ICA,11 17 and pathological evidence of cerebral
emboli originating from ICA plaques has been demonstrated in this
context.18 In addition, clinically undetectable
cerebral embolism has been shown with cerebral angiography and
indirectly with cerebral blood flow studies and electroencephalography
in patients with transient retinal
ischemia.19 However, these studies are
not routinely obtained because of their invasiveness or lack of
sensitivity. Therefore, the frequency of in vivo cerebral embolism in
patients with retinal ischemia is presently unknown.
High-intensity transient signals detected by TCD have been identified
in the intracranial vasculature of asymptomatic individuals
as well as in patients with symptoms of cerebral ischemia
associated with cardiac lesions20 21 22 or
high-grade ICA stenoses.23 24 25 Laboratory
models and pathological studies show that these signals can correspond
to microemboli composed of thrombus, platelet-rich aggregates,
atheromatous material, cholesterol, fat,
and gaseous material.26 27 The ability to detect
these signals provides a means to monitor cerebral microembolism in
vivo.
In this study we investigated the frequency of cerebral microembolism
in patients with clinical evidence of retinal ischemia,
including TMB, central and branch retinal artery infarction, and
ischemic oculopathy, and assessed its correlation with carotid
artery stenosis.
Clinical evidence of retinal ischemia included patients with
TMB and patients with retinal infarction (central and branch retinal
artery occlusion and ischemic oculopathy). TMB was defined as
painless, transient, monocular visual loss, with complete resolution
usually within minutes of symptom onset. Its diagnosis was clinically
verified by one of the authors, a neurologist. Cases of central or
branch retinal artery occlusion were confirmed by formal ophthalmologic
examination. Chronic ocular ischemia was diagnosed in one
patient, who had complete loss of vision ipsilateral to an occluded ICA
with associated neovascular glaucoma, clouding of the cornea, and
neovascularization of the iris. Patients with clinically silent retinal
emboli were not included in this study.
Patients with evidence of cardioembolic disease were excluded from the
study. All patients were routinely evaluated for the presence of
cardioembolic disease by history, physical examination, and admission
electrocardiograms; results of
transthoracic or transesophageal
echocardiograms were available in 23 patients with retinal
ischemia and in 94 patients with asymptomatic MCAs.
Six MCAs were excluded in the group of patients with retinal
ischemia and 57 in the group of asymptomatic MCAs.
Reasons for exclusion were atrial fibrillation (n=29),
prosthetic heart valves (n=13), akinetic segments (n=11) or
aneurysms (n=6) of the left ventricle, cardiac thrombi (n=2),
recent myocardial infarction (n=1), and right-left shunts (n=1). The
presence of aortic arch atheroma with debris was not
considered a criterion for exclusion.
After patients with cardioembolic disease were excluded, the study
group consisted of 45 MCAs in 44 patients with ipsilateral retinal
ischemia. All patients had unilateral symptoms except for one
patient who had episodes of TMB affecting one eye or the other at
different times. Thirty MCAs were in patients with TMB alone, 7 in
patients with TMB and transient ipsilateral hemispheric
ischemia or infarction, 7 in patients with central or branch
retinal artery occlusion, and 1 in a patient with ischemic
oculopathy as described above. The 141 asymptomatic MCAs in
119 patients who had no symptoms of the ipsilateral retina or cerebral
hemisphere and who had no cardioembolic lesions served as controls.
Antithrombotic medications prescribed during TCD testing were reviewed.
In the group of symptomatic MCAs, 40 were in patients
receiving antiplatelet agents or anticoagulants, 4 were in patients
on no antithrombotic agents, and data were not available in 1 case. The
corresponding figures for the asymptomatic group were 104,
34, and 3, respectively. An analysis of these data showed that
the study sample was too small to determine the potential effects of
these drugs on the prevalence of microembolic
signals.
TCD Studies
Cerebrovascular Imaging Studies
Statistical Methods
Microembolic signals were detected in 40.0% of MCAs in
the study group and in 9.2% of MCAs in the control group. The
difference between the two groups was significant (Table 1
The median time interval between onset of symptoms and TCD testing was
9 days (range, 0.16 to 250 days). Microembolic signals
occurred in 13 of 21 study group MCAs (61.9%) that were tested within
a week from symptom onset (Table 1
Twelve of the patients in the study group with
microembolic signals underwent
microembolic detection studies of both MCAs, one on the
side of the symptomatic eye, and one on the
asymptomatic side. Of these, 10 (83.3%) had
microembolic signals only in the MCA ipsilateral to the
symptomatic eye, 1 (8.3%) in both MCAs, and 1 (8.3%) in
only the asymptomatic MCA. Thus, when present,
microembolic signals usually occurred only on the side
of the symptomatic eye.
The incidence of microembolic signals correlated with
the degree of ICA stenosis proximal to the corresponding MCA.
Microembolic signals occurred in 19 of 75 MCAs (25.3%)
distal to ICAs with luminal stenosis of 70% or more and in
only 12 of 107 MCAs (11.2%) distal to ICAs with luminal
stenosis of 69% or less (P=0.013; OR, 2.7; 95% CI,
1.2 to 5.9). (Note that n=182 because there are
no data available on the severity of ICA stenosis for four MCAs
in the control group.)
Severe ICA stenosis occurred more frequently in the study group
than in the control group (Table 2
ICA stenosis of less than 30% was observed in five study group
patients, three without and two with microembolic
signals in the MCA ipsilateral to the symptomatic retina.
Transesophageal echocardiography
showed moderate to severe aortic arch plaques exceeding 5-mm thickness
in the latter two patients.
Because of the significant correlation between the frequency of
microembolic signals and the degree of ICA
stenosis and the higher frequency of severe ICA
stenosis in the study group, the interaction between these
variables was analyzed further. After we controlled for the
degree of ICA stenosis, the MCAs in the study group were still
five times more likely to have microembolic signals
than the MCAs in the control group (P<0.001; OR, 5.0; 95%
CI, 2.2 to 11.4).
Retinal ischemia is associated with ICA
disease.5 6 7 In this study a high-grade ICA
stenosis of 70% or more, or occlusion, was detected in 73% of
cases with retinal ischemia. The frequency of cerebral
microembolism correlated significantly with the severity of ICA
stenosis, suggesting that the source of the microemboli was
often the ipsilateral carotid artery. However, after we controlled for
the degree of ICA stenosis, microembolism was still five times
more likely to occur in the arteries of symptomatic
patients, indicating that additional factors contributed to the
association of cerebral microembolism with retinal ischemia.
One of these factors might be the level of activity of the ongoing
process in the vascular wall. It is of interest that embolism to the
retina may arise from embolic sources other than the ICA. Two patients
with cerebral microembolism and without ICA stenosis had severe
atheromatous aortic arch plaques.
The majority of our patients with retinal ischemia and cerebral
microembolism did not have symptoms of cerebral ischemia, nor
did the cases with cerebral microembolism in the
asymptomatic control group. These microemboli were
"silent" in that they did not result in clinically detectable
neurological deficits. Indirect evidence of clinically undetectable
cerebral involvement in patients with retinal ischemia has been
reported previously,19 and pathological studies
have shown incidental emboli in the lumen of cerebral arterioles
without evidence of ischemic changes in the adjacent brain
tissue.34 35 In addition, silent retinal embolism
is frequently detected by routine funduscopic examination in patients
without clinical evidence of retinal
ischemia.15 36 Thus,
asymptomatic retinal and cerebral emboli occur and do not
always precede clinically or radiologically evidenced ischemia
or infarction. Conversely, longitudinal follow-up studies suggest that
retinal emboli as detected on funduscopic
examination2 15 and cerebral
microemboli24 as detected by TCD are both
associated with an increased risk for subsequent cerebral infarction.
However, the magnitude of this risk is presently unknown.
This study has several limitations. First, the patient population
consisted predominantly of elderly white men, many with multiple risk
factors for cerebrovascular disease referred to a tertiary care medical
center. The relevance of our findings to an unselected population
should be interpreted in this context. Other mechanisms of retinal
ischemia may occur more frequently in different patient
populations. Second, data were collected retrospectively and should be
interpreted within the limits of such a study design. Third, the
effects of antiplatelet agents and anticoagulant drugs on cerebral
microembolism remain undetermined and may have affected the study's
results. Lastly, various imaging modalities were used to determine the
severity of ICA stenosis, thereby introducing a certain degree
of inconsistency in the methods.
In summary, in patients without cardioembolic sources, cerebral
microembolism is frequently present on the side of retinal
ischemia, particularly during the week after onset of
symptoms. It is often associated with severe stenosis or
occlusion of the ipsilateral ICA.
Received January 22, 1998;
revision received March 17, 1998;
accepted March 17, 1998.
© 1998 American Heart Association, Inc.
Original Contributions
Cerebral Microembolism in Patients With Retinal Ischemia
Abstract
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Background and Purpose—We
investigated the frequency of cerebral microembolism detected by
transcranial Doppler ultrasonography in patients with
clinical evidence of retinal ischemia, including transient
monocular blindness, central and branch retinal artery infarction, and
ischemic oculopathy, and assessed its correlation with carotid
artery stenosis. 
70%) carotid
stenosis or occlusion was more frequent in the study group
(P<0.001). Microembolic signals were
detected in 25.3% and 11.2%, respectively, of MCAs distal to carotid
arteries with 70% to 100% and 0% to 69% stenosis
(P=0.013).
Key Words: cerebral embolism • retina • ultrasonography, Doppler
Introduction
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Transient monocular
blindness is recognized as a warning sign for retinal and cerebral
infarction1 2 3 4 and is associated with ICA
disease.5 6 7 Common mechanisms of TMB are
embolism, causing transient occlusion of retinal arterioles, and
retinal vascular insufficiency secondary to a
hemodynamically significant stenosis of feeding
arteries.8 Other less common etiologies include
vasospasm, states of altered coagulability, and
thrombocytosis.9 Since treatment strategies may
be influenced by the mechanism of TMB, identification of the cause in
an individual patient is of considerable clinical importance.
Subjects and Methods
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Subjects
The records of consecutive patients who were examined
between March 29, 1993, and February 21, 1997, at the Neurovascular
Laboratory of this tertiary care medical center were reviewed. During
this time period, 453 intracranial arteries in 331 patients were
examined for the presence of microembolic signals. Of
these, all MCAs in patients with clinical evidence of ipsilateral
retinal ischemia with or without associated symptoms of
cerebral ischemia (n=51) and all asymptomatic MCAs
(n=198) were selected for the purpose of this study. MCAs on the side
of a recent carotid endarterectomy (n=10), arteries
with incomplete data sheets or missing medical records (n=6), and
arteries in patients with other concomitant cerebrovascular diagnoses
(n=7) were excluded.
Transcranial Doppler studies were performed on
either a TC-2000 or a TC-2020 instrument (Nicolet/EME) equipped with
software for microemboli detection. The Neurovascular Laboratory's
methods and criteria for identification of microembolic
signals have been described previously.23 28 The
criteria for identification of microembolic signals are
similar to the ones established by the Consensus
Committee.29
The presence of ICA disease proximal to the corresponding
MCA was determined by cerebral angiography, duplex ultrasound, or MRA.
An Ultramark 9-HDI instrument (Advanced Technology Laboratories) was
used for duplex imaging, and a 1.5-T Signa unit (General Electric
Medical Systems) was used for MRI. The original films were reviewed to
determine the degree of ICA stenosis. In patients with multiple
imaging modalities, the cerebral angiogram was used whenever it was
available, and the duplex ultrasound was preferred over the MRA. The
degree of ICA stenosis was determined by contrast cerebral
angiography in 63, duplex ultrasound in 101, and MRA in 18 arteries in
the study. The distribution of these three imaging modalities was not
different between the study and the control groups. Films were not
available for review for seven arteries in the study group and 19
arteries in the control group, and the radiology reports were used
instead. No imaging studies were obtained in four
asymptomatic arteries. The methods of determining the
degree of stenosis by contrast cerebral angiography and MRA
have been described in an earlier report.23 The
severity of stenosis by duplex ultrasound was based on the
criteria of Faught et al30 and Hood et
al.31 The degree of extracranial ICA was divided
into one of four categories: mild stenosis (0% to 29%),
moderate stenosis (30% to 69%), severe stenosis (70%
to 99%), and occlusion.
All data were stored on a personal computer with the use of
Microsoft Excel software (version 5.0). Group comparisons were made
with the 
2 test, Fisher's exact test
(two-tailed), or the Wilcoxon rank sum test. Adjusted group
comparisons were made with the Mantel-Haenszel
2 test. All statistical analyses were
performed at the Data Coordinating Center of the Boston University
School of Public Health with the use of the SAS System for Windows,
Release 6.12.
Results
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
The mean ages of the study and control populations were 70.0 years
(range, 51 to 91 years) and 67.1 years (range, 30 to 91 years),
respectively. Age distribution was not significantly different between
the two groups. All MCAs in the study group were from male patients, as
were all but four MCAs in the control group. 
). In the subgroup of patients with TMB
(with or without associated cerebral ischemia),
microembolic signals were detected in 16 of 37 MCAs
(43.2%). Again, there was a significant difference with the
asymptomatic group (P<0.001; OR, 7.5; 95% CI,
3.2 to 17.8).
View this table:
[in a new window]
Table 1. Frequency of Microembolic Signals in MCAs of
Patients With Ipsilateral Retinal Ischemia and in Asymptomatic
MCAs 
) and in only 5 of 24 MCAs (20.8%)
that were tested more than a week after the onset of symptoms. The
difference between the two groups was significant (P=0.005;
OR, 6.2; 95% CI, 1.6 to 23.1). ).
Severe ICA stenosis or occlusion was present proximal to 33
of 45 MCAs (73.3%) in the study group and 42 of 137 MCAs (30.7%) in
the control group (P<0.001; OR, 6.2; 95% CI, 2.9 to
13.2).
View this table:
[in a new window]
Table 2. Degree of ICA Stenosis Proximal to MCAs of Patients
With Ipsilateral Retinal Ischemia and Asymptomatic
MCAs
Discussion
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
In this study we determined the frequency of in vivo
cerebral microembolism in patients with retinal ischemia.
Microembolic signals were detected in 40% of MCAs in
patients with ipsilateral symptoms of retinal ischemia and in
62% of those tested within a week from symptom onset. These rates are
more than four and six times as high, respectively, as the rate of
cerebral microembolism in the asymptomatic controls. In
addition, microembolic signals were usually detected
only on the side of the symptomatic eye, indicating that
symptoms of retinal ischemia coincide with the presence of
cerebral microembolism. Despite this high frequency of cerebral
microembolism in the study group, it is likely that we underestimated
the true frequency of cerebral microembolism. Half-hour
microembolic signal detection studies may be too short
to identify all patients with cerebral
microembolism.32 Furthermore, the frequency of
microembolic signals peaks between 4 and 6
AM,32 a test period that is outside
the regular hours of our laboratory. In addition, although the
microembolic signal criteria that we used were similar
to those established by the Consensus
Committee,29 "small" particles less than 9 dB
in intensity were excluded, possibly resulting in an underestimation of
the true frequency of microembolism. Nevertheless, our findings support
the notion of an embolic etiology33 in
the majority of cases with retinal ischemia. Embolism may
either be the sole or one of several causal factors.
Selected Abbreviations and Acronyms
CI
=
confidence interval
ICA
=
internal carotid artery
MCA
=
middle cerebral artery
OR
=
odds ratio
TCD
=
transcranial Doppler ultrasonography
TMB
=
transient monocular blindness
References
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
This article has been cited by other articles:
 |
|
 |
|
  S. Sarkar, S. Ghosh, S. K. Ghosh, and A. Collier Role of transcranial Doppler ultrasonography in stroke Postgrad. Med. J., November 1, 2007; 83(985): 683 - 689. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. A.C. Wijman, J. A. Gomes, M. R. Winter, B. Koleini, I. C.A. Matjucha, V. E. Pochay, and V. L. Babikian Symptomatic and Asymptomatic Retinal Embolism Have Different Mechanisms Stroke, May 1, 2004; 35(5): e100 - e102. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Rapp, X. M. Pan, B. Yu, R. A. Swanson, R. T. Higashida, P. Simpson, and D. Saloner Cerebral Ischemia and Infarction From Atheroemboli <100 {micro}m in Size Stroke, August 1, 2003; 34(8): 1976 - 1980. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J. Tegos, E. Kalodiki, M. M. Sabetai, and A. N. Nicolaides Stroke: Pathogenesis, Investigations, and Prognosis: Part II of III Angiology, November 1, 2000; 51(11): 885 - 894. [Abstract] [PDF]
|
|
|
|
|
|
S. Meairs and M. Hennerici Four-Dimensional Ultrasonographic Characterization of Plaque Surface Motion in Patients With Symptomatic and Asymptomatic Carotid Artery Stenosis Stroke, September 1, 1999; 30(9): 1807 - 1813. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. M. Ahuja, S. Chaturvedi, D. Eliott, N. Joshi, J. E. Puklin, and G. W. Abrams Mechanisms of Retinal Arterial Occlusive Disease in African American and Caucasian Patients Stroke, August 1, 1999; 30(8): 1506 - 1509. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||