Case Report

Multiple Cerebral Infarctions From Nonbacterial Thrombotic Endocarditis
Mimicking Cerebral Vasculitis
ROBERT VASSALLO, MD; ELLEN D. REMSTEIN, MD; JOSEPH E. PARISI, MD; JOHN HUSTON III, MD; AND
ROBERT D. BROWN, JR, MD

Primary vasculitis of the central nervous system (PVCNS)
is an uncommon disorder that can present with a variety of
symptoms, making diagnosis and management difficult.
We describe a case of cerebral infarction that occurred
from nonbacterial thrombotic endocarditis (NBTE) and
presented with clinical and radiologic imaging features
that suggested PVCNS. The patient was a 58-year-old
woman with left hemiparesis, aphasia, and episodic confusion. Magnetic resonance imaging of the brain demonstrated multifocallesions consistent with infarction involving both cerebral hemispheres, and cerebral angiography
showed changes consistent with vasculitis. Although brain
biopsy findings were normal, the patient was treated for
presumed vasculitis with cyclophosphamide and prednisone. Four months later respiratory failure secondary to
polymicrobial pneumonia and adult respiratory distress
syndrome developed, and she died. Autopsy revealed
multiple infarcts in the heart, lungs, right kidney, spleen,

and brain. Multiple thrombotic platelet-fibrin vegetations consistent with NBTE were found on all cardiac
valves. Examination of the brain revealed no evidence of
active or healed vasculitis. Cerebral angiography may
show findings that suggest vasculitis, but it is not diagnostic, as several other conditions may cause similar changes.
Nonbacterial thrombotic endocarditis may cause multiple
cerebral infarctions and can be difficult to distinguish
from vasculitis, as specific diagnostic tests for PVCNS are
lacking.
Mayo Clin Proc. 1999;74:798-802

CNS = central nervous system; CSF = cerebrospinal fluid;
CT = computed tomographic; MRI = magnetic resonance
imaging; NBTE = nonbacterial thrombotic endocarditis;
PVCNS = primary vasculitis of the central nervous system;
TEE = transesophageal echocardiography; TTE = transthoracic echocardiography

N

resonance imaging (MRI) of the brain demonstrated multifocal lesions with abnormally increased T, signal involving
the left basal ganglia as well as the white matter and cortex
in both cerebral hemispheres consistent with cerebral
infarction (Figure 1). Some of thecortical lesions demonstrated a gyriform pattern. A few of the areas of T, signal
abnormality were enhanced after administration of gadolinium contrast. Cerebral angiography showed focal areas
of narrowing involving both large- and small-caliber vessels, and the findings were believed to be consistent with
primary vasculitis of the central nervous system (PVCNS)
(Figure 2). Examination of the cerebrospinal fluid (CSF)
yielded normal findings except for a slightly elevated protein count (57 mg/dL), and there were no oligoclonal
bands. Stereotactic brain biopsy specimen showed 2 subacute infarcts without evidence of vasculitis. An MRI of
the head with gadolinium performed for stereotactic guidance during cerebral biopsy (3 days after the first MRI)
revealed multiple foci of T, signal abnormality, as noted in
the previous MRI, and also new lesions involving both
cerebral hemispheres and both occipital lobes (Figure 3),
indicating progression of the disease process. Transthoracic and transesophageal echocardiography (TTE and
TEE) revealed normal valves without evidence of intracardiac thrombi.

onbacterial thrombotic endocarditis (NBTE) is an unusual cause of systemic embolization that is most
frequently associated with malignancy and hypercoagulable states. Diagnosis is difficult and is most commonly
made at postmortem examination.
Described herein is a case of cerebral and systemic
embolization that probably resulted from NBTE that was
not detected on extensive clinical evaluation but was found
at autopsy. There was no associated malignancy, hypercoagulable state, or evidence of disseminated intravascular
coagulation.
REPORT OF A CASE
A 58-year-old woman presented in October 1996 with left
hemiparesis, aphasia, and episodic confusion. Magnetic
From the Thoracic Diseases Research Unit, Division of Pulmonary
and Critical Care Medicine and Internal Medicine (R.V.), Division of
Anatomic Pathology (E.D.R., J.E.P.), Department of Radiology (J.H.),
and Department of Neurology (R.D.B.), Mayo Clinic Rochester, Rochester, Minn.
This study was supported in part by National Institutes of Health
grant NS 06663.
Address reprint requests and correspondence to Robert D. Brown,
Jr, MD, Department of Neurology, Mayo Clinic Rochester, 200 First
sr 5W, Rochester, MN 55905.

Mayo Clin Proc. 1999;74:798-802

798

© 1999 Mayo Foundation/or Medical Education and Research

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

Mayo Clin Proc, August 1999, Vol 74

Figure 1. Magnetic resonance image of the brain demonstrates
abnormally increased T2 signal involving the left caudate nucleus
and basal ganglia.

Serologic studies, including antinuclear antibodies,
lupus anticoagulant, anticardiolipin antibodies, rheumatoid
factor, cryoglobulins, hepatitis serologies, and extractable
nuclear antigens, were negative. A comprehensive hemostatic survey including procoagulant factor levels, proteins
C and S, antithrombin III, and factor V Leiden levels
showed results within normal limits.
Based on the clinical picture of multiple neurologic
deficits, the MRI findings, and the cerebral angiographic
findings, the patient was treated for presumed PVCNS with
cyclophosphamide, 2 mg/kg per day, and prednisone, 1
mg/kg per day.
In April 1997, she was admitted to our hospital because
of shortness of breath and cough, which progressed in spite
of treatment with broad-spectrum antibiotics. She had not
been receiving prophylaxis for Pneumocystis carinii pneumonia because of an allergy to sulfa drugs. The prednisone
dose had been tapered to 10 mg/d.
Her medical history included treated essential hypertension, left carotid endarterectomy in June 1996 for
amaurosis fugax, mild hyperlipidemia, steroid-induced
diabetes, and fibromyalgia. The medications included cyclophosphamide, prednisone, ceftriaxone, and fluconazole.
Physical examination revealed bilateral coarse crackles
in the lower lung fields, a soft systolic murmur at the base,
and an unremarkable abdominal examination. Neurologic
examination revealed mild left-sided weakness but was
otherwise unremarkable.

Cerebral Infarction From Nonbacterial Thrombotic Endocarditis

799

On the day after admission, bronchoscopy was performed, and P carinii organisms were seen on bronchoalveolar lavage. Following bronchoscopy she required
intubation and mechanical ventilation because of worsening respiratory failure. Intravenous treatment with
pentamidine was started because of her allergy to sulfonamides. In spite of maximal ventilatory support, treatment
with pentamidine, and broad-spectrum antibiotic coverage,
her condition progressively deteriorated with worsening
respiratory failure. A computed tomographic (CT) scan of
the head, which was performed because of excessive
drowsiness and altered mentation, demonstrated unchanged, multiple, small cortical and subcortical infarcts.
Echocardiography was not performed during the hospital
stay because the CT scan showed no evidence of recurrent
cerebral infarction and also because a cardiac source of
emboli was not suspected on clinical grounds.
Two weeks after admission, air was noticed under the
diaphragm on chest x-ray examination, and further evaluation showed diverticulitis with diverticular abscess and
bowel perforation. At this point the family wished to withdraw support, and the patient died.
AUTOPSY FINDINGS
At autopsy, examination of the heart demonstrated plateletfibrin vegetations extensively involving the aortic, mitral,
pulmonary, and tricuspid valves (Figure 4). The subjacent
valvular tissue showed minimal reactive changes only.
Special stains (Gram, Grocott methenamine silver) used on
sections of the vegetations from all 4 valves were negative

Figure 2. Cerebral angiogram demonstrates diffuse changes involving both large-caliber (arrow) and small-caliber (curved arrow) vessels consistent with vasculitis.

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

800

Cerebral Infarction From Nonbacterial Thrombotic Endocarditis

Mayo Clio Proc, August 1999, Vol 74

Figure 3. Repeat magnetic resonance image (MRI) of the head performed 3 days after the first MRI showed
significant progression with multiple new areas of T 2 signal abnormality in both cerebral hemispheres (A) and
both occipital lobes (B).

for bacterial or fungal organisms, supporting the diagnosis
of NBTE. Embolic infarcts of variable age with associated
platelet-fibrin thrombi, presumably originating from the
platelet-fibrin vegetations found on the cardiac valves,
were identified involving multiple organs, including heart,
lungs, right kidney, spleen, and brain. The heart showed
acute infarcts of the posteroseptalleft ventricle and posterior papillary muscle, the lungs contained multiple peripheral acute and chronic infarcts involving all lobes except
the right middle lobe, the right kidney contained multiple

Figure 4. Extensive involvement of the cardiac valves with platelet-fibrin vegetations was seen at autopsy.

acute and chronic infarcts, and multiple chronic infarcts
were found in the spleen. Parenchymal vessels showed no
evidence of acute or healed vasculitis, and further examination of mesenteric vessels using an elastic (Verhoeff-Van
Gieson) stain also showed no acute or healed vasculitis.
Examination of the brain revealed multiple acute, subacute, and chronic ischemic infarcts involving white and
gray matter, also associated with platelet-fibrin thrombi.
These infarcts involved multiple sites, including bilateral
superior and middle frontal, superior temporal, inferior
temporal, and superior occipital gyri, the left putamen, the
right dentate nucleus and deep cerebellar white matter, and
the left peripheral cerebellar hemisphere. The findings of
the previous left frontal brain biopsy in October 1996 also
included multiple subacute infarcts, but no definite platelet-fibrin thrombi were identified. The bilateral temporal
arteries, the left middle cerebral artery, and the parenchymal vessels from the autopsy brain specimen, as well as the
parenchymal vessels from the October 1996 brain biopsy,
were completely free of acute or healed vasculitis. In addition, transmission electron microscopy performed on the
brain tissue from both the autopsy specimen and the prior
left frontal biopsy showed that the cerebral vessels contained focal deposits of phagocytosed lipid and cell membrane debris (nonspecific findings), but cerebral autosomal
dominant arteriopathy with subcortical infarcts and leukoencephalopathy was not evident. Generalized, mild (grade

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

Mayo Clin Proc, August 1999, Vol 74

1 of 4) atherosclerosis was present in the right and left
carotid arteries and cranial vessels from the autopsy brain.
The patient ultimately died of respiratory failure, with
evidence of diffuse alveolar damage (clinically adult respiratory distress syndrome) and polymicrobial pneumonia at
autopsy. The diffuse alveolar damage was manifest by
marked diffuse hyaline membrane formation. Grocott
methenamine silver staining of sections from the right
lower lobe and the left lung showed rare organisms consistent with P carinii, cytomegalovirus in situ hybridization
demonstrated occasional cytomegalovirus inclusions in the
left lung, and postmortem lung cultures were positive for
cytomegalovirus and coagulase-negative staphylococcus,
providing evidence of polymicrobial pneumonia. In addition, sigmoid diverticulitis and focal acute fibrinous peritonitis with rare cytomegalovirus inclusions on in situ hybridization were identified. No occult malignancy was
found.

COMMENT
This unusual case illustrates a number of important insights
into the diagnosis of PVCNS and NBTE.
An uncommon condition, PVCNS is difficult to diagnose because of protean clinical manifestations and a lack
of specific noninvasive markers. Analysis of CSF in
PVCNS may show nonspecific abnormalities such as an
increase in protein or pleocytosis. In some studies, CSF
examination yielded normal findings in approximately
50% of cases of PVCNS diagnosed without a brain biopsy.'> However, in a larger study involving 70 cases of
histopathologically proven vasculitis, the CSF was abnormal in 90% of cases,' suggesting that normal CSF findings
argue against but are not inconsistent with a diagnosis of
vasculitis.
The MRI is abnormal in most patients with PVCNS. 4 ,5
Common findings include supratentorial infarcts involving
both gray and white matter, which are often bilateral in
distribution. Harris et aIS described 70 patients who underwent MRI examination of the head as part of a work-up to
exclude vasculitis. Of 7 patients eventually diagnosed as
having vasculitis, Harris et al found abnormalities on all 7
MRI examinations. The diagnosis of PVCNS was established by angiography in 5 of these patients and by cerebral
biopsy in the other 2 patients. Nineteen patients had completely normal MRI examinations, and of these, none was
diagnosed as having vasculitis after the complete evaluation. They concluded that a normal MRI examination of the
head makes the diagnosis of intracranial vasculitis unlikely. However, none of the patients with normal MRI
examinations underwent cerebral biopsy, and no data in the
literature compare the sensitivity of MRI with brain biopsy.
Although the sensitivity ofMRI in the diagnosis ofPVCNS

Cerebral Infarction From Nonbacterial Thrombotic Endocarditis

801

appears to be high, it lacks specificity in that there is no
characteristic pattern of vasculitis on MRI, and the test is
poor in differentiating vasculitis from other conditions."
The most important tests in establishing the diagnosis of
PVCNS have traditionally been cerebral angiography and
cerebral biopsy (the "gold standard"). However, the sensitivity, specificity, and predictive value of these tests are not
known with certainty. In up to 38% of histologically confirmed cases of PVCNS, cerebral angiography is normal,'
and conversely, even the classic picture of alternating
stenosis and dilatation ("beading") on cerebral angiography is not specific for CNS vasculitis and may be seen in
other conditions,' as this case illustrates. The findings
may be seen with a variety of neoplastic, degenerative, infectious, and vasospastic vascular disorders." In
addition, recanalized emboli can also cause similar findings. The cerebral angiogram in our patient was interpreted
by 2 expert neuroradiologists in a tertiary care center
with considerable experience in vasculitis. Although not
pathognomonic of vasculitis, the findings on angiography
were believed to be consistent with this diagnosis. One can
speculate that the changes on angiography may have been
caused by recanalized emboli that originated from the vegetations on the cardiac valves.
Brain biopsy is considered the gold standard for the
diagnosis of PVCNS, but because of the patchy nature of
the disease, the result may be a false negative. Calabrese et
al' reported that of 43 premortem brain biopsies, 32 were
interpreted as positive and 11 were nondiagnostic or
negative. The latter 11 cases were later confirmed at autopsy to be vasculitis, giving biopsy a 74.4% sensitivity in
the diagnosis of vasculitis. In addition, biopsy is essential
to rule out other disorders, such as neoplastic conditions
(especially lymphoma), chronic infections, and inflammatory processes that may mimic vasculitis clinically.
Although the sensitivity of brain biopsy is only moderate,
a suggestive biopsy markedly raises the threshold for
treatment with aggressive immunosuppressive therapy. If
the biopsy is negative or nondiagnostic, a presumptive
diagnosis of cerebral vasculitis may be considered on the
basis of an appropriate clinical picture together with consistent MRI and angiographic findings, after other diagnoses have been excluded. Because untreated cerebral
vasculitis can lead to devastating and permanent neurologic deficit, the clinician may be compelled to treat in
these circumstances.
The optimal treatment for PVCNShas not been established by means of prospective randomized trials but is
rather based on anecdotal experience and case reports.
Treatment usually includes prednisone at a dose of 1 mg/kg
per day, sometimes in conjunction with cyclophosphamide
at 2 mg/kg per day." It is important to remember that the

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

802

Cerebral Infarction From Nonbacterial Thrombotic Endocarditis

treatment of the condition itself is associated with considerable morbidity and mortality, as seen in our patient who
died of infection caused by opportunistic pathogens.
An unusual cause of systemic embolization, NBTE occurs most frequently in patients with malignancy and
sepsis." It is difficult to diagnose clinically as there are no
pathognomonic signs or symptoms, and even cardiac murmurs may be absent.'? Echocardiography is at present the
most useful diagnostic modality available for the diagnosis
of NBTE,9.12 even though there are no data on the sensitivity of echocardiography in the diagnosis of NBTE before
death. Lopez et al" reported 6 cases ofNBTE diagnosed at
necropsy; all 6 patients had had TTE performed within 45
days of death. Valvular vegetations were demonstrated in 4
of 6 patients with pathologic evidence of NBTE. All leftsided valvular vegetations larger than 4 mm were accurately detected in this study. However, TEE is more effective in the detection of valvular vegetations in infective
endocarditis," and it may be more sensitive than TTE in the
diagnosis of NBTE. There is only 1 case report in the
English-language literature in which the diagnosis of
NBTE was established by TEE, which showed vegetations
on all 4 cardiac valves, while TTE showed only a questionable mitral valve vegetation." If the clinical suspicion of
NBTE is high and TTE is normal, it is reasonable to perform TEE to achieve optimal visualization of the cardiac
valves, although TEE may still give a false-negative result.
Multiple brain infarcts have been described in the setting of NBTE.15 In this case we speculate that NBTE was
the original cause of the patient's multiple cerebral infarcts
although we cannot implicate a specific cause for the
NBTE. The NBTE probably preceded the systemic infections that led to respiratory failure and eventually death in
this patient. The TEE and TTE performed at the time of
diagnosis of the "cerebral vasculitis" did not show vegetations, but unfortunately these tests were not repeated prior
to her death 4 months later because there was no suspicion
of recurrent embolic events.

Mayo Clin Proc, August 1999, Vol 74

that resulted in angiographic findings suggestive of cerebral vasculitis. Cerebral angiography must be interpreted
with caution as a number of conditions can cause changes
on angiography that mimic vasculitis. Cerebral biopsy is
the preferred technique to establish the diagnosis of
PVCNS, but it is not always abnormal in PVCNS. Nonbacterial thrombotic endocarditis should be considered in the
evaluation of patients with multiple cerebral infarcts and an
abnormal arteriogram, particularly when the cortical!
meningeal biopsy findings are normal.

REFERENCES
I.

2.

3.
4.
5.
6.

7.
8.
9.
10.
11.
12.

CONCLUSION
Primary vasculitis of the CNS is difficult to diagnose with
noninvasive studies, and even aggressive evaluation with
brain biopsy may not yield diagnostic findings. Accurate
diagnosis requires careful clinical, radiologic, and pathologic correlation. This case represents the first report we
are aware of multiple emboli to the brain likely arising
from occult nonbacterial vegetations on the cardiac valves

13.
14.
15.

Scolding NJ, Jayne DRW, Zajicek JP, Meyer PAR, Wraight EP,
Lockwood CM. Cerebral vasculitis-recognition, diagnosis and
management. QJM. 1997;90:61-73.
Stone JH, Pomper MG, Roubenoff R, Miller TJ, Hellmann DB.
Sensitivities of noninvasive tests for central nervous system
vasculitis: a comparison of lumbar puncture, computed tomography, and magnetic resonance imaging. J Rheumatol. 1994;
21:1277-1282.
Calabrese LH, Furlun AJ, Gragg LA, Ropos TJ. Primary angiitis of
the central nervous system: diagnostic criteria and clinical approach. Cleve ClinJ Med. 1992;59:293-306.
Greenan TJ, Grossman RI, Goldberg HI. Cerebral vasculitis: MR
imaging and angiographic correlation. Radiology. 1992;182:65-72.
Harris KG, Tran DD, Sickels WJ, Cornell SH, Yuh WTC. Diagnosing intracranial vasculitis: the roles of MR and angiography. AJNR
Am J Neuroradiol. 1994;15:317-330.
Miller DH, Ormerod lEC, Gibson A, du Boulay EPGH, Rudge P,
McDonald WI. MR brain scanning in patients with vasculitis:
differentiation from multiple sclerosis. Neuroradiology. 1987;29:
226-231.
Ferris EJ, Levine HL. Cerebral arteritis: classification. Radiology.
1973;109:327-341.
Cohen BA, Biller J. Hemorrhagic stroke due to cerebral vasculitis
and the role of immunosuppressive therapy. Neurosurg Clin N Am.
1992;3:611-624.
Joffe II, Jacobs LE, Owen AN, Ioli A, Kotler MN. Noninfective
valvular masses: review of the literature with emphasis on imaging
techniques and management. Am Heart J. 1996;131:1175-1183.
Lopez JA, Ross RS, Fishbein MC, Siegel RJ. Nonbacterial
thrombotic endocarditis: a review. Am Heart J. 1987;113:773-784.
Edoute Y, Haim N, Rinkevich D, Brenner B, Reisner SA. Cardiac
valvular vegetations in cancer patients: a prospective echocardiographic study of 200 patients. Am J Med. 1997;102:252-258.
Blanchard DG, Ross RS, Dittrich HC. Nonbacterial thrombotic
endocarditis: assessment by transesophageal echocardiography.
Chest. 1992;102:954-956.
Lopez JA, Fishbein MC, Siegel RJ. Echocardiographic features of
nonbacterial thrombotic endocarditis. Am J Cardiol. 1987;59:478480.
Heinle SK, Kisslo J. The clinical utility of transesophageal
echocardiography in patients with left-sided infective endocarditis.
Am J Card Imaging. 1995;9:199-202.
Fujishima S, Okada Y, Irie K, et al. Multiple brain infarction and
hemorrhage by nonbacterial thrombotic endocarditis in occult lung
cancer-a case report. Angiology. 1994;45:161-166.

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.