0148-396X /80/0705-0488802.00/0
NEUROSURGERY
Copyright © 1980 by the Congress of Neurological Surgeons

Vol. 7, No. 5. 1980.
Printed in U.S.A

Subarachnoid Hemorrhage—the Initial Manifestation of
Bacterial Endocarditis

Report of a Case with Negative Arteriography and Computed Tomography

Frederick M. Vincent, M.D., J. Eric Zimmerman, M.D., Thomas C. Auer, M.D., and David B. Martin, M.D.

Munson Medical Center, Traverse City, and Departments of Medicine and Pathology, Michigan State University College of Human Medicine, East
Lansing, Michigan

The case of a patient who suffered a nontraumatic subarachnoid hemorrhage (SAH) associated with normal angiography
is reported. Three weeks later he developed an embolic stroke secondary to a nonhemolytic Staphylococcus epidermidis
endocarditis of the mitral valve; thus, the SAH was the initial manifestation of bacterial endocarditis. Bacterial endocarditis
should be considered a possible cause of SAH, especially in the 7% of patients with angiographically negative SAH.

( Neurosurgery, 7: 488-490, 1980)

Key words: Bacterial endocarditis, Cerebral aneurysm, Cerebral angiography, Cerebral embolism, Embolic stroke, Mitral

valve, Subarachnoid hemorrhage

Neurological complications remain a significant problem in
bacterial endocarditis (BE), although rarely will patients pres-
ent with nervous system (NS) complications (4, 10). We are
presenting the clinical features of a patient with a nontraumatic
subarachnoid hemorrhage (SAH) associated with negative pan-
cerebral arteriography and computed tomography (CT), 3
weeks after the SAH he developed an embolic stroke secondary
to a nonhemolytic Staphylococcus epidermidis endocarditis of
the mitral valve.

CASE REPORT

A 24-year-old, right-handed man developed the sudden on-
set of headache, nausea, and transient right upper extremity
paresis. There was no history of trauma, and he was not
receiving any medication. His past history was unremarkable
except for an episode of rheumatic fever at age 11, which had
left him with a persistent heart murmur. There was no history
of antecedent infection or dental work. His general examina-
tion was normal except for the presence of a pansystolic heart
murmur. He was afebrile, and his neurological examination
was normal, except for mild nuchal rigidity. A CT scan was
normal and a lumbar puncture revealed bloody cerebrospinal
fluid (CSF) with 70,000 red blood cells (RBCs)/mm* and 37
white blood cells (WBCs)/mm’; the protein content was 48
mg/dl, the glucose content was 46 mg/dl, and the CSF was
xanthochromic after centrifugation; routine smears for bacteria
and fungi were normal, as were cultures. A diagnosis of SAH
was made and the patient was treated with bedrest, fluid
restriction, phenobarbital, and epsilon-aminocaproic acid (36
g/day). Six days after admission pancerebral arteriography was
performed and was normal. The patient was discharged from
the hospital 6 days later without any complaints.

Three days after his discharge from the hospital he acutely
developed right arm paresis and numbness, dysphasia, and
partial seizures involving the right upper extremity. On exam-
ination he was afebrile with normal vital signs, and his neck
was supple. Osler’s nodes were present on the fingers and his
spleen was enlarged. A Grade IV/VI pansystolic murmur was
heard at the apex with radiation to the axilla; it was louder
than on previous examination. A mild right hemiparesis and
hemianesthesia were present; muscle stretch reflexes were in-
creased on the right, and a right Babinski sign was present. He

488

also had a slight expressive dysphasia. CT revealed a hemor-
rhagic infarct in the left parietal area, His hemoglobin value
was 11.8 g, and his WBC count was 8400 with 73% polymor-
phonuclear leukocytes. An erythrocyte sedimentation rate was
44 mm/hour. Because of the evidence of an embolic stroke, a
heart murmur that had increased in intensity, and evidence of
systemic emboli, the diagnosis of BE was considered strongly.
Eleven blood cultures were obtained and the patient was then
started on treatment with tobramycin, 90 mg tid, and oxacillin,
12 g/day. Echocardiography of the heart revealed left atrial
enlargement, but no evidence of vegetations or tumor. Seven
days after being drawn, the blood cultures (10 of 11) became
positive, as did a culture obtained at the time of biopsy of an
Osler’s node; the offending organism was a coagulase-negative
Staphylococcus epidermidis. Ten days after admission, the pa-
tient developed left upper quadrant pain; a liver-spleen scan
was positive for an infarct in the spleen, which had not been
present on the study performed on admission to the hospital.
He remained afebrile throughout his hospital stay and was
discharged after 6 weeks of intravenous antibiotic therapy. A
CT scan at this time was normal, except for a slight lucency
where the hemorrhagic stroke had occurred. Blood cultures
made after the discontinuation of antibiotics revealed no
growth of organisms. The following laboratory tests gave nor-
mal results during his hospital stay: antinuclear antibody test,
monospot, cryoglobulins, rheumatoid factor, creatine phospho-
kinase, lactic dehydrogenase, serum glutamic-pyruvic trans-
aminase, prothrombin time, partial thromboplastin time, blood
urea nitrogen, antistreptolysin O titer, CSF analysis (including
cultures), bone scan, and gallium scan. The patient remains
well 1% years later, without recurrent NS symptoms or signs.

DISCUSSION

Fifteen to 30% of patients with BE will develop neurological
signs at some time during their illness, and 6 to 15% will present
with a neurological chief complaint (4, 5, 10). In Osler’s series
the incidence of neurological signs was 12.5%; 3% of his patients
presented with primary neurological problems (8). A higher
percentage of patients with BE with NS complications will die
than will those without NS involvement (5, 6, 10). In a review
of 218 patients with BE by Pruitt et al., 84 of them (39%) had
neurological complications; 49 (58%) of these patients died, in

November 1980

contrast to 27 (20%) of those who did not suffer NS complica-
tions (10). Of these 218 patients with BE, 36 (16.5%) presented
initially with NS manifestations, and a SAH was the initial
manifestation of BE in 0.6% (10). Jones et al. found that the
mortality rate in their patients with NS complications was 50%,
with an overall mortality rate of 14% among all of their patients
with BE (6). Harrison and Hampton noted mortality rates of
67% in their patients with NS complications and 44% in patients
with BE without NS involvement (5).

Cerebral embolism is the most common NS complication of
BE, having been reported in 6 to 31% of patients with BE (4,
6, 10, 16). Cerebral emboli occur more commonly in patients
with miiral valve infections and in those with infections caused
by more virulent organisms. Microscopic emboli may occur in
approximately 11% of patients with BE and may be detected
only at postmortem examination of the brain (10). Some 3% of
all patients who develop cerebral emboli will have an infected
heart valve as a source. The morbidity and mortality rates are
significantly increased in patients with BE who suffer embolic
stroke. Thirty of the 37 patients (80%) of Pruitt et al. with an
embolic stroke from BE died (10). Pearce and Guze noted an
early mortality rate of 67% in their patients with BE with
embolic strokes (9), and Rabinovitch et al. noted that the early
mortality rate for their patients with embolic strokes was 30%,
with a late mortality rate of 50% (11).

Focal seizures occur more commonly in endocarditis patients
with acute embolic disease; generalized seizures are usually
secondary to metabolic derangements or are medication-related
(e.g.. lidocaine or penicillin toxicity) (4, 10). Cerebral mycotic
aneurysm has been noted to occur in 2 to 10% of patients with
BE (1, 10, 12, 14). Three to 6% of all intracerebral aneurysms
are mycotic in origin; they tend to occur more frequently
during the course of acute BE, rather than late in the course of
subacute BE (10).

SAH not of aneurysmal or embolic origin apparently occurs
rarely in patients with BE. No mention is made in the Coop-
erative Study of Intracranial Aneurysms and SAH of nonmy-
cotic aneurysm-related SAH in BE (7). Five of the 218 patients
reported by Pruitt et al. suffered a SAH during their bout with
BE, although only 2 (0.6%) presented with a SAH apparently
not related to an aneurysm or embolism (10).

In the past, 18 to 27% of patients with verified nontraumatic
SAH have had a normal cerebral arteriogram (7, 15). However,
with modern angiographic techniques (femoral catheter studies
allowing easier and better visualization of the cerebral arteries
and subtraction and magnification techniques), the incidence
of angiographically negative SAH is about 7% (15).

The etiology of our patient’s SAH is a bit obscure. A likely
explanation is that it was caused by a ruptured aneurysm that
spontaneously occluded after the hemorrhage. This may well
be a significant factor in those 7% of cases with negative
arteriograms. However, focal vasculitis has been reported as a
cause of SAH (3), and in BE a proliferative endarteritis can
occur, which can lead to vessel rupture and occlusion. Angio-
graphically negative SAH has been reported after streptococcal
glomerulonephritis in children, and a focal vasculitis (infectious
or immune complex-related) is a likely cause of the hemorrhage
in these patients (2).

The patient reported here had transient right-sided weakness
at the time of his SAH, and this recurred at the time of the
embolic stroke. Thromboembolism can occur from arterial
cerebral aneurysms (13) and it is possible that, if he did have
an aneurysm that clotted spontaneously, an embolus could
have arisen from it and caused symptoms in the same arterial
distribution. However, in our patient there was evidence of
systemic embolization at the time of his embolic stroke, so a
new embolus from the heart to the brain likely occurred. Two
of the patients reported by Pruitt et al. who had mycotic

SAH; INITIAL MANIFESTATION OF BACTERIAL ENDOCARDITIS 489

aneurysms experienced transient neurological symptoms con-
sistent with an embolism to the cerebral vessel in which the
aneurysm was located (10).

The importance of this case lies in the fact that BE is a rare
but treatable cause of SAH. In the 7% of cases of SAH with a
normal arteriogram, BE should be considered as a possible
cause of SAH, especially if a heart murmur is present. Prompt
recognition of this disorder will lead to the appropriate therapy
and, although the prognosis is worse in those patients with NS
complications of BE, favorable results will hopefully be ob-
tained.

Received for publication, June 3, 1980; accepted, July 13, 1980.
Reprint requests: Frederick M. Vincent, M.D., 211 Beaumont Place,
Traverse City, Michigan 49684.

REFERENCES

1, Cantu, R. C., LeMay, M., and Wilkinson, H. A. The importance
of repeated angiography in the treatment of mycotic-embolic in-
tracranial aneurysms. J. Neurosurg., 25: 189-193, 1966.

2. DeBeukelaer, M. M. and Young, G. F. Subarachnoid hemorrhage
complicating acute poststreptococcal glomerulonephritis. Arch.
Neurol., 35: 473-474, 1978.

3. Edwards, K. R. Hemorrhagic complications of cerebral arteritis.
Arch, Neurol., 34: 549-552, 1977.

4. Greenlee, J. E. and Mandell, G. L. Neurological manifestations of
infective endocarditis: A review. Stroke, 4: 958-963, 1973.

5. Harrison, M. J. G. and Hampton, J. R. Neurological presentation
of bacterial endocarditis. Br. Med. J., 2: 148-151, 1967.

6. Jones, H. R., Jr, Siekert, R. G., and Geraci, J. E. Neurologic
manifestations of bacterial endocarditis. Ann. Intern. Med., 71:
21-28, 1969.

7. Locksley, H. B. Report on the Cooperative Study of Intracranial
Aneurysms and Subarachnoid Hemorrhage, Section V, Part II:
Natural history of subarachnoid hemorrhage, intracranial aneu-
rysms and arteriovenous malformations: Based on 6368 cases in
the Cooperative Study. J. Neurosurg., 25: 321-368, 1966.

8. Osler, W. Gulstonian lectures on malignant endocarditis. Lancet,
1: 415-418, 459-464, 505-508, 1885.

9. Pearce, M. L. and Guze, L. B. Some factors affecting prognosis in
bacterial endocarditis. Ann. Intern. Med., 55: 270-282, 1961.

10. Pruitt, A. A., Rubin, R. H., Karchamer, A. W., and Duncan, G.
W. Neurologic complications of bacterial endocarditis. Medicine
(Baltimore), 57: 329-343, 1978.

11. Rabinovich, S., Evans, J., Smith, I. M., and January, L. E. A long-
term view of bacterial endocarditis: 337 cases 1924 to 1963. Ann.
Intern. Med., 63: 185-198, 1965.

12. Roach, M. R. and Drake, C. G. Ruptured cerebral aneurysms
caused by micro-organisms. N. Engl. J. Med., 273: 240-244, 1965.

13. Stewart, R. M., Samson, D., Diehl, J., Hinton, R., and Ditmore, Q.
M. Unruptured cerebral aneurysms presenting as recurrent tran-
sient neurologic deficits. Neurology (N. Y.), 30: 47-51, 1980

14. Toone, E. C., Jr. Cerebral manifestations of bacterial endocarditis.
Ann, Intern. Med., 14: 1551-1574, 1941.

15. West, H. H., Mani, R. L., Eisenberg, R. L., Tuerk, K., and Stucker,
T. B. Normal cerebral arteriography in patients with spontaneous
subarachnoid hemorrhage. Neurology (N. Y.), 27: 592-594, 1977.

16. Ziment, I. Nervous system complications in bacterial endocarditis.
Am, J. Med., 47: 593-607, 1969.

COMMENTS

This is a neat and scholarly review of an increasingly rare
cause of subarachnoid hemorrhage. To our knowledge. we
have not seen this mechanism, and it is interesting to speculate
as to the precise pathogenesis of hemorrhage falling upon an
embolic event and, in particular, a septic embolic event. The
possibility of an aneurysm formation with rupture and subse-
quent thrombosis is attractive, but equally likely is a possibility
of direct ischemic and infectious damage to the wall by the
embolus, which then breaks up and moves along, allowing high

490 VINCENT etal.

pressure arterial blood to leak from the damaged segment. This
paper is a valuable reminder for all to search for and identify
the cause of every subarachnoid hemorrhage.

S. J. Peerless, M.D.
London, Ontario

The authors have presented an interesting case of subarach-
noid hemorrhage in a patient with subacute bacterial endocar-
ditis. They have called our attention to yet another unusual
cause of bloody spinal fluid and one that should be kept in

Neurosurgery, Vol. 7, No. 5

mind. I would differ in managing the case in that I think a
repeat angiogram, when the patient again presented with neu-
rological deficit, should have been carried out. The possibility
of finding an aneurysm at that time would probably be greater.
Even though the patient has done well for the past year and a
half, if an aneurysm is still present, it should be dealt with on
an elective basis, rather than awaiting another subarachnoid
hemorrhage.

Eugene S. Flamm, M.D.
New York, New York