J Neurosurg 55:590-603, 1981

Neuropathological and computerized tomographic

findings in experimental brain abscess
RICHARD H. BRITT, M.D., PH.D., DIETER R. ENZMANN, M.D., AND
ANNE S. YEAGER, M.D.
Department of Surgery, Division of Neurosurgery, Department of Radiology, Division of Neuroradiology,
and Department of Pediatrics, Division of Pediatric Infectious Disease, Stanford University School of
Medicine, Stanford, California

Vt The neuropathological progression of brain abscess formation was studied experimentally at sequential
stages in dogs, and the findings correlated with the appearance on computerized tomographic (CT) brain
scans. The evolution of brain-abscess formation was divided into four stages based on histological criteria:
early cerebritis (Days 1 to 3); late cerebritis (Days 4 to 9); early capsule (Days 10 to 13); and late capsule
formation (Day 14 and later). The cerebritis stage was characterized by prominent perivascular cuffing by
inflammatory cells in the area adjacent to the developing necrotic center. However, the early elements of
capsule formation appeared with the presence of fibroblasts by Day 5. The CT scans showed ring-shaped
contrast enhancement by Day 3. Delayed scans at 30 minutes revealed diffusion of the contrast material
into the developing necrotic center, forming a solid lesion. In lesions that were well encapsulated (14 days
and older), five distinct histological zones were apparent: 1) a well formed necrotic center; 2) a peripheral
zone of inflammatory cells, macrophages, and fibroblasts; 3) the dense collagenous capsule; 4) a layer of
neovascularity associated with continuing cerebritis; and 5) reactive astrocytes, gliosis, and cerebral edema
external to the capsule. The CT appearance of well encapsulated abscesses showed a typical ring-shaped
contrast-enhancing lesion. On delayed scans, the "ring" did not fill in with contrast enhancement. The
diameter of the ring correlated best with the presence of cerebritis (perivascular infiltrates in the adventitial
sheaths of vessels surrounding the abscess). The discussion focuses on the relevance of this study to the
current management of patients with brain abscess.

KEY WORDS 9 brain abscess 9 cerebritis 9 computerized tomography 9

ring-like contrast e n h a n c e m e n t 9 capsule formation

F
AILURE to identify and accurately localize brain
abscesses is a major factor that contributes to
the high mortality and morbidity rates associ-
ated with these lesions. 4'12'2~ The introduction
of computerized tomographic (CT) brain scans has
been a significant advance in localizing suspected
intracranial infection. 12'14'27'29'39'43-47'51'59'61'62'68'71'75-77It
has been assumed that the cerebritis stage of abscess
formation correlates with the CT finding o f a low-
density lesion with little or no contrast enhance-
ment, 12'27'~''4~'76 and the appearance of ring-shaped
contrast enhancement represents an encapsulated ab-
scess. 12'27'29'43'44'61'62'76'77 Based on these assumptions
and the accuracy o f CT scanning, a new conservative
approach to brain abscess management has emerged
using antibiotics alone. 3,21'24,3x'33'~4'~G'G5Precise criteria
as to which lesions will respond to antibiotics alone
and which will require surgical aspiration or excision
for cure have not been established.
We have developed an animal model for studying
the CT findings associated with the evolution of brain
abscesses. The histological progression o f experimen-
tal brain abscess in this model is analyzed from the
stage of early cerebritis through the stage of well
encapsulated lesions. The dynamic aspects o f con-
trast enhancement have been reported in a previous
paper. 17
M a t e r i a l s and M e t h o d s
In 19 dogs, intracranial abscess formation was in-
duced by intracranial injection of a bacterial strain of
alpha streptococcus. For both the surgical procedures

590 J. Neurosurg. / Volume 55 / October, 1981

Evolution of experimental brain abscess
and the subsequent CT scanning procedures, the dogs
were tranquilized using acepromazine maleate, 10 mg
/15 kg injected intramuscularly, and then anesthetized
with intravenous pentobarbital (60 rag/2.5 kg). The
animal's head was fixed to a standard stereotaxic
apparatus. The shaved scalp was prepared with a
5-minute lavage with Betadine and was aseptically
draped. A scalp incision was made in the midline, and
the left temporalis muscle was reflected. A single burr
hole was placed using a Hudson brace with perforator
and burr. The bone margin was waxed, and the wound
flushed of any debris using normal saline. An in-
oculum of alpha streptococci was drawn into a 1-cc
syringe, and 0.2 cc (approximately 6 • 10 7 colony-
forming units in 0.5% agarose) was injected 5 mm
deep into the left parietal white matter with a No. 25
needle. The needle was withdrawn and the wound
flushed with saline. Immaculate hemostasis was
established to prevent an epidural hematoma, and
the wound was closed in layers. The outer skin was
sprayed with Aeroplast.
Two dogs served as controls. In the first, 0.1 cc of
saline was injected into the posterior parietal lobe of
the left hemisphere, and 2 months later 0.2 cc was
injected into the same location in the right hemi-
sphere. In the second animal, 0.1 and 0.2 cc of 1%
agarose without organisms was injected at similar
times and locations as the normal saline.
The alpha streptococcus was grown in Todd-Hewitt
broth in an overnight culture at a temperature of
36~ In each experiment, the exact number of col-
ony-forming units inoculated was determined by mak-
ing serial dilutions of an aliquot of the broth culture
and inoculating 0.1 cc of each 10-fold dilution on a
blood agar plate.
Computerized tomographic brain scans were ob-
tained in the coronal plane by using a custom-made
Plexiglas holder and polystyrene rings to position the
head. Scans were obtained both before and after
injection of 66% diatrizoate meglumine and 10% dia-
trizoate sodium (4 ml/kg) (Renografin 76). Serial
scans were made at the following time intervals: 0 to
5, 10 to 15, 20 to 25, 30 to 35, 45 to 50, and 60 to 65
minutes. Serial scans were obtained in each animal at
2- to 3-day intervals during the first 2 weeks, and then
every 5 to 7 days in lesions older than 2 weeks.
Brains were removed on the day of sacrifice, and
placed in 10% formalin solution for a minimum of 3
weeks. The brains were cut in 1-cm coronal sections,
photographed, and representative sections were taken
for processing and embedding in paraffin. Routine
hematoxylin and eosin (H & E) staining was per-
formed for general morphology. The Gram stain was
J. Neurosurg. / Volume 55 / October, 1981
used to evaluate the presence of bacteria. To evaluate
the progression of encapsulation, the reticulin stain
was used to evaluate reticulin or pre-collagen deposi-
tion 6~ and Masson's trichrome and hematoxylin van
Giesen (HVG) stains were used to determine the
presence and extent of mature collagen. Phospho-
tungstic acid hematoxylin (PTAH) and glial fibrillary
acid protein (GFAP) stains were used to evaluate the
presence of gliosis and reactive astrocytes. The gross
dimensions of the lesions were compared with the size
of the lesions on CT scans, assuming a 17% shrinkage
artifact of the tissue due to formalin fixation. Twenty-
one animals were sacrificed at the following time
intervals: 1, 3, 5, 6, 7, 8, 9, 10, 14, 19, 21, 22, and
38 days.
Results
Classification of Findings
Based on the detailed histological evaluation of the
material from the 19 experimental animals, four stages
of brain abscess formation were defmed: early cere-
britis (1 to 3 days); late cerebritis (4 to 9 days); early
capsule formation (10 to 13 days); and late capsule
formation (14 days and later). To some extent this is
an arbitrary division, since the evolution of brain
abscess formation is a continuing process. However,
there were salient features to be discussed that justified
this division into four stages. In each of the four
stages, the histological criteria for each of five zones
present in all well developed abscesses were evaluated.
The five zones were as follows: Zone 1, the necrotic
center; Zone 2, an inflammatory infiltrate mixed with
macrophages and fibroblasts surrounding the necrotic
center; Zone 3, collagenous capsule formation; Zone
4, cerebritis: perivascular infiltration of inflammatory
cells in the adventitial sheaths of blood vessels sur-
rounding the lesion; neovascularity; and Zone 5, re-
active astrocytes, gliosis, and cerebral edema. Table 1
summarizes the salient features of each stage de-
scribed below.
Early Cerebritis (1 to 3 Days)
Twenty-four hours after inoculation with alpha
streptococcus organisms, a marked inflammatory in-
filtrate comprised of polymorphonuclear cells, lym-
phocytes, and plasma cells was seen interspersed
among the residual agarose (Zone 1, Fig. 1A). A
Gram stain showed Gram-positive cocci scattered
throughout this developing necrotic center. At this
early stage, there had not been time for a significant
inflammatory response to be induced in the brain
peripheral to the developing necrotic center (Zone 2).
However, there was early infiltration of acute inflam-
591
 R. H. Britt, D. R. Enzmann and A. S. Yeager

TABLE 1
Histological and computerized tomography (CT)findings in sequential stages of brain-abscess development

Early Cerebritis Late Cerebritis Early Capsule Formation Late Capsule Formation
Findings (Days 1-3) (Days 4-9) (Days 10-13) (Day 14 and later)
Histological Features
Zone 1: necrotic inflammatory cells inter- necrotic center enlarged & necrotic center began to necrotic center continued
center spersed among residual reached its maximal size decrease in size to decrease in size with
agarose; bacteria present further encapsulation of
on Gram stain abscess

Zone 2: acute inflammatory cells inflammatory cells mixed increased numbers of fi- number of fibroblasts con-
inflammatory seen between the develop- with macrophages & fi- broblasts & macrophages tinued to increase in rela-
border ing necrotic center & broblasts appeared during this stage tion to inflammatory cells
brain & macrophages

Zone 3: collagen no evidence of reticulin fibroblasts appeared with mature collagen evolved a collagen capsule was
capsule formation until Day 3 rapid formation of reticu- from reticulin precursors; completed by end of 2nd
lin surrounding necrotic forming capsule was less week; during 3rd week,
center developed on ventricular capsule increased in den-
side of lesion sity of collagen deposition
& thickness

Zone 4: cerebritis & perivascular infiltration of
neovascularity polymorphonuclear leuko-
cytes, plasma cells, &
mononuclear cells (cere-
britis) developed rapidly
cerebritis reached its max-
imal extent during this
stage; new vessel forma-
tion rapidly increased
around developing abscess
cerebritis was less acute;
maximal degree of neo-
vascularity developed at
this stage
cerebritis was restricted to
area immediately outside
collagen capsule; perivas-
cular infiltrate contained
fewer inflammatory cells
& more fibroblasts; neo-
vascularity was less
marked

Zone 5: reactive marked cerebral edema cerebral edema continued cerebral edema started to surrounding cerebral
gliosis & cerebral surrounded lesion at this to be prominent in sur- regress; number of reac- edema regressed as cap-
edema stage rounding white matter; re- tive astrocytes increased sule developed; in 3rd
active astrocytes began to week, marked gliosis
appear late in this stage developed outside the
capsule; large numbers
of reactive astrocytes
surrounded abscess

Computerized Tomography
CT scan findings partial ring contrast en- ring enhancement with ring enhancement with ring enhancement without
hancement on Day l diffusion of contrast me- less diffusion of contrast diffusion of contrast mate-
evolved to well developed dium into the lucent cen- material into lucent center rial into lucent center was
ring enhancement by ter on delayed scans was was seen in incompletely salient feature of well en-
Day 3 salient CT finding of encapsulated abscesses capsulated abscess
cerebritis stage

m a t o r y cells in the adventitial sheaths o f vessels sur-
r o u n d i n g the i n o c u l u m (Fig. 1B). P o l y m o r p h o n u c l e a r
leukocytes, small lymphocytes, a n d a few large m o n o -
nuclear cells were present. T h e reticulin stain showed
the splitting o f the vessel walls caused by the perivas-
cular cuffmg (Zone 4, Fig. 1C). A G F A P stain showed
no evidence o f reactive astrocytes at this early stage
(Zone 5). Meningitis was p r o m i n e n t at D a y 1 and
extended d o w n the V i r c h o w - R o b i n spaces near the
site o f injection (Fig. 1D).
By D a y 3, there was a m a r k e d increase in the
perivascular infiltration o f i n f l a m m a t o r y cells in the
vessels surrounding the developing necrotic center
(Fig. IE). A few blood vessels began to show sprouting
o f reticulin, indicating the early presence o f fibroblasts
in or adjacent to the vessel walls (Fig. IF). There was
m a r k e d e d e m a o f the white matter surrounding the
lesions during this phase o f abscess formation.
In the early cerebritis stage, C T brain scans at
D a y 1 generally demonstrated low-density lesions
with some degree o f partial ring-like enhancement.
By D a y 3, there was a p r o m i n e n t ring enhancement,
the diameter o f which best correlated with the diam-
eter o f cerebritis (vessels with perivascular cuffing o f
i n f l a m m a t o r y cells) present surrounding the develop-
ing necrotic center.
Late Cerebritis (4 to 9 Days)
T h e most significant histological changes began to
take place on D a y 4 and extended t h r o u g h D a y 9. A

592 J. Neurosurg. / Volume 55 / October, 1981

Evolution of experimental brain abscess

FIG. 1. Photomicrographs showing early cerebritis (Days 1 to 3). A: On Day 1, an inflammatory

infdtrate consisting of polymorphonuclear leukocytes, lymphocytes, and plasma cells was seen interspersed
among the residual agarose (Zone 1). H & E, • 300. B: Early cerebritis on Day 1 consisting of a modest
perivascular infdtrate as seen in this section around two vessels (Zone 4). H & E, x 300. C: On Day 1,
there was no evidence of new reticulin formation in vessels adjacent to the developing necrotic center.
Reticulin stain, x 300. D: There was prominent meningitis at the site of needle entry for inoculating the
organisms into the brain during the early cerebritis stage. H & E, • 300, E: By Day 3, the degree of
perivascular infdtrate or cerebritis markedly increased compared with Day 1 (B). H & E, x 300, F: By
Day 3, there was early sprouting of reticulin, extending from the sides of a few vessels near the developing
necrotic center. Reticulin stain, x 300.
593
 R. H. Britt, D. R. Enzmann and A. S. Yeager

Fro. 2. Photomicrographs showing late cerebritis (Days 4 to 9). A: The necrotic center had developed
and consisted of inflammatory cells and debris (Zone 1). H & E, x 300. B: Along the margin of the
necrotic center, a mixture of inflammatory ceils, large foamy macrophages, and fibroblasts (spindle-shaped
celts) was seen (Zone 2). H & E, x 300. C: New vessels formed around the necrotic center and had
perivascular inflammatory infiltrates containing fibroblasts. H & E, x 300. D: In places along the
periphery of the developing necrotic center, the reticulin stain demonstrated a marked proliferation of
reticulin in areas comparable to photomicrographs B and C. x 300. E: Only a minimal amount of
maturing collagen was associated with the neovascularity. Masson's trichrome, x 300. F: During the
cerebritis stage, a few reactive astrocytes were seen in the white matter overlying the developing abscess.
GFAP, x 300.
594
Evolution of experimental brain abscess
well formed necrotic center developed and enlarged,
consisting of a mixture of tissue debris and inflam-
matory cells (Zone 2). The agarose was nearly de-
pleted (Fig. 2A). At the periphery of the developing
necrotic center, there was a zone of inflammatory
cells, macrophages (foam cells), and scattered fibro-
blasts (Zone 2, Fig. 2B). The fibroblasts were char-
acterized by oval nuclei with spindle-shaped cyto-
plasm. Cerebritis (Zone 4) was maximal during this
time. In addition, the microscopic features of the
perivascular cuffing changed with the appearance
of fibroblasts in addition to inflammatory cells. The
rapid increase in new vessel formation during this
stage at Zones 3 to 4 is shown on H & E stain (Fig.
2C), reticulin stain (Fig. 2D), and Masson's trichrome
stain (Fig. 2E). By Day 7, there was a significant
amount of reticulin bridging the neovascularity (Fig.
2D). Masson's trichrome and HVG stains showed
only modest amounts of maturing collagen in
the immediate vicinity of blood vessels. There was
still a considerable amount of white matter edema
beyond Zone 4. Beginning on Day 7, GFAP-stained
sections showed the presence of early reactive astro-
cytes in the area just outside the zone of cerebritis
(Zone 5, Fig. 2F).
Late cerebritis was characterized on CT scan by a
well formed ring 10 minutes after infusion of contrast
medium (Fig. 3). The contrast medium gradually
diffused into the lucent center and gave the appear-
ance of a homogeneous lesion. This was the salient
feature of cerebritis on CT in this experimental model.
Early Capsule Formation (10 to 13 Days)
At 10 to 13 days, the necrotic center (Zone 1)
decreased slightly in size. There was a marked increase
in the number of fibroblasts in the zone of inflam-
matory cells surrounding the necrotic center (Zone 2,
Fig. 4A). By this time, a well developed layer of
fibroblasts had been established. There was a signifi-
FtG, 3. Computerized tomography findings of late cerebritis. Sequential scans in a 5-day-old lesion
after contrast infusion showed a gradual diffusion of the contrast medium into the lucent center forming
a solid lesion at 0 to 5 minutes (left), 10 to 15 minutes (center left), 20 to 25 minutes (center right), and 45
to 50 minutes (righO.
J. Neurosurg. / Volume 55 / October, 1981
cant difference in the amount of reticulin deposition
on the cortical (Fig. 4C) and the ventricular (Fig. 4B)
sides of the lesions. Mature collagen was seen bridging
the spaces between vessels on Masson's trichrome and
HVG stains (Fig. 4E, Zone 3). Outside the developing
capsule, there was a region of continued cerebritis and
neovascularity (Fig. 4D, Zone 4). The number of
reactive astrocytes in Zone 5 had increased by this
time (Fig. 4F). The degree of edema in the white
matter gradually began to diminish.
The CT appearance during the early stages of cap-
sule formation was similar to that previously described
for the cerebritis stage. The necrotic center seen on
the CT scan was smaller, and this was verified histo-
logically. Serial scans continued to show some diffu-
sion of contrast material into the necrotic center.
Late Capsule Stage (14 Days and Later)
Histologically, the late capsule stage showed a con-
tinuation of the process already in progress. The
necrotic center continued to get smaller in the majority
of cases (Fig. 5A, Zone 1). In one lesion, however, a
multiloculated abscess with daughter abscesses devel-
oped. In Zone 2, adjacent to the necrotic center, the
percentage of inflammatory cells and macrophages
decreased while fibroblasts increased (Fig. 5B). The
capsule was well developed and was comprised of
fibroblasts, reticulin, and mature collagen (Zone 3).
The increased thickness of the capsule was seen on
both reticulin (Fig. 5D) and Masson's trichrome (Fig.
5E) stains. Mature collagen was present in abundance
beginning in the 3rd week of abscess formation. The
capsule increased in thickness due to the migration of
fibroblasts migrating from new vessels adjacent to the
capsule (Zone 4, Fig. 5F). Cerebritis, although pres-
ent, continued to show fewer infammatory cells (Zone
4, Fig. 5C). The number of reactive astrocytes contin-
ued to increase both in numbers and in the intensity
595
 R. H. Britt, D. R. Enzmann and A. S. Yeager

FIG. 4. Photomicrographs showing early capsule formation (Days 10 to 13). A: At the periphery of
the necrotic center, the number of fibroblasts increased significantly (top right) (Zones 1 to 2). H & E,
x 300. B and C: Reticulin increased greatly, but there was an asymmetry in the majority of lesions, with
more dense deposition on the cortical surface (C) compared with the ventricular surface (B) (Zone 3).
Reticulin stain, x 300. D: Cerebritis, although still present on the margin of the developing capsule,
consisted of fewer inflammatory cells and more fibroblasts (Zone 4). H & E, • 300 E: A dense deposition
of mature collagen extended between new vessels which have developed (Zones 3 to 4). Masson's trichrome,
x 300. F: There was an increase in the number of reactive astrocytes in the surrounding white matter
compared with the late cerebritis stage (Fig. 2F) (Zone 5). GFAP, X 300.

596
Evolution of experimental brain abscess

Flo. 5. Photomicrographs of late capsule formation (Day 14 and on). A: The necrotic center was
filled with increasing amounts of acellular debris and was diminished in size (Zone 1). Inflammatory cells
were still scattered throughout. H & E, • 300. B: At the periphery of the necrotic center, the number of
fibroblasts continued to increase (top left, Zones 2 to 3). Foamy macrophages and a few inflammatory ceils
were present close to the necrotic center (Zone 2). H & E, x 300. C: Cerebritis was present outside the
capsule associated with the neovascularity which had evolved (Zone 4). H & E, • 300. D: Reticulin
increased in both density and amount in the well developing capsule. Reticulin stain, x 300. E: A thick,
loosely woven collagen capsule matrix can be seen surrounding the necrotic center (Zone 3). Masson's
trichrome, • 300. F: The reticulin stain demonstrated the significant neovascularity present outside the
capsule. It also demonstrated the process by which the capsule thickens: the neovascularity allowed
increased numbers of fibroblasts to migrate into the area, which in turn generated reticulin and mature
collagen. Reticulin stain, • 190. 597
 R. H. Britt, D. R. Enzmann and A. S. Yeager

F1o. 6. Photomicrographs showing mature capsule for-

mation. A: The degree and extent of reticulin formation
continued to increase and, by the end of the 3rd week, a
solid mature capsule was present surrounding a small ne-
crotic center. Reticuhn stain, x 300. B: A thick mature
collagen capsule can be seen. Masson's trichrome, x 300.
C: Reactive astrocytes proliferated and were most nu-
merous in the overlying gray and subcortical white matter.
Note the astrocytic foot-processes lining the blood vessel.
GFAP, • 300.

of staining with GFAP. The white matter edema
lessened as encapsulation increased.
By the end of the 3rd week, the capsule was very
thick and was intensely reactive to both reticulin (Fig.
6A) and collagen (Fig. 6B) stains. The reaction of the
surrounding brain was easily detected by the end of
the 3rd week, and was well developed by 38 days.
Staining with PTAH demonstrated an increase in the
number of astrocytes associated with positive-reacting
glial filaments in the area immediately outside the
developing capsule. By 38 days, thick glial fibrils were
easily identified; they were maximal immediately ad-
jacent to the capsule, and were intertwined with the
neovascularity (Zone 4, extending into Zone 5). The
G F A P stain demonstrated the presence of glial fila-
ments adjacent to the capsule, but also revealed large
numbers of reactive astrocytes that were concentrated
in the cortical gray matter and the subcortical white
matter overlying the lesion (Fig. 6C).
The CT appearance of well encapsulated brain
abscesses continued to show ring-like contrast en-
hancement 10 minutes after infusion. The necrotic
center did not fill in by diffusion of contrast medium
as seen in the lesions in the cerebritis stage on delayed
scans (Fig. 7). Unfortunately, the majority of these
experimental lesions had necrotic centers that were
microscopic in size and therefore were below the
spatial resolution of the CT scanner.
Discussion
Experimental Models of Brain Abscess
Experimental brain abscess models have been de-
veloped in the primate, 73'74 dog, 17'25'38'4~ cat, ls'22 rab-
bit, 19'23'25'5~guinea pig, 25 and rat. 69'7~The majority of
authors have inoculated organisms in a suitable me-
dium directly into the brain, 17-ta'22,23,25,3s'50,67'69,70,T3but
two studies utilized septic emboli 4~ to induce

598 J. Neurosurg. / Volume 55 / October, 1981

Evolution of experimental brain abscess

FIG. 7. Sequential computerized tomography (CT) findings from late cerebritis to a well encapsulated
brain abscess. These serial CT scans were performed l0 to 15 minutes after contrast infusion. A: Day 5.
The ring enhancement was thick in the late cerebritis stage and diffused into the lucent center. B: In the
early capsule stage (Day 13), the ring enhancement was thinner and contrast material did not diffuse into
the lucent center. C: In the late capsule stage (Day 19), the lucent center was much smaller, but it did not
fill in with contrast material.

metastatic brain abscess formation. Histological eval-
uation of sequential stages of brain-abscess formation
has been documented in relatively few of these
studies. 17-19,25,69,70
In the present study, the dog was used to develop
this model for two reasons: 1) the brain size was large
enough to image with the resolution of current CT
scanners, and 2) the response of the dog's central
nervous system to infection is comparable to that of
man. Although classical evolution of brain-abscess
formation with encapsulation can occur in any of the
animal models cited, there are species differences. In
1913, Hom~n 25 used a mixture of streptococci, staph-
ylococci, and Bacillus perfringens (Clostridium welchii)
to obtain brain abscesses in dogs, rabbits, and guinea
pigs. In the rabbit, all components of capsule forma-
tion were more poorly formed and appeared later
than in the dog. It was Hom6n's conclusion that, in
the dog, abscess formation approximated most closely
the disease as it occurs in humans.
Species differences are of even more importance
when it comes to manipulating an experimental model
to study the effects of different treatment modalities
or the influence of steroids on encapsulation. 72 Long
and Meacham, a8 in a study using dogs, found that
dexamethasone given at the time of inoculation with
Staphylococcus aureus or Proteus mirabilis retarded,
but did not totally inhibit, capsule formation. Dexa-
methasone started a few weeks after inoculation had
no apparent effect on further encapsulation. 38 Quar-
tey, et al., 5~ studied the effect of dexamethasone and
antibiotics administered at the time of inoculation
with Streptococcus pyogenes or Staphylococcus aureus
on brain abscess formation in rabbits. In the rabbits
treated with dexamethasone and antibiotics, no evi-
dence of encapsulation was present at 10 days, and
only necrotic areas could be found. The control ani-
mals had well encapsulated lesions. The difference in
these two studies can be explained by immunological
studies which suggest that rats, mice, and rabbits are
extremely sensitive to corticosteroids with respect to
lymphoid depletion, circulating antibody production,
and cell-mediated immunity, whereas man, guinea
pig, 11and probably dogs are more resistant to steroids.
The formation of the collagen capsule in a devel-
oping abscess is the single most important response
that limits the spread of infection in the brain. The
biosynthesis of collagen has largely been worked out
biochemically,4a'4a and is known to originate from
fibroblasts derived from the connective tissue ele-
ments associated with blood vessels. 1~ With respect
to brain abscess, two particular features of collagen
formation are of interest. The first is that encapsula-
tion is more rapid on the cortical side of the lesion
compared with the ventricular side. The second is the
relatively limited encapsulation seen with metastatic
brain abscesses caused by septic emboli, as compared
with those caused by direct extension or trauma. Both
of these features of encapsulation can be explained by
the fact that oxygen is required for pro-alpha chains
to form the triple-helix strands of collagen. 1~ The
increased vascularity of normal cortical gray matter
allows for faster proliferation of new capillaries, which
in turn release more fibroblasts in a relatively oxygen-
rich environment compared with the more poorly
perfused deep white matter. The thinner capsules of
metastatic brain abscesses 73'74 can also be explained
by this theory. The vegetative embolus initially causes
an infarct. This area of nonviable tissue is hypoxic
and impedes new vessel formation and the collagen-
forming process of the fibroblasts.
The glial reaction of the brain to abscess formation

J. Neurosurg. / Volume 55 / October, 1981 599

 R. H. Britt, D. R. Enzmann and A. S. Yeager
has not previously been studied in detail. In this
model, two stains were used to study this response:
Mallory's phosphotungstic acid hematoxylin (PTAH)
and the immunocytochemical localization of the glial
fibrillary acidic protein (GFAP). The GFAP first
described by Eng, et al., TM has been used extensively
to identify astrocytes, both of neoplastic 13'1~and non-
neoplastic origins, is The GFAP is found in normal
fibrillary astrocytes, and questionably in normal pro-
toplasmic astrocytes, Bergmann glial cells, reactive
astrocytes, and reactive fibrillated ependymal ceUs. It
is not found in fibroblasts, microglial cells, or macro-
phages. ~3 This study represents its first use to study
intracranial infection and also its first use to demon-
strate reactivity in dog brain. This study confirms the
previously reported finding that there is not always a
direct relationship between GFAP-positivity and the
demonstration of glial fibers by traditional neurohis-
tological stains (PTAH and Holzer crystal violet
stains). '3 Whereas GFAP does not stain well the
intricate and closely textured network of glial cell
processes and fibers of white matter, it does stain
intensely the perikarya and cell processes of the re-
active astrocytes in cortical gray matter. ~3 It has been
hypothesized that the soluble form of GFA protein is
a subunit or precursor of glial fibers. Indirect evidence
suggests the insoluble form of GFAP is associated
with glial filaments, ~3 and perhaps this explains the
difference in the results obtained with the two stains.
In this study, the reactive astrocytes that stained with
GFAP were most marked in the cortical gray and
subcortical white matter, although gliat fibrils were
prominently stained in some lesions at 3 weeks in
areas adjacent to the capsule. In contrast, the PTAH
demonstrated gliosis and cells of astrocytic origin
maximally in the area immediately adjacent to the
outer edge of the capsule. By Day 7, GFAP was
weakly reactive in a few astrocytes, but it was not
until the end of the 2nd week that significant numbers
of reactive astrocytes were observed. By the end of the
3rd week, this process was even more evident using
both GFAP and PTAH stains.
Clinical Correlates of this Experimental Study
Brain abscesses have been associated with a high
mortality (29% to 64%) and morbidity 1'2'4'6"9'2~
30,32,34-37,42.57,64despite the advances of antibiotic ther-
apy? '26"a4'36'37Lack of clinical suspicion and poor lo-
calization were often cited as reasons for this high
mortality. 4'9'28'3~ With the advent of CT brain scan-
ning, earlier diagnosis and accurate localization of
intracranial suppuration have resulted in a decreased
mortality resulting from brain abscess. 27'43'53'62'68'75
Management of brain abscess cases in recent years
600
has focused on nonsurgical treatment in which anti-
biotics alone are used. a'21,24,3t'33,54,56,63'65The accuracy
of CT scanning has aided in this approach, allowing
the size of the lesions to be followed. Although class-
ical "ring" enhancement has been equated with
capsule formation, 12'27'29'43'44'61'62'76'77this study conclu-
sively demonstrated ring enhancement during the
cerebritis stage prior to any evidence of capsule
formation. Ring enhancement correlated closely with
the extent of cerebritis (perivascular infiltration of
inflammatory cells), and is likely due to altered
permeability of the normal blood-brain barrier? 6
Rapid proliferation of new vessels associated with
cerebritis was seen in this study in the area surround-
ing the developing abscess. Studies have shown that
proliferating capillaries leak protein-containing fluid
due to incomplete sealing of the endothelial lining. $8
In a study of regeneration of brain capillaries after
local freezing injury, it was demonstrated that the new
vessels had endothelial cells which lacked intercon-
necting tight junctions?
Some preliminary clinical evidence in the literature
supports the findings in this experimental study. A
recent case report of fatal cerebritis in man due to
Clostridiurn septicum showed a ring-like contrast-en-
hancing lesion with no evidence of capsule formation
histologically? 2 Whelan and Hilal 6s also suggested
that the results of this study are applicable to man.
In a series of 20 cases of brain abscess, 13 patients
required surgery because of failure to improve on
antibiotics alone or because of clinical deterioration.
However, a firm capsule was encountered at surgery
in only eight patients. In all eight cases, the symptoms
had been present for longer than 2 weeks. In the five
patients not having capsules at surgery, all had ring-
like contrast-enhancing lesions which the authors felt
represented cerebritis. Since delayed scans were not
performed in these studies, it is not known whether
the results would have been similar to those of our
experimental model. Some of the reports purporting
to have cured brain abscess with antibiotics only may
have been due to the fact that the lesions were in the
cerebritis stage, since the CT scans showed significant
diffusion of contrast into the necrotic center. TM
Clinically, the cerebritis and well encapsulated
stages may require different treatment. One clinical
study demonstrated that patients with well encapsu-
lated lesions, who had adequate killing levels of an-
tibiotic in the abscess fluid, continued to deteriorate
clinically and required aspiration. ~ It was possible to
grow out the offending organisms in each case. ~ Whe-
lan and Hilal 6s reported that despite administration of
intravenous antibiotics for a minimum of 2 days, eight
of 13 surgical specimens had positive cultures.
J. Neurosurg. / Volume 55 / October, 1981
Evolution of experimental brain abscess
Ideally, one would like precise criteria that separate
the cerebritis stage from the encapsulated stage. Clin-
ically, the first step is treatment with appropriate
antibiotics. If a presumptive organism is not known,
then one should seriously consider aspirating the le-
sion(s), under CT guidance if necessary, 71 to obtain
material for culture. This is particularly true in the
immunocompromised host, since the offending orga-
nism may be either bacterial, fungal, or protozoan in
nature. 7 On the basis o f this study, we suggest the use
of delayed CT scans at least 30 minutes after contrast
infusion so that the degree of contrast diffusion into
the lucent center can be evaluated. If the contrast
material diffuses completely into the necrotic center,
then one can presume that the lesion is in the cerebritis
stage and continue the antibiotics alone, providing the
patient is stable neurologically. If the contrast material
does not diffuse into the center, a well encapsulated
lesion is probably present. The approach will then
depend on the size o f the abscess and the clinical
condition of the patient. Abscess size may be the
determining factor whether antibiotics alone will be
effective or whether surgical aspiration or excision
will be required. Two studies have shown that lesions
greater than 4 cm in size usually require surgical
decompression) 4'~ whereas lesions of less than 2 cm
will respond to conservative therapy. 54 Unfortunately,
data are not available in these studies correlating the
size of the lesion and the degree of encapsulation. It
is likely that some lesions in the developing stages of
encapsulation will require surgical decompression be-
cause of clinical deterioration. Future work on this
experimental model will concentrate on developing
additional methods for separating the cerebritis and
encapsulated stages, and will hopefully lead to precise
criteria for determining when surgical intervention is
required or when antibiotic therapy alone will succeed
in curing brain abscesses.
Acknowledgments
We should like to thank Lucien Rubinstein, M.D., and
Mary Hermann, M.D., Department of Pathology (Neuro-
pathology) for reviewing selected cases; Elizabeth Miller-
man, Kaye Jenkins, and David Furgassa for histological
preparation; Robert McGowan for performing the GFAP
stains; L. F. Eng, M.D., for providing the GFAP antisera;
Phil Home for microphotography; Charles Sheldon and
Bernard Lyons for technical assistance; and Cynthia Holl-
man for manuscript preparation.
References
1. Ballantine HT Jr, White JC: Brain abscess. Influence of
the antibiotics on therapy and mortality. N Engl J Med
248:14-19, 1953
2. Belier AJ, Sahar A, Praiss I: Brain abscess. Review of
J. Neurosurg. / Volume 55 / October, 1981
89 cases over a period of 30 years. J Neurol Neurosurg
Psychiatry 36:757-768, 1973
3. Berg B, Franklin G, Cuneo R, et al: Nonsurgical cure
of brain abscess: early diagnosis and follow-up with
computerized tomography. Ann Neurol 3:474-478, 1978
4. Bhatia R, Tandon PN, Banerji AK: Brain abscess--an
analysis of 55 cases. Int Surg 58:565-568, 1973
5. Black P, Grayhill JR, Charache P: Penetration of brain
abscess by systemically administered antibiotics. J Neu-
rosurg 38:705-709, 1973
6. Brewer NS, MacCarty CS, Wellman WE: Brain abscess:
a review of recent experience. Ann Intern Med 82:
571-576, 1975
7. Britt RH, Enzmann DR, Remington JS: Intracranial
infection in cardiac transplant recipients. Ann Neurol 9:
107-119, 1981
8. Cancilla PA, Robert C, Frommes SP: Regeneration of
cerebral capillaries after local freezing injury. J Neu-
ropathol Exp Neurol 33:182-183, 1974 (Abstract)
9. Carey ME, Chou SN, French LA: Experience with
brain abscesses. J Neurosurg 36:1-9, 1972
10. Chapman JA: Morphological and chemical studies of
collagen formation. I. The fine structure of guinea pig
granulomata. J Biophys Biochem Cytol 9:.639-651, 1961
11. Claman HN: Corticosteroids and lymphoid cells. N
Engl J Med 287:388-397, 1972
12. Claveria LE, du Boulay GH, Moseley IF: Intracranial
infections: investigation by computerized axial tomog-
raphy. Neuroradiology 12:59-71, 1976
13. De Armond SJ, Eng LF, Rubinstein LJ: The application
of glial fibrillary acidic (GFA) protein immunohisto-
chemistry in neurooncology. A progress report. Pathol
Res Praet 168:374-394, 1980
14. Danziger A, Price H, Schechter MM: An analysis of
113 intracranial infections. Neuroradiology 19:31-34,
1980
15. Eng LF, Rubinstein LJ: Contribution of immunohisto-
chemistry to diagnostic problems of human cerebral
tumors. J Histochem Cytochem 26:513-522, 1978
16. Eng LF, Vanderhaeghen J J, Bignami A, et al: An acidic
protein isolated from fibrous astrocytes. Brain Res 28:
351-354, 1971
17. Enzmann DR, Britt RH, Yeager AS: Experimental
brain abscess evolution: computed tomographic and
neuropathologic correlation. Radiology 133:113-122,
1979
18. Essick CR: Pathology ofexperimental traumatic abscess
of the brain. Arch Neurol Psychiatry 1:673-703, 1919
19. Falconer MA, McFarlan AM, Russell DS: Experimen-
tal brain abscesses in the rabbit. Br J Surg 30:245-260,
1943
20. Garfield J: Management of supratentorial intracranial
abscess: a review of 200 cases. Br J Med 2:7-11, 1969
21. George B, Roux F, Pillon M, et al: Relevance of anti-
biotics in the treatment of brain abscesses. Report of a
case with eight simultaneous brain abscesses treated
and cured medically. Acta Neurochir 47:285-291, 1979
22. Groff RA: Experimental production of abscess of the
brain in cats. Arch Neuroi Psychiatry 31:199-204, 1934
23. Hassler O, Forsgren A: Experimental abscesses in brain
and subcutis. A microangiographic study in the rabbit.
Acta Pathol (Microbiol Scand) 62:59-67, 1964
24. Heinman HS, Braude AI, Osterhoim JL: Intracranial
suppurative disease. Early presumptive diagnosis and
601
 R. H. Britt, D. R. Enzmann and A. S. Yeager
successful treatment without surgery. JAMA 218:
1542-1547, 1971
25. Hom6n EA: Experimentelle und pathologische Beitr/ige
zur Kenntnis der Hirnabszesse. Arb Pathol Inst Hel-
singfors 1:1-80, 1913
26. Jooma OV, Pennybacker JB, Tutton GK: Brain abscess:
aspiration, drainage, or excision? J Neuroi Neurosurg
Psychiatry 14:308-313, 1951
27. Joubert M J, Stephanov S: Computerized tomography
and surgical treatment in intracranial suppuration. Re-
port of 30 consecutive unselected cases of brain abscess
and subdural empyema. J Neurosurg 47:73-78, 1977
28. Karandanis D, Shulman JA: Factors associated with
mortality in brain abscess. Arch Intern Med 135:
1145-1150, 1975
29. Kaufman DM, Leeds NE: Computed tomography (CT)
in the diagnosis of intracranial abscesses. Brain abscess,
subdural empyema, and epidural empyema. Neurology
27:1069-1073, 1977
30. Kerr FWL, King RB, Meagher JN: Brain abscess. A
study of forty-seven consecutive cases. JAMA 168:
868-872, 1958
31. Kirmani N, Tuazon CU, Ocuin JA, et al: Extensive
cerebral nocardiosis cured with antibiotic therapy alone.
Case report. J Neurosurg 49:924-928, 1978
32. Kiser JL, Kendig JH: Intracranial suppuration. A re-
view of 139 consecutive cases with electron-microscopic
observations on three. J Neurosurg 20:494-511, 1963
33. Kottas M, Smith LG: A possible new approach to the
management of brain abscesses. Infection 6:81-83, 1978
34. Krayenbiihl HA: Abscess of the brain. Clin Neurosurg
14:25-44, 1967
35. LeBeau J, Cressard P, Harispe L, et al: Surgical treat-
ment of brain abscess and subdural empyema. J Neu-
rosurg 38:198-203, 1973
36. Liske E, Weikers NJ: Changing aspects of brain ab-
scesses. Review of cases in Wisconsin 1940 through
1962. Neurology 14:294-300, 1964
37. Loeser E Jr, Scheinberg L: Brain abscesses. A review of
ninety-nine cases. Neurology 7:601-609, 1957
38. Long WD, Meacham WF: Experimental method for
producing brain abscess in dogs with evaluation of the
effect of dexamethasone and antibiotic therapy on the
pathogenesis of intracerebral abscesses. Surg Forum 19:.
437438, 1968.
39. Lott T, El Gammal T, Da Silva R, et al: Evaluation of
brain and epidural abscesses by computed tomography.
Radiology 122:371-376, 1977
40. Molinari GF: Septic cerebral embolism. Stroke 3:
117-122, 1972
41. Molinari GF, Smith L, Goldstein MN, et al: Brain
abscess from septic cerebral embolism: an experimental
model. Neurology 23:1205-1210, 1973
42. Morgan H, Wood MW, Murphey F: Experience with
88 consecutive cases of brain abscess. J Neurosurg 38:
698-704, 1973
43. Moussa AH, Dawson BH: Computed tomography and
the mortality rate in brain abscess. Surg Neurol 10:
301-304, 1978
44. New PFJ, Davis KR, Ballantine HT Jr: Computed
tomography in cerebral abscess. Radiology 121:641-646,
1976
45. Newton TH, Norman D, Alvord EC, et al: The CT scan
in infectious disease of the CNS, in Norman D, Korob-
602
kin M, Newton TH (eds): Computed Tomography. San
Francisco: University of California Printing, 1977, pp
317-337
46. Nielsen H, Gyldensted C: Computed tomography in
the diagnosis of cerebral abscess. Neuroradiology 12:
207-217, 1977
47. Paxton R, Ambrose J: The EMI scanner. A brief review
of the first 650 patients. Br J Radiol 47:530-565, 1974
48. Prockop D J, Berg RA, Kivirikko KI, et al: Intracellular
steps in the biosynthesis of collagen, in Ramachandran
GN, Reddi AH (eds): Biochemistry of Collagen. New
York: Plenum Press, 1976, pp 163-273
49. Prockop D J, Kivirikko KI, Tuderman L, et al: The
biosynthesis of collagen and its disorders. Part I. N Engl
J Med 301:13-23, 1979
50. Quartey GRC, Johnston JA, Rozdilsky B: Decadron in
the treatment of cerebral abscess. An experimental
study. J Neurosurg 45:301-310, 1976
51. Reinhard B: CT-Befunde bei Hirnabszessen. Fortschr
Geb Rontgenstr Nuklearmed 131:644-648, 1979
52. Roeltgen D, Shugar G, Towfighi J: Cerebritis due to
Clostridium septicum. Neurology 30:1314-1316, 1980
53. Rosenblum ML, Hoff JT, Norman D, et al: Decreased
mortality from brain abscesses since the advent of com-
puterized tomography. J Neurosurg 49:658-668, 1978
54. Rosenblum ML, Hoff JT, Norman D, et al: Nonoper-
ative treatment of brain abscesses in selected high-risk
patients. J Neurosurg 52:217-225, 1980
55. Ross R, Everett NB, Tyler R: Wound healing and
collagen formation. VI. The origin of the wound fibro-
blast studied in parabiosis. J Cell Bio144:645-654, 1970
56. Rotheram EB Jr, Kessler LA: Use of computerized
tomography in nonsurgical management of brain ab-
scess. Arch Neurol 36:25-26, 1979
57. Samson DS, Clark K: A current review of brain abscess.
Am J Med 54:201-210, 1973
58. Schoefl GI: Studies on inflammation. IlL Growing
capillaries: their structure and permeability. Virchows
Arch (Pathol Anat) 337:97-141, 1963
59. Shaw MDM, Russell JA: Value of computed tomogra-
phy in the diagnosis of intracranial abscess. J Neurol
Neurosurg Psychiatry 40:214-220, 1977
60. Snodgrass MJ: Ultrastructural distinction between re-
ticular and collagenous fibers with an ammoniacal sil-
ver stain. Anat Rec 187:191-206, 1977
61. Stephanov S: Experience with multiloculated brain ab-
scesses. J Neurosurg 49:199-203, 1978
62. Stevens EA, Norman D, Kramer RA, et al: Computed
tomographic brain scanning in intraparenchymal py-
ogenic abscesses. Am J Roentgenol 130:111-114, 1978
63. Tremonti LP, Dart LH: Focal encephalitis due to Pseu-
domonaspseudomallei. JAMA 215:112-113, 1971
64. Van Alphen HAM, Dreissen JJR: Brain abscess and
subdural empyema. Factors in influencing mortality
and results of various surgical techniques. J Neurol
Neurosurg Psychiatry 39:481-490, 1976
65. Viroslav J, Williams TW Jr: Nocardial infection of the
pulmonary and central nervous system: successful treat-
ment with medical therapy. South Med J 64:1382-1385,
1971
66. Waggener JD: The pathophysiology of bacterial men-
ingitis and cerebral abscesses: an anatomical interpre-
tation. Adv Neurol 6:1-17, 1974
67. Wallenfang T, Bohl J, Schurmann K: l~tude exp6rimen-
J. Neurosurg. / Volume 55 / October, 1981
Evolution of experimental brain abscess
tale de l'6volution des abc~s cerveau. Oed~me c6r6bral
et corr61ation neuropathologique. Neuroehirurgie 26:
145-152, 1980
68. Whelan MA, Hilal SK: Computed tomography as a
guide in the diagnosis and follow-up of brain abscesses.
Radiology 135:663-671, 1980
69. Winn HR, Mendes M, Moore P, et al: Production of
experimental brain abscess in the rat. J Neurosurg 51:
685-690, 1979
70. Winn HR, Rodeheaver G, Moore P, et al: A new model
for experimental brain abscess in rats and mice. Surg
Forum 29:500-502, 1978
71. Wise BL, Gleason CA: CT-directed stereotactic surgery
in the management of brain abscess. Ann Neurol 6:457,
1979
72. Wood JH: Brain abscess model. J Neurosurg 52:603,
1980 (Letter)
73. Wood JH, Doppman JL, Lightfoote WE II, et al: Role
of vascular proliferation on angiographic appearance
and encapsulation of experimental traumatic and met-
astatic brain abscesses. J Neurosurg 48:264-273, 1978
74. Wood JH, Lightfoote WE II, Ommaya AK: Cerebral
abscesses produced by bacterial implantation and septic
J. Neurosurg. / Volume 55 / October, 1981
embolisation in primates. J Neurol Neurosurg Psychia-
try 42:63-69, 1979
75. Wood MW, Cardenas R, Moore C: Influence of CAT
scanning on diagnosis and mortality rate of brain ab-
scess. J Term Med Assoc 73:325-327, 1980
76. Zimmerman RA, Bilaniuk LT, Shipkin PM, et al:
Evolution of cerebral abscess: correlation of clinical
features with computed tomography. A case report.
Neurology 27:14-19, 1977
77. Zimmerman RA, Patel S, Bilaniuk LT: Demonstration
of purulent bacterial intracranial infections by com-
puted tomography. Am J Roentgeno1127:155-165, 1976
Manuscript received January 28, 1981.
Accepted in fmal form May 1, 1981.
This work was supported by NIH/NINCDS Grant
1R01NSI6452, NIH Biomedical Research Support Grant
5S07RR05353-18, and the Stanford Neurosurgery Research
Fund.
Address reprint requests to: Richard H. Britt, M.D., Ph.D.,
Division of Neurosurgery R 155, Stanford University School
of Medicine, Stanford, California 94305.
603