Christopher Raymond

Acute

A. Meyer, K. Thompson,

Capt, MC, USAR #{149} Stuart MD #{149} Mario A. Gutierrez,

Traumatic

Experimental

Midbrain

and

Traumatic brain stem hemorrhage (TBH) after blunt head impact is an uncommon injury and has historically been associated with high mortality. Retrospective clinical review identified 64 patients with TBH admitted during a 5-year penod. Complete imaging and clinical records for 45 of these patients demonstrated that TBH could be categorized into three groups. The most frequent site of hemorrhage, in 31 (69%) of 45 patients (group 1), was the midline rostral anterior brain stem, posterior to the interpeduncular cistern, and this injury was associated with a 71% survival rate. This pattern was also associated with a predominantly anterior site of head and/or face impact. Eight (18%) patients (group 2) had miscellaneous foci of acute brain stem hemorrhage, with seven (88%) surviving. Six (13%) patients (group 3) had brain stem hemorrhage associated with transtentorial herniation and brain stem compression, with 100% mortality. Experimental findings in a canine model and clinical results indicate that the anterior rostral midbrain is a common site of TBH and appears to arise from sudden craniocaudal displacement of the brain at impact. Survival is unexpectedly high with this location of traumatic midbrain hemorrhage. Index terms: Brain, CT, 10.1211 #{149} Brain, hemorrhage, 14.434, 15.434 #{149} Brain, infarction, 14.435, 15.435 #{149} Brain, injuries, 10.43, 14.43, 15.43 #{149} Meninges, injuries, 13.43 Radiology

E. Mirvis, MD

Clinical

T

MD

#{149} Aizik

brain stem hemorrhage (TBH) after closed head trauma is an uncommon injury, which occurs in 0.75%-3.6% of all patients admitted to emergency centers with significant closed head injury (1-3). Clinically, TBH typically resuits in coma, decerebrate posturing, and autonomic nervous system dysfunction (4). Both computed tomographic (CT) and magnetic resonance (MR) imaging can be used to identify brain stem hemorrhage in posttrauma patients (1-10). Primary TBH result

tortion during hemorrhage

of the impact;

of direct

mechanical

brain stem secondary develops

MD

Hemorrhage: Observations

UTIC

is the

L Wolf,

dis-

that occurs brain stem

at some time and results

after the initial injury from diffuse cerebral edema, hypoxia, posttraumatic vasospasm, and transtentorial herniation (11-14). Secondary brain stem hemorrhages

that result from herniation

may occur

within 30 minutes after initial injury (15) and thus may be difficult to distinguish from primary mechanical injury. The biomechanical cause of TBH and its influence on ultimate neurologic outcome is controversial (16). The proposed mechanisms of TBH include direct impact of the brain stem against the rigid tentorium; shearing lesions due to differences in acceleration and deceleration of connected tissues, which produce diffuse axonal disruption; and pontomedullary junction tears due to hyperdistraction (12,17,18). Another proposed

1991; 179:813-818

I From the Department of Radiology, Walter Reed Army Medical Center, Washington, DC (CAM.); the Departments of Radiology (S.E.M.) and Neurosurgery (A.L.W.), Maryland Institute for Emergency Medical Services Systems, 22 5 Greene St. Baltimore, MD 21202; and the Division of Neurosurgery, University of Maryland Medical Center, Baltimore (R.K.T., M.A.G.). Received Dccember 3, 1990; revision requested January 25, 1991; revision received February 22; accepted February 25. From the 1990 RSNA scientific assembly. Address reprint requests to S.E.M. The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense. C RSNA, 1991

CT’

with

mechanism perforating terior rostral ilar

and

of TBH is disruption arteries that enter brain stem from

proximal

posterior

arteries; this disruption sudden craniocaudal the brain stem at the

of

the anthe bas-

cerebral

results from displacement of time of impact

(14,19,20).

To our knowledge, most of the literature to date suggests that hemorrhaged brain stem lesions almost uniformly portend a dismal prognosis. Previous studies have shown that patients with brain stem injury after trauma have a 31%-45% concurrent chance of having other supratentorial

lesions

(21,22).

In addition,

stud-

ies of primary traumatic brain stem injury have indicated mortality rates of 83%, with up to one-half of the surviving patients remaining in a persistent vegetative state (3). Our series revealed that TBH is less ominous than was previously believed. The rostral anterior midbrain is the most frequent site of TBH in our experience; a proposed mechanism of this lesion, developed on the basis of results in a canine model and

findings with

at patient

cranial

evaluation

CT, is suggested.

SUBJECTS

AND

Experimental

METHODS

Model

Two groups of mongrel to the experimental

dogs were subdesign described previously by Thompson and Sabcman (19). Four dogs weighing from 10 to 20 kg were anesthetized; then, an jected

intracranial pressure mino Laboratories,

San

in the

lobe.

right

monitor

Abbreviations: hemorrhage, time.

TE

(Ca-

was placed

In two dogs the between the carotid

frontal

vascular connections and basilar systems base of the skull. the carotid-basilar

bolt

Diego)

were

severed

at the

In the other two dogs vascular connections

TBH echo

=

traumatic time, TR

=

brain stem repetition

813

RANCHO

Functional

Level I II III IV V VI VII VIII

were

Scores

Clinical

Correlate

No response Generalized response Localized response Confused-agitated Confused-inappropriate Confused-appropriate Automatic-appropriate Purposefuland appropriate

left

intact,

thus

anchoring

the

vas-

cular

supply of the brain stem to the base of the skull (Fig 1). By using a twist drill, a balloon catheter was placed in the frontal epidural space flated at a rate of

for a total this

of 9 mL and

volume

for

maximum The balloon

the

and

balloon of intracranial

was

Medical

Systems,

Ti-weighted echo

of

means

MR were

registry

and

trauma

at

of stem

studies data

registry

to

base

of the

(University Center,

of

Baltimore).

period of time, 1,783 to the Shock-Trauma

ter with evidence ties at CT. During (3.6% of all those

juries) cranial

injection

reviewed. The from a radiology

Medical this admitted

sequences gap. Immedidogs were

of patients with from July 1985

Center

Maryland

macc,

500/17)

=

Evaluation

images obtained

Shock-Trauma

use of

Gross pathologic brain and brain

the

inthe

of 500

TR/TE

of a lethal

1990 were retrieved

During were

time

Imaging

CT and TBH that July were

NJ), with

(2,500/90) with a 20% imaging the

potassium chloride. specimens of the were obtained.

Clinical

3

of craniocerebral injuthis period 64 patients with craniocerebral in-

within these

ly obtained mission.

Of

imaging available

studies and medical records were for 45 (70%), and these patients

64

at least patients,

1 hour of adcomplete

constituted our study population. This group comprised 13 women 32 men, with an age range of 14-92 33.1 years). (a) motor

(automobile, 32, (b) falls from assault mined

814

in

pedestrian)

in six, (c) blunt in five.

#{149} Radiology

two, The

and years

Mechanisms of injury vehicle accidents

motorcycle, and

in

head trauma (d) undeter-

distribution

a.

b.

Figure

2. (a) Coronal T2-weighted (2,500/90) MR image of canine brain (gross specimen) reveals low signal intensity in the anterior rostral midbrain between the cerebral aqueduct and the ventral brain stem (arrows). Low signal intensity is believed to be due to the predominance of intracellular deoxyhemoglobin. (b) Corresponding pathologic specimen of rostral brain stem confirms distribution of blood.

patients Cen-

were found to have TBH at initial CT examination, which was usual-

(mean, included

Figure 1. Experimental design that produced craniocaudal displacement of brain in canine. Top illustration shows normal position of canine brain with no stretching of perforating vessels in the anterior rostral brain stem. Lower left illustration shows caudal displacement of the midbrain by inflation of balloon catheter in the frontal region with stretching of perforating vessels, which produced midline rostra! midbrain bleeding (cross section). Lower right illustration demonstrates lack of stretching of anterior brain stem perforators after caudal displacement of the midbrain due to severing of the basilar-carotid connections. Cross section shows lack of resulting hemorrhage.

of the canine (Siemens

Iselin,

17 msec,

and T2-weighted 4-mm intervals ately after MR by

within

all animals. Anesuntil MR imaging

(repetition

time

killed

in

and axial planes a 1.5-T Magnetom

on

pressure

8 hours after balloon images were acquired in

MR

coronal brain

was a

removed.

documented

of deflation was maintained

flation.

pressure reached

was

measurements

was performed

at bal-

Hg in all animals. was suddenly re-

Normalization minutes thesia

in-

maintained During

intracranial each dog

of 70 mm pressure

and

was

2 minutes.

loon inflation, monitored for

leased,

of each dog and was 1 mL every 15 seconds

of mecha-

nisms group

of closed is consistent

head

injuries with other

closed head injury reported ture (23). All CT scans were acquired

in in

this reports the

of litera-

with

use

pratentorial fied, including tricular hematoma,

of

findings at CT subarachnoid

hemorrhage, cerebral

shearing

injury,

were identior intraven-

subdural cortical

or epidural contusion,

diffuse

cerebral

edema

of basilar

cistern

compres-

a Hi-Q or DRH unit (Siemens) or a 9800 unit (GE Medical Systems, Milwaukee), with use of 8- or 10-mm-thick sections

with

evidence

sion,

brain

and CT

of Gentry shearing

et a! (2), diffuse axonal injuries were identified

multiple

punctate

section intervals scans were obtained

ters and ensuring

study tissue um; the

brain

windows brain

and

stem

10 mm. All use of cen-

that were optimal bone detail. Each

was reviewed or bone injury this evidence probable site

of 8 or with

for evidence to the face

provided of impact.

hemorrhage

stentorial

for

of softor calvari-

insight into The site(s) of

and

associated

su-

stem

distortion,

herniation.

at the of the

gray-white overlying

and/or

corpus

and/or By using

lesions junction, cortex,

callosum.

sions were identified presence of mixed ma that produced

tranthe

were

criteria and when

observed

with sparing deep gray nuclei,

Cerebral

on the hemorrhage local mass

contu-

basis

of the and edeeffect with-

June

1991

stem hemorrhage. Cranial CT scan of 38year-old man involved in motor vehicle trauma reveals midline hemorrhage in the anterior rostra! brain stem at the level of the interpeduncular cistern (arrow). Note the contusions in the bases of the frontal lobes (arrowheads).

.

a.

b.

Figure 4. CT scans of anterior old man who sustained severe large area of contusion in the

rostra! midbrain hemorrhage. (a) Cranial CT scan of 20-yearleft frontal impact in motor vehicle accident demonstrates left frontal lobe. Bone fragments from the left orbital roof are displaced into the frontal lobe. (b) More inferior CT image reveals faint attenuation of blood that occupies anterior rostral brain stem in midline behind the interpeduncular cistern (ar-

row)

and

pneumocephalus.

i’!k sion,

RESULTS

and

(Fig Canine

Experimental

Model

Brain onstrated

stem hemorrhage was demin the two animals with in-

tact carotid-basilar

vascular connections at MR imaging and pathologic examination (Fig 2), while in the pair of dogs with transected vascular connections to the carotid arteries, no hemorrhages

sults tions

were

support the of Thompson

Hemorrhages

&1J,-

Figure 5. CT scan of anterior rostral midbrain hemorrhage. Cranial CT scan of 28year-old woman reveals typical pattern of anterior midline rostra! brain stem hemorrhage. Contusions are noted in the frontal and anterior temporal areas, suggesting sagittal deceleration force.

out the bution

anatomic typical

A medical tients was Glasgow

configuration of a hemorrhagic

chart conducted Coma

at the

time

pathe

and behavioral an eight-level

function scale (Table)

distributed

Imaging Outcome

179

Number

#{149}

3

in

vascular

the

scans

divided

rostral

groups

on

into the

anteri-

midline

and

obtained

ani-

systems

in the

Evaluation

re-

observaSalcman (19).

in a primarily

were

Clinical

at admission

one

basis

of

of

in other

the brain six patients “Duret-type” in

as

of three the

site

sagittal

rhages

and

parasagittal

and

accompanied

edema,

locations

basilar

Group with an jected

below

extended

with-

inferior

pontine hemorby diffuse cerecistern compres-

the

caudally

typically posterior

observed perforated

in the

aqueduct.

and

as far as the anlesions

to extend substance

interpeduncular

often The

pro-

ventricle

The

pons.

teriorly,

hemorrhages that

third

terosuperior

lie

herniation

1 patients had ovoid configuration

were

from the and to fossa

pos-

as far as the cerebral group comprised eight

women

and 23 men who were 14-92 of age (mean, 29.3 years). Of these 31 patients, 14 had evidence of frontal calvarial impact-six, facial; four, occipital calvarial; four, temporal calvarial; and three, undetermined impact sites. Eighteen (58%) of these 31 patients had one or more associated cerebral cortical contusions, including 21 frontal, 10 anterior temporal, and one parietal in location.

years

Eleven of the

stem. Group 3 comprised (13%) who exhibited hemorrhages, defined

mesencephalon bral

(24).

Volume

were

hemorrhage

of admis-

sion (23), prinicpal findings at neurologic examination, and ultimate outcome, including RANCHO functional score at discharge. The RANCHO score quantitates the cognitive a patient on

intact

These

brain stem hemorrhage. Group 1 comprised 31 patients (69%) with antenor rostral midbrain hemorrhage (Figs 3-5). Group 2 comprised eight patients (18%) with varying sites of

or distriinfarct.

review of these to determine

Score

with

CT

previous and

observed

mals

or midbrain location.

c:;

seen.

transtentorial

6).

(35%)

associated

of the

31 patients

intraventricular

arachnoid Glasgow

or

hemorrhage. Coma

Score

had sub-

The overall at admission,

rounded to the nearest whole number, was 7, with a range of 3-13. Seven (22%) of these 31 patients had associated multiple punctate hemorrhagic foci in the gray-white junction, compatible with rotational shearing forces. The survival rate for group

1 was

tional

recovery

the from

71%,

RANCHO

with

score scale,

a mean

func-

of V to VI on with

a range

II to VIII. Radiology

#{149} 815

The eight patients in group 2, including seven men and one woman, had brain stem hemorrhages that varied in location. The hemorrhages

were

not located

tenor

rostra!

primarily

midbrain

in the an-

nor

were

they

associated tion. 15-79

with transtentorial herniaThe age range of this group was years (mean, 41.9 years). Glas-

gow

Coma

mission

Scores

ranged

in this from

group

8). The survival rate was 88%. At CT, evidence

was present cerebral present (50%)

current

at ad-

4 to 15 (mean, for this group

of shearing

in one (12%)

patient, and cortical contusion was in two others (25%). Four of the eight patients had con-

intraventricular

noid hemorrhage. ial cerebral injury

or subarachSome supratentorwas noted in seven

(88%). As expected, in general, the lower the Glasgow Coma Score at admission, the less favorable the functional recovery, with a mean recovery score of RANCHO V and a range of Il-VIII. The six group 3 patients had brain stem hemorrhage associated with diffuse cerebral edema and transtentorial herniation, as evidenced by brain stem compression and basilar cistern obliteration. These six patients ranged in age from 21 to 69 years (mean, 44.6 years). Three women and three men were in this group. Mean Glasgow Coma Score at admission was 4.5, with a range from 3 to 9. By definition, 100% of these patients had supratentorial abnormalities (ie, diffuse cerebral edema with transtentorial herniation). In two (33%) of these six patients, subarachnoid hemorrhage was observed. No patient in

this

group

survived.

One

been injured by means pact and one by means impact, but the impact terminate in the other on the basis of findings

the clinical Finally, tients were

fuse

patient

had

of frontal imof occipital site was indefour patients at CT and in

record. eight (18%) identified

axonal

and/or

of all 45 paas having dif-

shearing

injuries.

Only three (38%) of these patients survived; at discharge, their mean functional RANCHO score was III.

DISCUSSION Our results demonstrate that acute traumatic brain stem hemorrhage resulting from blunt head trauma most frequently produces lesions localized to the anterior rostral midbrain in proximity to the midline. Findings from our canine experimental model and clinical imaging data suggest that frontal or occipital impact with sudden craniocau-

816

#{149} Radiology

a.

b.

Figure 6. CT scans of secondary “Duret” hemorrhage. (a, b) Cranial CT images through the midbrain of a 47-year-old man injured in a motor vehicle collision reveal compression of the basal cisterns due to transtentorial herniation. Contusions exist in the anterior temporal lobes.

Hemorrhage

blood

is present

dal

(sagittal

brain

most

sions.

As

is noted

plane)

produces

proposed

(19),

posterior

midbrain

displacement

likely

Salcman

in the

and

in the circummesencephalic

by

the

these Thompson

basilar

cerebral

vessels

of the

with

le-

tears,

and

and

pons

(arrows

the

(ie,

proximal

injury)

are tethered

trary

to

comitant

caudal displacement tension on vessels the anterior rostra! sults

in

sudden

cranial

of the brain places that penetrate into brain stem and re-

hemorrhage

from

(Figs 1, 7). Disconnection id-basilar vascular system, cally

performed

or

in

our

that

imaging, the

laceration

would injuries

experi-

at

and

was of

con-

forces

in some

with

brain

of

that

not

performed

patients,

nonhemorrhagic with

most

types

our

consistent

shearing

are

without

It is possible

majority

reveal

observed

shearing

which

vast

tients

CT.

of

injuries

observed of

stem Con-

have stem

commonly

(82%) injury

of the carotas was surgicanine

very

brain moment

isolation.

we brain

evidence

brain in

in

occur

opinion,

occur

the

not

primary

frequently

MR

to

at the

this of

pontomedullary

occurring

does

brain

in

of damage

damage

a class

results

b). Subarachnoid

exception

primary

to the skull base by their connections to the carotid vessels, which are rigidly fixed in the skull base. Distortion of the that

in a and

cistern.

brain rotational

of these

stem

pa-

hemorrhage

in

to move with the displaced brain stem without causing tension on and nipture of these penetrating vessels. It has been demonstrated that prima17 brain stem injury may result from shearing injuries due to accelerationdeceleration in various axes of rotation

the

At CT, confined

an area of high attenuation to the region of the interpe-

(17).

duncular

cistern

ments,

allows

the

Rotational

basilar

forces

artery

produce

in

demonstrated

eral quadrant (4,12,16). In a series impact brain

in

of the

rostral

the

stem

hemorrhage fossa

sign)

in

the

state of

onal

injury

brain

jury

led

with

Gentry

primary et al

to conclude

stem that,

However,

imaging

has,

in

the

No

rived

high

from

of

ventral the

mid-

of secthat

some

subarachnoid

fossa

hemorrhagic

pathologic

may the

plane

suggest

interpeduncular

the

some

attenuation

of the

would

may

brain rostral

stem

midbrain.

evidence

exists

to

original

article

and,

of these a trauma

in-

authors

into

in the

in

fact, the

the

deeply

basis

We

the

In

hemorrhage

cases

contrary

fact,

sign,

reported

of

(25).

describing of

project

on the (25).

lesion

cases

article area

to

interpeduncular

many

the

represent

in-

(ie, in

attributed subarachnoid

fossa

to

the

been

isolated

subarachnoid

appear

blood

ax-

has of

original that

the

tion

of 40 patients with bluntinjury who were studied by Gentry et al, diffuse axonal injury was the most common primary injury

of diffuse

MR

terpeduncular

of

(48%) (2). The association

experience,

presence

brain

dorsolat-

brain

midline.

trauma.

shear

the

our

anterior

general, not revealed additional nonhemorrhagic injuries when performed after CT scanning for acute blunt head

stresses at the gray-white junction interface, in the deep gray nuclei, in the corpus callosum, and in the brain stem, with certain foci of lesions predominating, depending on the major axis of rotation (2). The lesions are discrete and often hemorrhagic. Brain stem injuries produced by rotational shear are most typically

rostra!

patients

were

population.

the in

deIn

June

our

1991

The vector of force required for the traumatic midbrain hemorrhage is clearly distinct from the mechanism required to produce shearing injury. Genarelli et al (17) described an animal model, the Penn II device, in which inertial loading in the coronal plane reliably produced grade 3 diffuse axonal injury.

Analysis

of

the

sites

of

impact

in

our group of 31 patients suggests that the more benign rostral midbrain lesion is produced by acceleration-deceleration in the sagittal plane, with only 22% of group 1 patients having cvidence of shearing injury at CT. Review of Generelli’s experience with the Penn II device is not inconsistent with this hypothesis (16). In animals subjected to a purely sagittal force, none had grade Figure

7.

Proposed

mechanism

of rostra!

anterior

midbrain

hemorrhage

in acute

trauma. Illustration of axial cross section at level of the interpeduncular cistern ma! length of perforating arteries and the “stretched” position that accompanies dal shift of the midbrain. Insert indicates bleeding from lacerated arteries that or rostra! midbrain.

patients blood

with adjacent

cistern,

we

parenchymal brain to the interpeduncular have

noted

stem

a tendency

for

cranial

hemorrhages associated with cerebral edema and transtentorial

diffuse

niation

had

(Duret

3 diffuse

depicts norcraniocausupply anteri-

her-

hemorrhages)

a

slow resolution of blood attenuation, persistent localization of the focal attenuation on serial CT scans, and, in the majority of cases, no other evidence

mortality rate of 67% in one series (1) and of 100% in our series. The survival rate in our group with primary traumatic rostra! midbrain hemorrhage was

of

71%

supra-

or

infratentorial

subarachnoid

hemorrhage or intraventricular hemorrhage. Several articles in the neurology and neurosurgery literature support the existence of acute TBH in the anterior rostral midbrain in a midline location (12,18,25,26). Ropper and Miller described

the

clinical,

brain-stem-audi-

tory-evoked-potentials CT findings, and, pathologic

abnormalities, available,

where

correlation

of

five patients (26). They cal signs that included fixed pupils; diminished impaired

horizontal

eye

Neuropathologically,

shown from duct

to

the

TBH

in

lesion

the rostral midbrain border of the aque-

interpeduncular

extension

fossa

to

the

jury

an

be

more

nonfatal initially

Why

do

observations brain stem

common

in

head injury appreciated patients

sustain

of tissue ondary

not

found

damage lesions

planes

to

that characterizes (18). In the former,

are

dissected

a well-defined

and

results

in

that,

a relatively

substantiated

tients

making

with

only

a favorable

10% recovery

(6).

As well, in our series, survival after shearing injuries was only 37.5%, with poor functional recovery in the surviving patients. Patients with brain stem

Volume

179

#{149} Number

3

of blood

various

locations

Conversely, hemorrhage

in

hemorrhagic

injury.

were in

in

of

Regardless

posterior

the

of

location

of

the the

the vector of acceleration-decelerremains in a sagittal plane distinct

from

is effective

and/or

in

shear this

the

in

most tenor

exclusively temporal

our

Further

that

is

rotational support is that

hemorrhagic

group

im-

force

mechanism

of the

sions

coronal

producing

injury.

proposed

locations

the

contu-

1 patients

were

a!-

in the frontal and antip regions. Lindenberg temporal and parietal im-

states

that

pacts

would

brain

infarction

small

by

Thus,

evidence

into

(5).

stem

brain stem an area of

due

distortion,

(13,27).

et a!

which

injected the

a secondary is essentially

pression,

sion

of pa-

and

or

on

by Cooper

studies

with

factor,

findings

result

in

acceleration-dc-

Thompson

been experimental

herniation,

prognostic

autopsy

(28,29).

small This

axonal

injury

the

panied

in

worst

of

TBH in the rostral ventral midbrain is not associated with the poor prognosis of the secondary Duret hemorrhage or primary brain stern hemorrhage accorn-

has

cavity.

reported

stem

fact,

celeration parallel to the tentorium that may produce direct contusion against the free edge of the tentorium but is ineffective in displacing the brain stem relative to the tentorial incisura (18). The last factor to take into consideration is the movement of the brain stem relative to the surrounding vascular skeleton. The mesencephalon delives its blood supply from the median and paramedian perforators of the proximal 7 mm of the posterior cerebral artery and more inferiorly from the superior cerebellar and basilar perforators (30,31). These small perforators penetrate the rostral floor of the interpeduncular fossa through the posterior perforated substance (Figs 1, 7).

cystic

fuse

single

sectis-

displaced

hematoma

was

brain

le-

extent

the vascular encephalon devastating

the

TBH

pact, ation

the

In

displacement

stem

anterior in

of

caudocranial

has

these

cause

and disruption supply to the caudal and rostra! pons due supratentorial injury

be

in

for

examination

are

cases

brain

neuropathologic sions

aliquots

to

1,367

clearly

cephalic junction. Lindenberg confirms this entity and clearly draws the distinction between the primary midline hemorrhage, which occurs in the upper half of the midbrain and appears just below the floor of the third ventricle, and the more inferiorly located secondary or “Duret” hemorrhages (18). In one series of 78 patients with difinjury,

of in-

patients

than (8).

who

hemosiderin-lined

with

mean

RANCHO

of

injury.

closed head injury (15), that impacts to the vertex or forehead most commonly produce lesions in the midbrain, basal ganglia, and upper pons. Brain stem hemorrhage may also result from predominantly occipital blows that result

the rostra! ventral midbrain tend to recover without devastating neurologic sequelae? Lindenberg speculates that at

resorption,

pontomesen-

impressive score

This supports the et al that primary may

with been

by

was

with recovery

V-VI. Gentry

sue

movement.

the

to occupy the ventral

inferior

acute

described clinicoma; dilated, limb tone; and

overall,

functional

axonal

13 animals subjected to a purely sagittal force, 1 1 had a good recovery and two had a moderate recovery. The neuropathologic series of Crompton, Lindenberg, and Freytag support the acceleration-deceleration model in a sagittal plane as the most frequent cause of TBH (15,18,28). Freytag states, in her autopsy review of

subsequent

to cornof mesto and

reperfu-

demonstration

of

diffuse

of

axonal

in

their

and

canine

Salcman

model

demonstrated

that

it is the

Radiology

teth#{149} 817

ering

of the

basilar

system

by the

suited

circle

of Willis and specifically the internal carotid arteries that prohibits the inferior movement of the vascular supply to accompany the rostrocaudal displacement of the brain stem (19). In their original experience, 32 (67%) of 48 dogs with intact internal carotid arteries experienced brain stem hemorrhage, while only nine (19%) of 46 animals with sectioned carotid arteries sustained

this

This

lesion

effect

(19).

has also

Weintraub,

been

Crompton,

(14,20,28).

and

As the brain

oriy it pushes

the

described Lindenberg

moves

midbrain

by

posteriand

pens

inferiorly through the incisura, buckling and foreshortening this region, while the medulla and cervical cord limit

the

inferior

migration

and

experi-

ence relatively little injury (27). Tethering of the basilar system occurs not only by means of the internal carotid artery but also by means of the posterior cerebral artery and the third cranial nerve (28). Microangiographic studies demonstrate that the median and paramedian basilar perforators enter the brain stem at varying angles and are predisposed to extreme stretching at certain brain stem levels after displacement (30). For Duret-type hemorrhages it has been well established that this mechanism results in vascular disruption of median perforators at the pontomesencephalic

mary”

junction.

traumatic

Acute

midbrain

“pri-

ity of patients

sult of failure orrhage

(9).

to detect In patients

stem hem-

with

strong

brain

findings at CT, MR imaging should be performed whenever feasible to assess better the integrity of the brain stem. Our review of 45 patients with primary brain stem hemorrhages that re-

818

#{149} Radiology

likely

9.

to be associated

with

orrhages;

this

ly better ature stem

these

for patients hemorrhage

Brain companied

stem hemorrhages

15.

16.

from

17.

ventral

or brain

The authors thank Thomas Stevenson and Mary Donnhauser of the University of Maryland Medical System IlDepartment

experimental

Freytag

E.

We also in our canine

75:402-413. Adams JH, Graham

5.

Mitchell

Graham

DI. Doyle

Gennarelli

PR, Maravilla K, Kirkpatrick J, et al. Traumatically induced brain stem hemorrhage and the computerized tomographic

clinical, observations.

pathological Neurosurgery

Cardobes

F, Lobato

traumatic

diffuse

RD.

axonal

and

experimental

1979; Rivas

JJ,

4:115-124. et al. Post-

brain injury: computed

of 78 patients studied with raphy. Acta Neurochirurgica

1986;

analysis tomog-

injuries

Lab Med 1963;

Pathol

DI, Gennarelli

TA, Thibault

DI, Thompson

TA.

LE, Adams

Con-

JH, Gra-

CJ, Marcincin

RP.

20.

Weintraub nerve and

osurg

Dif-

1966;

Significance from blunt

in the

of the tentorium forces. Clin Neur-

12:129-142. axial in the role 1988;

22:629-632. CM. Bruising the pathogenesis

of the third of midbrain

cranial hem-

21.

orrhage. Espersen

2.2.

mography in patients with head injuries. Neurochirurgica 1981; 56:201-217. French BN, Dublin AB. The value of com-

BrJ Surg 1960; 48:62-68. JO. Petersen OF. Computerized

puterized tomography 1000 consecutive head 1977; 7:171-183. 23.

Clifton

GL.

Traumatic

RN, ed. The clinical New York: Churchill Cope brain

to-

Acta

in the management injuries.

Surg

lesions. neurosciences.

of

Neurol

In: Rosenberg

Vol 2. 1983;

Livingstone.

DN. The rehabilitation of traumatic injury. In: Kottke FJ, Lehmann JF, eds.

D.

25.

Saunders, Yeakley

26.

Ropper

27.

28.

29.

impact

Cooper

scan: 6.

DE.

brain damage of immediate Brain 1977; 100:489-502.

in head

and Exp

1269-1272.

Tsai FY, TealJS, Quinn MF. et al. CT of brainstem injury. AJR 1980; 134:717-723. Gentry LR, Codersky JC, Thompson B. MR imaging of head trauma: review of the distribution and radiopathologic features of traumatic lesions. AJR 1988; 150:663-672. Zuccarello M, Fiore DL, Trincia C, DeCaro R, Pardatscher K, Andrioli CC. Traumatic primary brain stem haemorrhage: a clinical and experimental study. Ada Neurochirurgica 1983; 67:103-113.

type.

Arch

Thompson RK, Salcman M. Dynamic brain stem distortion as a mechanism production of brain stem hemorrhages: of the carotid arteries. Neurosurgery

24.

References

Diffuse

findings

forces.

19.

studies.

AdamsJH,

Pediatr

14:272-276.

Autopsy

blunt

handbook

rehabilitation.

4.

syn-

trauma.

Klintworgh GK. The pathogenesis of secondary brainstem hemorrhages as studied in an experimental model. Am J Pathol 1965; 47:525-536. Lindenberg R. Compression of brain arteries

Krusen’s

3.

brainstem

Lindenberg R. in head injuries

for contributing

Figures 1 and 7 of this manuscript. thank Joseph Kelly, who assisted

Neurol-

An unusual

18.

U

Services

1988;

JA.

fuse axonal injury and traumatic coma primate. Ann Neurol 1982; 12:564-574.

Acknowledgments:

lustrative

of pediatric

ham

hemorrhage

disrupted

stem perforators.

ex-

system trauma status report. Bethesda. Md: National Institute of Neurological Communicative Disorders and Stroke, 1985; 65-77.

that are ac-

subarachnoid

Ra-

temporary neuropathological considerations regarding brain damage in head injury. In: Becker DP, Polishock J, eds. Central nervous

by other supratentorial roshearing injury and hemorthat result from downward transtentorial herniation are less commonly seen clinically in our experience and have a much poorer prognosis. We speculate that many patients with socalled isolated interpeduncular cistern blood may actually have true parenchymay have that arises

hemorrhages.

DD. Winfield

from

brain

stem hemorrhages

RR. Autopsy tomography:

Galyon

tational rhages

brain

Jacobs L, Kinkel WR. Heffner correlation of computerized

as pathogenetic factor for tissue necrosis their areas of predilection. J Neuropathol Neurol 1955; 14:223-243.

in the liter-

with traumatic in general.

BH.

MR imaging.

12.

14.

significant-

that reported

JC. Thompson

of

1 1.

stem hem-

was

outcome

than

LR, Godersky

Neurosci

sagittally

brain

trauma.

Traumatic brain stem injury: diology 1989; 171:177-187.

drome

and deceleration and is frequently associ-

with

CT in the evaluation AJNR 1988; 9:91-100.

perience with 6,000 CT scans. Neurology 1976; 26:1111-1118. Sahuquillo J, Vilalta J. Lamarca J, Rubio E, Rodriquez-Pazos M, Salva JA. Diffuse axonal injury after severe head trauma. Acta Neurochirurgica 1989; 101:149-158. Friede U, Roessmann U. The pathogenesis

acceleration

in patients

Gentry

B, Dunn of inter-

MR and

of secondary midbrain ogy 1966; 16:1210-1216.

ated with contusions in the frontal and anterior temporal lobes. A fairly good ultimate neurologic outcome was seen

the re-

brain

clinical evidence of traumatic stem injury but with negative

the most

of the cranium

that one-third were

that

oriented

2.

CT scans

8.

mid-

13.

with closed head injury. Noneto our knowledge, the previous has emphasized the limitaof CT evaluation of brain stem

of false-negative

rostral

LR, Codersky JC. Thompson Prospective comparative study

mediate-field closed head

exhibited

in the ventral

Gentry VD.

pears

tients

indicating

7.

10.

1.

hemorrhage,

series

mdi-

distinct The major-

mechanism to account for TBH is stretching and tearing of ventral brain stem perforators (due to fixation of the basilar and posterior cerebral arteries by the carotid system) with concurrent displacement of the brain stem. The injury ap-

suggest

pa-

tions

impact

brain adjacent to the interpeduncular cistern. Correlation between results of animal experiments and findings at CT

hemorrhage

literature

in this

hemorrhage

stem

in multitrauma

cranial

hemorrhages have and prognoses.

mal ventral

hemor-

rhages occur more superiorly, perhaps reflecting differences in the relative movement of the posterior cerebral arteries in relation to the incisura when brain stem hemorrhage results from diffuse cerebral edema (secondary) or from acute traumatic displacement (14). Finally, an element of brain stem laceration is believed to occur in primary acute brain stem injury as the brain moves relative to the fixed arterial perforators (20). CT remains the modality of choice for the recognition of brain

theless,

from blunt

cates that mechanisms

of physical

medicine

and

4th ed. Philadelphia:

1990; 1217-1222. JW, Patchall LL, Lee

KF. Interpeduncular fossa sign: CT criterion of subarachnoid hemorrhage. Radiology 1986; 158:699700. AH,

Miller

DC.

Acute

traumatic

midbrain hemorrhage. Ann Neurol 1985; 18:80-86. Tominson BE. Brainstem lesions after head injury. J Clin Pathol 1970; 23 (suppl 4):154165. Crompton

MR.

Brainstem

lesions

due

to

closed head injury. Lancet 1971; 1:669-673. Thompson RK, Salcman M. Brainstem hemorrhage: historical perspective. Neurosurgery 1988; 22:623-628.

30.

Hastier

31.

stem. Neurology 1967; 17:368-375. Margolis MT. Newton TH, Hoyt

0.

Arterial

WF.

and

In: Newton

roentgenographic

TH, Potts D, eds. brain: Angiography.

pattern

of human

anatomy.

Radiology Vol

brain-

Gross

of the skull 2. St Louis:

and

Mosby.

1974.

81:27-35.

June

1991