See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/14517381

Mechanisms of Basilar Skull Fracture

Article in Journal of Neurotrauma · September 1995

DOI: 10.1089/neu.1995.12.669 · Source: PubMed

CITATIONS READS

28 830

4 authors, including:

Roger Nightingale
Duke University
75 PUBLICATIONS 3,059 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Pediatric Neck Biomechanics View project

All content following this page was uploaded by Roger Nightingale on 18 June 2015.

The user has requested enhancement of the downloaded file.

JOURNAL OF NEUROTRAUMA
Volume 12, Number 4, 1995
Mary Ann Lieben, Inc.

Mechanisms of Basilar Skull Fracture

JAMES H. McELHANEY, ROBERT H. HOPPER, JR., ROGER W. NIGHTINGALE, and

BARRY S. MYERS

ABSTRACT

Basilar skull fractures comprise a broad category of injuries that have been attributed to a variety
of causal mechanisms. The objective of this work is to develop an understanding of the biomechanical
mechanisms that result in basilar skull fractures, specifically focusing on mandibular impact and
neck loading as potential mechanisms. In the characterization of the injury mechanisms, three ex-
perimental studies have been performed. The first study evaluated the response of the base of the
skull to midsymphysis loading on the mental protuberance (chin) of the mandible. Five dynamic im-
pacts using a vertical drop track and one quasi-static test in a servohydraulic test frame have been
performed. In each test, clinically relevant mandibular fractures were produced but no basilar skull
fractures were observed. The second study assessed the fracture tolerance of the base of the skull
subject to direct loading on the temporomandibular joint in conjunction with tensile loading im-
posed locally around the foramen magnum to simulate the effect of the ligaments and musculature
of the neck. Among four specimens that sustained either complete or incomplete basilar skull ring
fractures remote from the sites of load application, the mean load at fracture was 4300 ± 350 N.
Energy to fracture was computed in three of those tests and averaged 13.0 ± 1.7 J. Injuries pro-
duced were consistent with clinical observations that have attributed basilar skull ring fractures to
mandibular impacts. In the third series of experimental tests, loading responses resulting from cra-
nial vault impacts were investigated using unembalmed human cadaver heads and ligamentous cer-
vical spines. Multiaxis load cells and accelerometers, coupled with high-speed digital video, were
used to quantify impact dynamics. The results of these experiments suggest that while there is a
greater probability of cervical spine injury, basilar skull ring fractures can result when the head is
constrained on the impact surface and the inertia of the torso drives the vertebral column onto the
occiput.
Key words: base of skull; fracture; head trauma; injury; mandible

INTRODUCTION injuries may occur, the principal cause among industri-

alized societies is motor vehicle accidents. One-third of
society, head
injuries rank among the lead-
I ing
N modern
causes of death and permanent disability. Across
the United States, more than two million head injuries
all motor vehicle injuries and 28% of all fractures re-
suiting from motor vehicle accidents involve the head and
face (National Safety Council, 1991). Among school age
occur each year and head injury is the leading cause of children, the greatest number of accidental deaths are due
death and disability among children and young adults to head injury and 71% of these injuries result from road
(National Institute of Neurological Disorders and Stroke, traffic accidents (Graham et al., 1989). Given these sta-
1989). While there are a variety of situations where head tistics, it becomes apparent that automotive occupants are

Department of Biomédical Engineering and Division of Orthopaedic Surgery, Duke University, Durham, North Carolina.

669
 McELHANEY ET AL.
particularly vulnerable to the potentially fatal and debil-
itating consequences of brain trauma resulting from skull
fractures. While there are many varieties of skull frac-
tures, certain types are most commonly associated with
automotive accidents. One particular injury whose fre-
quency has dramatically increased with the advent of the
automobile is the basilar skull ring fracture. In general,
ring fractures are considered to be the consequence of vi-
olent trauma to the head and face, usually in the form of
a blow to the chin or forehead. While most frequently as-
sociated with automotive accidents, these fractures have
also been observed among motorcyclists and victims of
falls. In a prospective study of 100 patients admitted be-
tween July 1986 and October 1988 in Malaysia who were
diagnosed with basilar skull fractures, half were motor-
cyclists and 22% were pedestrians hit by motor vehicles
while 20% were the victims of falls (Chee and Ali, 1991).
In a retrospective study of patients under 18 years of age
at the University of Southern California Medical Center,
62 patients with basilar skull fractures were admitted dur-
ing a 7.5-year period (Liu-Shindo and Hawkins, 1989).
The most common etiologies were pedestrians struck by
vehicles (42%), falls (27%), vehicle accidents (23%), and
impacts from an object (8%).
Defined as any fracture involving the floor of the an-
terior, middle, or posterior cranial fossae, basilar skull
fractures comprise a broad category of injuries. Owing
to the complex interactions of the local anatomical fea-
tures, the irregular geometry, and the anisotropic, non-
homogenous material properties, a variety of both local
and remote loading mechanisms are thought to induce
basilar skull fractures (Harvey and Jones, 1980). Local
loading mechanisms consist of forces applied directly to
the base of the skull as evidenced by superficial external
trauma. Remote loading mechanisms result when the pri-
mary site of force application is removed from the base
of the skull. Injuries stemming from these mechanisms
result from the applied load acting in conjunction with
the constraints and secondary loading imposed by liga-
ments and muscles in the basilar skull region. Inertial
loading contributions may also play a significant role in
these injury mechanisms.
In the clinical literature, ring fractures have been tra-
ditionally attributed to both tensile and compressive load-
ing mechanisms (LeCount and Hockzema, 1934; Vance,
1927). Tensile mechanisms, such as hyperextension, tend
to pull the head away from the spine. In these scenarios,
several authors have hypothesized that because the
strength of the atlantooccipital ligaments and muscula-
ture exceeds that of the occipital bone, failure of the oc-
cipital skull base results instead of separation between
the cervical spine and the base of the skull (Moritz, 1954;
Werne, 1957). Conversely, compression type ring frac-
670
tures about the foramen magnum are thought to result
from cranial vault impacts that drive the skull downward
onto the cervical spine. Several investigators have pos-
tulated that cranial vault deformation results in high
stresses away from the impact site. In this scenario, the
relatively elastic and smoothly curved cranial vault can
withstand sudden distortion without fracturing, transmit-
ting the impact forces to the rigid, irregularly shaped base
of the skull that sustains a fracture.
Voigt and Skold (1974) reviewed 77 cases involving
ring fractures and investigated causal mechanisms
through accident reconstruction. They concluded that a
variety of mechanisms can cause similar ring fractures
running through the floor of the middle fossa of the base
of the skull and the squama of the temporal and occipi-
tal bone. These causal mechanisms include the traditional
mechanisms of traction caused by retroflexion and an-
teroflexion in addition to torsion and wedging effects.
Torsional fractures are attributed to oblique impacts
against the temple or against the posterolateral portion of
the head. Wedging effects result from trauma against the
forehead, which is proposed to lead to deformation of the
skull base by increasing the transverse diameter of the
skull at the level of the anterior fossa. Among the popu-
lation that Voigt and Skold examined, the most common
type of ring fracture at the base of the skull was attrib-
uted to torsion. Among the ring fractures caused by hy-
perextension, the most common trauma was a blow to the
face, usually against the chin or forehead. Most remark-
ably, they found that blows to the chin resulting in ring
fractures rarely caused mandibular fractures and that ring
fractures were found rather than impression fractures in
the mandibular fossae. A characteristic transverse skin
laceration was seen at the tip of the chin evidencing chin
trauma. Widespread hemorrhages in the dorsal muscula-
ture of the neck were noted in some cases although they
were unable to assess whether the ligaments between the
skull base and the atlas and axis had been torn.
Huelke et al. (1988) have also examined case histories
involving basilar skull fractures in addition to a review
of the literature. They conclude that while basilar skull
fractures have been traditionally attributed to cranial
vault impacts, facial impacts can also cause basilar skull
fractures, especially impacts to the anterior mandible.
While the condition and direction of most skull fractures
correlate well with the site of impact, Harvey and Jones
(1980) have shown, based on case studies, that "typical"
basilar skull fractures (Rawling, 1912) extending along
both petrous bones may result from impacts to any point
of the head, including the chin. Vondra (1965) has also
implicated forces generated at the temporomandibular
joint as one mechanism resulting in "typical" basilar skull
fractures. Based on studies of craniofacial fracture pat-
 BASILAR SKULL FRACTURE
terns, Cooter and David (1990) observed that hospital-
ized motorcyclists who had worn open-face helmets, or
full-face helmets with flexible face-bars, had sustained
facial fracturing but minimal brain injury. In contrast,
motorcyclists killed from anterior craniofacial impacts
while wearing full-face helmets with rigid face-bars had
sustained fatal skull base fracturing in the absence of sig-
nificant facial trauma. These observations led to their pos-
tulate that impacts to the face-bar of a full-face helmet
may be transmitted via the chin strap through the
mandibular rami and mandibular condyles to the skull
base with subsequent fracturing of the middle cranial
fossa.
While well-documented laboratory experiments that
have induced basilar skull fractures are relatively rare,
Merignargues et al. (1975) have reported five clinical
cases of ring fractures and produced two experimen-
tally. In one case, a cadaver impacted a vehicle instru-
ment panel with a linear head velocity of 58 km/hr and
negligible angular velocity. A basilar skull fracture was
observed in addition to a fractured mandible. In the sec-
ond test, a lap-belted cadaver struck the top edge of the
16 Channel Data
Acquisition
System
Drop Track
FIG. 1. Experimental apparatus for evaluating head response
to mandibular impact.
671
steering wheel and hub with a linear head velocity of
50 km/hr. Impact accelerations ranged between 100 and
125 g over a duration of 8 ms and resulted in a basilar
ring fracture. While investigating tensile neck response
to axial loading, Sanees et al. (1981) observed basilar
skull ring fractures without any apparent cervical liga-
mentous damage at 1780 N quasi-statically in an iso-
lated cadaveric spine and at 3780 N in a dynamically
loaded cadaver consisting of an intact head, neck, and
torso.
In summary, the literature on basilar skull fractures is
rather sparse despite the frequency and significance of
these injuries. Yet speculation and generalizations re-
garding the mechanisms of these injuries abound.
Unfortunately, few quantitative data are available re-
garding basilar skull tolerances and specific mechanisms
that result in basilar skull ring fractures. Consequently,
the purpose of this study is to investigate remote loading
mechanisms that result in basilar skull ring fractures,
specifically focusing on mandibular impacts in addition
to tensile and compressive neck loading.
MATERIALS AND METHODS
To evaluate remote loading mechanisms resulting in
basilar skull fractures, a series of three experimental pro-
tocols were developed. The first series of tests sought to
evaluate Cooter's hypothesis that mandibular loading re-
sults in basilar skull fractures by directly transferring load
through the temporomandibular joint to the base of the
skull and to examine the effects of loading rate on
mandibular fracture tolerance. To isolate only mandibu-
lar loading, specimens consisted of whole, unembalmed
cadaver heads disarticulated from the neck at the base of
the occiput. A series of five dynamic impact tests were
performed using a vertical drop track (Fig. 1). To vary
the loading rate in the dynamic tests, the impact surface
was covered with various attenuators. To compare ca-
daveric response at a significantly slower loading rate, a
single quasi-static test was undertaken using a servohy-
draulic test frame (Fig. 2). Prior to testing, cadaver
pathologies and head radiographs were reviewed to en-
sure that specimens were not predisposed to mandibular
or basilar skull fractures.
In the dynamic tests, impact velocity was determined
using an MTS optical sensor and the entire event was
recorded at 1000 Hz with a Kodak Ektapro EM-2 high-
speed digital imaging system. The impact surfaces for
these tests consisted of a steel plate covered with various
attenuators (Table 1) having different stiffnesses that af-
forded variable loading rates. Impact velocities varied
from 6.08 to 7.15 m/s and specimens were oriented so
that the impacts occurred on the mental protuberance with
 McELHANEY ET AL.

Table 1. Mandibular Impact Data Summary

Head Impact Peak Time to Energy to

Impact mass velocity force peak force fracture
Test attenuation (kg) (m/s) (N) (ms) (J) Bony injuries
None 3.87 6.08 4670 0.688 11.4 Bilateral Fx of condyles
Fx of right coronoid process
Comminuted Fx of left side
Fx along alveolar process
2 0.635 cm butyl 4.52 7.15 6740 1.45 27.2 Bilateral Fx of condyles
rubber Fx around mental
protuberance
3 5.08 cm-thick 3.81 6.76 4460 10.7 65.9 Bilateral Fx of condyles
polystyrene
4 5.08 cm-thick 4.42 6.83 4960 11.3 66.9 Bilateral Fx of condyles
polystyrene Fx along the left interior
surface of the mandible
5 2.54 cm polystyrene 5.43 7.09 6100 13.2 119 Fx inferior to the
2.54 cm expanded head of the left
polystyrene mandibular condyle
0.635 cm butyl rubber
6 Styrene-based 4.82 0.0156 4690 414 12.4 Bilateral Fx of condyles
potting compound Fx of symphysis

the normal to the impact surface directed through the protuberance were assessed with a three-axis strain gage

specimen's center of gravity. A Kistler model 9067 three- load cell that measured two planar forces, denoted by x and
axis piezoelectric load cell was used to measure x-, y-, v, and the associated moment, M (Fig. 2). The planar forces
and z-axis impact forces. between the teeth, y and z, were quantified by a three-axis
In the quasi-static test, mandibular loads at the mental piezoelectric load cell. Like the dynamic tests, the contact
surface at the mental protuberance was oriented so that the
16 Channel Data normal to the contact surface was directed through the spec-
Acquisition imen's center of gravity. The constraint imposed by the sur-
System face supporting the posterior aspect of the head was de-
signed to simulate the inertia of the head in the dynamic
tests. Based on the test geometry and mandibular mea-
surements made after the test in conjunction with the force
histories supplied by the transducers, the reactions at the
temporomandibular joint, denoted by Fx and Fy, could be
computed using a static analysis.
In both environments, transducer signals were ampli-
fied and acquired using an IBM PC-based, RC-Electronics
digital data acquisition system that had an aggregate sam-
pling rate of 1 MHz. After acquisition, the data were up-
loaded to a Sun workstation for analysis. All dynamic im-
pact data were digitally filtered using a 50th-order finite
impulse response filter with a cutoff frequency of 1000
Hz as specified by SAE standard J21 lb. Palpation, radi-
ography, and dissection were used to evaluate injury. For
the purposes of comparing the static and dynamic tests,
3-Axis Piezoelectric the total energy associated with the fractures was com-
Transducer
puted. In this context, the energy to fracture represents the
work done to deform both the head and impact surface.
FIG. 2. Experimental apparatus for evaluating head response In the dynamic tests, the energy at failure was calculated
to quasi-static mandibular loading. based on the principles of impulse and momentum. In the

672
 BASILAR SKULL FRACTURE

quasi-static test, energy was determined by evaluating the 16 Channel Data

work done by the hydraulic ram.
In the second series of experimental tests, the same un- Acquisition
embalmed cadaver heads from the first series of tests were System
subject to direct temporomandibular loading in conjunction
with tensile neck loading. To simulate the inertia of the
head in a frontal impact and to provide additional constraint,
the specimens were also supported posteriorly on the oc-
cipital bone. Specimens were tested in a quasi-static envi-
ronment using an MTS servohydraulic test frame (Fig. 3).
Since the specimens did not include necks, the ligaments
and musculature of the neck were emulated by potting a
steel post in the base of the skull which extended through
the foramen magnum. The post was secured with a styrene-
based potting compound that conformed to the topography
of the base of the skull and was confined to the local re-
gion around the foramen and the anatomical insertions of
the atlanto-occipital ligaments and prevertebral muscula-
ture. A 2000-lb single axis GSE load cell mounted below Single Axis
the lower U-joint quantified the resultant force. Among the Strain Gage
six specimens, five were supported anteriorly beneath the Load Cell
temporomandibular joint and one specimen was supported
beneath the maxilla.
Basilar skull fractures were obtained by imposing a
constant velocity displacement at rates specified in Table
FIG. 3. Experimental apparatus for evaluating basilar skull
2. Incorporating the specimen and frame geometry de-
fracture tolerance by imposing a tensile load locally around the
rived from the high-speed video with the force history foramen magnum. Specimens were supported anteriorly be-
and employing an equilibrium analysis predicated on neath the temporomandibular joint or maxilla and posteriorly
sagittal plane symmetry, the reactions at both the ante- on the occiput.
rior and posterior supports could be determined. Since
the frame and fixturing significantly exceeded the stiff-
ness of the skull, the energy to failure could be deter-

Table 2. Basilar Skull Tolerance Data

Displacement Time to Peak neck Anterior Support Energy to

rate peak force force support reaction fracture
Test (cm/s) (sec) (N) location (N) (J) Bony injuries
1 0.819 0.603 4720 TMJa 4880 11.0 Complete basilar
skull ring Fx
2 6.48 0.0599 2740 TMJ 2590 4.33 Ring Fx originating
in the dorsum sellae
and radiating into the
foramen magnum
0.811 0.913 4360 TMJ 4640 14.0 Complete basilar
skull ring Fx
0.825 0.984 3390 Maxilla 4430 13.6 Facial Fx
1.90 2.56 4240 TMJ 3650 b Incomplete basilar
skull ring Fx
6.51 0.101 3860 TMJ 3480 14.0 Incomplete basilar
skull ring Fx

TMJ, temporomandibular joint.

''Unknown.

673
 McELHANEY ET AL.

Table 3. Cranial Vault Impact Data Summary"

Impact Surface Peak head Peak neck

velocity Impact angle resultant resultant Injury
Test (m/s) surface (deg) (N) (N) type

1 3.26 PTFE 0 8111 3010 Cervical spine

2 2.43 PTFE 0 7938 4189 None
3 3.23 PTFE 30 8790 1971 Cervical spine
4 3.28 PTFE -15 5615 2079 None
5 3.14 PTFE -15 11621 2891 None
6 3.20 PTFE 0 8566 2643 Cervical spine
7 3.26 PTFE 15 7498 2533 Base of skull;
cervical spine
8 3.13 2.5 EPS 30 1759 1762 Cervical spine
9 3.03 3.8 OPU 30 3623 2128 None
10 3.51 3.8 OPU 30 3857 2595 Cervical spine
11 3.07 3.8 OPU -15 2604 1289 Cervical spine

"PTFE, teflon; EPS, expanded polystyrene foam; OPU, open cell polyurethane foam.

mined by evaluating the work done by the ram. Ram dis- transducers included a PCB 302A02 accelerometer to
placement was recorded by an internal LVDT. Experi- measure torso acceleration, and two PCB 306A06 triax-
mental data were uploaded to a Sun workstation for ial accelerometers on the head to quantify sagittal plane
analysis and filtered at 250 Hz. Additional details re- head kinematics. An MTS optical sensor was used to
garding the experimental methods utilized in the first and record the impact velocity. All transducer output was
second studies may be found in Hopper et al. (1994). sampled at 62.5 kHz using a PC-based digital data ac-
In the third series of tests, an experimental apparatus quisition system. Each impact test was imaged using a
was designed to model head and cervical spine trauma Kodak Ektapro EM-2 digital camera at 1000 frames per
due to a cranial vault impact with a following torso (Fig. second. A flat steel plate measuring 15.25 cm in diame-
4). In these tests, intact cadaver heads and necks were ter and 4 cm in thickness was used as an impact surface.
drop tested with a simulated torso mass. Specimen prepa- It was mounted on a locking clevis, which allowed vari-
rations were rigidly mounted to a 6-axis GSE 6607-00 ation of the incident impact angle about the y-axis (nor-
load cell in an inverted posture, and the initial positions mal to the sagittal plane).
of the head and neck were established to maintain rest- Eleven tests were conducted on unembalmed human
ing lordosis. A Kistler 9067 3-axis piezoelectric load cell heads with intact spines. Muscular tissues were removed
was used to measure the head impact forces. Acceleration but all ligamentous structures were kept intact, with the

8000

6000

Z 4000

£ 2000

0.00 0.01 0.02 0.03 0.04

Time (sec)

FIG. 5. Mandibular impact resultant and component forces

for Test 2. Force components are labeled according to the axis
FIG. 4. Dynamic test configuration utilized to simulate head convention illustrated in Figure 1. The peak force reported in
and cervical spine trauma resulting from cranial vault impacts. Table 1 is based on the first local maximum.

674
 BASILAR SKULL FRACTURE

exception of the ligamentum nuchae. The bottom two tho- 8000

racic vertebrae of each specimen were cast in aluminum Resultant Force
Y-Axis Component
cups with reinforced polyester resin. The most rostral un- 6000
X-Axis Component
cast vertebra was free of the resin and had a full range
of motion. The C7-T1 intervertebral disk was oriented at Z 4000
+25° to the transverse plane to preserve the resting lor-
dosis of the cervical spine. A strip of scalp was removed
from the parietal bone, superior to the parietomastoid su-
ture and posterior to the coronal suture to install an ac-
celerometer array, which was attached to the skull on a
bed of dental acrylic using bone screws.
0.00 0.25 0.50 0.75 1.00 1.25 1.50
Tests were conducted using two contact surfaces. An
unconstrained head end condition was simulated by a rigid Displacement (cm)
steel surface covered with a 5-mm sheet of lubricated FIG. 7. Mandibular resultant and component force-displace-
teflon (n 7). The remaining tests (n = 4) used a more
=
ment data for Test 6. Force components are labeled according
constrained head end condition, which was either an ex- to the axis convention illustrated in Figure 2. The peak force
panded polystyrene foam (E 2100 kPa, ay 206 kPa,
= =
reported in Table 1 is based on the first local maximum.
p 0.0284 g/cm3) or a less stiff, open cell polyurefhane
=

foam (E 159 kPa, cry 7.0 kPa, p 0.0277 g/cm3).

= = =

The angle of the impact surface was varied between —15°
(posterior head impact) and +30° (anterior head impact).
A drop height of 0.53 m was selected for these tests.
Injuries to both the heads and cervical spines were docu-
mented by radiography and dissection. Additional details
regarding the experimental methods utilized in the third
study may be found in Nightingale et al. (1995).
RESULTS
In the first series of tests, the mandible consistently
fractured without any injury to the base of the skull.
Among five dynamically loaded specimens, the mean
peak force for mandibular fracture was 5390 ± 990 N
and the duration of the impacts, measured from initial
contact to the instant of peak force, varied from 0.688 to
13.2 ms (Table 1, Figs. 5 and 6). Impact velocity among
these specimens averaged 6.78 ± 0.43 m/s. Total energy
to fracture depended heavily on the type of impact sur-
face, ranging from 11.4 J with a rigid impact surface to
119 J with a compliant surface. Bilateral fractures of the
necks of both mandibular condyles occurred in four of
five specimens. The remaining specimen suffered a uni-
lateral fracture of the condylar neck. Fractures occurring
in the neighborhood of the coronoid process, symphysis,
and alveolar processes were also noted.
Force-displacement data from the quasi-static test re-
vealed a peak resultant force of 4690 N at the mental pro-
tuberance with a peak resultant of 4620 N transmitted
through the temporomandibular joint to the base of the
skull (Fig. 7). The first fracture occurred after 0.644 cm
of ram displacement. Bilateral fractures of the condylar

8000
Resultant Force 5000
6000 Z-Axis Component
X-Axis Component
4000
Y-Axis Component
Z 4000
M
U
z 3000
J 2000
<u
u
E
o 2000
Um

1000

0.00 0.01 0.02 0.03 0.04

Time (sec) 0.00 0.30 0.60 0.90 1.20 1.50
Displacement (cm)
FIG. 6. Mandibular impact resultant and component forces
for Test 5. Force components are labeled according to the axis FIG. 8. Basilar skull neck post force history for Test 3. The
convention illustrated in Figure 1. The peak force reported in peak force reported in Table 2 is based on the first local max-
Table 1 is based on the first local maximum. imum.

675
 McELHANEY ET AL.

5000 350 N. The sagittal plane component of the anterior sup-

0.00 0.30
0.60 0.90 1.20
Displacement (cm)
FIG. 9. Basilar skull neck post force history for Test 6. The
peak force reported in Table 2 is based on the first local max-
imum.
necks and in the region of the symphysis were noted
which were similar to the fractures observed in the dy-
namically loaded specimens. Although static and dy-
namic loading rates varied by several orders of magni-
tude, peak mandibular loads and bony injuries were
similar. Among all six specimens, the mean mandibular
fracture tolerance was found to be 5270 ± 930 N.
In the second series of experimental tests, basilar skull
fractures were observed in all specimens supported beneath
the temporomandibular joint (Figs. 8 and 9). The single
specimen supported beneath the maxilla sustained a facial
fracture without any accompanying basilar skull fracture.
Among the population of five basilar skull fractures, four
were consistent with clinically observed injuries and were
remote from the loading sites, indicating that stress con-
centrations introduced by the fixturing did not significantly
influence the failure mechanism (Table 2). Among those
four specimens, the mean neck load at fracture was 4300 ±
2000
"T<
-2000
® -4000 | the work done on both the specimen and the impact sur-
Head Force
Neck Force
-6000
-8000
0.00 0.02 0.04
Time (sec)
FIG. 10. Axial head and neck force histories for Test 7 in
which a cranial vault impact resulted in an impression type basi-
lar skull fracture.
port reaction at the temporomandibular joint averaged
4160 ± 700 N. Energy to fracture was available in three
of those tests and averaged 13.0 ± 1.7 J.
In the third experimental series, cervical spine injury
occurred in 7 of 11 drops at a mean axial compressive
neck force of 1727 ± 387 N. A basilar skull fracture was
noted only in Test 7 (Table 3). In that test, the head struck
a teflon-covered impact surface but, because the orienta-
tion of the impact surface matched the contour of the top
of the skull, the head did not translate. The axis of the
1.50
neck remained aligned with the direction of the velocity
of the simulated torso mass, resulting in an impression
fracture as the vertebral column impacted the base of the
skull. The injury was characterized by bilateral fractures
extending from the posterior edge of the foramen mag-
num around both occipital condyles and along the infe-
rior petrosal sinus terminating in the region of the dor-
sum sellae. Data acquired during the impact included
head forces and neck forces (Fig. 10) in addition to neck
moments and head accelerations.
DISCUSSION
Although the specimen population in the mandibular im-
pact study is small, the data suggest several observations.
The absence of basilar skull injuries indicates that the
mechanism involved in producing clinically observed
ring fractures involves more than just the transmission of
forces through the temporomandibular joint to the base
of the skull. In addition to providing insight into basilar
skull fractures, the experiments also afford an assessment
of mandibular fracture tolerance. Since the variance in
peak load to mandibular fracture among the tests is not
large, the fracture tolerance does not appear to be a func-
tion of impact surface or pulse duration (p > 0.50), at
least over the range of loading rates considered in this
study. The apparent insensitivity of fracture tolerance to
loading rate has also been observed by Nahum et al.
(1968). Although peak load is insensitive to impact sur-
face, the total energy associated with fracture depends on
the impact surface. Since the energy at fracture represents
face, it follows that as the compliance of the impact sur-
face increases, the energy to fracture should also increase.
This trend is observed in the experimental data.
While the first experimental series is quite similar to
0.06
mandibular impact studies undertaken by Schneider and
Nahum (1972), the average tolerance of 5270 N is sig-
nificantly higher than the average tolerance of 2840 N
based on their work. This variation in mean fracture tol-
erance may owe in part to impactor size (Hodgson, 1967).

676
 BASILAR SKULL FRACTURE
The available contact area was 127 cm2 in the present
study while it was 6.45 cm2 in their work.
It is important to note that a mandibular fracture tol-
erance in excess of 5000 N is within the reported load to
failure tolerances of other head and neck structures
(McElhaney and Myers, 1993). Consequently, both head
and neck injuries remote from the mandible may occur
as a consequence of mandibular loading. Incorporating
the failure tolerances of other head and neck structures
into computational models may provide insight into in-
juries resulting from mandibular loading.
Among the five specimens tested in the basilar skull tol-
erance study with anterior fixation beneath the temporo-
mandibular joint, four basilar skull fractures were noted
that exhibited the same features as clinically observed frac-
tures. The single basilar skull tolerance test undertaken
with anterior fixation beneath the maxilla, resulting in a
facial fracture, suggests that the tolerance of the facial
bones is less than that of the base of the skull. The absence
of a basilar skull fracture also supports the contention that
the facial bones may serve to absorb impact energy and
reduce the probability of basilar skull fractures. It also in-
dicates that the orientation of the impact force may be im-
portant to the injury mode. If the mandibular impact is pre-
dominantly inferior-superior in its orientation, facial
fractures may result instead of basilar skull fractures.
In an attempt to correlate the results of the first two
experimental series, the resultant mandibular force from
the first series was compared with the computed tem-
poromandibular resultant force in the second series of
tests. The goal of this analysis was to assess whether the
mandibular fracture tolerance was sufficient to induce
basilar skull fractures when the tensile contribution from
the neck was included in the injury mechanism. Although
the magnitude of the computed anterior support reaction
was quite sensitive to test frame geometry, values for the
sagittal plane component of the temporomandibular joint
reactions slightly exceeded mandibular tolerances in
Tests 1 and 3 but were substantially less than the mandi-
bular tolerances in Tests 5 and 6. On average, the mandi-
bular tolerance exceeded the peak resultant force at the
temporomandibular joint required to induce basilar skull
ring fractures, indicating that mandibular loading is a vi-
able mechanism for basilar skull fractures. Yet individ-
ual anatomical variations, in conjunction with other im-
pact conditions, suggest that some people may be
predisposed to mandibular fractures while others are at
greater risk for basilar skull fractures.
The third experimental series illustrates that cervical
spine injuries are more probable than basilar skull frac-
tures in the event of low velocity cranial vault impacts.
Several unique circumstances contributed to the basilar
skull fracture observed in Test 7. First, the orientation of
677
the impact surface matched the contour of the cranial
vault, effectively constraining the head during the initial
phase of the impact even though the rigid impact surface
was covered with lubricated teflon. Second, the cervical
spine remained aligned with the skull instead of buck-
ling, as in the other tests, resulting in a link that trans-
ferred the inertial torso loads directly to the base of the
skull.
Comparing the temporal nature of the head and neck
force histories, the neck forces clearly lag behind the head
forces. Consequently, it can be concluded that the initial
spike in the head force data serves to decelerate the head
mass. Since the head and neck are not rigidly coupled,
this spike is not apparent in the neck force data and neck
response lags behind the head response. Although the
force at the base of the skull was not determined exper-
imentally, it must lie within the envelope defined by the
head and neck load curves illustrated in Figure 10. In
fact, the force exerted by the vertebral column at the base
of the skull should closely follow the measured neck force
plus some inertial component associated with the accel-
eration of the neck mass.
Although the incidence of basilar skull fracture appears
to be far less than the occurrence of cervical spine in-
juries for this series of drops from 0.53 m, it should be
noted that the neck musculature would provide additional
neck stabilization in vivo, decreasing the propensity for
buckling and increasing the probability that inertial torso
loading would be transferred to the base of the skull.
It is also interesting to note that the ring fractures as-
sociated with tensile and compressive neck loading dif-
fer anatomically. The compressive ring fracture was con-
fined to the dorsum sellae and the region around the
occipital facets while the tensile ring fractures tended to
propagate at a higher level in the occiput.
CONCLUSIONS
In the experimental series consisting of isolated
mandibular loading, basilar skull fractures were not ob-
served in any of the specimens, indicating that chin im-
pacts alone, without loading contributions from the neck,
are not sufficient to induce basilar skull fractures. The
basilar skull tolerance tests revealed that temporo-
mandibular loading in conjunction with neck tension is
sufficient to induce basilar skull fractures. Complete or
incomplete basilar skull fractures, consistent with clini-
cally observed fractures, were noted in four of five spec-
imens supported beneath the temporomandibular joint.
These fractures occurred at a mean neck load of 4300 ±
350 N. The energy associated with the injury was found
to be 13.0 ± 1.7 J. Comparing mandibular fracture tol-
 McELHANEY ET AL.
erance with tolerances of other head and neck
reported
structures supports the hypothesis that mandibular load-
ing is a mechanism for both head and neck injury. It re-
mains, however, reproduce this mechanism in cadav-
to
ers with intact heads and necks and to confirm these
conclusions with greater statistical significance by addi-
tional experimental testing. In the third experimental se-
ries, cranial vault impacts to inverted cadaveric heads and
cervical spines revealed that if the head remains con-
strained on the impact surface and the spine does not
buckle, inertial torso loading communicated to the base
of the skull may result in basilar skull fractures. Cervical
spine injuries, however, appear to be a more probable
outcome than basilar skull fractures.
ACKNOWLEDGMENTS
This work was supported by Grant R49/CCR402396-
09 from the Centers for Disease Control, and by the
Virginia Flowers Baker Chair at Duke University.
REFERENCES
CHEE, C.P., and ALI, A. (1991). Basal skull fractures: A
prospective study of 100 consecutive admissions. Aust. N.
Z. J. Surg. 61, 597-602.
COOTER, R.D., and DAVID, D.J. (1990). Motorcyclist cran-
iofacial injury patterns. Int. Motorcycle Safety Conf. Proc. I,
3.1-3.13.
GRAHAM, D.I., FORD, I., HUME ADAMS, J., DOYLE, D.,
LAWRENCE, A.E., McLELLAN, D.R., and NG, H.K. (1989).
injury in children. J. Clin. Pathol. 42, 18-22.
Fatal head
HARVEY, F.H., and JONES, A.M. (1980). Typical basilar skull
fracture of both petrous bones: An unreliable indicator of
head impact site. J. Foren. Sei. 25, 280-286.
HODGSON, V.R. (1967). Tolerance of the facial bones to im-
pact. Am. J. Anat. 120, 113-122.
HOPPER, R.H., McELHANEY, J.H., and MYERS, B.S. (1994).
Mandibular impact as a mechanism for basilar skull fracture.
Proc. 38th Stapp Car Crash Conf. SAE Paper #942213.
HUELKE, D.F., SMOCK, W.S., FULLER, P.M., and
NICHOLS, G.R. (1988). Basilar skull fractures produced by
facial impacts—case histories and a review of the literature.
Proc. 32nd Stapp Car Crash Conf. SAE Paper #881711.
LECOUNT, E.R., and HOCKZEMA, J. (1934). Symmetrical trau-
matic fractures of the cranium; symmetrical fragmentation—
Comments on their mechanism. Arch. Surg. 29, 171-226.
LIU-SHINDO, M., and HAWKINS, D.B. (1989). Basilar skull
fractures in children. Int. J. Pediatr. Otorhin. 17, 109-117.
678
View publication stats
McELHANEY, J.H., and MYERS, B.S. (1993). Biomechanical
aspects of cervical trauma, in: Accidental Injury. Springer-
Verlag: New York, pp. 311-361.
MERIGNARGUES, G, GOT, C, TARRIERE, C, et al. (1975).
Les Fractures Circulaires de la Base du Crane au Cours des
Accidents de la Route. Nouvelle Presse Med. 4(31), 2245-
2248.
MORITZ, A.R. (1954). The Pathology of Trauma. Henry
Kimpton: London.
NAHUM, A.M., GATTS, J.D., GADD, C.W., and DAN-
FORTH, J. (1968). Impact tolerance of the skull and face.
Proc. 12th Stapp Car Crash Conf. SAE Paper #680785.
NATIONAL INSTITUTE OF NEUROLOGICAL DISOR-
DERS AND STROKE, NATIONAL INSTITUTES OF
HEALTH, PUBLIC HEALTH SERVICE, and DEPART-
MENT OF HEALTH AND HUMAN SERVICES. (1989).
Interagency Head Injury Task Force Report. U.S.
Government, Department of Health and Human Services
Publication: Washington, DC.
NATIONAL SAFETY COUNCIL. (1991). Accident Facts.
National Safety Council: Chicago, p. 75.
NIGHTINGALE, R.W., McELHANEY, J.H., RICHARDSON,
W.J., and MYERS, B.S. (1995). Dynamic responses of the head
and cervical spine to axial impact loading. J. Biomech., in press.
RAWLING, L.B. (1912). Fracture of the skull, in: The Surgery
of the Skull and Brain. Oxford University Press: London, pp.
68-134.
SANCES, A., MYKLEBUST, J., CUSICK, J.F, et al. (1981).
Experimental studies of brain and neck injury. Proc. 25th
Stapp Car Crash Conf. SAE Paper #811032.
SCHNEIDER, D.C., and NAHUM, A.M. (1972). Impact stud-
ies of facial bones and skull. Proc. 16th Stapp Car Crash
Conf. SAE Paper #720965.
VANCE, B.M. (1927). Fractures of the skull—Complications
and causes of death. A review of 512 necropsies and of 61
cases studied clinically. Arch. Surg. 14, 1023-1092.
VOIGT, GE., and SKOLD, G. (1974). Ring fractures of the
base of the skull. Trauma 14(6), 494-505.
VONDRA, J. (1965). Fractures of the Base of the Skull. Charles
C. Thomas, Springfield, IL.
WERNE, S. (1957). Studies in spontaneous atlas dislocation.
Acta Orthop. Scand. Suppl. 23, 1-150.
Address
reprint requests to:
Barry S. Myers
Dr.
Department of Biomédical Engineering
Duke University
Box 90281
Durham, NC 27708-0281