In the ancient Egyptian papyrus, Traumatic
quadriplegia and paraplegia was considered a condition not to be
treated. Paul of Aegina (AD 625-690) was the first surgeon to
do laminectomy for this condition. When the vertebral column
injury occurs some degree of neurological deficit is present in
significant percentage of cases. Neurological injury occurs in
40% in the cervical vertebral injury and 15% in thoraco lumbar
injury. The outlook for patients with incomplete thoracolumbar
neurological injuries has improved and can be enhanced if an
optimum environment for neurological recovery is provided. The
long-term survival and the possibility of functional
rehabilitation of patients with spinal cord injuries has led to
increased interest in treatment of these patients. Although
corticosteroids have been shown to improve neurological recovery
of patients with spinal cord injuries16, the primary treatment
of patients with thoracolumbar injuries is surgical reduction,
decompression, and stabilization. Controversies regarding the
classification of fractures, timing of surgical intervention and
type of surgical approach exist.
Spinal fracture classifications systems
have been developed to aid in the understanding of fracture
stability and help direct treatment. Denis and McAfee et al
described fractures based on their morphologic features.
Holdsworth and Whitesides used combinations of mechanistic and
morphologic systems to classify injuries. The McAfee system
identifies six fracture types (Table 1) based on CT scan
findings is currently the best classification.
Table 1: McAfee Classification of
thoracolumbar spine fractures
Isolated anterior column failure
Stable burst fracture
Anterior and middle-column compression
failure, posterior column intact
Unstable burst fracture I
Compressive failure of anterior and
middle columns, disruption of posterior column
Horizontal vertebral avulsion injury
with center of rotation anterior to vertebral body
Compressive failure of anterior column,
tensile failure of posterior column. The center of rotation
is posterior to anterior longitudinal ligament
Disruption of spinal canal alignment in
transverse plane, shear mechanism common
The American Spinal Injury Association
(ASIA) defines a complete neurologic lesion as an absence of
sensory and motor function below the level of injury including
the lowest sacral segment2. Although patients can be in “spinal
shock”, when all reflex activities are lost, Stauffer et al have
demonstrated that a bulbocavernosus reflex returns within 24
hours in 99% of patients, thus indicating the end of spinal
shock. Complete injuries have a less than 3% chance of
functional motor recovery if no neurological return is seen
within 24 hours and have no chance of neurological recovery
after 24 to 48 hours.
A lesion is incomplete if sensory, motor,
or both functions are partially present below the neurological
level of injury. Sacral sensation at the mucocutaneous junction
and the presence of voluntary contraction of the external anal
sphincter on digital examination should be carefully elicited as
this may be the only sign of preserved function. Patients with
incomplete neurological injuries are expected to improve, with
many regaining the ability to ambulate.
The level of neurological injury is graded
at the lowest nerve root level that has at least antigravity
strength. Clinically recognized syndromes are given in Table
2. Overall gross function of patients with spinal cord injuries
is assessed by the Frankel classification. This scale has
recently been modified by ASIA (Table 3).
Table 2: Incomplete cord syndromes
A lesion that produces variable loss of
motor function and of sensitivity to pain and temperature,
while preserving proprioception
A lesion that produces relatively
greater ipsilateral proprioceptive and motor loss and
contralateral loss of sensitivity to pain and temperature
A lesion, occurring almost exclusively
in the cervical region, that produces sacral sensory sparing
and greater weakness in the upper limbs than in the lower
Dorsal cord (posterior cord)
A lesion occurring almost in the dorsal
sensory column mainly affecting proprioception
Injury of the sacral cord (conus) and
lumbar nerve roots within the neural canal, which usually
results in an areflexic bladder, bowel, and lower limbs.
Sacral segments may occasionally show preserved reflexes.
E.g., bulbocavernosus and micturition reflexes.
Injury to the lumbosacral nerve roots
within the neural canal resulting in areflexic bladder,
bowel and lower limbs.
Table 3: ASIA impairment scale
Complete: No sensory or motor function
below level of neurologic deficit level. Sacral sparing is
Incomplete. Sensory but not motor
function is preserved below the neurologic deficit level
Incomplete. Motor function is
preserved below the neurologic deficit level, and the
majority of key muscles below the neurologic deficit level
has a muscle grade lower than 3
Incomplete. Motor function is
preserved below the neurologic deficit level, and the
majority of key muscles below the neurologic deficit level,
has a muscle grade higher or equal to 3.
Sensory and motor function is normal
Neurological injury of spinal cord
Primary injuries – which occur at
the time of injury.
Secondary injury – which occur later
as the result of
Swelling that accompany all spinal injuries
Delayed neurological deterioration due to
spinal deformity, residual spinal compression, scar formation
and post traumatic Syringomyelia
Neurologic injuries result from two basic
mechanisms, compression or traction. Traction injuries in the
thoracolumbar area frequently arise from seat-belt-type
accidents Traction injuries to neurologic structures have a
poor prognosis for recovery. Compressive injuries cause
neurologic injury from crushing and then from persistent
compression on the cord and nerve roots.
The effects of the duration and magnitude
of compression and contusion on neurologic recovery have been
studied in animal experiments. Bohlmann et al developed an
anterior cervical spinal cord injury model in dogs. He
demonstrated differences in neurologic recovery between
compression and contusion injuries. Animals that had contusion
injuries recovered fully over time; however, those animals with
persistent compression achieved full recovery only after
decompression. No correlation between the neurologic function
of the animals and the extent of histopathologic damage was
found. Using a rat model, Rivlin and Tator found that the
clinical consequence of cord compression was inversely related
to the duration of cord compression. Similarly, Dolan et al
found that increasing the force and the duration of spinal cord
compression was directly proportional to a reduction in
neurological recovery. Delamarter et al showed that a 50%
compression of the Cauda equina led to significant
neurophysiologic and electrodiagnostic changes in dogs. A 75%
compression of the Cauda equina resulted in permanent
More recently, Delamarter et al created
spinal cord injuries at L5 and L6 just above the conus in dogs
using constriction bands. The animals underwent “decompression”
immediately, and at 1,6, and 24 hours. Those animals that had
decompression immediately and at 1 hour were clinically and
electrophysiologically normal at 6 weeks, although
histologically gray matter necrosis was noted. The animals
“decompressed” at 6 hours or later did not make significant
The intrinsic mechanisms occurring during
cord injury have been extensively investigated. The initial
phase is characterised by hemorrhage into the cord and formation
of edema at the injured site and surrounding region. Local
spinal cord blood flow is diminished and can be further
compromised by extension of edema. A secondary tissue
degeneration phase begins within hours of injury due to release
of membrane-destabilizing enzymes, mediators of inflammation,
and disturbance of electrophysiological coupling by disruption
of calcium channel pathways. Lipid peroxidation and hydrolysis
appears to play a major role in the secondary phase of spinal
cord injury. Current pharmacologic intervention, such as
methylprednisolone, gangliosides, and lazaroids, aims to
suppress the secondary cell destruction phase on a biochemical
basis. Bracken et al16 reported the results of a randomized
double blind study in patients with spinal cord injuries. High
doses of methylprednisolone, naloxone, or a placebo were
administered to 487 patients. The patients given
methylprednisolone had statistically significant improvement of
motor and sensory scores compared with the two other groups. To
be effective, the drug had to be administered within 8 hours of
injury. Methylprednisolone was administered as a 30mg/kg
loading dose and as a continuous 5.4mg/kg/hr dosage for 23
Recovery of neural function may be further
influenced by the vascular supply to the spinal cord. The
thoracic spinal cord has relatively poor vascular supply as
opposed to the cervical spinal cord and the conus medullaris,
which have a more abundant collateral network. Additionally,
after injury, the spinal cord loses its ability to autoregulate,
and, therefore, spinal cord blood flow becomes directly related
to blood pressure.
These studies support the basic principles
of immobilization, fracture reduction, stabilization and
decompression in patients with spinal cord injuries. The
concept of early decompression is supported by human and animal
studies; however, some neurologic recovery can be gained by late
Resuscitation of patients with spinal cord
injuries follows the principles established by the American
College of Surgeons. Further diagnostic and therapeutic
measures are dictated by the presence of concomitant injuries,
which occur in 47% to 60% of patients with spinal injuries.
Hypertension associated with bradycardia may be secondary to
neurogenic shock and often is present in patients with
thoracolumbar spinal cord injuries. The presence of protracted
neurogenic shock in the absence of other sources of volume loss
should be treated with vasopressors to avoid over infusion and
possible development of pulmonary edema.
After resuscitation, the patient is rolled
to allow inspection and palpation of the spinal column. Areas
of localized tenderness, gaps between spinous processes,
swelling, and gibbous deformities are elicited. Persistent
localized tenderness despite normal radiographs is associated
with occult fractures in more than 30% of our patients.
Muscle strength is measured by systematic
evaluation of all muscle groups of the extremities and the anal
sphincter. Sensory examination includes testing of
proprioceptive and pain temperature pathways. Presence and
quality of deep tendon reflexes and long tract signs are
evaluated in all extremities. Because the prognosis of a
patient with an incomplete neurologic deficit may be
considerably altered by an appropriate intervention, it is
imperative that this neurologic examination is recorded in the
patient’s chart. Preservation of perianal pin prick sensation,
voluntary rectal tone, and great toe flexion are considered
signs of sacral sparing and may be the only remaining neurologic
function in some patients with incomplete spinal cord injuries.
The presence of sacral sparing implies at least some structural
continuity of the long tracts and should lead to timely attempts
at optimizing chances for neural recovery.
Anteroposterior and lateral radiographs are
obtained in patients with suspected spinal injuries. If one
spinal fracture is detected, a complete spine series is
required, because noncontiguous spine fractures occur in 10% to
30% of patients.
CT has enhanced the understanding of
mechanisms of neurologic injury and fracture morphology. The
middle column and subtle posterior injuries can easily be
diagnosed with. To allow accurate fracture classification and
to help direct treatment, we recommend CT examination in the
majority of patients with thoracolumbar fractures.
Magnetic resonance imaging (MRI) can
demonstrate spinal cord pathology and the presence of neural
compression. Other soft tissue injuries and the state of the
intervertebral disk can be identified. Cotler et al reported
that MRI can predict neurologic recovery based on T2 weighted
images. MRI is indicated in patients with progressive
neurologic deterioration, incongruous neurologic and skeletal
injury, and unexplained neurologic deficit. Also, MRI can be
used to assess the status of the posterior ligamentous complex.
Myelography has little role in the evaluation of patients with
acute neurologic injury, unless MRI is not available.
Prior to definitive treatment, patients are
rolled every 2 hours to prevent decubitus ulcers. Considerable
fracture site motion can occur with this technique; therefore,
we believe truly unstable fractures require surgical
All patients with spinal cord injuries
presenting with in 8 hours from injury should receive high doses
of methylprednisolone as specified previously.
We currently recommend early surgical
treatment for patients with thoracolumbar fracture for complete
and incomplete spinal cord injury.
Goals of surgery
The goals of surgery are to reduce fracture
and dislocations, stabilize the injured segment and decompress
the neural elements. Injuries of the cervical spine can usually
be reduced quickly using skeletal traction. Fracture reduction
results in an indirect decompression of the neural elements and
may lead to improved neurologic recovery. Because nonoperative
means usually fail to achieve an anatomic fracture reduction in
the thoracolumbar spine, we, therefore, recommend early surgical
treatment in patients with partial defects. Unfortunately,
surgical treatment for complete paraplegic patients does not
result in significant neurologic recovery. Early surgical care
in patients with complete paraplegia does decrease the rate of
complications, hospitalization time, and overall costs,
however. Other indications for emergent surgical care are
patients with progressive neurologic deficits, open spine
fractures, and burns of the torso with concomitant unstable
Timing of surgery
Timing of surgical intervention and the
effect on neurologic outcome remain controversial. The only
accepted indication for emergent surgical treatment in a
patient with a thoracolumbar fracture is progressive neurologic
deterioration. This complication is rare, seen in 1% t0 2% of
cases and may be secondary to fracture displacement, expanding
epidural hematoma, spinal cord edema, or infarction. Immediate
spinal realignment and decompression of the neural elements has
been shown to consistently reduce permanent neurologic sequelae
in several animal investigations. There appears to be a small
window of time below 4 hours, when spinal cord decompression can
result in reversal of neurologic deficits. We have observed
this phenomenon in three patients with bilateral cervical facet
dislocations who had reduction within 2 hours of injury. After
reduction, these patients felt an immediate return of sensation,
and the rapid return of motor function over the next 24 hours.
Krengel et al evaluated efficacy of
surgical decompression within the first 24 hours in patients
with incomplete spinal cord injury secondary to fractures
between T3 and T11. Thirteen patients presented with Frankel B
or C paraplegia. The average improvement in Frankel score was
2.2 grades, with 11 of 13 regaining the ability to ambulate.
These results exceeded outcomes reported for other treatment
methods, including nonoperative treatment, delayed posterior
instrumentation, or late anterior decompression. No patients
developed neurologic worsening as a result of early treatment.
Non operative management
Postural reduction for the treatment of
spinal cord injuries was developed by L.Guttman at the Stoke
Mandeville Hospital. The treatment methods consisted of
rotating beds supplemented by pillows and rolls placed under the
patients’ backs to support a reduced position. Duration of
recumbency averaged 10 to 13 weeks followed by a supportive
plastic jacket for several weeks.
A long term outcome of closed treatment in
neurologically intact patients with thoracolumbar burst
fractures was recently reported by Mumford et al. They found
that 66% of patients had good or excellent results at a
follow-up an average of 2 years later. Only one patient
developed neurologic deterioration (2.1%). Radiagraphic
follow-up of these patients revealed an average increase of
kyphosis of 30 and an improvement of the canal compromise from
an initial average of 37% (16% to 66%) to 14% (3% to 40%)
secondary to remodelling.
Jacobs et al compared recumbent treatment
and posterior instrumentation in patients with incomplete
thoracolumbar cord lesions. They documented a 44% incidence of
improvement in the non operative treatment group compared with a
53% improvement in the Harrington rod group. The group treated
with Harrington rods had better fracture reduction, fewer
complications, and spent less time for rehabilitation than the
recumbent treatment group. Burke and Murray, however, found no
significant differences in their comparison of nonoperatively
and operatively treated patients with injuries between T2 and
T11 and initial Frankel grades B and C.
In the majority of comparison studies
between surgical and non surgical treatment, it appears that
neurologic recovery is enhanced in surgically treated patients.
The choice of surgical approach
and instrumentation requires a thorough understanding of
fracture type, injury level, and degree of neural injury.
Patients who have a distractive injury of the posterior elements
that occurs in flexion-distraction injuries, Chance-type
injuries, and fracture-dislocations are best treated with
posterior instrumentation. Patients with unstable burst
injuries and incomplete paraplegia associated with high-grade
spinal canal stenosis may benefit in the long term from
immediate anterior decompression. Patients with unstable burst
fractures and lesser degrees of canal stenosis are treated by
Principles of surgical care
For posterior approaches, the patient is
placed prone using a Stryker frame. A transverse pelvic roll
increases lordosis, thus reducing the deformity. The use of
succinylcholine is contraindicated in patients with neurologic
injury because it can lead to an uncontrolled release of
potassium and ventricular fibrillation.
Early surgical treatment of thoracolumbar
fractures may be associated with extensive blood loss; however,
induced hypotension is contraindicated because this can lead to
a decrease in spinal perfusion because of the loss of spinal
cord autoregulation. To limit transfusion requirements, we
routinely collect blood in the cell saver device.
After surgery, the patient is casted in the
operating theatre for a custom-molded orthosis. This step
allows for earlier patient ambulation and decreases
hospitalization time. The orthoses is worn for 3 to 4 months.
Posterior Surgical Treatment of Unstable
In unstable burst-type fractures, the
spinal deformity consists of a loss of vertebral body height,
kyphotic angulation, and retropulsion of bone into the spinal
canal. Posterior instrumentation restores vertebral body height
via distraction forces. Additionally, the anterior and middle
columns are restored to normal length during correction of the
kyphosis. The spinal column assumes the contour of the rods,
resulting in correction of the kyphotic deformity. This
correction is facilitated by using Harrington rods with
square-ended hooks or rod sleeves. Newer systems reduce
deformity by in situ rod contouring or allowing rotation of
hooks and pedicle screws about vertical bars or rods.
The techniques of posterior instrumentation
for thoracolumbar burst fractures have been studied
extensively. In simple rod-hook systems at least two levels
above and below the injury should be spanned to provide
sufficient leverage to achieve and maintain fracture reduction.
The use of instrumentation systems that allow for rod contouring
is essential. Indirect spinal canal decompression is thought to
be achieved by development of tension in the posterior
longitudinal ligament, which then pushes the retropulsed bone
fragments forward. This process has been termed ligamentotaxis.
Crutcher et al found posterior distraction instrumentation to
reduce canal compromise by only 50% of the initial occlusion,
leaving an average of 32% canal compromise.
Edwards et al developed rod sleeves that
are centered over the pedicle of the fractured vertebrae. They
act to push the vertebrae forward, thereby reducing the kyphotic
deformity. To use these rod sleeves, the posterior neural arch
must be intact, otherwise iatrogenic neurologic injury can
result. Edwards et al found excellent maintenance of alignment
in terms of kyphosis, vertebral body height, and translation in
their review of 122 patients. The spinal canal area was
improved by 32% (from 55% patency to 87%) if the rod-sleeve
construct was inserted within 2 days of injury. Between 3 and
14 days, they found a 23% improvement in canal deviance (range,
53% to 76%). Little improvement was found with surgery after 14
days. An average of 64% recovery of original neurologic
function occurred in incomplete thoracolumbar fractures.
Dick described the “fixateur interne”,
which uses 5mm Schantz pins linked to 7mm threaded rods. The
Schanz pins are placed into the pedicles of the vertebrae above
and below the injured level. The long lever arms of the Schanz
pins allow excellent control of the unstable spine.
Satisfactory reduction of kyphosis and neurologic recovery
occurred in 52% of patients with a Frankel score of A through
C. All patients with Frankel D had at least partial recovery.
Short segment fixation using
Cotrel-Du-bousset (CD) instrumentation has had poor outcome as
reported by McLain et al. They reviewed 19 patients and found
vertebral collapse, vertebral translation, or hardware failure
in 10 patients. The primary cause for failure was attributed to
the fixation device. Outcome studies using newer systems such
as the Texas Scottish Rite Hospital or ISOLA® Spinal Systems in
trauma applications have been good. Moss Miami posterior spinal
instrumentation is a new spinal instrumentation introduced,
which is a hybrid system using pedicular screws, rods, laminar
hooks and pedicular hooks. It is easy to apply.
An alternative to the arthrodesis of long
segments is to limit the fusion to the level of injury and
remove the rods about 1 year later to regain motion of the
instrumented but not fused segments. Kahanovitz et al found
evidence of articular cartilage degeneration of unfused but
instrumented levels in animal and human studies. Dekutoski et
al documented good clinical results and radiographic return of
motion in unfused levels in patients with the rod long-fuse
Posterior treatment of
flexion-distraction and Chance-Type Injuries
Flexion-distraction and Chance-Type
injuries are associated with neurologic deficit in 10% of
cases. Anderson et al reported on the complete neurologic
recovery in all patients with incomplete impairment who were
treated with posterior instrumentation. Patients may present
with complete spinal cord injuries at higher levels than the
bony injury because of traction spinal cord injuries.
Unfortunately, the prognosis for neural recovery is extremely
Posterior instrumentation is indicated in
patients with Chance-type injuries and flexion distraction
injuries who have ligamentous injuries, kyphosis of more than
100, or associated abdominal trauma. Reduction is achieved with
lordotic positioning and placement of an interspinous wire.
Flexion-distraction injuries require neutralization using either
a rod-hook sleeve or pedicle screw system. Chance-type injuries
can be treated with compression instrumentation. During
reduction, patients should be assessed for displacement of the
intervertebral disk into the canal as has been reported by
Levine et al and others.
Posterior Treatment of Fracture
Patients with facet fracture-dislocations
and anterior vertebral body translation usually present with
complete paraplegia, although they may rarely present with
incomplete neurologic impairment or intact. These patients are
treated posteriorly with segmental instrumentation. Unchecked
distraction can result in excessive lengthening of the spinal
column because these injuries may be associated with complete
ligament disruption. Krengel et al reported anatomic reduction
after Harrington rod distraction instrumentation in 9 of 10
patients with fracture dislocations of the thoracic spine. For
patients with fracture-dislocations of the lumbar spine, we
recommend the use of pedicle screw systems.
Laminectomy is indicated for depressed
fractures of the posterior elements, epidural hematoma, and
dural lacerations with cerebrospinal fluid leak.
Flesh et al described the posterolateral
decompression technique that allows forward impaction of the
retropulsed bone and disk fragments, thus decompressing the
ventral side of the neural tissues. This technique is
facilitated by removal of the facets or pedicles. Manipulation
and medial retraction of the neural elements in the injury zone
and copious epidural bleeding are potential drawbacks of this
Anterior Decompression and Fusion
Spinal cord injury results in permanent
loss of cell nuclei and axonal tracts; however, some viable
neural tissue may remain, which may be prevented from recovery
because of persistent cord compression. Anterior decompression
has been shown to increase axoplasmic flow, decrease ischemia,
and lead to improvement of neurolgic function.
Anterior decompression and strut grafting
of the acute burst fracture causes a significant increase of
instability and the possible development of a progressive
deformity and graft dislodgement. Anterior instrumentation has
been developed to minimize this complication. The Kaneda device
uses threaded rods that rigidly connect to screws placed
transversely across the vertebral body. Distraction or
compression can then be applied.
A potential cause of anterior device
failure is the settling of the anterior strut graft that occurs
during healing. This settling increases bending on the screws
and couplings. Carl et al found that 3mm of bone graft settling
occurred if a strut graft was placed on the bleeding subchondral
with keyed-in grafts. Based on their findings, the authors
recommended a dynamized anterior CD instrumentation. This
implant system, however, allowed an average of 80 loss of
1. Aebi M, Etter C, Kehl T et al:
Stabilization of the lower thoracic and Lumbar spine with the
internal skeletal fixation system: Indications, techniques, and
first results of treatment. Spine 12: 544-551, 1987.
2. American College of Surgeons: Advanced Trauma Life Support
Manual. Chicago, American College of Surgeons, 1992.
3. Anderson PA, Bohlmann HH: Anterior decompression and
arthrodesis of the cervical spine: Long-term motor improvement:
II. Improvement in complete traumatic quadriplegia. J Bone
Joint Surg 74A: 671-682, 1992.
4. Anderson PA, Henley MB, Rivara FP, et al: Flexion
distraction and chance injuries to the thoracolumbar spine. J
Orthop Trauma 5: 153-160, 1991.
5. Bohlmann HH, Anderson PA: Anterior decompression and
arthrodesis of the cervical spine: Long-term motor
improvement: I. Improvement in incomplete traumatic
quadriparesis. J Bone Joint Surg 74A: 659-670, 1992.
6. Bradford DS, Akhbarnia BA, Winter RB et al: Surgical
stabilization of fracture and fracture dislocations of the
thoracic spine. Spine 2: 185-196, 1977.
7. Breig A: Adverse mechanical tension in the central nervous
system. Uppsala, Almquist and Wiksell, 1978.
8. Denis F: The three column spine and its significance in the
classification of acute thoracolumbar spinal injuries. Spine 8:
9. Dick W: The “fixateur interne” as a versatile implant for
spine surgery. Spine 12: 882-899, 1987.
10. Ducker TB, Solomon M, Daniel HB: Experimental spinal cord
trauma: III. Therapeutic effect of immobilization and
pharmacologic agents. Surg Neurol 10: 71-76,1978.
11. Flesh JR, Leider LL, Erickson DL, et al: Harrington
instrumentation and spine fusion for unstable fractures and
fracture-dislocations of the thoracic and lumbar spine. J Bone
Joint Surg 59A: 143-153, 1977.
12. Gorio A, Di Guillo AM, Young W, et al: GM1 effects on
chemical, traumatic and peripheral nerve induced lesions to the
spinal cord. In Goldberger ME, Gorio A, Murray M (eds):
Development and Plasticity of the Mammalian Spinal Cord, vol.3.
Padua, Italy, Liviana Press, 1986, pp2270242.
13. Goris RJA, Gimbere JSF, van Niekirk JLM et al: Early
osteosynthesis and prophylactic mechanical ventilation in the
multitrauma patient. J Trauma 22: 895-903, 1982.
14. Grundy D, Swain A, Russell J: ABC of spinal cord injury:
Early management and complications. Br. Med K 292:44-47, 1986.
15. Gurr KR, McAfree PC, Chi-Ming S: Biomechanical analysis of
anterior and posterior instrumentation systems after corpectomy.
J Bone Joint Surg 70A: 1182-1191, 1988.