FRACTURES, DISLOCATIONS, AND FRACTURE-DISLOCATIONS

OF THE SPINE*

F. W. HOLDSWORTH, SHEFFIELD, ENGLAND

1 have chosen the subject of injuries to the vertebral column because it is one that has
interested Watson-Jones for many years, and one upon which he has written extensively.
Indeed it was his writings which stimulated me to study vertebral injuries, especially those
associated with paraplegia. I am pleased to be able in this paper to pay tribute to him as a
great surgeon who is also a great personal friend.
Thirty years ago Watson Jones (1931), following the work of Davis (1929), published a
classification of fractures of the spine together with a method of treatment of what were
considered to be pure flexion fractures, whereby the fracture was reduced by hyperextension,
position being maintained by fixation ofthe spine in plaster in hyperextension for many weeks.
He stated that accurate reduction was almost always possible, that consolidation occurred
without deformity, and that provided exercise in plaster was properly performed the functional
results were excellent.
My experience of the treatment of vertebral fractures did not confirm this. I found that
in a considerable proportion of patients the displacement could not be reduced by
hyperextension, and that even if good position was obtained it could not be maintained by
a plaster jacket in hyperextension. Moreover, the final functional results were far from
satisfactory. Many patients suffered from persistent pain, not so much at the fracture site
as over the whole back.
This opinion was confirmed by Nicoll (1949), who in a classic paper based upon the study
of a large number of fractures of the thoraco-lumbar spine found the results of treatment by
hyperextension to be unsatisfactory, and suggested a new classification of thoraco-lumbar
fractures into stable and unstable types. In stable fractures the intraspinous ligaments
remained intact, whereas in the unstable types these ligaments were ruptured. He stated that
in stable fractures reduction and immobilisation were unnecessary-indeed if the spine was
left free and exercises were carried out from within a week or two of the time of injury the
results were much better than those after plaster fixation. Unstable fractures and fracture-
dislocations, however, required plaster fixation but not reduction. The best results were
obtained when the affected vertebral bodies fused spontaneously in the deformed position.
His investigation was confined to thoraco-lumbar fractures.
At this time I had begun to be interested in traumatic paraplegia, particularly at the
thoraco-lumbar level, and in order to investigate the exact nature of the damage to the spinal
cord, the nerve roots and the vertebral column, I explored a number of fracture-dislocations
with paraplegia, and was struck by the fact that almost invariably the posterior ligaments
were completely ruptured. This observation led to a long investigation into vertebral injuries.
As a result of clinical, radiological, post-mortem and operative investigations on a large number
of injured spines, a classification has now been evolved which forms, in my opinion, a rational
basis for treatment. This classification is an elaboration of Nicoll’s original classification into
stable and unstable injuries extended to the whole spine.

ANATOMY
With the exception of the first two cervical and the sacral vertebrae all the vertebrae
articulate together in the same way-by the intervertebral disc and by the postero-lateral

* Third Watson-Jones Lecture delivered at the Royal College of Surgeons of England on January Il, 1962.

6 THE JOURNAL OF BONE AND JOINT SURGERY

 FRACTURES, DISLOCATIONS AND FRACTURE-DISLOCATIONS OF THE SPINE 7

joints. The synarthroses between the vertebral bodies rely for their stability upon the immensely
strong annulus fibrosus. The diarthrodial apophysial joints are stabilised by the capsule, by
the intraspinous and supraspinous ligaments and the ligamenta flava. This group of ligaments
I call the “posterior ligament complex.” It is upon this complex that the stability of the spine
largely depends (Fig. I).

Poefertor Lonit&*1jrtaL -,

Postertor

Iriter5firtous

FIG. 1
Diagram of the vertebral articulations with the principal ligaments. The
intraspinous and supraspinous ligaments, the capsules of the lateral joints
and the ligarnentum flavum constitute the “posterior ligament complex.”

FIG. 2
The articular processes of the cervical, thoracic and lumbar vertebrae.

The shape of the articular processes varies in the different parts of the spine (Fig. 2).
In the cervical spine the articular processes are small and flat, those of the upper vertebra
pointing downwards and forwards, those of the lower upwards and backwards. In the thoracic
spine the processes are also flat, but point almost directly backwards and forwards. In the

VOL. 45 B, NO. 1, FEBRUARY 1963

8 F. W. HOLDSWORTH

lumbar spine they are large and curved and point almost directly inwards and outwards, the
upper facets of the lower vertebra embracing the lower facets of the upper vertebra. The
joints of the thoracic spine are further stabilised by the rigidity of the thoracic cage.
The spine may be subjected to four types
of violence: 1) flexion, 2) flexion and rotation,
3) extension, and 4) compression. The type of
fracture, dislocation or fracture-dislocation which
results from each of these types of violence will depend
upon whether or not the posterior ligament complex
is ruptured and, because of the shape of the bones,
upon the part of the spine involved.
The posterior ligament complex is seldom if ever
ruptured by pure flexion violence; it is always or
almost always ruptured when the spine is subjected
to forcible rotation or flexion-rotation, and the
greater the rotational element the greater the liability
to rupture.
If the posterior ligament complex remains intact
then the violence is expended upon the vertebral
body and a wedge compression fracture results
(Fig. 3). This type of fracture results from pure
FIG. 3 flexion and is most common in the thoracic and
lumbar spines, for in the thoracic spine stability is

ligament complex is intact, increased by the rib cage and in the lumbar spine the
ligament complex is extremely strong. The resulting
wedge compression fracture is stable, for the fragments of the body are firmly impacted
and the articular processes and the posterior ligaments are intact. Further displacement will
not occur unless the spine is subjected to greater violence than that which caused the fracture.

FIG. 4
Diagrams to show dislocation of the cervical spine and of the lumbar spine. The
posterior ligament complex is ruptured in each case. More flexion is necessary to
dislocate the lumbar spine than to dislocate the cervical spine.

With flexion-rotation violence the posterior ligaments rupture, the posterior tie is lost,
and if the amount of flexion is sufficient to disengage the articular processes, then pure
dislocation results (Fig. 4). The vertebral body cannot be compressed because there is no

THE JOURNAL OF BONE AND JOINT SURGERY

 FRACTURES, DISLOCATIONS AND FRACTURE-DISLOCATIONS OF THE SPINE 9

fulcrum for the lever. Pure dislocation is common in the cervical spine, for the amount of
flexion necessary to disengage the articular processes is slight. It occurs but rarely in the
lumbar spine, where the amount of flexion necessary to disengage the large articular processes

FIG. 5
Diagram of rotational fracture-dislocation of the lumbar spine. The
posterior ligament complex is ruptured and the vertebral body is fractured.
This is a very unstable injury.

is great. It never occurs ill the thoracic spine because of the supporting rib cage. Pure
dislocations are unstable.
Rotation or flexion-rotation in the lumbar spine results in a fracture-dislocation, for the
flexion is not sufficient to disengage the articular processes
(Fig. 5). The posterior ligaments rupture, one or other
or both the articular processes fracture and the upper
vertebra swings upon the lower, carrying with it the disc
alid a wedged slice of the upper border of the lower
vertebral body. This is the classical rotational fracture-
dislocation of the thoraco-lumbar and lumbar spine,
which is so constantly associated with paraplegia. This
fracture-dislocation is clearly very unstable. It corre-
sponds to Nicoll’s fracture-dislocation and lateral wedge
fracture.
Extension violence produces the reverse of flexion
injury. In pure extension the ligaments usually remain
intact and the posterior part of the vertebra, that is the . FIG. 6
Diagram of extension injury. The
lamina and pedicles, fractures. This is a rare injury, anterior common ligament is ruptured.
because more frequently the anterior common ligament The posterior complex is intact.

ruptures and an extension dislocation occurs. This

is conimon in the cervical region as a result of “whiplash” injuries, but extremely rare in
the lumbar spine. These injuries are stable in flexion (Fig. 6).
The fourth type of violence is a compression force-that is, a force transmitted directly
along the line of the vertebral bodies. This can only occur in those parts of the spine that can
be held straight, the lumbar and cervical spines. With the neck slightly flexed the cervical
spine is straight and violence applied to the vertex of the skull will be transmitted along the
line of the vertebral bodies (Fig. 7). This force ruptures one of the end plates and the disc
is forced into the body of the vertebra, which bursts (Fig. 8). All the ligaments remain intact

VOL. 45 B, NO. 1, FEBRUARY 1963

10 F. W. HOLDSWORTH

and the spine is stable. This is the compression “burst” fracture and corresponds to the
comminuted wedge fracture of Nicoll and Watson-Jones. It is often associated with tetraplegia
in the cervical region, the cord being damaged by the posterior fragments of the exploding
body. This type of violence also produces the characteristic “burst” fracture of the atlas.

I-__. I
The cervical spine is straight when the head is in slight flexion.

The mechanism of production of the burst fracture as well as the tearing of the
posterior ligament complex by rotational violence has recently been confirmed by the
ingenious experimental work of Roaf (1960).
Thus the basic classification of stable alid unstable
fractures which Nicoll applied to the thoraco-lumbar
spine can be extended to cover the whole spine and
elaborated in order to describe more accurately the
exact type of injury.
DIAGNOSIS
Early diagnosis of the type of injury is of crucial
importance because treatment and prognosis depend
upon an accurate assessment of the damage to the spine.
Diagnosis is easy if it is appreciated firstly that
stability after injury depends upon whether or not
the posterior ligament complex remains intact,
and secondly that each type of fracture and
fracture-dislocation has a characteristic radiographic
appearance.
Clinical examination is of great importance.
FIG. 8 Inspection of the back will often give a clue to the
Diagram of “burst” fracture. All the exact type of violence which has caused the injury.
ligaments are intact.
For example, abrasions and contusions over one
scapula show that the force has connected at that spot, producing flexion and rotation at the
thoraco-lumbar level (Fig. 9); or an abrasion on the vertex would suggest vertical compression
force to the cervical spine. Palpation of the back is of great value. Rupture of the

THE JOURNAL OF BONE AND JOINT SURGERY

 FRACTURES, DISLOCATIONS AND FRACTURE-DISLOCATIONS OF THE SPINE 11

supraspinous and intraspinous ligaments can readily be detected. This physical sign is perhaps
the best indication of instability of the Spine, and together with careful examination of
good antero-posterior and lateral radiographs which include the spinous processes will
usually enable the injury to be placed in the appropriate category.

FIG. 9
Abrasion over shoulder
one indicating flexion-rotation
violence. This patient sustained a rotational fracture-
dislocation of the twelfth thoracic vertebra upon the first
lumbar.

Radiographs of the wedge fracture show a compression of the anterior part of the body
with minimal separation of the spinous processes (Fig. 10). Often more than one vertebra is
involved and the end plates are usually intact. _______
With compression or “burst” fractures 1’ ft.
the vertebral body is shattered and fragments
are displaced outwards from the centre; again
the spinous processes are not separated (Figs.
11 and 12). ‘K-
With flexion-rotation force and rupture of
the ligament complex there is separation of the
spinous processes and, if the articular processes
disengage, the pure dislocation is obvious
(Figs. 13 and 14).
If there is more rotation than flexion and
the articular processes fail to disengage then a
rotational fracture-dislocation results. Separ-
ation and lateral displacement of the spinous
processes, together with fracture of one or both
articular processes, can be seen in the antero-
posterior view and the characteristic slice wedge

:
of the upper part of the body of the lower
vertebra can be seen in both views (Fig. 15). .-‘ #{149}‘-‘ -.

It is this slice wedge of the body which can lead

to confusion, for it can easily be mistaken in ______________
the lateral view for a wedge compression with FIG. 10
comminution. The wedge fracture is never Wedge, fracture of the seventh thoracic vertebra.
There is no separation of the spinous processes.
extensively comminuted.
Great care must be taken to recognise this fracture-dislocation. The displacement seen
in the radiographs is often minimal, for the fracture is so unstable that the spine falls into
alignment when the patient is laid supine, or when the pelvis is held in line with the shoulders.
aIld conversely, it may completely redislocate with torsion. Injudicious handling can easily
damage the cord or nerve roots. Almost all paraplegia in the lumbar and thoraco-lumbar
regions is associated with this fracture-dislocation.

VOL. 45 B, NO. 1, FEBRUARY 1963

12 F. W. HOLDSWORTH

FiG. II FIG. 12
Figure 1 l-’ Burst ‘ fracture of two lumbar vertebrae. The typical shattering of the vertebral body
is well illustrated. The spinous processes are not separated. Figure 12-” Burst “ fracture of a
cervical vertebra.

FIG. 13 FIG. 14
Figure 13-Dislocation of the cervical spine. Figure 14-Dislocation of the lumbar spine. In both
cases the posterior ligament complex is ruptured. Note the separation of the spinous processes.

THE JOURNAL OF BONE AND JOINT SURGERY

 FRACTURES, DISLOCATIONS AND FRACTURE-DISLOCATIONS OF THE SPINE 13

FIG. 15
Rotational fracture-dislocation of the twelfth thoracic vertebra upon the first lumbar.
Note the “slice” fracture of the vertebral body and the displacement of the spinous
processes.

FIG. 16 FIG. 17
Figure 16-Extension dislocation in ankylosing spondylitis. The deformity has remained.
Figure 17-Extension dislocation. The lower anterior margin of the body of the second
cervical vertebra has been avulsed. This is an unusually large fragment.

VOL. 45 B, NO. 1, FEBRUARY 1963

 14 F. W. HOLDSWORTH

Extension dislocations can be difficult to recognise. They are common in the cervical
region. The dislocation usually becomes reduced spontaneously and the reduction is stable
in flexion. The only radiographic abnormality is often a small avulsion fracture of the front
edge of the vertebral body indicating a rupture of the anterior common ligament. This small
avulsion fracture can be recognised with good quality radiographs including oblique views
(Figs. 16 and 17).
To recapitulate: fractures, dislocations and fracture-dislocations of the whole spine can
be divided into stable and unstable types. The diagnosis of each type of fracture by clinical
examination and by radiography is simple.

FIG. 18 FIG. 19
Figure 18-” Burst” fracture of the third lumbar vertebra. Figure 19-Fusion between the bodies
of the second, third and fourth lumbar vertebrae after immobilisation for four months.

TREATMENT
The classification of fractures is only of real value if the separation into distinct types is
of value in treatment. In this respect the importance of distinguishing between stable and
unstable fractures of the spine cannot be over-emphasised. In addition to this, however,
treatment depends upon a number of additional factors.
The object of treatment of any fracture, dislocation or fracture-dislocation of the spine
is to restore a painless, powerful, mobile back. Nicoll has shown that this objective is best
achieved by obtaining stability at the affected segments rather than by the restoration of
anatomical position, and most surgeons with experience of vertebral fractures will agree
with this.
When the posterior ligament complex has been ruptured, healing such as to restore the
original strength does not occur. For example, it is common knowledge that after accurate
reduction and prolonged immobilisation of dislocations of the cervical spine late redisplacement
often occurs. Thus after unstable injuries to the vertebral column permanent restoration of
stability cannot be achieved by healing of the ligaments alone.

THE JOURNAL OF BONE AND JOINT SURGERY

 FRACTURES, DISLOCATIONS AND FRACTURE-DISLOCATIONS OF THE SPINE 15

When the vertebral body is fractured

and the intervertebral disc destroyed, heal-
ing of the fracture almost always results
in fusion of the fractured body to that of
the adjacent vertebra, and this is even
more certain if the articular processes have
also been broken. The result of this spon-
taneous inter-body fusion is the restoration
of complete stability to the injured segment
of the spine, and is independent of healing
of the ligaments. It does not usually occur,
however, if the vertebra is not fractured.
If these facts are remembered rational
treatment of each type of vertebral injury
becomes clear.

STABLE INJURIES FIG. 20 FIG. 21

Wedge compression fracture-The simple Figure 20-Dislocation of cervical spine. Figure 21-
Fusion by posterior grafts. The spinous processes are
wedge fracture, often occurring in more tied together with wire.
than one vertebra in the same patient, does
not require reduction or immobilisation. All that is necessary is rest in bed for two to three
weeks, followed by active exercises for six to eight weeks. The fracture unites with some
deformity, but this seldom causes any disability, for since the posterior ligament complex is
intact the spine is stable. The patients are usually able to return to all but the most arduous
forms of physical work, and even this is often possible. Nicoll found that over 80 per cent of
his patients were able to return to hard manual work underground in the coal pits.
Compression burst fractures-Although the
ligaments remain intact and the spine is
therefore stable, the body of the vertebra is
comminuted, and since the bone fragments
are not impacted any movement causes COn-

siderable pain. It is therefore necessary to

immobilise the spine by a plaster applied in
the neutral position. Reduction of deformity
is not important. Since the body of the
vertebra is fractured and the disc disrupted.
healing can confidently be expected to lead
not only to consolidation of the fracture
but to fusion of the affected vertebra with
the vertebra above, with resulting complete
stability of the spine (Figs. 18 and 19).
In both the cervical and lumbar spines
a plaster is therefore applied with the spine
in the neutral position, and retained until
union and fusion have occurred-usually a
matter of three to four months.

UNSTABLE INJURIES

Dislocations-In dislocations the posterior

Anterior fusion by grtGohird fourth and fifth ligament complex is always ruptured and
cervical vertebrae, healing sufficiently strong to restore the full

VOL. 45 B, NO. 1, FEBRUARY 1963

16 F. W. HOLDSWORTH

stability of the spine cannot be depended upon even after months of immobilisation. In all
dislocations, therefore, stability must be restored by bone grafting of the affected vertebrae.
It is only necessary to fuse the two affected vertebrae.
In the cervical spine reduction of the dislocation is effected by manipulation under
anaesthesia, with the additional aid of skull tongs or calipers. After reduction position is
maintained for some days by weight traction upon the tongs and then fusion is performed
either posteriorly by grafts along the laminae or anteriorly by grafts sunk into the vertebral
bodies (Figs. 20 to 22). We now prefer the anterior route because fusion appears to be more
Certain by this method. The same programme is followed whether there is cord damage or not.
In the rare lumbar dislocations, operative reduction and fusion by bone grafts bolted to
the spinous processes and reinforced by small grafts applied to the roughened laminae is
carried out at once (Figs. 23 and 24).

FIG. 23 FIG. 24
Figure 23-Dislocation of the lumbar spine. Figure 24-After reduction of the dislocation the spine is fused
by tibial grafts bolted to the spinous processes.

Extension fractures and dislocations-All these injuries are stable in flexion, and stability can
be restored by holding the spine in the neutral position for two to three months. Fusion is
unnecessary.
Rotational fracture-dislocations-These fracture-dislocations only occur at the thoraco-lumbar
junction and in the lumbar spine. The horizontal fracture of the body, the fracture of the
articular processes and rupture of all the major ligaments make this fracture the most unstable
of all vertebral injuries. lt is so unstable that reasonable alignment of the spine can be obtained
merely by laying the patient supine. The instability of the spine is so great that the cord and
roots are in grave danger of damage-indeed 95 per cent of all paraplegias at the thoraco-
lumbar level are associated with this fracture-dislocation.
Because of the fractures of the body and of the articular processes, healing constantly
results in fusion of the affected vertebral bodies and complete stability of the spine. In the
absence of paraplegia and sensory loss over the sacrum, buttocks and legs, the spine is easily

THE JOURNAL OF BONE AND JOINT SURGERY

 FRACTURES, DISLOCATIONS AND FRACTURE-DISLOCATIONS OF THE SPINE 17

Rotational fracture-dislocation without paraplegia.

FIG. 26
After three months’ immobilisation in plaster, fusion is taking place between the adjacent bodies.

VOL. 45 B, NO. 1, FEBRUARY 1963

18 F. W. HOLDSWORTH

restored to satisfactory alignment by laying the patient supine, and satisfactory immobilisation
can be secured by nursing the patient in a plaster bed with a turning case. The plaster bed
immobilisation should be continued for six weeks, by which time the fracture is sufficiently

-. - 1 ‘- - -

-
I-

- :;“,P.C,
‘- I.

FIG. 27
Flexion-rotation violence. (Reproduced by permission of the Yorkshire Post.)

stable to allow the application of a plaster jacket without risk of damage to the spinal cord or
nerve roots. After the jacket has been applied mobilisation and exercise must be insisted upon.
The jacket is removed when fusion is firm, usually at twelve to fourteen weeks (Figs. 25 and 26).
When at the original injury the cord or nerve roots
or both have been damaged and there is loss of sensation
over the sacrum, buttocks and legs, plaster must never
be used because it invariably causes serious skin ulceration.
Skin ulceration can be prevented only by turning the
patient two-hourly day and night. In this region of the
spine the paraplegia is mostly due to damage to the nerve
roots of the cauda equina, and recovery is always poss-
ible; therefore the vertebral fracture must always be
accurately reduced in order to free the roots from
pressure and firmly fixed to prevent movement at this
extremely unstable fracture during the necessary turning.
This stability can best be achieved by internal fixation
of the unstable spine by plates bolted to the spinous
processes (Figs. 27 to 34). The fracture-dislocation is
exposed and the spinous processes immediately above
and below are gripped by powerful forceps. The

FIG. 28
Case 1-The rotational fracture-dislocation of the twelfth thoracic
vertebra upon the first lumbar produced by the violence shown in
Figure 27. The patient sustained a transection of the cord at the
first sacral neurological segment and damage to all the lumbar nerve
roots with paraplegia complete from the first lumbar neurological
FIG. 28 segment. In these circumstances internal fixation is indicated.

THE JOURNAL OF BONE AND JOINT SURGERY

 FRACTURES, DISLOCATIONS AND FRACTURE-DISLOCATIONS OF THE SPINE 19

displacement is then reduced by manipulation under direct vision. At the same time the
damage to the dura and cord can be inspected through the tear in the ligamentum flavum.
Plates are laid along each side of the spinous processes and firmly fixed by bolts passed
through holes drilled in the processes. Because the fracture-dislocation is the result of
rotational violence full stability can be achieved only by fixing at least four spinous
processes, two above and two below the fracture-dislocation. The plates firmly fix the spine
until spontaneous interbody fusion occurs, usually in three months. It must be emphasised

FIG. 29 FiG. 30
Case 1. Figure 29-Appearance after incision through the skin. Note the tear in the fascia and the supraspinous
and intraspinous ligaments. Figure 30-After clearance of the spinous processes the complete tear of the
posterior ligament complex and wide separation of the spinous processes is clearly seen. The deformity is
reduced by manipulation using the powerful forceps seen holding the spinous processes.

FIG. 31 FIG. 32
Case I. Figure 31-The deformity reduced. The tear of the posterior ligament complex can still be seen.
Figure 32-The plates in position. The spine is securely fixed.

that this operation is only indicated in unstable fracture-dislocations associated with paraplegia
where there is some hope of recovery in cord or root function. The results so far as the spine
is concerned are excellent provided the operation is properly performed.

Twenty years ago fractures of the spine were almost entirely the result of accidents
occurring in heavy industry, particularly coal mining, and were therefore geographically
restricted. Now the incidence is almost equal throughout the country, for whereas spinal

VOL. 45 B, NO. 1, FEBRUARY 1963

B81
20 F. W. HOLDSWORTH

FIG. 33 L1.J.

Case 1. Figure 33-Radiograph of the fixed spine. The position is anatomical and the fixation firm. Figure 34-
One year after plating there was complete recovery of the lumbar roots on the left side, but no recovery
on ihe right. Sacral paraplegia was complete. Note free back movements.

injuries in heavy industry are decreasing, those from road accidents are greatly increasing.
We must all be familiar with these serious spinal injuries. This is my excuse for inflicting this
dissertation upon you.

REFERENCES

DAvIS, A. G. (1929): Fractures of the Spine. Journal of Bone and Joint Surgery, 11, 133.
DAvis, A. G. (1938): Tensile Strength of the Anterior Longitudinal Ligament in Relation to Treatment of
132 Crush Fractures of the Spine. Journal of Bone and Joint Surgery, 20, 429.
NicoLL, E. A. (1949): Fractures of the Dorso-lumbar Spine. Journal of Bone and Joint Surgery, 31-B, 376.
ROAF, R. (1960): A Study of the Mechanics of Spinal Injuries. Journal of Bone and Joint Surgeri, 42-B, 810.
WATSON JONES, R. (1931): Manipulative Reduction of Crush Fractures of the Spine. British Medical Journal,
1, 300
WATSON JONES, R. (1934): The Treatment of Fractures and Fracture Dislocations of’ the Spine. Journal of
Bone and Joint Surgery, 16, 30.
WATSON-JONES, R. (1940): Fractures and Other Bone and Joint Injuries. First edition, pp. 201-2 17. Edinburgh:
E. & S. Livingstone.

THE JOURNAL OF BONE AND JOINT SURGERY