|
 |
|||||||||||||||||||
C. ARGENSON,
F. DE PERETTI, A. GHABRIS, P. EUDE, J. LOVET, I. HOVORKA
Department of Orthopaedics and Spinal Surgery - F-06202 Nice, France
| A | I. Anterior compression | II. Comminuted fracture | III. Teardrop fracture |
| B | I. Moderate sprain dislocation |
II. Severe sprain | III. Bilateral fracture |
| C | I. Unifacet fracture (UFF) | II. Fracture separation of the articular pillar (FSAP) |
III. Unilateral dislocation (UD) |
The scheme proposed in this paper follows on from previous classifications devised by Allen [1], Harris [14], and Senegas [24]. All the previous schemes were based upon the assumption that injuries to the lower cervical spine (LCS) are produced by multiple simultaneous forces, with one force vector predominant. In our scheme, we have assumed compression, flexion-extension-distraction, and rotation to be the main force vectors, which occur with equal frequency. Within each of these vector groups, the injuries produced are graded into three subgroups, as a function of the intensity of the trauma-producing force.
The scheme is intended to provide a basis for the practical management of LCS trauma cases. It derives from a study of 255 trauma patients admitted, between 1980 and 1994, to the Department of Orthopaedics and Traumatology of Nice University Hospital. The patients were aged between 16 years and 85 years. The causes of trauma were road accidents (60%), sports accidents, including diving accidents (12%), and falls from a height (28%). There was a total of 306 severe injuries. Ninety-five per cent of the patients underwent surgical treatment. Minor injuries (mainly sprains) were not taken into account in this study. Such injuries occur about six times more frequently than do severe ones; they are almost always treated conservatively. Sixty-three per cent of the patients had neurological lesions; of these, 40% had spinal cord involvement, while 60% had nerve root compromise.
All the patients were investigated with conventional radiography; 148 underwent computed tomography (CT), and 40 magnetic resonance imaging (MRI). The patients in whom neither MRI nor CT was performed had been examined with tomography. The patients' records were analyzed by three orthopaedic surgeons and by a radiologist specializing in CT.
Injuries were classified under the following headings:
A - Compression injuries (33% of cases)
Compression causes mainly bony damage. Depending on the amount of disruption of osseous and neural elements, three patterns may be distinguished:
I - Anterior compression (3%)
II - Comminuted fractures (7%)
III - Teardrop fractures (23%).
Teardrop fractures constitute the transition into the following group:
B - Flexion-extension-distraction injuries (28%)
These injuries involve chiefly the discs and the ligaments. Hyperflexion results in compressive trauma of the anterior column of the spine, and in a distraction injury of the posterior part of that column; while hyperextension will have the opposite effect. Often, these two movements will occur one after the other. This makes it difficult to decide which effects are caused by which movement.
Different degrees of intensity of the main vector force will produce different degrees of severity of LCS lesions:
I - Moderately severe sprains (whiplash injury) are very common. They have not, therefore, been considered in our study of 306 severe injuries, except for the 5% of the cases that had neurological deficits.
II - Severe sprains (14%);
III - Bilateral fracture-dislocations (9%).
C - Rotation injuries (39%)
In this type of injury, we consider rotation to be the dominant vector force. This is why we use the general term "traumatic rotatory displacement" to include the three asymmetrical joint lesions listed below:-
I - Unifacet fractures (20%);
II - Fracture separation of the articular pillar (10%);
III - Unilateral dislocations (9%), which cause the greatest amount of neurological damage.
A - Compression injuries
Compression as the dominant injury mechanism produces very specific anatomical injury patterns; all of these patterns are marked by bone trauma. However, as a result of the great mobility of the C-spine, compression is often associated with a flexion force, which will cause trauma to the discs and ligaments. Different degrees of force will produce different degrees of trauma severity:-
I - Anterior vertebral body compression (wedge fractures)
At the cervical level, this lesion occurs much less frequently than in the thoracic spine: We found it in only 3% of our cases. In these cases, the posterior part of the vertebral body and the posterior ligament complex are left intact. Frequently, the lesions form part of a wider injury pattern; isolated wedge fractures with an intact posterior wall on CT scans are stable lesions.
II - Comminuted fractures (7%)
These fractures result from pure axial compression. This accounts for their low incidence. There are several fracture lines, detaching a number of fragments. The fragments may be pushed back into the spinal canal. As in the lumbar spine, where these lesions affect the most caudal vertebrae, the main site of C-spine comminuted fractures is at C7. The typical radiographic pattern is that of double-headed top ("diabolo"), with a void at the centre of the vertebral body; the anterior portion of the vertebra is beaked forward of the vertebra above, whose anterior inferior margin is embedded in the fractured vertebra. In three of our cases, the spinous process was seen to be displaced on the a.p. view, and there was a unilateral facet lesion. This enabled us to diagnose a "burst-rotation" injury, which we have found to occur at the same rate as that reported by F. Denis at the thoracolumbar level. Whether or not such injuries are unstable depends on the extent to which the posterior wall is pushed back into the spinal canal. It is this retropulsion that produces the neurological complications seen in 50% of the cases. If there is no backward displacement of fragments, cervical burst fractures should not be automatically considered to be unstable; in particular, if the posterior longitudinal ligament (PLL) is intact, it will (as in the lumbar spine) reduce the encroachment by "ligamentotaxis".
III - Teardrop fractures (23%)
Schneider and Kahn [23] were the first to draw attention to the fracture of the anterior inferior corner of the vertebral body, and to assess the neurological implications of this lesion. They described the mechanism as one of acute flexion. Since then, a number of experimental studies have been performed; teardrop fractures have been found in shallow-water diving accidents where the crown of the victim's head hits the bottom of the pool; tackling accidents of American football players, involving head-on axial impact, have been retrospectively analyzed from game films [27]. The evidence thus obtained indicates that the typical C-spine lesion is produced by axial compression that is suddenly exerted on a spine that is flexed to a greater or lesser extent. Schneider and Kahn did not mention the bony, vertebral body lesions shown by tomography and, later, by CT; equally, nothing was said about the damage to discs and ligaments that has been demonstrated by French authors [15, 22].
Teardrop lesions may comprise a number of features (listed here from anterior to posterior):-
° In the vertebral body, a small triangular fragment is detached from the anteroinferior corner of the body; however, the fragment remains attached to the anterior corner of the subjacent disc. Our observations differ from those reported by Torg [27], in that, in our series, the fragment very rarely involved the full height of the anterior border of the vertebral body. The characteristic feature of the teardrop fracture is the detached bone fragment (regardless of size).
° A sagittal fracture line running through the vertebral body, as described by Fuentès [11] was seen in two thirds of our cases. This pattern has been seen more frequently since the advent of CT. We have not found such fracture lines to threaten the neurological outcome. In 1% of the cases, the suprajacent vertebra was also seen to have a sagittal fracture line. Its detection is of the utmost importance: If it is missed, anterior plating with only one midline screw would give poor stability, since (without the surgeon realizing it) the screw would be in the fracture line.
° A lesion of the disc and of the PLL, which would be suggested by the lateral film, on which the posterior part of the vertebral body is seen to be displaced backwards. In 50% of the cases, the lesion was at the level of the disc subjacent to the teardrop fracture (mean displacement: 4.3 mm); in only 2% of the cases did we see it at the level of the suprajacent disc (mean displacement: 1.1 mm); while in 26% of the cases, there was symmetrical but moderate (3 mm) displacement of the posterior walls into the spinal canal, showing that two segments were involved. In 22% of the cases, there was no displacement. This typical backward displacement of the posterior wall in teardrop fractures accounts for the neural damage caused by these lesions (80%).
° Damage to the posterior ligamentous structures produces a number of patterns: Involvement of the facet joint capsules leads to the facets no longer being parallel; while damage to the supra/interspinous ligaments makes the spines gape. We found this lesion in three quarters of our cases, most commonly (in one out of two cases) at the level of the disc lesion; less often (one out of four), it was seen at the higher level, or at both levels. Instead of the ligamentous lesions, there may a fracture of the neural arch (15% of the cases).
In conclusion, the analysis of the different lesions involved in teardrop injuries allows the injury to be classified as a "half-way" lesion, produced by compression associated with a greater or lesser degree of flexion. Where there is much flexion, discoligamentous injuries will predominate; where the compressive force exceeds that of flexion, the pattern will be mainly one of bony trauma. This dual nature of teardrop injuries accounts for the different opinions expressed in the literature as to the stability of such lesions. We think that where there is discoligamentous disruption of the middle and posterior columns, as shown by retropulsion of the posterior wall, displacement of the facet joints, and interspinous gaping, the lesion should be considered to be unstable, and surgical treatment would be mandatory. This goes for the overwhelming majority of teardrop fractures, the most typical of which is undoubtedly that sustained when diving head first into shallow water. In that case, the fracture usually involves the C5 vertebra; and the neurological status may be aggravated by the respiratory complications produced by near-drowning. On the other hand, some chiefly bony teardrop lesions, such as a sagittal fracture line running through the vertebral body, or a neural arch fracture, may be considered to be stable, and should heal with properly conducted conservative treatment.
B - Flexion-extension-distraction injuries (28%)
In our scheme, this category encompasses all anatomical lesions resulting from either forced flexion or forced extension movements, since flexion and extension invariably involve distraction in the column opposite to the movement: If one considers flexion and extension taking place in the sagittal plane, around a transverse axis that goes through the posterior part of the vertebral body [15], it will be seen that any forced flexion will compress the anterior column (vertebral bodies and discs) and distract the posterior structures (supra/interspinous ligaments and facet joint capsules); while extension will have the opposite effect.
The physiological movements of flexion and extension are controlled by the posterior anulus fibrosus and by the PLL, which is broad and strong in the cervical spine. Following other authors' line of human research, we have shown, in monkeys [2], that the PLL acts as a stabilizer in the same way as the cruciate ligaments do in the knee joint. Like the cruciates, the PLL is centrally placed, and does not heal spontaneously. The PLL is, thus, a "middle column", like the one described by Denis in the thoracolumbar spine.
In a recent experimental study, McLain [16] was able to show, in detail, the essential stabilizing function of this "middle column": He measured the elongation of the neural axis after the progressive destabilization of the different columns that make up the spine. The results, summarized in the Table, show that if, in forced flexion, the posterior and the middle column are dissociated, elongation will be twice the value observed when only the posterior column is severed off the spine. In extension, the stability-providing function is even more obvious, since the destabilization of the middle column after dissociation of the anterior column results in a fourfold increase in elongation.
| Elongation of the neural axis as a result of the successive division of the 3 columns | ||
| Divided column | % elongation | |
| Flexion |
PC + MC + AC |
52 110 175 |
| Extension | PC + MC + AC |
13 52 140 |
In the light of these results, three groups, of increasing severity, may be established, as a function of the degree of involvement of the PLL:
° A moderate sprain, or whiplash injury, in which the PLL is intact. The transverse axis remains unaltered. Around this axis, flexion and extension follow each other, in such a way as to make it impossible to identify the specific consequences of either movement. These sprains are very common, which is why it was decided not to include them in this study. The only exception to this rule were the 15 cases that showed no destabilizing bone or discoligamentous lesions, but still had neurological deficits.
° Severe sprains (14%) are defined by the existence of a PLL lesion as a result (usually) of flexion-distraction or (sometimes) of extension-distraction. These two types may be readily distinguished on plain radiographs.
° Bilateral fracture-dislocations (9%), as a result of the progression of the flexion or extension movement: Here, the distinction between the two types is less easy.
I - Moderate sprain, or whiplash injury
The most common cause is sudden extension of the head, during a road traffic accident involving a rear-end collision. The wearing of a seat belt does not prevent the occurrence of this injury. Hyperextension is usually limited by the head restraint; it is, however, usually followed by a greater or lesser degree of flexion. In these very common injuries, the PLL is left intact, so that there is no disc- or ligament-related instability. Anterior compression may produce disc trauma, either in the form of micro-trauma that may lead to disc degeneration or (much less often) in the form of "acute disc herniation", as documented in the literature and observed by ourselves. Equally, posterior distraction may result in microscopic tears in the supra/interspinous ligaments, as seen in cadaver studies [12]. However, as shown by dynamic studies, these partial lesions do not affect stability.
The clinical picture in whiplash injury is very varied, ranging from simple "headaches" to neck pain, paraesthesias in the hands, visual and hearing problems, dizziness, chest pain, and even low back pain. There is evidence from large studies [18] with a long follow-up to show that 50% of the patients will recover within six weeks; others may take four months; while, in some, clinical sequelae may still exist three years after the accident, even where insurance claims have been settled in the meantime.
Within the overall group of "whiplash injury," mention should be made of a particular group of patients who have no demonstrable compressive or destabilizing lesions of bone, discs, or ligaments, and who yet have a spinal sensory and motor deficit. This syndrome was first described by Schneider [23], who called it a central spinal cord injury. The condition is characterized by paralysis, mainly of the upper limbs and the sphincters. It usually results from hyperextension trauma, with a typical history (a fall downstairs on the face) and typical associated facial lesions. This assumption of a hyperextension mechanism agrees well with the findings of Panjabi, Louis, and Penning, whose physiological studies showed that the narrowing of the spinal canal in extension is three times greater than that produced by flexion.
The clinical picture and the course of the condition will vary with patient age: In young subjects, there will be simple transient "neurapraxia", from which the patient will recover within a few minutes to a few hours; at worst, there will be a slight residual spasticity when walking.
In a group of 32 American football players, Torg [26] showed that subjects with developmental spinal stenosis (12.6 mm) had a higher rate of such neurological loss than did a group of control subjects with a normal canal diameter (19 mm). The spinal canal stenosis was assessed using a ratio method (width of canal to width of vertebral body). The ratio was found to be 0.80 in patients with neurological problems, and was taken to be 0.95 in the general population.
In adults and in the elderly, the clinical picture tends to be more complete and more severe [3]; neurological loss may amount to complete quadriplegia. Usually, patients recover from their initial deficits. The deficits start in the lower limbs and ascend to the hands, where paralysis of the interosseous muscles is a typical and often permanent feature. However, improvement is not universally seen, and elderly patients may have deterioration of their respiratory and general status, and eventually die. These subjects, too, will be seen to have narrowing of the spinal canal. In these cases, though, the condition will be an acquired one, resulting from osteophytes on the vertebral bodies and/or the facet joints. Neural damage is caused by anterior contusion of the spinal cord and its blood vessels during sudden hyperextension [3]. Penning (Fig. 2) postulates a pincers mechanism, in which the cord is nipped between the posterior inferior border of the suprajacent vertebra, and the neural arch of the subjacent vertebra. Taylor and Blackwood [25] (Fig. 2) think that the compression occurs at the level of one and the same vertebra. In these classical concepts, the only reason why the vertebral body can tilt is that the disc has been completely sheared off the vertebral body. However, routine MRI studies have failed to corroborate this view. It would appear that the elasticity of the discs and ligaments is such as to cause contusion of a spinal cord that is tightly squeezed in a narrow spinal canal.
Figure 2 - Mechanism of cord lesions in B I sprains
(a) Postulated by Penning (b) Postulated by Taylor
II - Severe sprains (14% of cases)
In severe sprains, there is damage to the PLL, as a result of forced flexion-distraction or extension-distraction, around the transverse axis in the vertebral body.
In flexion
Severe sprains are extremely unstable lesions caused by the failure of the middle column (discs and ligament). The condition can be diagnosed when at least three of the features described by René Louis [15] are seen:-
- anterolisthesis (forward displacement) of the vertebral body, by > 3.5 mm above C4, and > 2.5 mm below that level;
- angulation of the endplates > 10°;
- loss of the parallel alignment of the facet joints;
- uncovering of > 50% of the superior facet of the vertebra below the lesion;
- abnormal distance between the spines. This gape is evidence of damage to the supra/interspinous ligament; instead, there may be a horizontal avulsion fracture of the spinous process, which is of great diagnostic value.
In our series, the mean kyphotic angulation was 15°, and the mean anterolisthesis, 4 mm. In 25% of the cases, radiological evidence was not obtained until several days after the accident, because reflex muscle contraction had initially held the spine straight. "Stress" films taken on Day 6 may, very occasionally, show the lesion. Our treatment policy is single-level anterior interbody fusion. In our series, we were surprised to see that this was the only lesion to affect males and females in equal numbers; one particularly astonishing finding was the high rate of severe sprains in patients over the age of 50. In adults and in the elderly, these sprains may occur after comparatively minor trauma. In these cases, the lesion will most commonly be seen at a level immediately above an arthrotic disc lesion, which had probably limited mobility at this level. The displacement was always perfectly reduced by traction, confirming the acute nature of the disc lesion. Three quarters of our patients with severe sprains had only this anatomical pattern. In the remaining 25%, however, there were other lesions that were invariably found to be due to a flexion force acting on the spine. Thus, in the early days of our work, we had two cases of severe sprain that were missed on the initial films, and were discovered after anterior surgical fixation of a teardrop fracture, in one case, and of a bilateral dislocation in the other case. If this associated pattern is suspected, either MRI should be performed, to look for an associated severe sprain; or an intraoperative dynamic study should be obtained after the fixation of what is thought to be the main lesion, especially if the spinous process of the vertebra above the "main" lesion is fractured.
We only had ten cases of severe sprains with neurological complications (usually nerve root compromise).
In extension
Forced extension, by itself or following forced flexion, causes distraction at the anterior column and compression at the posterior column.
The column and the facet joints are very sturdy structures, which is why there are so few anatomical lesions at this site. Another reason is the control of the backward tilt by the middle column, and its further restriction by the constraint of the travel of the facets, whereas in hyperflexion the facets can freely glide forwards. However, if the hyperextension force continues to be applied, the anterior discoligamentous structures will give, and the PLL will be stripped off, giving rise to a severe sprain in extension. The transverse movement axis may shift backwards, to the point of contact between the suprajacent and the subjacent neural arch. This axis shift will intensify the extension movement (Fig. 3).
Figure 3 - Backward displacement of flexion-extension axis following PLL tear
The rate of these severe sprains in extension (2%) is less than that of severe sprains in flexion (12%); however, these sprains cause major instability, and substantial neurological loss. Sometimes, the lesion can be seen on radiographs, where it shows up in the form of an anterior gaping between the anterior edges of the vertebral bodies that becomes worse on stress views in hyperextension; equally, there may be a small anterior bone fragment from the vertebral body still attached to the disc, while the vertebral body has tilted backwards - what might be termed a "reverse teardrop fracture" (Fig. 4).
Figure 4 - Severe sprain in extension C2-C3 (B II)
MRI has made it possible to detect the typical injury pattern of the anterior soft tissues: haematoma, as well as rupture of ligaments, discs, or even the dural sac.
III - Bilateral fracture-dislocations (9%)
These lesions are caused by the same mechanism as the severe sprains described above; however, the traumatizing force goes on acting after the PLL has torn.
In flexion
These lesions are a progression from a neglected severe sprain; equally, they may occur after violent trauma in hyperflexion. According to Roaf [20], the flexion required to rupture the PLL is such that the vertebral body will always become crushed before the ligament tears. The radiological diagnosis of the injury is straightforward: The vertebral body will be seen to be displaced forward by more than one third of its sagittal width. The routine use of CT has shown that the ends of the joint facets are often broken. This finding justifies the use of the term bilateral fracture-dislocations, rather than bilateral dislocations, for this group of lesions. In our series of 23 patients, spinal cord complications were observed in only half the cases.
In extension
These lesions are rare ( 2% of cases), but should be correctly diagnosed. In front, the ligament and the disc are avulsed by the hyperextension movement; at the back, the two articular pillars will dislocate or fracture.
Bilateral fracture-dislocation in extension (B III) |
 Figure 5b - vertebral body incarceration (CT scan) |
In two of our cases, the cervical spine was in extension on the lateral film (Fig. 5); the facets were dislocated as a result of major spondylosis in one case, and of an incarcerated vertebral body fragment in the other case. These two patients were quadriplegic. In another two cases, the diagnosis of bilateral fracture-dislocation was less easy to confirm, since the spine was in flexion (Fig. 6), making the radiographic pattern resemble that of bilateral fracture-dislocation in flexion. In these two cases, the joint lesion was one of bilateral fracture separation of the articular pillar, rather than of dislocation. One can imagine an initial forced hyperextension movement causing the anterior lesions by distraction, and bilateral posterior articular pillar fractures by pure compression without any rotation; this pattern being followed by spontaneous forward flexion under the weight of the completely destabilized head. Hyperextension as a causative mechanism may be suspected if there are facial skin lesions, severe facet joint damage, multiple spinous process fractures, evidence of posterior disc avulsion on the MRI scans (Fig. 6), and difficulty when reduction of the lesion by traction is attempted. Reduction by traction is, of course, pointless in this type of extension injury: The two articular pillars are isolated by the fracture line, and cannot realign themselves, since they are not attached to any discs or ligaments. This is why, in our two cases, reduction was performed from a posterior approach, which was possible only after removal of a completely incarcerated joint facet. For stabilization, this facet had to be replaced by a Roy-Camille tile plate. This example shows the need for correctly diagnosing this lesion before surgery, in order to choose the appropriate posterior approach straight away.
Fig. 6 Bilateral fracture-dislocation (B III) in extension, with bilateral fracture separation of the articular pillars (subject in flexion) |
 D: At one year: one articular mass has healed; the other was replaced by a plaque. |
C - Rotation injuries (39%)
The dominant vector is rotation. However, as shown in the biomechanical studies performed by Penning and White & Panjabi, this rotatory movement will always be associated with lateral flexion. This produces asymmetrical lesions of the LCS.
The rôle of rotation was stressed by Roaf [20], and confirmed in a number of experimental studies. Recently, Myers [17] showed that while the atlantoaxis was the weakest joint in torsion in the C-spine, it was possible to produce unilateral injury in the lower cervical spine by applying torque directly to the LCS.
In our study, the importance of the rotation vector was confirmed by the observation of a unifacet fracture in a young woman who had been involved in a shipboard accident, when the sudden release of a winch had thrown out rope that had wrapped itself forcefully around her neck. Apart from being half-strangled by the rope, she also suffered major skin lesions around the neck.
Unlike Allen [1] and Harris [14], whose schemes do not make special allowance for injuries caused by a rotatory mechanism, we think that rotation is the only consistently observed trauma-producing vector: It may be associated with flexion or extension, to produce a unifacet fracture (UFF), fracture separation of the articular pillar (FSAP), or unilateral dislocation (UD).
We have grouped these three unilateral lesions under the heading of "rotation injuries". Most of them cannot occur unless, during the initial trauma, the disc has been stressed beyond its elastic limit.
These three anatomical lesions have typical radiographic patterns, which reflect the vertebral rotation underlying the injuries (Fig. 7):
- on the a.p. view: a spinous process is displaced towards the lesion;
- on the lateral view: anterolisthesis by ca. one third of the width of the vertebral body; three-quarter view of all the vertebrae above the disc lesion;
- on the oblique view, uncovertebral "gaping" (Dosch [8]), as evidence of a gap between the vertebral bodies and, hence, loss of ipsilateral vertebral stacking;
- displacement in the transverse plane, on the CT scan.
Figure 7 - Common radiographic pattern of Rotatory Displacements (type C lesions)
Where the radiological patterns listed above are seen, a search should be made, on ascending oblique views, axial scans or sagittal reconstructions of CT scans (spiral CT!), or MRI, for one of the joint lesions described above.
For the group of injuries caused by a predominant rotatory vector, we have used the term "traumatic rotatory displacement" (TRD) [4]. In a similar descriptive study, Dosch [8] uses the term "lateralized anterolisthesis." We find these terms preferable to the term "subluxation" used in the English-language literature. Subluxation denotes a displacement that may occur in a variety of flexion or flexion-rotation injuries.
I - Unifacet fractures (20% of injuries)
The mechanism that causes these injuries is one of flexion-rotation in the direction opposite to the lesion [21]. As we were able to confirm in our experimental study [4], the force is applied at a high velocity. These were the most frequently seen injuries (accounting for 50% of all traumatic rotatory displacements); in 15% of the cases, more than one level was involved. These injuries are frequently associated with other lesions, which is why their diagnosis may be delayed, as it was in 40% of Rorabeck's [21] cases. Late detection is also due to the fact that the radiologically demonstrable displacement often does not show up until several days after the initial accident (25% of our cases). The mean anterolisthesis was 3 mm; the mean kyphotic angulation, 8°. The superior articular process of the vertebra below the forward slipped one was more frequently involved (two thirds of the cases) than was the inferior one of the vertebra above. In three cases, the distal ends of both articular processes were involved; this pattern could be termed "unilateral biarticular fracture."
We performed a detailed analysis of the CT scans, working with a radiologist experienced in spinal trauma. This exercise was undertaken in order to improve our understanding of the bony lesions underlying rotatory displacement. We have described [7] a "triple image" pattern (Fig. 8), with (from in front backwards) a broken facet fragment; an unaffected suprajacent articular process; and a decapitated subjacent articular process. This pattern provides evidence of a fracture of the superior articular process of the subjacent vertebra. Where there is a biarticular fracture, the triple image may become a quadruple one [7].
 Figure 8a - Triple image on a CT scan. Typical pattern of a unifacet fracture (C I) of the superior facet of C6 |
 Figure 8b - |
The CT scans of facet fractures are not easy to analyze; however, this skill will need to be mastered, since CT is rapidly replacing the more familiar lateral and oblique tomograms. Sagittal reconstruction should be used routinely.
Clinical signs of nerve root compression were observed in 64% of the cases; usually, the problem was a transient one. Apart from these "neurological" cases, instability is a function of the existence and the size of a disc lesion. This condition is readily diagnosed where there is clear anterolisthesis; where no forward displacement can be seen, MRI scans will need to analyzed. This analysis is not easy. In the light of the high rate of secondary displacements, we would regard the majority of these fractures as unstable.
II - Fracture separation of the articular pillar (10%)
This injury has been described by Judet and Roy-Camille [22]. Its characteristic feature is a double fracture line that sets an entire articular pillar free. The anterior line passes through the pedicle; the posterior one goes through the junction between the articular mass and the lamina. Once dissociated, the facet will horizontalize. This is the tell-tale radiological pattern of FSAP. In the English-speaking countries, the pattern has been recognized only fairly recently, although Forsyth [9] referred, as early as 1964, to the "horizontal facet". Since the introduction of CT, the lesion has been more frequently diagnosed.
Most authors consider the lesion to be the result of a hyperextension-in-rotation mechanism, which leads to the subjacent articular mass being fractured by the one above, under conditions of axial compression. However, none of the papers in the literature reports the experimental reproduction of FSAP, and the underlying mechanism can only be surmised.
We think that rotation is the only consistently encountered vector, whereas hyperextension is not the only associated injury mechanism in FSAP. In fact, the finding of associated features of flexion-compression injuries has prompted us to recognize this vector pattern as being potentially also involved in FSAP: In three of our patients, FSAP was associated with subjacent teardrop fractures, which are typical of flexion-compression trauma. We may, therefore, assume that hyperflexion or sudden flexion-compression may give rise to FSAP, if they occur in combination with rotation. Under these conditions, the articular mass of the vertebra above, acting like a hammer, will hit the anvil constituted by the articular mass of the vertebra below, detaching it from the rest of the vertebra and allowing it to drift into a horizontal position. However, this sequence of events is predicated upon the lower facet joint being "caught" with its facets apart, i.e. in slight flexion, since in complete extension, the facets are packed tight.
The a.p. film may show lateral displacement of the horizontalized articular mass, or
a fractured lamina. However, the important view is the lateral one: It shows the double outline produced by the lack of superimposition of the two articular processes, with the posterior border of the fractured process sitting posterior to the intact one. CT scans will confirm the diagnosis, by showing the two fracture lines framing the articular pillar.
Anterolisthesis is less pronounced than in the other forms of TRD (mean displacement: 2.75 mm). Fuentes [10] found evidence, in three of his cases, of FSAP without any forward displacement. Anterolisthesis was found by us in 70% of the cases. It is suggestive of a disc lesion, and is seen three times more often at the level of the disc below the articular process injury (mean displacement: 3 mm) than at the level of the one above (mean displacement: 0.5 mm).
Like Marie-Anne [22], we have seen, in two cases, a secondary anterior displacement (Fig. 9), which confirms that the unifacet lesion is associated with middle column involvement. It is this association that makes the lesion so unstable. A careful search with CT allowed us, in one case, to detect a particular form of FSAP in which the articular pillar was detached from the rest of the vertebra without the continuity of the neural arch being broken. These very rare forms of "extra-pediculolaminar" injuries had previously been described by Dosch [8].
Figure 9 - Combined (anterior + posterior) fixation of a severe sprain (B II) in a spondylotic patient
In 40% of the cases, the FSAP was found to be associated with neurological deficits.
III - Unilateral dislocation (9% of lesions)
These injuries are caused by a mechanism of progressive flexion-rotation on the side opposite to the lesion [4]. In our patients, the mean anterolisthesis was found to be 4 mm, while the mean kyphotic angulation was 9°. These values were the highest found in TRD; also, the displacement of the spinous processes on the a.p. film was more pronounced, and the rate of spinal cord complications was higher than that of nerve root complications.
The diagnosis is confirmed, on a strictly lateral view, by the finding of a "bow-tie" pattern, produced by the two superior articular processes of the dislocated vertebra.
On the oblique tomograms, the inferior articular process of the suprajacent vertebra is seen lying in front of the superior articular process of the subjacent vertebra. CT scans clearly show the way in which the inferior facet has gone forward, leaving the superior facet typically "bare". The facet of the inferior process is readily recognizable by its curved posterior border, which differs from the straight posterior border of the superior facet. This investigation will also reveal the following features: normal pattern of the contralateral facet joint, and gaping of the uncovertebral joint on the side of the lesion [8]. Sagittal reconstruction will confirm the diagnosis, and may show a comminuted fracture of the facet.
In five cases, the inferior articular process of the suprajacent vertebra had not moved fully forward, but had remained atop of the superior articular process of the subjacent vertebra. In this "perched" form, which is very unstable, anterolisthesis is slight, and the condition may be missed.
Recently, the existence of disc lesions in this type of injury has been stressed [13, 19]. According to some authors, such lesions may occur in up to 67% of all "dislocating" injuries. The studies cited here also show that sudden deterioration in the patient's neurological status may occur during attempts at closed reduction, as disc fragments become posteriorly displaced and encroach upon the spinal canal. This finding provides yet another argument in favour of our policy of removing the disc from an anterior approach (see remarks below).
In conclusion, we wish to draw attention to the high rate of associated lesions (20%, if the upper cervical spine is included in the consideration) (Fig. 11) shown by this analytical study. Also, we wish to emphasize the value of CT for the detection of all bony lesions (Fig. 12), and of MRI in the search for disc, ligament, and spinal cord trauma.
 |
 |
 |
| Figure 11 - Multi-segment injury: Fracture of dens with anterior displacement, and unifacet fracture of C5. Anterior approach: dens fixed with Vichard plate; fixation and fusion of C5-C6. | ||
 Figure 12a - Multi-segment injury: Teardrop fracture of C6 |
 Figure 12b - Rotatory displacement of C4 was missed (tomogram shows pedicle fracture). Fusion C5-C7 - postoperatively, aggravation of the displacement of this C4 FSAP. |
The choice of treatment is guided by the scheme outlined above, which classifies injuries in subgroups of increasing trauma severity. Type I lesions will be managed conservatively, in the first instance; for Type II lesions, one- or two-level fusion through a single approach is the treatment of choice; while Type III lesions are so unstable as to warrant combined (anterior and posterior) fixation.
In our Department, we use a variety of techniques [5]. Of these, we prefer anterior fixation and fusion after external reduction. This approach does not require the patient to be repositioned, neither is there a need for detaching muscles. Using this technique, we have had an extremely low rate of complications. The patients also benefit in terms of rapid recovery from surgery, without any residual stiffness or neck pain.
 |
A/I
Anterior compression
|
- For A I lesions (anterior vertebral compression), treatment is definitely conservative. The patients are given a cervicothoracic orthosis, which is applied with the patient seated, and includes an occipital and a chin support. The chin support must be shaped in such a way as to allow the patient to open his or her mouth wide. This can be achieved by putting a cork between the patient's teeth during the moulding of the support. The brace goes half-way down the sternum at the front, and half-way down the back. This protects the patient from hyperextension. Before the fitting of the orthosis, reducing traction may be applied for a few hours.
- B I lesions (sprains) These cases, too, are managed with a brace or, more commonly, a simple collar. The only question in this group arises in patients with acute sprains that do not involve destabilizing bony, disc or ligamentous lesions, but where there is neurological loss (5% of patients in our series):-
- transient neurapraxia in young subjects is managed with a foam collar and drug treatment. However, in these patients a search must be made for spinal stenosis. Any patients found to have this condition should be told to refrain from contact sports.
- In more permanent deficits, in older subjects who are not progressing any further in their neurological improvement, one might consider wide anterior decompression and stabilization of the narrow arthrotic spinal canal. However, this kind of surgery is long and bloody. In these elderly, debilitated patients, the technique has not come up to our expectations [3]. Perhaps these cases should be managed with decompression via a posterior approach, with a single, laterally hinged "door" as recommended by Japanese authors, or with a bilaterally hinged "French door". These techniques can be performed much faster, and would therefore be less stressful.
- C I lesions (unifacet fractures)
If there is no compression of nerve roots, the patient may be put in a cervicothoracic orthosis. However, for this treatment to be considered, it must be certain that there are no disc lesions, even in multi-level injuries, that could cause instability. The search for these discoligamentous injuries is made with MRI; unfortunately, this imaging technique is not, as yet, 100% accurate. There is also a risk of secondary displacement, which may damage the nerve root. In such cases, with malunion of the fracture, treatment may be difficult. This is why it is our policy to treat the majority of these cases with single-level anterior fusion. However, if nerve root pain persists after surgery and/or the facet is not properly reduced, a second, posterior approach may be considered. Some authors recommend the exclusive use of such a posterior approach. It should also be remembered that UFF may be associated with another "main" lesion.
 |
A/II
Comminuted fracture
|
The tell-tale feature usually is a discoligamentous lesion. These injuries require surgical stabilization.
- Conservative treatment is an option only in A II lesions (burst fractures). These patients could have traction, followed by the fitting of a cervicothoracic orthosis. However, surgical treatment would be preferable if there is neurological involvement; if the fragment in the spinal canal is not reduced by "ligamentotaxis"; or if an articular process is also damaged. Anterior corpectomy should be performed over the entire height of the comminuted vertebral body. This anterior decompression must be done with great care. In particular, it should be borne in mind that, laterally, a bone fragment may have staunched the bleeding from veins or even from the vertebral artery. The bleeding that results when this fragment is removed may be very difficult to control, even with packing, bone wax, or the use of cement. (There are reports according to which the neurological sequelae of arterial occlusion were not as disastrous as had been anticipated.) Fusion is done over two segments. Comminuted vertebral fractures sometimes occur in severely osteoporotic bone. In such cases, internal fixation should not be performed, since the screws would be unable to obtain a purchase in the bone. Instead, these patients should be managed with an anterior tricortical graft, with protective continuous traction for 21 days, followed by the wearing of a brace.
- Severe sprains (B II lesions) in flexion are treated, in our Department, with single-level anterior fusion, which allows complete removal of the disc and provides complete stabilization, provided that the uncovering of the facet joint has been completely reduced in lordosis. Also, the facets should be anatomically normal, rather than "blunted" by spondylosis, as we have observed in elderly subjects. In such cases, there will be degenerative spondylolisthesis (forward sliding). Therefore, these patients will need combined (anterior and posterior) fixation, with plates or clamps (Fig. 9).
Severe sprains in extension, which may be complicated by neurological problems, with anterior tears in the dura, are managed by us with anterior fusion using a C3-C2 transbody screw or a plate.
- For the management of fracture separation of the articular pillar (C II), we prefer anterior fusion of one level or two levels (depending on the extent of the displacement). This provides good stabilization (Fig. 10); however, the tilt of the articular process is not always perfectly corrected. This is why some surgeons prefer a posterior approach, which permits reduction and plating.
 |
 |
| Figure 10 - Rotatory displacement (C) Fracture separation of the articular pillar (C II) Secondary displacement of the lesion confirmed by CT; "orthopaedic" reduction and anterior fixation and fusion. Vertebral alignment restored; FSAP healed; but incomplete correction of facet horizontalization. No more nerve root signs and symptoms. |
 |
A/III
Teardrop fracture
|
These lesions are very unstable, since the disc and ligaments of the entire motion segment have been damaged. In many of these cases, combined (anterior and posterior) fixation should be considered.
- Teardrop fractures (A III)
Apart from certain rare, predominantly bony lesions that might be treated conservatively, most of the cases in this group should be managed with anterior fixation and fusion after CT-controlled corpectomy. We would always perform two-segment fixation: Since the fracture line in the anterior cortex of the vertebral body extends into the endplate, there would not be sufficient stability to allow single-segment interbody grafting, neither could the screws of the anterior plate get a good purchase. It should also be remembered that, often, the disc above the affected vertebra may also be involved.
- Bilateral fracture-dislocations (B III)
* In the flexion variety, reduction is readily obtained with traction, which may either be progressive, using tongs, or manual as an emergency procedure. Through an anterior approach, the disc may then be completely resected, and stabilization obtained. In a second stage of the procedure, fixation with a plate or hooks is applied through a posterior approach. In cases of "pure" dislocation, where the condition of the facet joints is thought to be correct, this posterior stage may be dispensed with.
* In the extension variety, diagnosed by radiology and confirmed by the fact that the lesion cannot be reduced with traction, a first posterior approach is used to effect reduction. Sometimes, reduction will necessitate the removal of a comminuted articular mass, which should then be replaced by a tile plate (Fig. 6). Bilateral posterior fixation is completed by anterior fusion, which should, if possible, be done under the same anaesthetic.
- Unilateral dislocation (C III)
This lesion, too, may be difficult to reduce. Reduction must be performed gradually, by tong traction using incremental weights (up to 10 or even 15 kg); as an emergency procedure (in patients with neural deficits), it may be applied manually, using the Galiber technique under general anaesthesia. (Large series investigated by other authors, and our own experience, have shown this technique to be perfectly safe.) After reduction, stabilization is carried out using single-segment anterior fixation and fusion, after careful and complete removal of the disc. If, in this severe anatomical lesion, the subjacent vertebra is anteriorly wedged, fusion will need to be over two levels (Fig. 13).
 |
 |
 |
| Figure 13 - Unilateral dislocation (C III) and anterior wedging (A I) - Frankel Grade C. External reduction under anaesthesia. Anterior fixation and fusion. Patient on way to full recovery. | ||
These dislocations will need to be examined with MRI, both before and after surgery, in order to obtain a complete picture of the disc damage. Only if reduction cannot be obtained, especially in patients not seen during the first 48 hours, do we use the posterior approach, which allows reduction under direct vision, using the well-known "tyre-lever" manoeuvre. Following this, bilateral posterior plating is performed. In such cases, postoperative MRI scans are all the more necessary, since, in the light of their findings, an additional anterior procedure may have to be performed.
Some authors have successfully reduced UDs from an anterior approach, using a distractor placed between the endplates, after disc removal. We do not use this technique ourselves, since we are concerned that it may not work or may, actually, damage neural structures.
It should also be borne in mind that vertebral artery damage may be caused by uni- or bilateral dislocations. If, in the first few minutes after the accident, there were any clinical manifestations suggesting reduced cerebral blood flow, the vertebral artery would be routinely investigated prior to surgery.
In conclusion, radiographs and CT scans should be done to detect bony lesions, while MRI should be used in the search for damage to the discs, ligaments, and the cord. These imaging techniques will show whether the injury is predominantly anterior (Group A), middle column (discs and ligaments - Group B), or in the facet joints (Group C). Stability should then be assessed in terms of the three subgroups of increasing severity proposed in this paper.
|
REFERENCES
2 - Argenson C, Dintimille H : Lésions traumatiques expérimentales du rachis chez le singe. Rev Chir Orthop, 1977 ; 63 ; 5 : 430-431. 3 - Argenson C, Frehel M, Lovet J, Griffet J, de Peretti F : Les contusions médullaires cervicales graves sans lésion ostéo-disco-ligamentaire traumatique. Rev Chir Orthop, 1990 ; 76 : 507-518. 4 - Argenson C, Lovet J, Sanouiller JL, de Peretti F : Traumatic rotatory displacement of the lower cervical spine. Spine, 1988 ; 13 : 767-773. 5 - Argenson C, De Peretti F, Boileau P : Chirurgie des Traumatismes du Rchis Cervical - Techniques Chirurgicales E.M.C. Orth. Traumat., 1994 ; 44 : 176 ; 15 p. 6 - Breig A, Turnbull I, Hassler O : Effects of mechanical stress on the spinal cord in cervical spondylosis. A study on fresh cadaver material. J Neurosurg,1966 ; 25 : 45-56. 7 - De Peretti F., Cambas P.M., Eude P., Ghabris A., Aboulker C., Argenson C. : “Triple et quadruple images” au scanner des fractures des apophyses articulaires du Rachis Cervical Inférieur. J. Radiol., 1994, t. 75, n°11, pp. 597-602, Masson, Paris. 8 - Dosch JC : Trauma conventional radiologic study in spine injury. Springer Verlag Ed., 1985. 9 - Forsyth H.F. : Extension Injuries of the Cervical Spine, J. Bone Joint Surg, 1964, 46 A, 1792 - 1796. 10 - Fuentes JM, Benezech J, Lussiez B, Vlahovitch B : La fracture-séparation du massif articulaire du rachis cervical inférieur. Ses rapports avec la fracture dislocation en hyperextension. Rev chir Orthop, 1986 ; 72 : 435-440. 11 - Fuentes JM, Bloncourt J, Vlahovitch B : La tear drop fracture. Neurochirurgie, 1975 ; 29 : 50-58. 12 - Halldor Jonson J., Rauschning W. : Soft tissue Injuries and Fractures of the Cervical Spine, These Uppsala University - Raven Press Edit - 1992. 13 - HarringtonJF, Likavec MJ, Smith AS : Disc herniation in cervical fracture subluxation. J Neurosurg,1991 ; Vol 29 ; 3 : 374-379. 14 - Harris JH Jr, Eideken-Monroe B, Kopaniky DR : A practical classification of acute cervical injuries. Orthop Clin North Am,1986 ; 17 : 15-30. 15 - Louis R : Traumatismes du rachis cervical.1. Entorses et hernies discales. Press. Méd., 1979 ; 8 ; 22 : 1843-1849. 16 - Mc Lain R.F., Aretakis A., Moseley T.A., Ser P., Benson D.R. : Sub-axial cervical dissociation. Anatomic and biomechanical principles of stabilization. Spine, 1994 ; 19 ; 6 : 653-659. 17 - Meyers BS, Mc Elhaney JH, Doherty BJ, Paver JG, Gray L : The role of torsion in cervical spine trauma. Spine,1991 ; 16 : 870-875. 18 - Norris SH, Watt I : The prognosis of neck injuries resulting from rear-end collisions. J Bone Joint Surg, 1983 ; 65B : 608-611. 19 - Rizzolo SJ, Piazza MR, Cotler JM, Balderston RA, Schaeffer D, Flanders A : Intervertebral disk injury complicating cervical spine trauma. Spine, 1991 ; 16 ; 6 ; 187-189. 20 - Roaf R : A study of the mechanics of spinal injuries. J Bone Joint Surg, 1960 ; 42B : 810-823. 21 - Rorabeck CH, Rock MG, Hawkins RJ, Bourne RB : Unilateral facet dislocation of the cervical spine. Spine, 1987 ;12 ;1 : 23-27. 22 - Roy-Camille R et coll : Rachis cervical traumatique non neurologique. 1ères Journées de la Pitié, Masson Ed., 1979, Paris. 23 - Schneider RC, Kahn : The syndrome of acute central cervical spinal cord followed by chronic anterior cervical cord injury (or compression) syndrome. J Bone Joint Surg, 1960 ; 42A : 253-260. 24 - Sénégas J, Vital JM, Barat M, Caille JM, Dabadie Ph : Traumatismes du rachis cervical. Encycl. Med. Chir. (Paris, France), Appareil Locomoteur, 1987, 15825 ; A10 : 9, 21. 25 - Taylor AR, Blackwood W : Paraplegia in hyperextension cervical injuries with normal radiographic appearances. J Bone Joint Surg,1948 ; 30B : 245-248. 26 - Torg JS, Pavlov H, Genuiario SE : Neuropraxia of the cervical spinal cord with transient quadriplegia. J Bone Joint Surg, 1988 ; 68A : 1354-1370. 27 - Torg JS, Pavlov H, O’Neill MJ, Nichols CE, Senett B : The axial load tear drop fracture. Am J Sports Med, 1991 ; 19 : 355-364. |