Transoral atlantoaxial reduction plate internal fixation for the treatment of irreducible atlantoaxial dislocation: a 2‐ to 4‐year follow‐up

Abstract

Objective:  To evaluate the mid‐term outcomes of transoral atlantoaxial reduction plate (TARP) internal fixation for the treatment of irreducible atlantoaxial dislocation.

Methods:  From April 2003 to April 2005, 31 patients with irreducible atlantoaxial dislocation were treated with TARP internal fixation. The average age was 37.9 years (range, 15–69 years). The subjective symptoms, objective signs, and neurological function of the patients were assessed. Radiography and magnetic resonance imaging (MRI) were performed and the results analyzed according to the Symon and Lavender clinical standard, Japanese Orthopaedic Association (JOA) score for spinal cord function and imaging standard for spinal cord decompression.

Results:  Complete or almost complete anatomical reduction was obtained in all 31 patients. No screw‐loosening or atlantoaxial redislocation was found in 29 cases. According to the Symon and Lavender clinical standard, 14 cases had recovered completely, 7 to mild, 6 to moderate, and 4 to severe type by final follow‐up, compared to the preoperative classifications of 4 as moderate, 15 as severe, and 12 as extra severe type. The outcome for 26 patients was evaluated as excellent and in 5 as adequate. The average postoperative improvement in spinal cord function was 73.3% and of decompression of the cervical cord 92.6%. The only complication was loosening of screws in two cases with senile osteoporosis. One case underwent TARP revision surgery and the other posterior occipitocervical internal fixation. Both of them were eventually cured.

Conclusion:  The TARP operation is a good choice for patients with irreducible atlantoaxial dislocation and has valuable clinical application.

Introduction

Patients with irreducible atlantoaxial dislocation generally undergo anterior release, decompression, and posterior atlantoaxial fixation or occipitocervical fixation ¹ . Though the techniques above‐mentioned are effective, they generally demand either a one‐stage operation via two approaches or staged operations. For the one‐stage operation via two approaches, the risk of spinal cord injury is increased by the necessity to turn the patient over during the procedure and inadequate reduction is often achieved, resulting in poor atlanto‐occipital function where occipitocervical fusion occurs. In addition, the operation time is prolonged because of the two different incisions. Staged operations prolong the hospital stay and increase the the patient's suffering because two separate operations are required. To avoid the above‐mentioned disadvantages, we designed transoral atlantoaxial reduction plate (TARP) in 2003 and applied it to patients with irreducible atlantoaxial dislocation. The short‐term effect was excellent ² , ³ . For additional evidence on the mid‐term effect of the TARP operation, 31 patients with complete clinical data were assessed after a follow‐up of 2–4 years.

Materials and methods

Clinical Data

From April 2003 to April 2005, 31 patients (16 male, 15 female) with irreducible atlantoaxial dislocation underwent TARP internal fixation. Among them, first generation plates (Fig. 1) were used in 13 patients and second generation plates (Fig. 2) in 18. The average age was 37.9 years (range, 15–69 years) and the history ranged from 2.5 months to 12 years. Seventeen cases were caused by traumatic fracture‐dislocation, 12 by congenital malformations, and 2 by rheumatoid arthritis. According to the Fielding classification ⁴ , 2 patients were classified as type I, 13 as type II, and 16 as type III.

Figure 1.

The first generation plate.

Figure 2.

The second generation plate with a more anatomically accurate shape.

MRI showed that all cases had ventral spinal cord compression in the atlantoaxial segment. All patients in the study had some signs and symptoms due to upper cervical cord compression. The average Japanese Orthopaedic Association (JOA) score ⁵ for spinal cord function was 11.2 ± 3.9 (range, 5–17). Four patients were classified as moderate, 15 as severe and 12 as extra severe according to the Symon and Lavender clinical standard ⁶ .

Surgical techniques

Exposure and release

After induction of general anesthesia with a nasal tracheal cannula or tracheostomy (four cases), the patient was placed in a supine position with skull traction. After the oral cavity had been cleansed, the face, oral cavity and upper pharynx were thoroughly sterilized with iodophor. The upper pharynx was exposed by a Codman mouth gag (Codman, Raynham, MA, USA) and a 4–6 cm long midline incision made in the posterior wall of the pharynx. The longus capitis and colli longus muscles were detached bilaterally to expose the prevertebral structure of the atlantoaxial joint. Any contractural scar tissue or hyperplastic osteotylus between the atlas and odontoid vertebrae, contractural articular capsule and scar tissue around the lateral mass and articular cartilage were removed. Once sufficient release had been achieved, the atlas would show some sign of loosening and some reduction could be obtained by skull traction in all patients.

Reduction and fixation

• 1A suitable plate was placed on C₁ and two screws anchored in the bilateral lateral masses of C₁ to fix the plate. The atlas and the plate then formed a complex of reduction.
• 2A reduction screw was immobilized in the axis vertebra with its head 2–3 mm higher than the surface of the plate. C₂ and the temporary fixation screw then formed another reduction complex (reduction fulcrum).
• 3With skull traction maintained, the downward‐dislocated atlas was braced upward by bracing the upper arm of the reductor that held the crossbeam of the plate and the inferior arm of the reductor that held the temporary reduction screw (the temporary reduction screw was able to pass the runner of the plate).
• 4The complex of atlas and plate was pushed back by revolving the haft in the superior arm of the reductor until reduction of the atlas was achieved.
• 5For rotational displacement in cases with Fielding classification types I and II, the atlas was reduced by prying the complex of atlas and plate transversally with a detacher after the plate had been fixed to the lateral mass of the atlas (before the temporary screw had been fixed in the axis). A temporary reduction screw was then passed through the plate runner and vertically fixed at the midline of the axis.
• 6C‐arm fluoroscopy was used to confirm that the sagittal diameter of the spinal canal had been restored and the cervical cord sufficiently decompressed.
• 7After the plate had been fixed and locked with the other two screws at C₂, the temporary reduction screw was removed. Once the above procedures had been completed, the atlas and axis were fixed in the reduced position. Sufficient strength of fixation was ensured by the use of two screws in the atlas leaning laterally and two screws in the odontoid vertebra leaning medially.

Bone graft

Autogenous iliac crest bone was harvested and, through the plate window, used to fill the joint space between C₁ and C₂ bilaterally.

Incision closure

The plate was covered with paravertebral muscle and the muscular layer and mucous membrane of the upper pharynx sutured.

Evaluation indexes at follow‐up

Clinical index

According to the Symon and Lavender clinical standard, the patients' functional states were classified into four grades: (i) mild, only slight dysfunction, normal work; (ii) moderate, obvious dysfunction and partial work; (iii) severe, unable to work and walk only indoors; and (iv) extra severe, unable to get out of bed, stand or walk. Improvement by one grade was defined as “adequate” and by two grades as “excellent”. Evaluation of neurologic function was based on the JOA standard (17 score method).

Radiographic imaging

X‐ray films and MRI images were taken of all patients to measure the extent of decompression ⁷ .

Complications

Any plate loosening, redislocation, sensation of a foreign body in the pharynx, difficulty in swallowing and dysfunction of cervical movement were assessed at follow‐up.

Statistical analysis

Statistical analysis was performed by the Department of Medical Statistics, Southern Medical University (Guangzhou, China) using the SPSS 13.0 software package. Student's paired t‐test was used for statistical comparison of the data, a P‐value of less than 0.05 being considered to indicate statistical significance.

Results

All the operations were performed well. The operative time was 2.5–4 h (average, 3.2 h) and average blood loss less than 100 ml.

Symptomatic outcome

The clinical symptoms of the 31 patients, which included neck pain, loss of limb sensation and muscle weakness all improved by varying degrees. According to the Symon and Lavender clinical standard, 14 patients recovered completely, 7 were classified as mild, 6 as moderate and 4 as severe by the final follow‐up. The outcome for 26 patients was evaluated as excellent and for 5 as adequate.

Muscle strength and limb sensation recovered by varying degrees and pathological reflexes disappeared in 14 patients (45.2%). The average JOA score was 15.5 ± 1.4 (range, 13–17) by final follow‐up, which was higher than preoperatively 11.2 ± 3.9 (range, 5–17) (t, 8.01, P, 0.000). According to the equation: improvement in neurological function = (postoperative JOA score–preoperative JOA score)/(17‐ preoperative JOA score), the average improvement in neurological function was 73.3% (range, 40–100%, only one patient was 40% and no others scored under 50%), which showed that the therapeutic effect was satisfactory.

Outcome according to radiological assessment

Radiographic outcome

Radiographs were taken preoperatively, immediately postoperatively, 3 months postoperatively and at the final follow‐up. Radiographs 1 week after operation showed that the atlantoaxial joint had achieved complete or almost complete anatomical reduction in 31 patients. The atlantoaxial joints had achieved fusion in 29 patients at 3 months after surgery. Plate loosening or redislocation occurred in two patients 2 months after surgery.

MRI outcome

MRI was performed before and after surgery to assess the degree of cord compression in the atlantoaxial segment. MRI showed that the improvement in spinal cord decompression was 88.3% immediately after surgery and 92.6% by final follow‐up. The increased signal intensity found in the spinal cord before surgery had reverted to normal by final follow‐up in 16 patients, the area of increased signal intensity had reduced or the signal intensity itself had decreased in 9 patients, and it remained unchanged in 6.

Imaging results for one typical case are shown in Fig. 3.

Figure 3.

A 32‐year man with irreducible atlantoaxial dislocation who had been injured in a traffic accident underwent TARP surgery. (A,B) Preoperative anteroposterior and lateral radiographs show dens fracture and atlantoaxial dislocation. (C) Preoperative MRI shows compression of the cervical spinal cord in the atlantoaxial segment. (D,E) Anteroposterior and lateral radiographs one week after surgery show satisfactory reduction and fixation of the atlantoaxial vertebrae. (F) MRI one week after surgery show sufficient decompression of the spinal cord in the atlantoaxial segment. (G,H) Radiographs 25 months after surgery show fusion at the atlantoaxial joint with no loosening of the plate. (I) MRI 25 months after operation shows no compression of the cervical cord.

Complications

Plate loosening or redislocation occurred in two patients, who also complained of a sensation of a foreign body in the pharynx due to screw loosening. These two patients were women over 65 years old with osteoporosis in whom first generation plates with unicortical screws fixation of C₂ had been used. After revision surgery with TARP utilizing different sized bicortical screws for fixation of C₂, the atlantoaxial joint of one case fused solidly without redislocation. Cervico‐occipital fusion was performed in the other patient because of severe osteoporosis.

For all 31 cases followed up, there was significant limitation of cervical rotation and slight restriction of flexion and extension in the early postoperative period. There was no obvious dysfunction after a minimum of two years.

No other complications, such as neurovascular injury or infection, occurred.

Discussion

The preliminary clinical effects of TARP surgery have been reported to be satisfactory ² , ³ . Based on a 2–4 year follow‐up of 31 patients who underwent TARP surgery, the principle of reduction, mid‐term effects, indications and contraindications, complications, and the prospects for TARP are discussed as follows.

Reduction principle and advantages of TARP

The purpose of atlantoaxial reduction is to decompress the spinal cord. Combined with use of the atlantoaxial reduction instrument (ZL2003 10117408.9, Wego Ortho, Weihai, Shandong, China), TARP can achieve reduction of the atlantoaxial joint after C₁‐C₂ joint release and cord decompression. During the reduction procedure, the quadrate notch on one arm of the reduction instrument engages the transverse portion of the TARP in the midline, while the crescent‐shaped notch on the other arm of the instrument engages the reduction screw. Closure of the handgrips imparts a local distraction force between C₁ and C₂. This maneuver is observed fluoroscopically to determine if and when suitable separation of the joint surfaces will permit the final phase of the reduction. During this maneuver the plate, together with C₁, will ascend through the open slot in the inferior portion of the plate based on the C₂ reduction screw. When the nut on the upper arm of the reduction instrument is turned, a posterior reduction force is applied to the upper portion of the plate, displacing C₁ posteriorly with respect to C₂. The progress of the reduction is followed fluoroscopically to determine when a satisfactory position has been achieved. The mechanical lever‐effects of brace and spiral‐compression play an important role in the reduction procedure, superior to the simple one‐dimensional reduction achieved by leverage. The successful achievement of complete or almost complete anatomical reduction in 31 cases shows the efficacy of this one‐stage approach to the surgical treatment of irreducible anterior atlantoaxial dislocation with spinal cord compression. It also has the advantages of reducing the risk of spinal injury during surgery and of subsequent dysfunction of the occipitocervical joints ⁸ .

TARP has another advantage in the reduction of C_1,2 rotational displacement, in that it solves the problem of C₁ resilience after simple reduction by leverage. By virtue of the lever‐effect and slot‐screw‐effect, rotational dislocation of C_1,2 will be corrected by the TARP procedure.

Intraoperative cervical canal fluoroscopy and postoperative MR scan demonstrated that restoration of the normal configuration of the upper cervical spine and brain stem‐spinal cord angle was achieved, and that ventral compression of the spinal cord was adequately relieved. Accordingly, in most cases there was no need for resection and decompression to the anterior arch of the atlas and the dens. In the first few cases, we tried to resect the anterior arch of C₁ and odontoid process of C₂ prior to attempting reduction. Although these procedures were carried out successfully, they made the operation more complicated and dangerous because of the associated high risk of injury to the spinal cord or dura matter, which can cause paralysis, intracranial infection, and even death. When we performed the reduction procedure without resection of the anterior arch of C₁ and the dens, we found that the effect was exactly the same. Once the compressive factor (the dens) had been reduced, compression of the spinal cord was relieved. As a result, it was unnecessary to also resect the anterior bony structure of C_1,2. The imaging and clinical results revealed that traditional ideas about the aggravation of spinal cord injury and compromised blood supply through the vertebrobasilar artery resulting from intraoperative reduction in the presence of scarring and adhesions in the soft tissue and intraspinal canal in patients with old atlantoaxial dislocation should be re‐recognized. Nevertheless, it is important to implement intraoperative spinal cord monitoring and complete release before reduction to ensure the safety of the reduction maneuver.

The mid‐term effect

After the preliminary results of the TARP operation were reported, although there was some support, some reservations and doubts about the small size of the sample and the fact that the procedure goes against some traditional ideas were also expressed. Now that 31 cases have been followed up for 2–4 years, not only have the preliminary clinical claims been adequately affirmed, but also gradual improvement in functions such as neurologic function, range of cervical movement and ability to work have become apparent. These improvements are due to restoration of the normal configuration of the atlantoaxial joint, decompression of the spinal cord and subsequent rehabilitative functional training. Along with increasing case numbers, with greater experience and a more proficient surgical technique, the surgical procedure has been simplified and the operation time (only 1.5–2 h) shortened. This has been helpful in increasing the curative effect and reducing, or even eliminating, the occurrence of complications. Mid‐term results, such as the successful rate of instant reduction of C_1,2 now being 100%, fewer complications and an improvement in the rate of spinal cord decompression to 92.6% show that the clinical application of TARP is fully warranted. The sensation of a foreign body during swallowing and dysphagia were not found in 29 of the follow‐up patients, which proves that the subjective supposition of some surgeons was incorrect.

Indications and contraindications

Owing to the characteristics of the technique and the clinical effects of TARP surgery, the following indications and contraindications are definitely proposed.

Indications: Fusion with TARP as described here is mainly indicated for patients with irreducible anterior atlantoaxial dislocation, especially for that caused by congenital basilar region malformation. Fusion with TARP is also well suited to cases with reducible anterior atlantoaxial dislocation or C_1,2 instability with myelopathy resulting from resection of the anterior arch of atlas and defects in the posterior stable structure. Therefore, fusion with TARP has wider indications than conventional methods.

Contraindications: (i) as with general surgery, inflammation in the oral cavity, such as periodontitis and tonsillitis, is considered to be an absolute contraindication; (ii) senile osteoporosis is a relative contraindication, fusion with TARP should be carefully considered; and (iii) patients with difficulty in opening their mouths for various reasons are also not suitable for TARP surgery.

Prevention of complications

There were two cases of atlantoaxial redislocation in the earlier period of plate application because of loosening of axis screws, some of the reasons being as follows:

• 1A phenomenon to do with suspension. In other words, the gradient between the surface of the arch/ lateral mass of the atlas and front of the axis body meant that there was an interspace between them which made it difficult to attach the straight plate securely to the surface of axis body. This eventually prejudiced the stability of the axis screw.
• 2The anti‐pull‐out strength of the axis screw was relatively low, because the path of the screw in the axis body was shorter than that in the lateral mass of the atlas.
• 3Both of the patients had osteoporosis.

With the aim of addressing the above‐mentioned problems, a second generation plate with a more anatomically accurate shape was designed for better attachment. Adoption of bicortical fixation of the axis resulted in enhancement of the anti‐pull‐out strength of the axis screw. In addition, elderly patients were not chosen for this operation. Therefore, from then on no screw loosening or redislocation occurred.

The possibility of infection in patients with fusion by TARP concerns some surgeons. We think that postoperative infection can be avoided by observing the above indications and contraindications, changing the contaminated oral operative field to an aseptic field, performing careful intraoperative dissection to avoid damage to the dura matter, paying careful attention to the oral cavity and airway postoperatively, enhancing anti‐infection and systemic support therapy, and so on. Our viewpoint is richly justified by the fact that there was no infection in this study.

Disclosure

The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.