One‐stage anterior release and reduction with posterior fusion for treatment of irreducible atlantoaxial dislocation

Abstract

Objective:  To analyze the approach and feasibility of one‐stage anterior release and reduction with posterior fusion for irreducible atlantoaxial dislocation.

Methods:  Ten male and 6 female patients, with an average age of 36 years, including 13 patients with old trauma, 2 with rheumatoid disease, and 1 with os odontoideum were studied. Anterior release and reduction was performed in the supine position. The atlas and vertebra dentate were fixed posteriorly and fused by one stage.

Results:  All patients were followed up from 15 to 40 months (mean, 23 months), and all gained anatomic reduction and bone fusion. Six months postoperatively, the Japanese Orthopaedic Association (JOA) score of the 12 patients with cord symptoms had improved from 8.3 preoperatively to 13.9, with a mean improvement of 87.5%.

Conclusion:  Treatment of irreducible atlantoaxial dislocation with one‐stage anterior release and reduction with posterior fusion is a reliable method.

Introduction

Irreducible atlantoaxial dislocation is not a rare disease. So far, it has been reported in the literature that this problem can be managed by internal fixation with anterior trans‐oral release and reduction and posterior atlantoaxial fusion, internal fixation with trans‐oral odontoid resection of posterior atlantoaxial fusion, and the posterior foramen magnum, the posterior arch of atlas decompression and occipito‐cervical fusion [2,4]. Since 2004, we have adopted a trans‐neck anterior release and atlantoaxial reduction, and one‐stage posterior atlantoaxial fusion and internal fixation to treat 16 patients with irreducible atlantoaxial dislocation. All patients achieved satisfactory results.

Materials and methods

General data

There were 16 patients in our study from January 2004 to March 2007, 10 male and 6 female patients, aged from 14 to 52 years, with an average age of 36 years. Thirteen of the lesions were caused by old trauma, two by rheumatoid disease, and one by os odontoideum. The duration of disease ranged from 5 months to 7 years, with an average of 2 years and 7 months. All patients suffered neck pain and limitation of activity, and 12 of them also had associated limb numbness and limitation of movement.

Preoperative preparation

On admission to hospital, immediate continuous skull traction with an initial weight of 3 kg was applied for 2 weeks to all patients with atlantoaxial dislocation. The weight was increased by 0.5 kg every day. Those patients in whom reduction had not been achieved by the time the weight had reached 8 kg were included in this group.

Surgical technique

Atlantoaxial reduction with anterior cervical release

After receiving general anesthesia, the patient was placed in a supine position with a soft pillow under the shoulders, and spinal cord detectors firmly connected. The head and neck were hyperextended and turned towards the left by about 30°. An oblique incision parallel with the mandible was made 2 cm below the angle of the jaw to the medial border of the sternomastoid muscle, and extended along the medial border of this muscle to the level of the fourth cervical vertebrae.

The hypoglossal nerve, which runs parallel to the digastric tendon in the deeper part, was separated and pulled upwards, the carotid sheath was pulled laterally and the retropharyngeal space exposed. The superior laryngeal nerve, which lies deep to the carotid artery, inclines to the center line to the hyoid arcturus and then enters the throat, was located. The genioglossus and hyoid were pulled to the midline, and the superior laryngeal nerve pulled out and down. Then sharp separation was applied to the retropharyngeal tissue space, in order to reach the anterior arch of the atlas and the front of the axis and third cervical vertebra. Along the anterior arch of the atlas, an electric knife was used to resect the anterior longitudinal ligament, the longus colli and the longus capitis, and to remove the contracting and hypertrophied joint capsules of both atlantoaxial joints, and scar tissue around the odontoid process and atlas, or epistropheus. The alar ligaments and odontoid ligaments were resected and then, maintaining hyperextension, reduction by traction with a force of 2 kg and monitoring by a C‐arm machine was attempted. If reduction was not achieved, the weight was gradually increased by 0.5 kg at 10 min intervals until electrophysiological change was observed. Once satisfactory reduction had been achieved, the posterior atlas and vertebra dentate were fixed and fused.

Posterior fusion and internal fixation

Under the protection of a neck collar, the patients were turned over to be monitored by a C‐arm machine. The head frame was adjusted to obtain a satisfactory atlantoaxial position, then traction with a weight of 1 kg was applied to the skull to maintain stability of the head and neck. A posterior median longitudinal incision was made to reveal the posterior arch of the atlas and axis, and tissue peeled back along the posterior arch of atlas to expose its roots. The nerve root of the axis and the venous plexus were pushed downwards to expose the area of extension between the posterior arch of the atlas and its lateral mass. The entry point was at the intersection of a line 18–20 mm from the mid‐point of the tubercle of the atlas and one 2 mm above the lower edge of the posterior arch. An electric drill was used to perforate the screw point of the posterior arch of the atlas. The trajectory was perpendicular to the coronal plane, and tilted headward at about 5° in the sagittal plane, with a depth control of about 25 mm. The same procedure was carried out on the contralateral side. A nerve dissector was used to retract the nerve root of the axis and the venous plexus to reveal the axial laminae, the bilateral lateral mass and the top and inner edge of the pars interarticularis. In order to avoid vertebral artery injury, the mid‐point of the articular process below the epistropheus was chosen as the screw entry point. An electric drill was used to drill through the cortical bone, advancing gradually until it penetrated the pedicle along the top and inner aspect of the interarticularis cortex, with an angle of about 15° between the screw route and sagittal plane, and an angle of about 30° between the screw route and the coronal plane. After tapping the screw route, an appropriate length of screw was inserted, the same procedure was then carried out on the contralateral side. A connecting rod of a suitable length was then selected, pre‐bent into a certain degree of arc, installed, and the nut screwed.

In two cases in this group anterior release and reduction did not produce a satisfactory result. They required a second attempt to remove osteophytes at the atlantoaxial joint, and further clearance of the tissue between the atlantoaxial joints, according to the degree of the previous reduction. A screw was implanted into the lateral mass of the atlas, slightly deeper than the depth to which the axial screw had been implanted. The difference between them depended on how much further forward degree the atlas was. The axial nut was fixed firstly, and then the atlas nut. Thus, through the pulling effect of the screw, the atlas was further reduced. A screw with a diameter of 4.0 mm was selected for all patients. A Vertex screw was used in ten patients and Axis in six (provided by Medtronic, Minneapolis, MN, USA). The cortex of the posterior arch of the atlas, and lamina and spinal process of the axis were removed, and iliac bone with cortical bone on one side, from the posterior superior iliac spine, was trimmed to a suitable size with a forked tail. The superior end was placed on the posterior arch of the atlas, and the inferior end attached to the spinal process of the axis, the mid hollow part being filled with cancellous bone. Fibrin glue was sprayed on to the area of surgery, and the incision closed layer by layer after placement of a drainage pipe. The whole process was performed with monitoring of spinal evoked potential (SEP).

Postoperative treatment

Postoperatively a neck collar was used for fixation. Twenty‐four h later the drainage pipe was removed, with an average drainage volume of 40 ml. For those with neurological symptoms, an intravenous infusion of 10 mg dexamethasone was administered daily, together with 125 ml mannitol, twice per day for three days, and intravenous injection of antibiotics for seven days. After discharge, the cervical collar was kept in place for 3 months.

Follow‐up

Postoperative follow‐up took place every 3 months in the first year and every 6 months in the second year. Radiological imaging was undertaken in all cases one week after the operation, with plain X‐rays, including lateral views, of the cervical spine and 3‐dimensional computerized tomography (3DCT). Comparisons were made between the preoperative and postoperative images to determine the degree of the reduction of the atlantoaxial joint. The pre‐ and post‐operative function of the cervical spinal cord was assessed according to the 17‐point Japanese Orthopaedic Association Scores for cervical myelopathy (JOA‐17). The postoperative improvement rate according to the JOA scores was calculated as follows:

Results

Sixteen patients achieved satisfactory reduction. In one patient, the traction resulted in injury of the superior laryngeal nerve, causing deepening of the voice and difficulty swallowing water, but these symptoms disappeared within 3 months. In another case, the venous plexus was injured while the screw route in the lateral mass of the atlas was being prepared, requiring hemostasis with gelatin sponge compression. By the time another lateral screw route had been drilled, the bleeding had stopped and a screw was then inserted. There was no injury to the vertebral artery or spinal cord in any case.

All patients were followed up from 15 to 40 months, with an average of 23 months, and all cases achieved bony fusion. In the 12 cases with spinal cord symptoms, the preoperative JOA score was 6.1–14.7 points, with an average of 8.3 points; six months after the operation, the postoperative score was 11.6–16.2 points, with an average of 13.9 points. The average improvement rate according to the JOA score was 87.5%. All implanted bone blocks fused without rupture or loosening of internal fixation (1, 2).

Figure 1.

Traumatic old atlantoaxial dislocation. (a) Preoperative X‐ray film showing atlantoaxial dislocation. (b) Intra‐operative X‐ray film showing atlantoaxial joint reduction with anterior reduction monitoring by C‐arm machine. (c) X‐ray film 1 week postoperatively showing atlantoaxial joint reduction and good internal fixation. (d) Three‐dimensional CT film 2 years after surgery showing bony fusion.

Figure 2.

Os odontoideum associated with atlantoaxial dislocation. (a) Three‐dimensional CT scan showing os odontoideum and atlantoaxial dislocation. (b) X‐ray film 3 years after surgery showing good atlantoaxial position and bony fusion. (c) Three ‐dimensional CT scan 3 years after surgery showing bony fusion.

Discussion

After atlantoaxial dislocation, with progression of the condition, forward and downward displacement of the atlas increases. The anterior longitudinal ligament, longus colli and longus capitis muscles, and joint capsules of the lateral mass in the atlantoaxial region develop contracture, and the surface of the lateral mass of the atlantoaxial joint becomes deformed, so that the forward and downward displacement of the atlas is fixed in a dislocated position, thus forming irreversible atlantoaxial dislocation. Two of our cases, who were positive for rheumatoid factor, and had a wide range of cervical hyperplasia, finger joint deformities, subcutaneous nodules, and an increase in C‐reactive protein, were diagnosed as rheumatoid atlantoaxial dislocation.

Surgical treatment of irreducible atlantoaxial dislocation and its advantages and disadvantages

Treating irreducible atlantoaxial dislocation is very difficult, and there are few reports in the literature ¹ . Some previous scholars have adopted a method of posterior decompression and occipito‐cervical fusion, but this operation does not relieve ventral compression from the upper end of the posterior axis and the odontoid process to the medulla oblongata and spinal cord. At the same time, removal of the posterior arch of the atlas, and occipital‐cervical fusion seriously damage the occipito‐cervical complex, resulting in serious loss of function, many complications, and unsatisfactory clinical results. For such patients, most domestic and foreign scholars claim that anterior trans‐oral decompression and removal of the odontoid process or part of the epistropheus should be performed to achieve direct decompression, following which, based on the result thus obtained, combined posterior decompression and fixation may be adopted ² , ³ , ⁴ , ⁵ , ⁶ .

We believe that trans‐oral decompression osteotomy has the following shortcomings ⁷ : (i) in anterior dislocation of the atlas, the dens or axial vertebral body is located deeper, and excision of the structures causing spinal cord compression is very difficult and dangerous; (ii) because most of the bones to be removed adjoin the dura mater, it is easily torn when resecting the deepest bones, and this can result in leakage of cerebrospinal fluid or spinal cord injury; (iii) for anterior dislocation of the atlas caused by an old odontoid process fracture or congenital dislocation of the dens, it is necessary to remove most of the centrum of the epistropheus to obtain thorough decompression, resulting in instability of the atlantoaxial joint. Once most of the axial body has been removed, the reduced bone base makes fixing of the pedicle screw difficult; (iv) in cases with severe anterior dislocation of the atlas, the lateral mass of the atlas slides forward and moves across the articular surface of the epistropheus to reach the front of the axial body, thus creating a severe kyphosis deformity of the atlantoaxial joint. In order to maintain a neutral head position, a compensatory anterior protrusion curvature of the lower cervical vertebrae progressively develops, leading to swan neck deformity. The intervertebral joints of the lower cervical vertebrae are subject to greater than normal stress, and are therefore prone to degeneration and slippage. Trans‐oral decompression osteotomy cannot correct a swan neck deformity. After this decompression operation, the atlantoaxial joint still has a kyphosis deformity, meaning there is a greater distance between the posterior arch of the atlas and that of the axis, which is not conducive to implementation of internal fixation and fusion; (v) other shortcomings such its deeper location, the limited size of the incision, and unavoidable heavy contamination at the incision site, which easily becomes infected, causing retropharyngeal dropsy and even abscess, necessitating long term postoperative nasal feeding, and causing an increase in the probability of patients' trauma and lung infection.

Jain et al. reported that, of 74 cases undergoing trans‐oral anterior decompression, 17 developed more serious spinal cord injury, and 5 had postoperative cerebrospinal fluid leakage ⁸ . Almost all irreducible atlantoaxial dislocations are anterior ones of the atlas, and the anatomical structures affecting reduction include the anterior longitudinal ligament, longus colli and longus capitis muscles and contracted joint capsule of the lateral mass. By resecting these structures, in the majority of cases reduction can be achieved with skull traction and instrument leverage. Reduction of the atlantoaxial joint can completely relieve compression on the spinal cord and restore the normal cervical curvature, thus facilitating posterior fixation by means of the posterior arch of the atlas and axis, and bone fusion to restore stability of the atlantoaxial joint. If the reduction is successful, there is no need to resect the bone adjoining the spinal dura matter at its superior aspect, thus greatly reducing the possibility of spinal cord injury and dural tear. Even if anatomical repositioning cannot be achieved and the odontoid process needs to be removed, the operation is relatively easy when the atlantoaxial joint has been partially repositioned, and when the odontoid process has been reset forward and downward. For these reasons, we believe that the atlantoaxial joint release and reduction approach has a better rationale than oropharyngeal decompression osteotomy.

Experience of the anterior release and reduction

The high anterior cervical pharynx approach can fully expose the joints from the anterior arch of the atlas to the third cervical vertebra and the lateral mass of the atlas and axis, providing enough room for the operation and at the same time avoiding the drawbacks of the trans‐oral approach ⁹ . We believe that attention should be paid to the following points: (i) In the submandibular triangle, the hypoglossal nerve superiorly, the outer carotid sheath and the external branch of the superior laryngeal nerve inferiorly form a frame structure, and retraction of these landmarks makes for a safe operative area; (ii) during surgery, an electrotome is used to resect, along the anterior arch of the atlas, the anterior longitudinal ligament, the longus colli and longus capitis muscles and to remove contractures of the joint capsules of the lateral mass of the atlantoaxial joint. At the same time, any scar tissue between the free odontoid process and centrum of the epistropheus is completely removed. In general, reduction can be obtained through traction. In those patients in whom satisfactory repositioning has not been achieved, further removal of the alar and odontoid ligaments should be performed. For patients with rheumatoid arthritis, in addition to gradually increasing the traction weight, the pannus around the odontoid process should be removed to facilitate better repositioning.

An anterior release and reduction approach was applied to all subjects in this group, reduction was satisfactory in 14 cases, and in the other two cases the 70% reduction initially achieved was further improved by traction of the screw in the posterior fusion, resulting in satisfactory outcomes without any incision complications.

Because of the potential risks, we do not advocate the use of various types of instruments to add leverage in an attempt to achieve reduction. Routine preoperative skull traction is recommended to release the contracture of the ligament and joint capsule to some extent, and to induce more adaptability to traction in the cervical spinal cord.

Posterior fusion fixation

After release and reduction, the atlas still has a spring recoil force which can create a tendency to dislocation. Therefore, posterior atlantoaxial fixation must be performed firmly and reliably, and three‐dimensional stability must be ensured. The commonly used methods of steel wire fixation (such as the Gallie and Brooks), and the laminar clamp fixation methods (such as the Halifax‐Clamps and Apofix), can not meet this requirement. Only with the application of screw‐rod or screw‐plate systems can the repositioning be maintained till the transplanted bones have fused. In practice, the latter two fixation systems not only provide rigid fixation, they also allow correction of unsatisfactory reduction of the atlantoaxial joint ¹⁰ . Posterior correction should focus on: (i) further resection of the osteophytes around the atlantoaxial joint and clearance of the joint space; (ii) careful control of the difference between the height of the atlas lateral mass screw and that of the axis vertebral pedicle screw, and the pre‐bending degree of the fixed plate (fixed rod); (iii) the nut of the axial pedicle screw should be fixed first, and then the nut of the atlas lateral mass screw. With this method, the two cases in this group with unsatisfactory reduction after anterior release achieved satisfactory results.