Abstract
We evaluated the results for fifteen patients in whom symptomatic osteoarthrosis of the atlanto-axial joints had been treated with an atlanto-axial or occipitocervical arthrodesis between 1973 and 1990. Thirteen patients had long-term follow-up (average duration, seven years and two months; range, four years and two months to nineteen years and two months). The two remaining patients had died: one, four days postoperatively, from a cardiopulmonary arrest, and the other, one year and eight months postoperatively, from complications related to bladder cancer.Preoperatively, all fifteen patients reported pain in the occipitocervical region that increased with any attempt at rotation of the neck and was unresponsive to immobilization with a collar and to analgesics. The average duration of the symptoms before the arthrodesis was three years. One patient had acute quadriparesis. All patients had radiographic evidence of osteoarthrosis involving the lateral atlanto-axial articulations. Four patients had atlanto-axial instability with an average of five millimeters (range, three to ten millimeters) of motion at the anterior atlanto-odontoid interval. Six patients had an associated spontaneous subaxial fusion, which was secondary to osteoarthrosis in five; three of the five also had atlanto-axial instability. Fourteen patients were managed with a posterior arthrodesis and one, with an anterior transoral arthrodesis. The procedures were performed to relieve pain, to stabilize the atlanto-axial joints, and to restore neurological function.Of the fourteen patients who were followed, thirteen had a solid fusion and one had a stable pseudarthrosis. The patient who had quadriparesis recovered. At the latest follow-up evaluation, thirteen patients had an excellent result and one had a fair result as determined with use of a modification of the criteria of Robinson et al. There were no poor results. Atlanto-axial arthrodesis can effectively relieve occipitocervical pain and correct atlanto-axial instability secondary to osteoarthrosis.
Osteoarthrosis of the atlanto-axial joints unrelated to trauma is a rare cause of pain in the occipitocervical region and an even more uncommon cause of atlanto-axial instability. Osteoarthrosis of the atlanto-axial articulations has only recently been described in the literature13. We are aware of no long-term follow-up studies of patients who were managed with arthrodesis for symptoms related to osteoarthrosis of the atlanto-axial joints and, to our knowledge, none of the previously reported cases of operatively treated atlanto-axial osteoarthrosis involved instability of the atlanto-axial joints6,12,14,18,21. The purpose of the current paper is to report the long-term results of arthrodesis in a series of patients who had occipitocervical pain secondary to osteoarthrosis, with some having atlanto-axial instability as well.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Read in part at the Annual Meeting of The American Academy of Orthopaedic Surgeons, Orlando, Florida, February 16, 1995.
‡Department of Orthopaedic Surgery, Loyola University Medical Center, 2160 South 1st Avenue, Maywood, Illinois 60153.
§Department of Orthopaedic Surgery, Campbell Clinic, 869 Madison Avenue, Memphis, Tennessee 38104.
¶Department of Orthopaedic Surgery, Case Western Reserve University, 11100 Euclid Avenue, Cleveland, Ohio 44106. Please address requests for reprints to Dr. Bohlman.
Fifteen patients who had pain in the occipitocervical region and osteoarthrosis of the atlanto-axial joints were managed by the senior one of us (H. H. B.) with an arthrodesis at either the University Hospitals or the Veterans Administration Medical Center in Cleveland between 1973 and 1990 (Table I). Of these fifteen patients, only thirteen had long-term follow-up, as one patient died four days postoperatively from a cardiopulmonary arrest and one died after one year and eight months of follow-up from complications related to bladder cancer. (Data for these two patients are included for discussion.)
The thirteen living patients were followed for four years and two months to nineteen years and two months (average, seven years and two months). All patients were examined preoperatively and followed by the senior one of us. Another one of us (A. J. G.) independently performed the latest follow-up examination for ten patients. Two patients could not return for the most recent evaluation, and the physical and radiographic examination was performed by a physician in their community. Radiographs made before the arthrodesis, immediately postoperatively, and at the most recent follow-up evaluation were available for all thirteen patients.
The primary indication for the arthrodesis was pain that had not responded to non-operative treatment in fourteen patients and pain that was associated with neurological compromise in one patient. Non-operative treatment consisted of analgesics and immobilization with a collar. Of the fifteen patients, four had atlanto-axial instability with at least three millimeters of anterior subluxation in addition to the pain.
Preoperatively, all fifteen patients had occipitocervical pain, which was well localized to the base of the occiput and the most cephalad portion of the posterior aspect of the neck. The pain was predominantly on the side of the osteoarthrotic joint in the eleven patients who had unilateral disease, and it increased with rotation of the head. The pain was not affected by flexion or extension of the neck, as would be expected in association with degenerative changes of the caudad part of the cervical spine. Two patients also had pain in the shoulder. One patient had severe pain in the posterior aspect of the oropharanx that increased with rotation of the head.
The average age of the fifteen patients was seventy-two years (range, fifty-four to eighty-six years) at the time of the arthrodesis. There were five men and ten women. Before the arthrodesis, the average duration of the symptoms (excluding the duration for the patient who had quadriparesis and for one patient for whom the duration was not known) was three years (range, nine months to ten years). Fourteen patients had been referred to us by their primary-care provider after a trial of non-operative treatment had failed.
One patient (Case 10), who had a very unstable atlanto-axial dislocation, had acute spastic quadriparesis when she was first seen by us three months after the development of the occipitocervical pain5. The other fourteen patients were neurologically intact.
Anteroposterior, lateral, and oblique radiographs of the cervical spine were made for all patients (Fig. 1-A). The diagnosis was not confirmed until an open-mouth radiograph of the odontoid process and the atlanto-axial joints had been made (Fig. 1-B). Anteroposterior and lateral tomograms were made to assess the atlanto-axial joints. Osteoarthrosis of the articulations of the lateral masses between the atlas and the axis is characterized by narrowing of the joint space, subchondral sclerosis, peripheral osteophyte formation, and erosions of the joint (Fig. 1-C). We did not make bone scans for diagnostic evaluation after the diagnosis of atlanto-axial osteoarthrosis had been confirmed with the open-mouth radiograph. Two patients had a bone scan before being seen at our institution; the scans revealed increased uptake at the osteoarthrotic joint.
For all but one patient, lateral radiographs were made with the cervical spine in flexion and extension to determine the presence of atlanto-axial instability and to measure the posterior atlanto-odontoid interval. In the exceptional patient (Case 10, Figs. 2-A and 2-B), who had an atlanto-axial dislocation and quadriparesis, the initial posterior atlanto-odontoid interval was measured on the lateral radiograph that had been made with the spine in the neutral position when she was admitted to the hospital. The posterior atlanto-odontoid interval was also measured postoperatively and at the latest follow-up evaluation.
Myelography with computerized axial tomography was performed preoperatively for two patients who had atlanto-axial instability (Fig. 2-C), one patient who had subaxial instability, and one patient who had a lytic lesion of the odontoid process. Magnetic resonance imaging was performed for two patients who had atlanto-axial instability.
The operative techniques that were used for the arthrodesis included the transoral approach as described by Southwick and Robinson, an occiput-to-axis technique as described by Wertheim and Bohlman, and a modification of the technique described by Gallie. In our modification of the Gallie technique, a 20-gauge wire is looped and then passed under the posterior arch of the atlas, and the free ends are then passed through the loop, which is tied snugly onto the arch of the atlas. A second 20-gauge wire is passed through and then around the inferior border of the spinous process of the axis. Two corticocancellous iliac-crest strut grafts are then secured to either side of the spine as the atlanto-axial wires are passed through the grafts and tightened over them. Additional cancellous bone graft is placed in the midline.
Lateral radiographs made with the cervical spine in flexion and extension and open-mouth radiographs of the odontoid process and the atlanto-axial joints were used to assess the fusion at the latest follow-up examination. Patients who had less than one millimeter of motion between the lamina of the atlas and the lamina of the axis were deemed to have a solid fusion.
Evaluation of the results was based on the status of the fusion, the degree of pain relief, the neurological status, and the activity status, with use of a modification of the criteria of Robinson et al.3,16. A result was considered excellent if the patient had no pain and was able to return to normal activities of daily living; good if there was only slight discomfort, occasional use of non-narcotic pain medication, and slight reduction in activity; fair if there was mild or moderate pain necessitating immobilization in a soft collar, use of non-narcotic pain medication, and limitation of activity; and poor if there was no relief of pain despite use of narcotic medication and there was severe limitation of activity.
Radiographic Results
Eleven of the fifteen patients had unilateral osteoarthrosis of the atlanto-axial lateral masses: seven, on the right and four, on the left. The four remaining patients had bilateral involvement of the lateral mass articulations.
Four patients had atlanto-axial instability. For three of them, lateral radiographs made with active flexion and extension of the cervical spine revealed three to ten millimeters of motion at the anterior atlanto-odontoid interval and a posterior atlanto-odontoid interval of ten to twenty millimeters (Table II). As mentioned, the patient who had acute quadriparesis did not have such radiographs made because of the acute dislocation; the preoperative posterior atlanto-odontoid interval was nine millimeters on the initial lateral radiograph of the cervical spine in flexion.
In six patients, the second cervical vertebra was seen to have spontaneously fused to the third cervical vertebra at the time of presentation. One of these patients had a congenital fusion with small vertebral bodies at these levels. The other five patients had otherwise normal morphological characteristics of the vertebral bodies. Three of the six patients had atlanto-axial instability in association with the spontaneous subaxial fusion.
Clinical Results
Twelve patients (Cases 1 through 4, 7, 8, 9, and 11 through 15) were managed with a posterior atlanto-axial arthrodesis for occipitocervical pain, and three of them also had atlanto-axial instability. Another patient (Case 10) was managed with an arthrodesis from the occiput to the axis. This patient had quadriparesis secondary to a non-traumatic atlanto-axial dislocation, and we believed that greater stability could be achieved with an arthrodesis to the occiput. An additional patient (Case 5) had a posterior atlanto-axial arthrodesis with extension to the fourth cervical level because of degenerative spondylolisthesis between the third and fourth cervical levels. The remaining patient (Case 6) had a lytic lesion of the odontoid process in addition to occipitocervical pain. Therefore, a transoral biopsy and an anterior atlanto-axial arthrodesis with autogenous iliac-crest bone graft was performed. Analysis of the histological section of the lesion revealed degenerative fibrous tissue.
Of the fifteen patients, thirteen were managed postoperatively with a two-post rigid cervical orthosis and two, both of whom had atlanto-axial instability, were managed with a halo vest. The immobilization was continued for eight to twelve weeks, until the fusion was determined to be solid on radiographs.
Of the thirteen patients who had long-term follow-up, twelve had a solid fusion (Fig. 1-D). One patient (Case 8) had a pseudarthrosis, as evidenced by the lack of a solid fusion mass and the finding of broken wires on radiographs, and she had continued discomfort in the occipitocervical region at six years and three months after the arthrodesis. No atlanto-axial instability was noted on lateral radiographs made with the spine in flexion and extension preoperatively and at the latest follow-up examination. Additional evaluation of the cervical spine or reoperation for the pseudarthrosis was not allowed by the patient or her family because of her deteriorating general health. She wore a soft cervical collar and was using acetaminophen for relief of pain at the latest follow-up evaluation.
One patient (Case 1) had cerebellar degeneration and was able to walk about the house only both preoperatively and at the one-year follow-up evaluation. At the latest (ten-year) follow-up evaluation, the underlying mobility disorder had progressed and the patient was restricted to transferring from a bed to a chair. He remained pain-free, with a solid fusion. The other eleven patients who had had normal neurological function preoperatively and long-term follow-up remained neurologically intact. The patient who had quadriparesis (Case 10) recovered all neurological function except that paresthesias in the fifth and sixth cervical dermatomes persisted bilaterally.
The patient who died after one year and eight months of follow-up (Case 7) had a solid fusion and normal neurological findings at the last examination.
Thus, of the fourteen patients who survived beyond the perioperative period, of whom thirteen had more than four years of follow-up, thirteen had an excellent result and one had a fair result. Each patient who had a solid fusion had an excellent result. The patient who had a pseudarthrosis had the only fair result.
Complications
One patient (Case 12) who had atlanto-axial instability died from a cardiopulmonary arrest four days after the arthrodesis; he had not had any neurological change during this brief postoperative period. Examination of the cervical cord at the time of the autopsy revealed no evidence of acute compression. The wire fixation from the atlas to the axis was intact, and the atlanto-odontoid interval was stable.
Another patient (Case 1) who had atlanto-axial instability lost the postoperative reduction and had a decrease in the posterior atlanto-odontoid interval, from nineteen millimeters postoperatively to thirteen millimeters at the time of union, but he had no neurological compromise or occipitocervical pain (Table II). He had been managed with a two-post cervical orthosis postoperatively. The other two surviving patients (Cases 10 and 15) who had atlanto-axial instability had maintenance of the posterior atlanto-odontoid interval postoperatively until union; they had been managed with a halo vest postoperatively. We now use a halo vest for all patients who have atlanto-axial instability and are managed with our modification of the Gallie technique of arthrodesis.
One patient reported persistent discomfort at the posterior iliac-crest-graft donor site.
Degenerative osteoarthrosis of the subaxial cervical spine is common in elderly patients2,7. Subaxial osteoarthrosis is typically characterized by neck, shoulder, and arm pain rather than occipitocervical pain2,7. Osteoarthrosis of the atlanto-axial joints may be overlooked when the patient has occipitocervical pain associated with degenerative changes in the subaxial spine and the physician fails to obtain an open-mouth radiograph of the odontoid process and the atlanto-axial joints. The current report addresses the operative treatment of this condition.
Halla and Hardin reported a 4 per cent prevalence of osteoarthrosis of the atlanto-axial lateral mass articulations in 705 consecutive outpatients who had peripheral osteoarthrosis or degenerative joint disease of the spine. Although these authors did not intend to determine the frequency of asymptomatic atlanto-axial osteoarthrosis, they noted that asymptomatic atlanto-axial osteoarthrotic changes were not present in patients who had symptomatic peripheral or subaxial osteoarthrosis. Most of their twenty-seven patients who had atlanto-axial osteoarthrosis had osteoarthrotic changes at sites other than the cervical spine. Harata et al. described thirty-one patients who had osteoarthrosis of the atlanto-axial joints. All but five patients (who had isolated osteoarthrosis of the atlanto-odontoid articulation) had involvement of the lateral masses. Their only patient who had been managed operatively had had a transoral atlanto-axial arthrodesis, which led to a marked decrease in the symptoms.
Harata et al. as well as Halla and Hardin reported on only one patient each (neither of whom had had atlanto-axial instability) who had been managed with an arthrodesis. Both patients had relief of the symptoms. The remaining patients in these series were managed with traction, trigger-point injections, analgesics, or non-steroidal anti-inflammatory medications, with varying degrees of pain relief. Halla and Hardin concluded that there is no apparent specific therapy for osteoarthrosis of the atlanto-axial facet joints.
To our knowledge, Whitesides and McDonald were the first authors to report successful treatment with arthrodesis for degenerative osteoarthrosis of the atlanto-axial lateral masses unrelated to trauma or rheumatoid arthritis. Ehni and Benner managed three patients operatively for unilateral atlanto-axial osteoarthrosis. Two of their patients had a posterior rhizotomy at the second cervical level, and one had a rhizotomy and an atlanto-axial arthrodesis; all three patients had satisfactory relief of pain.
Joseph and Kumar performed an arthrodesis with use of the Gallie technique on four patients, all of whom had a history of carrying heavy loads at work. Each patient had a solid fusion with total relief of the symptoms two years postoperatively. Star et al. reported on five patients who had been managed with an arthrodesis for unilateral atlanto-axial osteoarthrosis. Two of their patients had a Brooks arthrodesis and three, a Gallie arthrodesis. One of the patients who had a Gallie procedure needed a reoperation for a non-union and eventually had a solid fusion and a good result. The remaining four patients had an uneventful recovery and, at an average of twenty-two months postoperatively, each had an excellent result.
Many causes of atlanto-axial instability other than osteoarthrosis have been reported in the literature; these causes include rheumatoid arthritis, ankylosing spondylitis, chondrodysplasia, tumor, and trauma5,8,11,15,20. To our knowledge, the series reported by Halla and Hardin is the only one that included patients who had atlanto-axial instability secondary to osteoarthrosis. However, the only patient in their series who was managed operatively did not have atlanto-axial instability.
Fielding et al. found that the stability of the atlanto-axial joints depends greatly on the ligamentous structures8,9. When the anterior atlanto-odontoid interval is more than three millimeters, there is disruption of the transverse ligament9. As the anterior atlanto-odontoid interval increases further, additional ligamentous damage occurs9. We propose that, as osteoarthrosis of the lateral mass articulations progresses, the synovitis gradually involves the ligamentous structures, weakens them or their attachments, and renders them prone to rupture (Fig. 3). It is plausible that a torn transverse ligament could lead to atlanto-axial instability followed by osteoarthrosis. However, these patients would have had a traumatic event that resulted in the torn transverse ligament. None of the patients in our study could recall any trauma to the cervical spine. Thus, we believe that the atlanto-axial instability in our patients resulted from the osteoarthrosis rather than the osteoarthrosis having developed as a result of a traumatic tear of the transverse ligament. One of our patients (Case 10) had atlanto-axial instability diagnosed on the basis of her clinical history and serial physical examinations. Before being seen by us, she had had occipitocervical pain, followed by difficulty walking. During the next three months, myelopathy with hyperreflexia developed. She finally became quadriparetic, with a dislocated atlanto-axial joint. Although no trauma was involved, the course of this patient is explainable, as compression of the spinal cord can occur once there is atlanto-axial instability8.
The appropriate time to perform an atlanto-axial arthrodesis for problems related to osteoarthrosis of the lateral mass articulations is variable. For patients who have pain, instability, and neurological involvement, our findings and those of others5,8,11,12,20 support the need for urgent reduction and stabilization. For patients who have pain, instability, and normal neurological findings, we recommend early stabilization to avoid potential neurological complications and to relieve occipitocervical pain. Boden et al. reported that a posterior atlanto-odontoid interval of less than fourteen millimeters was a predictor of paralysis in patients who had rheumatoid arthritis of the cervical spine. Although this is an extrapolation to patients who have osteoarthrosis and instability of the atlanto-axial joints, a posterior atlanto-odontoid interval of less than fourteen millimeters may be a useful threshold when operative treatment for this condition is considered.
For patients who have occipitocervical pain only, we recommend a trial of non-operative treatment consisting of immobilization in a collar and use of analgesics, provided that instability does not develop. If non-operative therapy fails to provide adequate relief of pain, an arthrodesis should be considered. All but one of our patients had an excellent long-term result after an atlanto-axial arthrodesis for unremitting occipitocervical pain.
Boden, S. D.; Dodge, L. D.; Bohlman, H. H.; and |and |Rechtine, G. R.: Rheumatoid arthritis of the cervical spine. A long-term analysis with predictors of paralysis and recovery. J. Bone and Joint Surg.,75-A: 1282-1297, Sept. 1993.75-A1282
1993
Bohlman, H. H.: Degenerative arthritis of the lower cervical spine. In Surgery of the Musculoskeletal System, edited by C. McC. Evarts. Ed. 2, vol. 2, pp. 1857-1886. New York, Churchill Livingstone, 1990.
Bohlman, H. H.; Emery, S. E.; Goodfellow, D. B.; and |and |Jones, P. K.: Robinson anterior cervical discectomy and arthrodesis for cervical radiculopathy. Long-term follow-up of one hundred and twenty-two patients. J. Bone and Joint Surg.,75-A: 1298-1307, Sept. 1993.75-A1298
1993
Brooks, A. L., and |and |Jenkins, E. B.: Atlanto-axial arthrodesis by the wedge compression method. J. Bone and Joint Surg.,60-A: 279-284, April 1978.60-A279
1978
Dunbar, H. S., and |and |Ray, B. S.: Chronic atlanto-axial dislocations with late neurologic manifestations. Surg., Gynec. and Obstet.,113: 757-762, 1961.113757
1961
Ehni, G., and |and |Benner, B.: Occipital neuralgia and the C1-2 arthrosis syndrome. J. Neurosurg.,61: 961-965, 1984.61961
1984
[PubMed][CrossRef]
Emery, S. E., and Bohlman, H. H.: Osteoarthritis of the cervical spine. In Osteoarthritis. Diagnosis and Medical/Surgical Management, edited by R. W. Moskowitz, D. S. Howell, V. M. Goldberg, and H. J. Mankin. Ed. 2, pp. 651-668. Philadelphia, W. B. Saunders, 1992.
Fielding, J. W.; Hawkins, R. J.; and |and |Ratzan, S. A.: Spine fusion for atlanto-axial instability. J. Bone and Joint Surg.,58-A: 400-407, April 1976.58-A400
1976
Fielding, J. W.; Cochran, G. V.; Lawsing, J. F., III; and |and |Hohl, M.: Tears of the transverse ligament of the atlas. A clinical and biomechanical study. J. Bone and Joint Surg.,56-A: 1683-1691, Dec. 1974.56-A1683
1974
Gallie, W. E.: Fractures and dislocations of the cervical spine. Am. J. Surg.,46: 495-499, 1939.46495
1939
[CrossRef]
Griswold, D. M.; Albright, J. A.; Schiffman, E.; Johnson, R.; and |and |Southwick, W.: Atlanto-axial fusion for instability. J. Bone and Joint Surg.,60-A: 285-292, April 1978.60-A285
1978
Halla, J. T., and |and |Hardin, J. G., Jr.: Atlantoaxial facet joint osteoarthritis: a distinctive clinical syndrome. Arthrit. and Rheumat.,30: 577-582, 1987.30577
1987
[CrossRef]
Harata, S.; Tolmo, S.; and |and |Kawagishi, T.: Osteoarthritis of the atlanto-axial joint. Internat. Orthop.,5: 277-282, 1981.5277
1981
[CrossRef]
Joseph, B., and |and |Kumar, B.: Gallie's fusion for atlantoaxial arthrosis with occipital neuralgia. Spine,19: 454-455, 1994.19454
1994
[PubMed][CrossRef]
McKeever, F. M.: Atlanto-axoid instability. Surg. Clin. North America,48: 1375-1390, 1968.481375
1968
Robinson, R. A.; Walker, A. E.; Ferlic, D. C.; and |and |Wiecking, D. K.: The results of anterior interbody fusion of the cervical spine. J. Bone and Joint Surg.,44-A: 1569-1587, Dec. 1962.44-A1569
1962
Southwick, W. O., and |and |Robinson, R. A.: Surgical approaches to the vertebral bodies in the cervical and lumbar regions. J. Bone and Joint Surg.,39-A: 631-644, June 1957.39-A631
1957
Star, M. J.; Curd, J. G.; and |and |Thorne, R. P.: Atlantoaxial lateral mass osteoarthritis. A frequently overlooked cause of severe occipitocervical pain. Spine,16(6S): 71-S76, 1992.16(6S)71
1992
[CrossRef]
Wertheim, S. B., and |and |Bohlman, H. H.: Occipitocervical fusion. Indications, technique, and long-term results in thirteen patients. J. Bone and Joint Surg.,69-A: 833-836, July 1987.69-A833
1987
White, A. A., III, and Panjabi, M. M.: The problem of clinical instability in the human spine: a systematic approach. In Clinical Biomechanics of the Spine. Ed. 2, pp. 283-302. Philadelphia, J. B. Lippincott, 1990.
Whitesides, T. E., Jr., and |and |McDonald, A. P.: Degenerative arthritis of the occipitocervical region—signs and symptoms. Orthop. Trans.,2: 32, 1978.232
1978