Annals of the Rheumatic Diseases, 1985, 44, 809-816

Surgical treatment of cervical cord compression inrheumatoid arthritisH A CROCKARD,' W K ESSIGMAN,2 J M STEVENS,' J L POZO,1

A 0 RANSFORD,1 B E KENDALL'From 'the National Hospitals for Nervous Diseases and University College Hospital, London WCJ;and the 2Lister Hospital, Stevenage, Herts

SUMMARY Cervical myelopathy is a rare but potentially dangerous complication of rheumatoidarthritis and presents considerable therapeutic problems. A conservative approach carries highmortality and surgical intervention is not without serious risks. Reduction of subluxation andposterior fusion is widely practised but may require prolonged bed rest and continuous skulltraction, sometimes for many weeks. When anterior decompression has been attemptedprolonged immobilisation and external fixation have created problems. In this series 23rheumatoid patients with cervical myelopathy were investigated over a four-year period.Seventeen underwent anterior decompression of the cervical cord, of whom 14 had a transoralremoval of the odontoid peg and pannus and posterior occipitocervical fusion during the sameanaesthetic without mortality or serious postoperative complications; all but one have improved.The authors believe that early mobilisation after a combined cord decompression and internalfixation has reduced the mortality and morbidity. Management of cervical myelopathy inrheumatoid arthritis and indications for operation are discussed.

Key words: atlantoaxial subluxation, rheumatoid pannus, cervical myelopathy, transoral surgery,surgery - rheumatoid arthritis.

The involvement of the cervical spine in rheumatoidarthritis has become recognised increasingly overthe last decade. In a recent prospective study ofpatients with rheumatoid arthritis Windfield et al.'have shown that cervical subluxation may develop inassociation with peripheral erosive disease withintwo years of its onset, though remaining asymp-tomatic. In those with longstanding disease 25-36%have significant radiological changes and 2-5% ofthese develop myelopathy.2 3 About half of thosedeveloping this complication progress rapidly andsuccumb within a year despite conservativemeasures.4A review of the literature5 indicates that in the

cervical spine it is the atlantoaxial joint which isinvolved most commonly, with the joints below C3being affected evenly to account for the remainder.Atlantoaxial subluxation was recognised as aAccepted for publication 23 May 1985.Correspondence to Mr H A Crockard, The National Hospitals forNervous Diseases, Queen Square, London WCi.

dangerous complication of trauma (1824) andsyphilitic ulceration of the pharynx (1830) by SirCharles Bell,: but it was not until 1951 that Davisand Markley6 drew attention to it as a cause ofmedullary compression in rheumatoid arthritis. Thesubluxation backwards of the odontoid peg resultingin the compression of the craniocervical junction ofthe cord is considered to be the usual cause ofneurological abnormalities in this condition.The transverse ligament which has such a crucial

role in maintaining stability at the atlantoaxial jointmay be damaged by riding over the eroded surfaceof the odontoid peg or be directly affected by theinflammatory changes. The role of pannus formed atany or all of the synovial joints surrounding theodontoid peg, however, has been less emphasised,and the soft tissue mass produced by the inflamma-tory process at the craniocervical junction, thoughdemonstrated at autopsy, has been implicated rarelyin pathophysiology of the neurological changes insuch patients.

809

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from

810 Crockard, Essigman, Stevens, Pozo, Ransford, Kendall

Neurological assessment of a patient withrheumatoid arthritis may be difficult. Increasingdifficulty in walking, stiffness of the limbs, clumsi-ness of the hands, muscle weakness, may all be theresult of active synovitis, progression of the disease,or involvement of either the peripheral or centralnervous system, and it may not be easy to assesstendon reflexes in the presence of painful or

ankylosed joints. As the outcome of conservativetreatment of progressive cervical myelopathy hasbeen uniformly fatal,7 surgical intervention alonemay offer hope to these patients. The accuratelocalisation and knowledge of the character of thecompressive lesion are essential for successful sur-

gical planning; conventional radiology andmyelography have not always been sufficient for thispurpose. With the advent of the modern computer-ised tomography with three dimensional reformat-ting and the new non-ionic contrast media assess-

ment of such patients has been greatly simplifiedand can even be carried out on a day patient basis.The surgical treatment of cervical cord com-

pression-described in the literature includes anteriorfusion8 with or without decompression9 and pos-

terior fusion as separate operations.I) All requirepre- and postoperative skull traction and prolongedimmobilisation in bed or in a 'halo body' jacket. Therecognised complications of such treatment and theauthors' own experience led to the decision toperform a one stage procedure in which decom-pression of the cervical cord and internal metalfixation were carried out (unpublished data). Earlymobilisation has reduced the incidence of postop-erative complications. This paper describes theinvestigation and management of 23 patients over afour year period.

Patients and methods

PATIENTS

Since 1979 we have evaluated 23 patients (14female, nine male) with either classical or definiterheumatoid arthritis who presented a clinical picturesuggesting the involvement of cervical spine in therheumatoid process. Three patients were less than31 years of age, one was 45 years, and the remainderwere over 50 years (range 22-77 years). Seventeenwere seropositive, four seronegative, and in two theserology was not known. All had a longstandingdisease, the shortest being five years, the longest 50years (average 16x9 years), and all of them at somestage of their illness had been treated with steroidsor remission inducing agents, or both. Only one ofthe 20 patients with atlantoaxial subluxation had notbeen treated with steroids. Approximately half ofthe patients reported here had already had ortho-

paedic operations on one or more of their limbjoints.

SYMPTOMS AND SIGNSNeck pain, paraesthesiae, and limb weakness to avarying degree were the presenting complaints inthe 23 patients in this series (Table 1).

Nineteen patients had neck pain made worse byneck movement and in four it was associated withoccipital neuralgia. Although the systemic diseasewas long standing, the neck pain was of relativelyrecent onset (average 10 months) in all but onepatient who had suffered thus for 18 years. Itpreceded paraesthesia or weakness in nine patientsand occurred at the same time in six patients. Thelatter group had the most rapid clinical course.

Sensory symptoms were common and extremelyvaried (Table 1). There was no obvious correlationwith any radiological abnormality. In some therewas clinical evidence of a peripheral entrapmentneuropathy. In two there was a patchy sensory loss.Vibration sense was usually preserved, but jointposition sensation was difficult or impossible toassess at deformed or stiff joints.Motor weakness was very difficult to evaluate in

the presence of widespread arthritis, but it wasusually spastic in type. The patients' assessment oftheir condition in terms of reduction in performancewas a more valuable indicator of neurologicaldeterioration. For the same reasons reflex changesand plantar responses were difficult to assess. Inmost cases reflexes were increased, and in all butone plantar responses were either doubtful or

Table 1 Summary ofsymptoms and signs in the 23 patients

Symptoms and signs No A verageduration(months)

Neck pain 19 lo*Occipital pain 4Numbness and paraesthesia 16 6-6one arm 6two arms 3legs 5'glove and stocking' 2Lhermitte's sign 6

Weakness 18 41one arm 1two arms 6one arm and leg 1two legs 3all limbs 7

Sphincter disturbance 2Brain stem signs 4

*Excluding one patient with 18-year history of neck pain.Note the rapid progression of the myelopathy despite the long-standing history of arthritis.

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from

Surgical treatment of cervical cord compression in rheumatoid arthritis 811

extensor. Three patients were quadriplegic at thetime of presentation. Two had lost all sphinctercontrol and one had alteration in her voice andsuffered nasal regurgitation of fluids.One of the youngest patients presented with a

right sided hemiparesis, loss of gag reflex, andvertical nystagmus. He reported florid nightmaresand a terror of going to sleep for the two weeksbefore his admission (Ondine's curse)-indicatingprofound depression of the respiratory centre.

CERVICAL RADIOGRAPHSAll patients had lateral cervical radiographs inmaximum flexion and extension. In two whosepresenting symptom was neck pain alone, onlyminimal radiological changes were found and nofurther investigations performed (Tables 2 and 3).The commonest site of abnormality was at the

atlantoaxial level (19 patients) and was the main sitein 13. In six others there was evidence of involve-ment at other levels. In the 13 patients the maximumatlanto-odontoid movement was 5-14 mm. In thosewith disease at other sites it was less than 4 mm (sixpatients).

Table 2 Summary of radiological investigations

Investigation No

Lateral cervical spine (flexion and extension) 23Computed myelotomography ± myelogram 20Conventional myelogram alone 1

Table 3 Findings on computed myelotomography*

(a) Bone deformitytAtlantoaxial

Total 19Alone 13

SubaxialTotal 9Alone 3

No gross bone problem 2(b) Medullary or cord compression, or bothMedullary (infraolivary) 3Cord 'craniocervical junction' 16

Degree Due to pannusSevere 3 3Moderate 7 2Mild 6 4

Cord 'subaxial' 5No compression 4

*These are divided into (a) the bony abnormalities and (b) the CTevidence of brain stem and cervical cord compression. Note that innine out of 16 patients a significant proportion of the cordcompression was not due to bony subluxation but to pannus.tIn five patients multiple levels were seen on computedmyelotomography.

In addition to movement in the sagittal planethere was upward movement of the peg (upwardtranslocation of the dens) through the foramenmagnum in three patients who had the most ad-vanced neurological signs.

Subluxation below the level of the axis was notedin nine cases and was the only site in three. Twoshowed a 'staircase' phenomenon with sequentialsubluxation at all levels.

COMPUTED MYELOTOMOGRAPHYTwenty-one of the 23 patients had contrast studies,the first by conventional myelography alone. Withthe installation of a high resolution computedtomographic (CT) scan (GE 8800) 20 patients hadan intrathecal injection of water soluble contrastmedium and sequential high resolution axial sec-tions 1-5 mm in thickness over the affected area. Inthe last six patients the studies were repeated withflexion and extension of the cervical spine. Threedimensional reformatting allowed further visualisa-tion of the brain stem and cervical cord, thecerebrospinal fluid spaces, and the bony and softtissue deformations (Fig. 1).

Details of the technique are given elsewhere."The distribution of the pathology is given in Table 3.There was definite deformity of the medulla causedby odontoid peg and/or soft tissue pannus in threeand definite distortion at the craniocervical junctionin seven cases (Fig. 2). In half of the 10 (three severe,seven moderate) it was the soft tissue which wasresponsible for the spinal cord compression. In sixothers with little evidence of bony movement, fourhad associated soft tissue distorting the dural sac inthe region. There were five in whom the main siteof compression was below the axis. No cord com-pression was noted in four patients.

SURGICAL TECHNIQUEDetails of surgical technique are given elsewhere;'2a broad outline is provided here.

Transoral craniocervical decompression is carriedout through a vertical posterior pharyngeal incisionover Cl and C2. The anterior arch of Cl and theodontoid peg are removed, together with any softtissue pannus in the region, and the anterior aspectof the dura is decompressed. Exposure has been amajor problem. Initially, an elective tracheostomywas performed and the soft palate split, but withtechnical improvements an armoured nasotrachealairway and retraction of the soft palate has beensufficient. Sometimes the temporomandibular jointhas been involved in the arthritic process, and in onecase a midline mandibular split was required toeffect exposure.

Posterior occipitoatlantoaxial fusion is performed

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from

812 Crockard, Essigman, Stevens, Pozo, Ransford, Kendall

Fig. 1 A sagittal reconstruction ofa computed myelotomographshowing the brain stem andcervical cord. There is apseudofracture of the odontoid pegwith a soft tissue mass (pannus) infront and between the peg and thebody ofC2. The dura is compressedby extensivepannus posteriorly.

after the anterior decompression by carefully turn- and allowed out of bed on the third to fifthing the patient to the prone position. Fixation is postoperative day.effected by interlaminar wiring, fusion by cancellous Anterior decompression at lower cervical levels isbone chips. along conventional neurosurgical lines8 and hasThe patients are provided with a moulded collar been combined with suitable posterior fusion to

prevent angulation at the site of decompression.

OPERATIVE COMPLICATIONSNo patient has died after the transoral procedure.The most serious complication has been bleedingfrom a displaced vertebral artery during odontoidremoval. In this patient Cl was rotated as well assubluxed and the vertebral artery 'presented' in themidline anteriorly. Haemostasis was required, butthe decompression had to be abandoned, and shehas continued to deteriorate. The soft palate hasrequired resuturing in two patients early in theseries. Modification of technique has removed thenecessity to split the palate. There have been nocomplications with the posterior pharyngeal wound.

Skull traction pre- and postoperatively was usedin the first three patients and in all three there wereproblems associated with prolonged immobilisation;pulmonary emboli in two, and pneumonia in all. Noembolic phenomena have been detected since exter-nal splintage has been abandoned.

HistologyAt operation the soft tissue mass varied in consist-ency. On several occasions colourless oily fluidemanated from an incision of the anterior longitu-dinal ligament before bone removal. With this, wasassociated florid soft pink tissue.

In two patients the pannus formation was sowidespread that the Cl lamina was completelyeroded and the muscles in the area infiltrated withthe tissue which contained small white 'melon

Fig. 2 (a) A computed myelotomograph showing a seeds'. More usually the tissue was firm grey pinktransverse section ofthe craniocervical junction at the level and showed end-stage chromic inflammation ofof the base of the peg. The cervical cord is severely and sue. ere werovic palamm andcompressedfrom the front by soft tissue. (b) Higher, at the synovial tissue. There were synovial palisades andtop ofthe odontoid peg, the dura is distorted by the bone but patchy inflammatory cell infiltration but no lym-mostly by the soft tissue mass. phoid follicles. There was hyperplasia and prolifera-

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from

Surgical treatment of cervical cord compression in rheumatoid arthritis 813

tion of capillaries. Very rarely giant cells have beenseen.

Case studiesOur management policy is illustrated by the follow-ing three case histories.

CASE 1

A female aged 35 had been suffering from classicalseropositive rheumatoid arthritis since the age of 16and had intermittent pain in the neck for severalyears. Her treatment included steroids and remis-sion inducing drugs. For five months the neck painhad become more severe and neck flexion producedpain and paraesthesiae down both arms. Examina-tion showed brisk but equal reflexes and an equivo-cal extensor plantar response on the left. Cervicalspine radiography showed atlantoaxial subluxation.Her symptoms settled down after four months, withno change in the neurological state. No furtherinvestigations have been carried out but carefulfollow up has been arranged.

CASE 2A 21 year old female developed chronic seronega-tive juvenile arthritis of rheumatoid type at the age

of six and required treatment with steroids andremission inducing agents. When 20 years old sheunderwent synovectomy of the right knee and in her21st year of life bilateral hip replacement. Routinex-ray of the cervical spine before the secondoperation showed atlantoaxial subluxation (5 mm).Four months later she developed pain at the back ofthe neck radiating to the right shoulder and madeworse by movements of the neck. On examinationall reflexes were brisk but equal, with equivocalplantar responses but no sensory changes. A com-

puted myelogram showed erosions of the odontoidprocess, indentation of the dura but no evidence ofmedullary or spinal compression. She was providedwith a collar and carefully followed up. Nine monthslater she developed mild weakness of the lowerlimbs and repeated CT myelography showed thepresence of atlantoaxial subluxation with pannusformation around the odontoid peg, compressingthe cord. She underwent transoral excision ofodontoid and posterior occipitocervical fusion inone procedure without any problems.

CASE 3A male aged 54 with classical seropositive rheuma-toid arthritis for 24 years had never been treatedwith steroids or remission inducing agents.Poliomyelitis in his youth left him with weakness ofthe left leg and some trunk muscles, and recently he

developed a peptic ulcer. In August 1983 move-ments of the neck became limited and painful, andan x-ray of the cervical spine showed markedatlantoaxial subluxation and forward slip of C4 onC5. Three months later he began complaining ofpain in the neck associated with transient weaknessof the left arm on leaning forward. The weaknessgradually increased, he developed tingling andnumbness, and became clumsy with his fingers.Computed myelography showed erosions on theodontoid peg and pannus formation above it. Inaddition there was flattening of the cord at theC4-C5 level. Transoral removal of the odontoid pegwith its associated pannus, anterior decompressionat the C4-C5 level, and posterior occipitocervicalfusion were carried out in one surgical procedure.The patient was out of bed in five days and allowedhome three weeks after the operation. He lost allneck pain and regained full function of the arm.

Results

A summary of the management and outcome of the23 patients is given in Table 4. Two patients withminimal neck pain and paraesthesia were carefullyfollowed up as outpatients and showed no signs for14 months until there were signs of progressivemyelopathy in one case (case 2), for which she hadsurgery. One patient was unfit for surgery and tworefused despite progressive deterioration. Seven-teen patients were offered surgery and, apart fromour first case on whom we performed a conven-tional posterior fusion alone, no patient has died inthe perioperative period; she died of progressiverespiratory failure. Two have deteriorated sincesurgery, one at three years due to subsequent C6-C7subluxation, the other within weeks of surgery. Inthis latter case profuse bleeding from a displaced

Table 4 Management atnd outcome in the 23 patienits

Managemnenit No Oiitcomtie

Observation only 2 2. No changeUnfit for surgery 1 1. No chaingeRefused surgerv 2 J 1. Died

Refusedvsurgery 2 1. DetcrioratcdPosterior fusion alone I 1. Died

0. DiedTransoral + posterior fusion 14 2. Deteriorated*

12. ImprosedtAnterior subaxial + posterior fusion 3 { Improsed

*One patient suffered profuse bleeding and the operation witsabandoned: one patient had subsequent subluxation at C6-C7.tOne patient died of myocardial infairction at 8/12; one patticnt dicdof carcinomatosis at 22/12.

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from

814 Crockard, Essigman, Stevens, Pozo, Ransford, Kendall

vertebral artery during a transoral procedure forcedus to abandon the decompression; neurologicallyshe continued to deteriorate and is now quadri-paretic.

All the other patients who had transoral surgery

or anterior subaxial surgery reported cessation oftheir neck pain, and 15 out of the 17 patientsimproved. The quadriparetic patients walked out ofhospital with recovered sphincter function. Thosewith paraesthesia noted an improvement, and theLhermitte's sign disappeared in all cases in which itwas present before the operation.There were no perioperative deaths in those

subjected to transoral surgery, but as might beexpected in this age group some succumbed forother reasons. Two of the patients aged 68 and 72died eight months and 22 months after surgery ofmyocardial infarction and carcinomatosis respec-

tively.

Discussion

The neurological assessment of a patient withrheumatoid arthritis, particularly in its more severe

form, may be difficult. Failure to establish an earlydiagnosis of cervical myelopathy, therefore, is notunusual and there are reports of a six months' delayin appreciating the neurological deterioration,which in some patients may lead to, or coincidewith, a fatal outcome. In Marks and Sharp's series319 of 31 patients died due to a myelopathy, 15 ofthem within six months of onset, and Meijers etal.13 reported recently that of nine patients withcervical myelopathy who were not operated upon,four died within a year.Cord compression in rheumatoid arthritis was first

described by Garrett in 1890,3 and a recent estima-tion of the incidence of this complication in a long-standing disease is between 2 and 5%.5 There is a

small proportion of patients who improve spon-taneously, whereas in others myelopathy is static foryears. However, any progression of an establishedmyelopathy and deterioration in the neurologicalstatus should be taken as evidence of an unfavour-able outcome if left untreated.The clinical picture of cervical myelopathy has

already been well described, and our patients were

similar to those of other series. There was no

obvious correlation with sensory symptoms and thedistribution of radiological abnormalities. There was

some relationship between motor signs and CTevidence of cord compression. We have found thefollowing particularly useful in alerting the clinicianto the possibility of myelopathy: (a) neck pain or

occipital neuralgia of recent onset in a rheumatoid

patient; (b) paraesthesiae or numbness in the trunkor limbs or 'electric shocks' produced by movementsof the neck, such as reading, lifting the head off thepillow, or leaning forward (Lhermitte's sign); and(c) the patient's own account of his diminishedmotor ability or a documented change since the lastoutpatient visit.Any synovial joint in the cervical spine may be

involved in rheumatoid arthritis, but in a review of325 cases assembled from the literature Boyle foundan incidence of 47% involvement at the C1-C2level.5 The role of atlantoaxial subluxation wasemphasised by Davis and Markley in 19516 and ledto an increased awareness of the condition. Stevenset al. reported2 that 36% of the patients in theirseries had this complication and estimated that twothirds of such patients eventually developed cervicalmyelopathy. They also recorded that 16% of theircases with no gross bony subluxation had cervicalmyelopathy and emphasised that the bony subluxa-tion in itself may not necessarily be the wholeproblem.The role of the rheumatoid pannus in the produc-

tion of cervical myelopathy has not been given itsdue importance, though its description has beenfound in pathology texts. More severe grades ofcompression may be shown with conventionalmyelography, but the advent of computedmyelography with facility to reformat in any planehas allowed the more precise evaluation of therelative roles played by the bony element and softtissue, and the detection of milder grades ofcompression of the cervical cord. Some advocatecomputed tomography without intrathecal contrastfor these patients, 4 but in our opinion cord com-pression may not be apparent in some cases.The results of different forms of treatment have

been evaluated by Marks and Sharp,3 though therewas considerable difficulty in establishing compar-able groups. In their untreated group there was a100% mortality, while 50% of patients treated witha cervical collar also died. All subjected to skulltraction alone died. In a recent communicationMeijers et al.13 reported that in their series 'ninepatients with cervical myelopathy who were notoperated upon all died within a year, four of themdue to consequence of cord compression'. Reduc-tion and fixation by posterior fusion was the mostsuccessful procedure, though two out of eight inBoyle's series5 died in the immediate postoperativeperiod from respiratory failure. A recent report hasagain shown the respiratory hazards of thisapproach.'5 These procedures have- aimed at thecorrection of subluxation and have not taken intoaccount the compressive effect of pannus, which

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from

Surgical treatment of cervical cord compression in rheumatoid arthritis 815

may increase when the bones have been realigned.All the published accounts describe immediaterespiratory failure or further cord compression in aproportion of their operated cases. One has gone asfar as saying that sudden death was an unfortunatebut inevitable consequence in some cases. 16 It is ouropinion that attention to the compressive elementanteriorly before posterior fixation is an essentialpart of the treatment for patients with pannus.

Meijers et al. 13 treated their 25 patients withcervical myelopathy by posterior fusion, precededfor three days to 10 weeks (mean three weeks), andfollowed for a period of 2½/2-3 months by continuousskull traction and nursing on a circoelectric bed.Three of their patients died from the complicationsof immobilisation, e.g., pulmonary embolism.There have been other reports of a two stageprocedure'7 with transoral decompression initiallyand occipitocervical fusion at a later date, but thepatients required continuous skull traction betweenthe two operations. None of the authors digressfrom purely technical and academic assessment oftheir results to describe their experience in terms ofhuman suffering of the patients exposed to aprolonged bed rest on continuous skull traction.Because of this and the increased incidence ofcomplications associated with immobilisation for along period of time, we decided to perform anteriordecompression and posterior fusion as a one stagesurgical procedure. We have been surprised todiscover how well the patients have toleratedsurgery of this magnitude. They have been out ofbed within five days, starting on an active mobilisa-tion programme, and discharged from the hospitalafter a fortnight to be followed up from both theneurosurgical and rheumatological outpatient de-partments. Their only inconvenience was a stiffcollar, which they have had to wear for about threemonths.

Obviously the cord may be compressed subaxiallyand the usual finding is a forward displacement ofthe upper on the lower vertebral body. In six of ourpatients with an atlantoaxial problem there wasevidence of subaxial involvement. The surgicaldecision in such patients was to approach the site ofmaximum compression as judged on computedmyelotomography and follow up the patient care-fully postoperatively. For subaxial compression wehave been disappointed with a conventional anteriordecompression and fusion (Cloward procedure),because of angulation at the operative site within afew months. The approach which we have adoptedis a simultaneous anterior decompression and pos-terior interlaminar fixation to an 0 ring. Thisproduces good decompression with complete stabil-

ity and thus allows early mobilisation of thesepatients without prolonged bed rest in skull traction.

It is still very early to assess the overall outcomeof the patients treated by this technique but, giventhe poor prognosis for patients with myelopathy,our experience is encouraging. All the patients onwhom the combined operation was carried outshowed subjective and objective improvement clini-cally and there have been no operative deaths orimmediate postoperative complications. We do not,however, underestimate all the potential dangers ofthis operation and must emphasise that in our viewthe only definite indications, for it are: clinicalevidence of an established cervical myelopathy orprogressing neurological deficit indicating cord com-pression. The anterior route is particularly necessaryif there is compression by rheumatoid pannus.

Conventional radiological abnormality alone,even when extensive, is not a criterion on whichsuch an operation should be advised.

In conclusion, it is necessary to emphasise theimportance of early recognition of cervical myelo-pathy in patients with rheumatoid arthritis. Theinvestigation by computed myelotomography is lesstraumatic for the patient and provides the clinicianwith more information. A one stage anterior decom-pression and posterior internal fixation is welltolerated, produces clinical improvement, and hasbeen much less hazardous for the patients thanconventional forms of treatment.

References

1 Windfield J, Cook D, Brook A S, Corbett M. A progressivestudy of the radiological changes in the cervical spine in earlyrheumatoid disease. Ann Rheum Dis 1981; 40: 109-14.

2 Stevens J C, Cartledge N E F, Saunders M, Appleby A, Hall M,Shaw D A. Atlanto-axial subluxation and cervical myelopathyin rheumatoid arthritis. Q J Med 1971; 159: 391-408.

3 Marks J S, Sharp J. Rheumatoid cervical myelopathy. Q J Med1981; 199: 307-19.

4 Nakano K K. Neurologic complications of rheumatoid arthritis.Orthop Clin North Am 1975; 6: 861-81.

5 Boyle A C. The rheumatoid neck. Proc R Soc Med 1971; 64:1161-5.

6 Davis F W, Markley H E. Rheumatoid arthritis with death frommedullary compression. Ann Intern Med 1951; 35: 451-61.

7 Mathews J A. Atlanto-axial subluxation in rheumatoid arthritis:a five year follow-up. Ann Rheum Dis 1974; 33: 526-31.

8 Cloward R B. The anterior approach for removal of rupturedcervical discs. J Neurosurg 1958; 15: 602-17.

9 Sukoff M H, Kadin M M, Moran T. Transoral decompressionfor myelopathy caused by rheumatoid arthritis of the spine. JNcurosurg 1972; 37: 492-7.

10 Newman P, Sweetman D R. Posterior fusion for atlanto-axialsubluxation in rheumatoid arthritis. J Bone Joint Surg 1969;51B: 423-8.

11 Stevens J M, Kendall B E, Crockard H A. The spinal cord in thecervical myelopathy of rheumatoid arthritis: a correlatedcompbuted myelographic study. Neurol Neurosurg Psychiatr (inpress).

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from

816 Crockard, Essigman, Stevens, Pozo, Ransford, Kendall

12 Crockard H A. The transoral approach to the base of the brainand upper cervical cord. Ann R Col Surg Engl 1985; 67: 321-5.

13 Meijers K A E, Cats A, Kremer H P H, Lugendij K W,Onrlec G J, Thomeer R T W M. Cervical myclopathy inrheumatoid arthritis. Clin Exp Rheumatol 1984; 2: 239-45.

14 Braunstcin E M, Weissman B N, Seltzer S E, et al. Computcdtomography and conventional radiographs of the cranio-vertebral regions in rheumatoid arthritis: a comparison.Arthritis Rheum 1984; 27: 26-31.

15 Christophidis N, Huskisson E C. Misleading symptoms andsigns of cervical spine subluxation in rheumatoid arthritis. BrMed J 1982; 285: 364-5.

16 Hamblen D L. Postgraduate text book, of clinical orthopaedics.Bristol, London, Boston: Wright, 1983: 487-97.

17 O'Laoire S A, Thomas D G T. Transoral approach to thecervical spine: report of four cascs. J Neurol NeurosurgPsychiatry 1982: 45: 6(1-3.

copyright. on A

pril 9, 2021 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.12.809 on 1 D

ecember 1985. D

ownloaded from