Annals of the Rheumatic Diseases 1994; 53: 714-717

CASE STUDIES IN DIAGNOSTIC IMAGING Series Editor: V N Cassar-Pullicino

Neurological disturbances in ankylosingspondylitis

P N M Tyrrell, AM Davies, N Evans

Case historyA 56 year old man presented with a seven weekhistory of tingling in both buttocks associatedwith numbness of the right buttock whichradiated down the back of the leg to the levelof the knee. He also gave a four year history ofnumbness along the heel and sole of the rightfoot. There was no motor symptoms. Nospecific urinary or bowel symptoms wereevident. In the past he had experiencedintermittent low back pain, but he had beenasymptomatic in that regard for many years.On examination there was impaired

sensation along the right S2 dermatome andalso perianally in the S3 to S5 distribution onthe right side. Movement of the spine waslimited in all directions. No focal tendernesswas elicited. Examination was otherwisenormal. Plain radiographs of the lumbar spinewere obtained.

PLAIN RADIOGRAPHIC FINDINGS

Radiographic examination of the lumbar spine

The MRI Centre, TheRoyal OrthopaedicHospital, BirminghamP N M TyrrellAM DaviesN EvansCorrespondence to:Dr Davies,The MRI Centre, The RoyalOrthopaedic Hospital, BristolRoad South, BirminghamB31 2AP.

Accepted for publication4 August 1994

(fig 1) demonstrated bilateral sacroileitis withfusion of the sacroiliac joints. Syndesmophyteformation was present throughout, giving aclassic "bamboo" appearance. Calcification ofintervertebral discs and pronounced thickening/calcification of the anterior longitudinalligament were also apparent. The lateral viewalso demonstrated a marked widening of thespinal canal.

Differential diagnosisThe patient demonstrates the radiographicfeatures of advanced ankylosing spondylitis(AS) with a fused immobile spine. Widening ofthe spinal canal is not a feature ofuncomplicated AS but the finding may indicatean underlying neurological disorder. Wideningof the spinal canal as a result of dural ectasiamay occur in other conditions such as Marfan'ssyndrome, Ehlers-Danlos syndrome andneurofibromatosis. The changes in AS, how-ever, result from involvement of the posteriorelements rather than the posterior aspect of the

Figure 1 Anteroposterior (A) and lateral (B) radiographs of the lumbosacral spine.

714

on Decem

ber 7, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.714 on 1 N

ovember 1994. D

ownloaded from

m.iS;- " IFigure 2 Sagittal TlW (A) and T2W (B) images of the lumbosacral spine. B shows, in addition to details apparent inA, the intermediate signal nerve roots descending on the thecal sac.

vertebral bodies as is usually seen in these otherconditions. The clinical assessment of thispatient is that of a cauda equina syndrome,which is a well recognised, albeit rare, compli-cation of chronic AS. A similar symptomcomplex may occur as a result of a compressivelesion of the cauda equina such as a spinaltumour or other extradural mass lesion,including an epidural haematoma which mayoccur in relation to a spinal fracture, both ofwhich occur with increased frequency in thepatient with advanced AS compared with thenormal population. Imaging is thus alwaysrequired for further evaluation.A magnetic resonance imaging (MRI)

examination of the lumbosacral spine (figs2-4) confirmed the diagnosis. The sagittalimages (fig 2) demonstrate a widened thecalsac with dorsal sacculations filled withcerebrospinal fluid (CSF) and are particularlyprominent on the right side at the L3 and L4levels. The posterior aspect of the vertebralbodies is not affected. On the T2 weighted(T2W) image (fig 2B), intermediate signalnerve roots can be seen descending in the highsignal CSF. On the axial Ti weighted (T1W)image (fig 3A) there is marked erosion of theright lamina into which projects the dorsaldiverticulum. The axial T2W image (fig 3B)shows the intermediate signal nerve roots tobetter advantage deviating towards thedeverticulum. The nerve roots appear free.

The dorsal diverticula do not encroach on theintervertebral foramina, which are shown to bewidely patent with in addition no evidence ofany bony encroachment (fig 4).The diagnosis is that of a cauda equina

syndrome complicating longstanding ankyl-osing spondylitis.

DiscussionAnkylosing spondylitis may be completelyasymptomatic and diagnosed incidentallywhen a radiograph is performed for someunrelated reason. It may also re-present aftera period of quiescence, or present late withcomplications (table). The cauda equinasyndrome (CES) is one such rare compli-cation. First described by Bowie and Glasgow,'this occurs in longstanding disease, usually ofmore than 15 years duration when disease isinactive and the patient typically has beensymptom-free for many years. Symptoms are

Neurological complications of ankylosing spondylitis

Cause Complication

Instability Atlanto-axial subluxationFracture-dislocationPseudoarthritis

Inflammation Cauda equina arachnoiditisSingle root lesions

Compression Ossified intraspinal ligamentsGranulation tissueForaminal stenosis

Neurological disturbances in ankylosing spondylitis 715

*s

on Decem

ber 7, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.714 on 1 N

ovember 1994. D

ownloaded from

Tyrrell, Davies, Evans

.RI 10 -3o/I

Figure 3 Axial Tl Wimage showing marked thinning of the lamina ofL3 (arrow), with a dorsal diverticulum extendinginto the eroded area. B: Axial T2Wimage showing high signal CSFfilling the diverticulum, and the intermediate signalnerve roots deviating towards the diverticulum (arrowheads).

F,N.

Figure 4 Sagittal Tl Wimage through the rightintervertebralforamina, demonstrating that there is nodiverticular or bony encroachment.

' usually sensory in nature, affecting lumbo-sacral nerve roots and frequently includeunnary and bowel symptoms. Motorsymptoms, if present, are usually mild. Despiteits relative rarity, it is a condition which hasbeen well described, particularly as a result ofthe newer imaging techniques of computedtomography (CT) and magnetic resonanceimaging (MRI).2-'The hallmark of CES in AS is the presence

of dorsal dural diverticula associated with awidened thecal sac and scalloped erosions ofthe laminae and spinous processes of thelumbar spine. Erosions of the laminae may bedetected on plain radiographs, but this is oftenin retrospect as they are difficult to appreciateunless looked for specifically, as in this patientand that of Young et al' In the past, imagingassessment relied on myelography which isfraught with problems because of thedifficulties in puncturing the theca in thepresence of bony ankylosis and soft tissuecalcification, and difficulty mobilising the rigidpatient to visualise the dorsal diverticula. CTis undoubtedly the imaging modality of choiceto show bony fusions and, more specific to theCES, the scalloped and eroded laminae andspinous processes. CT will show the CSF filleddorsal diverticula but MRI, with its inherentsoft tissue contrast, is far superior in thisregard. While T1W images are satisfactory, theT2W scans will not only show the high signalCSF in the outpouchings, but also exquisitelydemonstrate the course of the nerve roots andwill show their location within the spinal canaland also their position relative to each other(figs 2B, 3B). Intravenous paramagneticcontrast medium (Gadolinium-DTPA) wasnot given at this MRI examination. Other cases

716

on Decem

ber 7, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.714 on 1 N

ovember 1994. D

ownloaded from

Neurological disturbances in ankylosing spondylitis

in which it was administered did notdemonstrate any abnormal enhancement.5 7The actual cause of the CES in AS has not

been specifically defined, although varioustheories have been proposed. The nerve rootsymptoms are presumably related to thestructural changes seen in the spinal canal. Anarteritis and demyelination have been suggestedas causes of the syndrome, but these seemunlikely to account for the laminar erosion anddural ectasia. Arachnoiditis, occurring early inthe disease, does, however, seem to play animportant role in problems that may laterdevelop. Tullous et al5 suggested that primaryligamentous inflammation in AS leads toinflammation of contiguous structures such asthe meninges and dorsal bone elements of thespine. This produces dorsal meningeal inflam-mation with adhesion formation. Sucharachnoid adhesions may lead to blind pouchesinto which CSF is forced by the pumpingaction of arterial pulsation, with consequentenlarging arachnoid cysts together withpressure erosion of the adjacent bone.10 Aresorption defect of the CSF was demonstratedby Confavreux et al. 1 CSF resorption normallyoccurs through arachnoid villi in epidural veins.If this does not occur, then this may contributeto thecal sac enlargement and other structuralchanges, and rapid CSF pressure fluctuationsmay thus not be damped efficiently and thiscould contribute to nerve root injury. Thestructural changes primarily affect the lumbarregion, but may also be present in the thoracicregion.'2 Inflammatory changes and arachnoidadhesions are likely to occur also in the cervicalspine, but here, pressure effects will be less thanin the lumbar region and hence diverticulaformation and bony pressure erosion areunlikely to occur. A postmortem study of thelower spinal canal with CES showed that whilemany nerve roots were normal, others showedfibrosis and loss of myelin. 0 Appearances of thenerve roots at operation include fibrosis andchronic arachnoiditis embedding atrophicnerve roots,'3 14 with small sacral roots splayedand adherent to the dura.'5 Fibrous tissue andscattered lymphocytes surrounding the nerveroots have also been described.'6 It does appear,therefore, that there is a varied picture of nerveroot damage in the cauda equina as a result offibrosis, variable demyelination and atrophy.The dorsal diverticula did not encroach on theintervertebral foramina in this patient, as hasbeen shown by others'3 1' (fig 4), or compressthe nerve roots, although such compression hasbeen reported in three cases7 18 19 andpostulated in one.20No specific management for these patients

has been defined. Surgery, with decompressionof the cysts, either at laminectomy7 18 21 22 orvia a lumbo-peritoneal shunt" 20 has met withvarying success. It seems likely, however, thatwhile the cysts may contribute to the patho-genesis of the condition by interfering withpressure gradients of the CSF, an equally andperhaps more important factor is the result ofan inflammatory process that became inactivesome long time ago, leaving a residuum ofirreversible change.

CT and MRI are complementary imagingtechniques in the evaluation of the patient withAS and neurological complications. WhileMRI has the advantages of superb soft tissuecontrast, the multiplanar facility and absenceof irradiation, problems may be encounteredeither where there has been previous surgicalinstrumentation giving rise to extensive signalvoid and artefactual change, or with thekyphotic patient who may not fit into thescanner. In the former case, CT will not beuseful either, and this may be the rare occasionwhen myelography may still be required,depending on the clinical question being asked.In the latter situation, imaging the patient withCT in the lateral decubitus position, togetherwith sagittal reconstruction, may provide therequired information.

1 Bowie E A, Glasgow G L. Cauda equina lesions associatedwith ankylosing spondylitis. BMJ 1961; 2: 24-7.

2 Abello R, Rovira M, Sanz M P, et al. MRI and CT ofankylosing spondylitis with vertebral scalloping.Neuroradiology 1988; 30: 272-5.

3 Rubenstein D J, Alvarez 0, Ghelman B, Marchisello P.Cauda equina syndrome complicating ankylosing spondy-litis: MR features. J7 Comput Assist Tomogr 1989; 13:511-3.

4 Sparling M J, Bartelson J D, Mcleod R A, Cohen M D,Ginsburg W W. Magnetic resonance imaging ofarachnoid diverticula associated with cauda equinasyndrome in ankylosing spondylitis. _ Rheumatology 1989;16: 1335-7.

5 Tullous M W, Skerhut H E I, Story J L, et al. Cauda equinasyndrome of long-standing ankylosing spondylitis. Casereport and review of the literature. J Neurosurg 1990; 73:441-7.

6 Grosman H, Gray R, St Louis E L. CT of long-standingankylosing spondylitis with cauda equina syndrome. AmJ7 Neuroradiol 1983; 4: 1077-80.

7 Shaw P J, Allcutt D A, Bates D, Crawford P J. Cauda equinasyndrome associated with multiple lumbar arachnoidcysts in ankylosing spondylitis: improvement followingsurgical therapy. J Neurol Neurosurg Psychiatry 1990; 53:1076-9.

8 Young A, Dixon A, Getty J, Renton P, Vacher H. Caudaequina syndrome complicating ankylosing spondylitis: useof electromyography and computerised tomography indiagnosis. Ann Rheum Dis 1981; 40: 317-22.

9 Kerslake R W, Mitchell L A, Worthington B S. Case report:CT and MRI of the cauda equina syndrome in ankylosingspondylitis. Clin Radiol 1992; 45: 134-6.

10 Matthews W B. The neurological complications of anky-losing spondylitis. J Neurol Sci 1968; 6: 561-73.

11 Confavreux C, Larbre J-P, Lejeune E, Sindou M, AimardG. Cerebrospinal fluid dynamics in the tardive caudaequina syndrome of ankylosing spondylitis. Ann Neurol1991;29:221-3.

12 Hassan I. Cauda equina syndrome in ankylosing spondylitis:a report of six cases. J Neurol Neurosurg Psychiatry 1976;39: 1172-8.

13 Gordon A L, Yudell A. Cauda equina lesion associated withrheumatoid spondylitis. Ann Int Med 1973; 78: 555-7.

14 Hauge T. Chronic rheumatoid polyarthritis and spondyl-arthritis associated with neurological symptoms and signsoccasionally simulating an intraspinal expansive process.Acta ChirScand 196 1; 120: 395-40 1.

15 Lee M L H, Waters D J. Neurological complications ofankylosing spondylitis. BMJ 1962; 1: 798.

16 Rosenkranz W. Ankylosing spondylitis: cauda equinasyndrome and multiple spinal arachnoid cysts. Casereport.J Neurosurg 197 1; 34: 241-3.

17 Thomas D J, Kendall M J, Whitfield A G W. Nervoussystem involvement in ankylosing spondylitis. BMJ 1974;1: 148-50.

18 Byrne E, McNeill P, Gilford E, Wright C. Indradural cystwith compression of the cauda equina in ankylosingspondylitis. Surg Neurol 1985; 23: 162-4.

19 Milde E-J, Aarli J, Larsen J L. Cauda equina lesions inankylosing spondylitis. Scand J Rheumatol 1977; 6:118-22.

20 Okada S, Hase H, Hirasawa Y, Ogawa H, Takahashi K,Shibata S. A case report of lumboperitoneal shunt forcauda equina syndrome in ankylosing spondylitis. Spine1992; 17 (suppl): 59-61.

21 Bartleson J D, Cohen M D, Harrington T M, Goldstein NP, Ginsburg W W. Cauda equina syndrome secondary tolong-standing ankylosing spondylitis. Ann Neurol 1983;14: 662-9.

22 Russell M L, Gordon D A, Ogryzlo M A, McPhedran R S.The cauda equina syndrome of ankylosing spondylitis.Ann Int Med 1973; 78: 551-4.

717

on Decem

ber 7, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.714 on 1 N

ovember 1994. D

ownloaded from