38ournal ofNeurology, Neurosurgery, and Psychiatry 1996;60:382-387

Clinical, CSF, and MRI findings in Devic'sneuromyelitis optica

J I O'Riordan, H L Gallagher, A J Thompson, R S Howard, D P E Kingsley,E J Thompson, W I McDonald, D H Miller

AbstractObjectives-Since Devic's original descrip-tion of neuromyelitis optica in 1894 therehas been much debate regarding its aetiol-ogy. A specific cause has been identified in aminority of cases but in most the questionhas arisen whether or not Devic's neu-romyelitis optica is a variant of multiplesclerosis. This study was undertaken tohelp clarify this issue.Methods-Neuromyelitis optica wasdefined as (1) a severe transverse myelitis;(2) an acute unilateral or bilateral opticneuropathy; (3) no clinical involvementbeyond the spinal cord or optic nerves, and(4) a monophasic or multiphasic illness.The clinical and autoantibody status wasdocumented. Patients underwent CSFexamination and MRI of brain and spinalcord.Results-Twelve patients, with a mean ageof presentation of 35-1 years, were seen.Eleven were women; vision was reduced tocounting fingers or worse in 10 patients andseven became confined to a wheelchair.Examination ofCSF showed local synthesisof oligoclonal bands in only two patientsand a neutrophil pleocytosis in two. A pos-sible aetiology was identified in five: a spe-cific connective tissue disorder (two),pulmonary tuberculosis (one), and possibleacute disseminated encephalomyelitis(two). Six had non-specific increases invarious autoantibodies. Eleven patientsunderwent MRI of the brain and spinalcord. In 10 there were diffuse abnormali-ties involving cervical and thoracic cordswith extensive swelling in the acute phase.Brain MRI was normal in five; in five therewere multiple deep white matter lesions,and one patient had minor age relatedchanges.Conclusion-It is proposed that Devic'sneuromyelitis optica is a distinctive disor-der with some clinical, CSF, and MRI fea-tures different from those found in classicmultiple sclerosis. In most cases a specificaetiology is not identified, but an immuno-logical mechanism of tissue damage seemslikely.

(3 Neurol Neurosurg Psychiatry 1996;60:382-387)

Keywords: neuromyelitis optica; multiple sclerosis

Sir Christopher Allbutt first alluded to anassociation between spinal cord disease and

visual loss in 1870.1 Eugene Devic, a physicianat L'Hopital de la Croix-Rousse in Lyon, sub-sequently described a case and inspired hisstudent Fernand Gault to write a thesis on thetopic.2-4 It was Gault who suggested a separateclinical entity termed neuroptico-myelites(neuromyelitis optica). The syndrome soonbecame known by Devic's name and is charac-terised by a severe acute transverse myelitisplus an acute or subacute optic neuropathywith or without recovery. It may follow amonophasic or multiphasic pattern. Despitethe many extensive reviews over the past cen-tury its status as a distinct entity has beenuncertain.5-8 In a proportion of cases a specificcause has been identified,9-'8 and in most casesof uncertain cause there has been much debateas to whether it is a variant of multiple sclerosisor a distinct clinical syndrome.We report theclinical, CSF, and MRI findings of 12 patientswith Devic's neuromyelitis optica and discusstheir importance.

Materials and methodsThe case records of patients attending theNational Hospital of Neurology andNeurosurgery, Queen Square and Moorfield'sEye Hospital between 1986 and 1994 withneuromyelitis optica were reviewed. The dis-ease was defined as (1) a complete transversemyelitis (an acutely developing and severeparaparesis or tetraparesis affecting motor andsensory pathways with or without sphinctericinvolvement; evolving over one to 14 days,with a sensory level, and in the absence of cordcompression); (2) an acute unilateral or bilat-eral optic neuropathy; (3) no clinical involve-ment beyond the spinal cord or optic nerves;and (4) the illness could be monophasic ormultiphasic. The clinical course of eachpatient was documented, and the eventualoutcome was graded as mild-minimal weak-ness and fully ambulatory; moderate-moder-ate weakness although still ambulatory withthe aid of unilateral or bilateral assistance; orsevere-severe paraparesis, confined to wheel-chair or bed.

In addition to the standard haematologicalinvestigations all patients had an autoantibodyscreen consisting of antinuclear antibody,rheumatoid factor, anti- double strandedDNA, thyroid thyroglobulin, thyroid micro-somal, gastric parietal, smooth muscle, mito-chondrial, reticulin, and antiphospholipidantibodies.The CSF was analysed as follows: a qualita-

Department of ClinicalNeurologyJ I O'RiordanH L GallagherA J ThompsonR S HowardD H MillerDepartment ofNeuroradiologyD P E KingsleyDepartment ofNeuroimmunology,The National HospitalofNeurology andNeurosurgery,Queen Square, LondonWC1N 3BG, UKE J ThompsonDepartment ofNeuro-ophthalmology,Moorfield's EyeHospital, City Road,London EC1V 2PD,UKW I McDonaldCorrespondence to:Dr D H Miller, NMRResearch Unit, The Instituteof Neurology, Queen Square,London WC1N 3BG, UK.Received 19 July 1995and in final revised form30 November 1995Accepted 8 December 1995

382 on January 10, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.60.4.382 on 1 A

pril 1996. Dow

nloaded from

Clinical, CSF, and MRIfindings in Devic's neuromyelitis optica

tive analysis involving electrophoresis on acry-lamide gel, light chain immunoblots of acry-lamide gel, and IgG immunoblots ofisoelectric focusing to identify the presence ofserum proteins, evidence of blood-brain bar-rier breakdown and the presence of oligoclonalbanding; a quantitative analysis of total pro-tein and components using a densitometerscan of acrylamide gel and total white cellcount, white cell differential, and cytology.

Magnetic resonance imaging was performedon a Picker 0 5 T or a GE Signa 1.5 T scan-

ner. Brain MRI was performed in the axialplane with 5 mm thick slices using Ti and T2weighted sequences. In addition three patientsreceived intravenous gadolinium DPTA. Thebrain was divided into seven areas: frontal,temporal, parietal, occipital, basal ganglia,brain stem, and cerebellum. Lesion sizewas graded as 1: < 5 mm, 2: 5-10 mm,3: > 10 mm or 4: confluent. The spinal cordwas imaged in the sagittal plane with axialimages through areas of abnormality. Thespinal cord was assessed with regard to cordswelling, cord atrophy, canal size, and signalintensity.

ResultsTwelve patients fulfilled the diagnostic criteriafor entry to the study. Four patients were ofAsian origin, one African, one Afro-WestIndian, five Caucasian, and one Mediter-ranean. There were 11 women and one man

with a mean age of 40 (range 16-65) years. Apossible aetiology was identified in five cases;two patients ( 10 and 12) developed a mono-

phasic transverse myelitis and bilateral opticneuritis several weeks after a non-specificinfective illness, and the probable diagnosiswas acute disseminated encephalomyelitis.Another, with a history of pulmonary tubercu-losis five years previously but with no clinical,radiological, or laboratory evidence of recur-

rence, developed bilateral optic neuropathyfollowed by transverse myelopathy after a

period of six months without further episodes.All other patients had multiphasic involvementof the optic nerve or spinal cord. Two patientshad a coexisting collagen vascular disease (sys-temic lupus erythematosus and mixed connec-

tive tissue disease). These two and six othershad raised concentrations of various autoanti-bodies. The interval between development of

Table 1 CSFfindings

Blood-brain Local synthesis Total Whitebarnier oligoclonal protein cell %

Patient no breakdown bands (mgldl) countlmm3 Polymorphs

1 + - 75 1 02 + - 140 4 943 - - 27 18 NA4 + - 60 5 05 NA - NA NA NA6 + - 38 1 147 + - 280 58 1 88 + + 52 1 149 + + 140 5 NA10 + - 208 38 01 1 + - 147 225 8012 + - 110 110 NA

NA = not available.

myelopathy and optic neuropathy was lessthan two years in 10 patients. In one there wasan interval of five years.

MYELOPATHYThe mean age of onset of myelopathy was35-1 (range 16-60) years. Four patients had asingle episode with poor recovery. Fourpatients had a single relapse; this followed aperiod of one year in one, two years in two,and 14 years in another. Four patients hadmore than one relapse. The myelopathyinvolved the cervical cord clinically in 10patients and response to treatment was vari-able but generally poor. All patients received acourse of high dose intravenous methyl pred-nisolone and three received cyclophos-phamide. Two patients exhibited steroiddependency with worsening of both myelopa-thy and optic neuropathy on reduction of thesteroid dosage. Another had an episode of pro-found bradycardia during methyl prednisoloneinfusion. Three patients required prolongedintensive care with mechanical ventilation, onenecessitating a tracheostomy, during the acutestages. Four patients made a good initialrecovery but developed persistent paraparesisafter subsequent relapses. At follow up fivepatients were ambulatory with the aid of uni-lateral or bilateral assistance; seven were con-fined to a wheelchair.

OPTIC NEUROPATHYThe mean age of onset of optic neuropathywas 35-2 (range 14-61) years. This was bilat-eral in 10 patients and followed a relapsing-remitting pattern in five. The prognosis forvisual symptoms was generally poor. Tenpatients had persistently diminished vision tocounting fingers or worse in one or both eyes.Of these, visual evoked responses were absentin nine. In one patient there was a normallatency but diminished amplitude; a patternmore suggestive of axonal damage thandemyelination. Visual evoked responses werenot available for one patient. Patient 11 under-went open biopsy of the right optic nerve nearthe chiasm, which showed reactive astrocyto-sis.

CSFAll patients had one or more CSF examina-tions. In 10 there was negative oligoclonalbanding. Of those who were positive, therewas a diffuse increase in r globulins with iden-tical oligoclonal banding in both CSF andserum in two indicating systemic synthesis ofIgG. Thus only two patients showed intrathecalsynthesis of oligoclonal bands. Patient 9 hadserial CSF which showed local synthesis ofIgG in the first sample plus a restricted anti-body response to an antigen outside the CNSin subsequent samples. The total proteincount was raised in nine. Seven patients hadfive or more white cells per mm3; differentialwhite cell count showed a predominantly poly-morphic leucocytosis in two; in three othersthere was an abnormal presence of poly-morphs; one patient had atypical or reactivelymphocytes on cytology.

383

on January 10, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.60.4.382 on 1 April 1996. D

ownloaded from

O'Riordan, Gallagher, Thompson, Howard, Kingsley, Thompson, et al

Table 2 Clinical and laboratory findings

Age at presentation Spinal Cord MRIPossible Brain MRI

Optic specific No abnormal Pattern ofPatient no Sex Myelopathy neuropathy Autoantibody causes Baseline Follow up segments Swelling illness

1 F 53 48 Thyroid thyroglobulin ND ND ND ND MultiphasicThyroid microsomal

2 F 33 32 Tuberculosis - - C7 to T6 + Monophasic3 F 35 37 Thyroid thyroglobulin - - C3 to T9 + Multiphasic

Thyroid microsomal4 M 19 19 - - C4 to T6 + Multiphasic5 F 42 41 ANA, smooth muscle

DS DNA,Gastroparietal SLE - - C5 to TI 1 + MultiphasicAcetyl choline rec abLow C3 & C4

6 F 21 22 Thyroid thyroglobulin + ND C 1/2 and - MultiphasicT1/2

7 F 18 17 ANA + ND C2 to T5 + Multiphasic8 F 56 58 Smooth muscle + ND C3 to C7 + Multiphasic

Rh factor9 F 60 61 ANA, Low C3

Thyroid thyroglobulin MCTD + No change Cl to TI0 + MultiphasicGastroparietal

10 F 36 36 Gastroparietal Probable + Partial Medulla to + MonophasicADEM resolution C7

1 1 F 16 14 Nil - - C3 to T5 + Multiphasic12 F 32 32 Nil Probable + Partial Extensive + Monophasic

ADEM resolution

ND = not done; ANA = anti nuclear antibody; DS DNA = double stranded DNA; acetylcholine rec ab = antiacetylcholine receptor antibody; SLE = systemiclupus erythematosus; MCTD = mixed connective tissue disease; ADEM = acute disseminated encephalomyelitis.

BRAIN MRIEleven patients underwent MRI of both brainand spinal cord. Four patients had serial brainand six serial spinal cord imaging. The intervalbetween scans varied between two months andfour years. the brain parenchyma was entirelynormal in five. In a sixth patient aged 60 there

was a confluent periventricular lesion and asingle separate lesion of less than 5 mm. Thiswas unchanged over three years and consid-ered to be a normal age related finding. Threepatients had normal serial brain imaging. Fivepatients exhibited multiple cerebral white mat-ter abnormalities considered to be abnormal

Figure 1 Sagittal T2weighted MRI ofspinalcord ofpatient 5 showingextensive swelling of thecervical and thoracicsegmnents with high signalintensity.

384

on January 10, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.60.4.382 on 1 April 1996. D

ownloaded from

Clinical, CSF, and MRIfindings in Devic's neuromyelitis optica

Figure 2 T2 weightedaxial MRI of brain ofpatient 10 showing multipledeep white matter lesions.

for their age. On follow up, there was partialreduction in the abnormalities in the twopatients with possible acute disseminatedencephalomyelitis. Abnormalities tended to besupratentorial and located predominantly inthe frontoparietal white matter. There was asingle lesion in the basal ganglia. Two patientshad signal change in the optic nerves; one withassociated swelling. Apart from involvement ofthe lower medulla in continuity with the cervi-cal cord abnormalities infratentorial lesionswere not seen.

SPINAL CORD MRITen patients had spinal cord swelling whichwas usually extensive during acute episodes ofmyelitis. Their length varied from a single levelonly to diffuse involvement of both cervicaland thoracic cords; usually multiple levelswere involved by the symptomatic lesion.There was an associated high signal intensityon the T2 weighted sequence. After the acutephase, swelling usually diminished and signalchange became less intense, but in twopatients there was some persistence of swellingand signal change for up to 18 months. Onepatient did not show cord swelling during theacute phase, but there were signal hyperinten-sities at the C1/2 and T1/2 levels. Diffuseenhancement over two segments of the spinalcord in one and seven in another was seenafter intravenous gadolinium DPTA.

CASE ILLUSTRATIONSPatient 5A 41 year old Caucasian woman with a previ-ous history of myasthenia gravis developedcomplete loss of vision in the left eye over fivedays without recovery. Four months later shehad a complete transverse myelitis withsensory level to T4 and a generalised lividoreticularis rash. She had an erythrocyte sedi-mentation rate of 60; low C3 and C4 concen-

trations; and a positive DNA binding antibodycompatible with a diagnosis of systemic lupuserythematosus. The myelopathy improvedafter treatment with steroids and cyclophos-phamide. Two years later there was recurrenceof myelitis with partial response to treatment.Brain MRI and CSF examination were nor-mal. In the spinal cord, however, MRI showeddiffuse cord swelling with signal change fromC3 to TI1 (fig 1). After three weeks theswelling had subsided and signal change waslimited to T2 to T9. Six months later she con-tinued to have a moderate paraparesis.

PATIENT 10A 36 year old patient of African origin devel-oped a non-specific gastroenteritis two weekspostpartum. Two weeks later there was apainful loss of vision in the left eye and con-comitant complete transverse myelitis withbulbar and respiratory involvement necessitat-ing mechanical ventilation. Six weeks laterthere was visual loss in the right eye. The CSFwhite cell count was 38 with a total protein of208 mg%. There was frank transudate ofserum proteins with blood-brain barrier break-down, but negative oligoclonal bands. BrainMRI (fig 2) showed 63 deep white matterlesions decreasing to 35 four months later.Spinal cord imaging showed diffuse swellingfrom the medulla to C7 and T3 to T5 (fig 3)which was less pronounced on follow up. Herclinical and MRI findings are consistent withprobable acute disseminated encephalo-myelitis. There was a slight response to intra-venous steroids; she continues to have amoderate paraparesis and has reduced bilat-eral vision to counting fingers.

PATIENT I 1A 14 year old girl of Mediterranean origindeveloped bilateral consecutive visual loss tocounting fingers over a four week period.

385 on January 10, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.60.4.382 on 1 A

pril 1996. Dow

nloaded from

O'Riordan, Gallagher, Thompson, Howard, Kingsley, Thompson, et al

Figure 3 Sagittal T2weighted MRI with axialslice through C4 ofpatient10 showing extensiveswellingfrom the medulla toC7 with involvement of thewhole diameter of the cord.

There was full recovery after intravenoussteroids. Eighteen months later there was arecurrence with bilateral progressive visualfailure unresponsive to steroids. Brainparenchyma MRI was normal; there wasintrinsic enhancement of the optic nerve withswelling at the chiasm. Biopsy showed a reac-tive astrocytosis. Six months later she devel-oped a complete transverse myelitis with asensory level to T6. Brain MRI was again nor-mal; MRI of the spinal cord showed diffuseswelling and signal change from C3 to T5.Her CSF showed a pleocytosis (white cellcount 225/mm3; 80% polymorphs) and a pro-tein count of 1-47 g/l. Oligoclonal bands werenegative. She was unresponsive to steroids andremains paraparetic and blind.

DiscussionSince Devic's original description in 1894 ofneuromyelitis optica there has been muchdebate regarding its aetiology.2 8 Withoutdoubt, a specific cause or associated diseasehas been identified in a minority of cases,9- 6

but in most it has not and in such patients thequestion has arisen whether or not Devic'sneuromyelitis optica is a variant of multiplesclerosis.6 8 7 Postmortem data in such caseshas been inconclusive-in some there havereports of typical demyelinating lesionsbeyond the spinal cord and optic nerves asseen in classic multiple sclerosis,7 whereasother reports describe more destructive fea-tures such as necrosis and cavitation withinthe spinal cord along with thickening of vesselwalls,820 features which are not expected inmultiple sclerosis.A difficulty in interpreting these and other

published studies is the variable definition ofthe clinical criteria used to diagnose Devic'sneuromyelitis optica. Involvement of thespinal cord and optic nerves without furtherqualification is insufficient-clearly manypatients with multiple sclerosis would beincluded. Conversely, restricting inclusiononly to those with a monophasic, simultaneousand severe affliction of the spinal cord andboth optic nerves would exclude patients whoseem to have a clinical pattern distinct from"classic" multiple sclerosis. Our classification

is a compromise between these two extremes,which seems to us to best describe a groupwho present a real challenge in terms of diag-nosis and management. The essential require-ments were: (1) a severe (more or lesscomplete) transverse myelitis, an uncommonfinding in multiple sclerosis; (2) an acute uni-lateral or bilateral optic neuropathy; (3) noclinical involvement beyond the spinal cord oroptic nerves, and (4) the illness could bemonophasic or multiphasic. We were particu-larly interested in evaluating the CSF andMRI findings for brain and spinal cord in thisgroup of patients.

In five of our 12 patients a probable aetiologywas identified. Two had probable acute dissem-inated encephalomyelitis, with a monophasicbilateral optic neuritis and transverse myelitisdeveloping within two weeks of a non-specificinfectious illness. Brain MRI showed multifocallesions of the cerebral white matter at presenta-tion, some of which had resolved at follow upafter four and 18 months without any newlesions developing, and spinal MRI showedswelling and signal change extending over manysegments of the cord in the acute phase. Theseimaging features are all characteristic of acutedisseminated encephalomyelitis and are atypicalof multiple sclerosis in which new lesions oftenappear at follow up and spinal lesions are usu-ally small.2125One patient had systemic lupus erythemato-

sus, another mixed connective tissue disease,and a third had a history of pulmonary tuber-culosis. The association of Devic's neuro-myelitis optica with acute disseminatedencephalomyelitis, systemic lupus erythemato-sus, and pulmonary tuberculosis has been pre-viously noted.'016 One patient with systemiclupus erythematosus had postmortem evi-dence of haemorrhage, vasculitis, extensivenecrosis, and arachnoidal fibrosis in the spinalcord and optic nerves.'0 The association ofpulmonary tuberculosis and necrotic myelitishas also been reported by Hughes et al, whosuggested that the pathogenesis may be animmune response of the spinal cord and opticnerves to the mycobacterial infection whethervia a shared antigen, non-specific adjuvanteffect, or other mechanisms.'5 An immunemediated response seems more likely in our

386

on January 10, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.60.4.382 on 1 April 1996. D

ownloaded from

Clinical, CSF, and MRIfindings in Devic's neuromyelitis optica

patients, and others described in the medicalliterature in whom the primary tuberculousinfection had been successfully treated at thetime of presentation with Devic's neu-romyelitis optica. An immunopathogenicmechanism also seems highly likely in ourpatients with systemic lupus erythematosus,mixed connective tissue disease, and acute dis-seminated encephalomyelitis, given that theseare accepted immune mediated conditions.Six other patients in our series exhibited avariety of organ specific autoantibodies whichagain points to an immunological mechanismfor Devic's neuromyelitis optica.

Whereas we accept in certain instances thatDevic's neuromyelitis optica may occur as aresult of demyelination from multiple sclero-sis, there are some clinical, CSF, and MRI fea-tures in our patients which would beunexpected in multiple sclerosis. Firstly, in theUnited Kingdom, multiple sclerosis is pre-dominantly seen in the Caucasoid ethnicgroup; in the present series, only five of 12were Caucasoid.5657 Secondly, patients wereoften left with permanent and severe deficitsafter the acute episode, such an outcome beingunusual in the early relapsing and remittingphase of multiple sclerosis.28 Thirdly, CSFoligoclonal bands and brain MRI white matterlesions were seen in only two of 12 and five of11 of our patients respectively-in clinicallydefinite multiple sclerosis such abnormalitiesare seen in over 90% of patients.29-31 Further-more, follow up MRI showed partial resolu-tion and/or no new lesions, an unlikely patternfor multiple sclerosis. Fourthly, the acutespinal cord lesion was invariably extensive,with swelling and signal change extending overmany spinal cord segments. Although swellingis often seen in acute cord lesions due to multi-ple sclerosis, it is rare for the lesion to extendbeyond a single segment.23

There have been some other recent reportsof CSF and MRI findings in Devic's neu-romyelitis optica.53233 The series of Mandler etal has several similarities to ours.8 Eightpatients were described, none of whom hadoligoclonal bands; brain MRI was normal inthree patients studied with this modality.Fazekas et al describe two patients in whomthere was swelling and enhancement in thechiasm during the acute illness (similar to ourpatient 1 1), but who had a normal brain MRIand no CSF oligoclonal bands.33

In conclusion, we believe that the presentseries of cases identifies Devic's neuromyelitisoptica as a distinct nosological entity.Although it may have multiple specific causes,an immunopathogenic mechanism seemslikely in all patients, and the clinical, MRI,and CSF features are distinct from those seenin relapsing-remitting multiple sclerosis.

1 Allbutt TC. On the ophthalmoscopic signs of spinal dis-ease. Lancet 1870;i:76-8.

2 Devic E. Myelite subaigue compliquee de nevrite optique.Bull Med 1894;8: 1033-4.

3 Devic E. Myelite aigue dorse-lombaire avec nevriteoptique, autopsie. Congress Francais Medicine (PremiereSession, Lyon) 1895;1:434-9.

4 Gault F. De la neuromyelite optique aigue. Lyon: Thesis 1894.5 Goulden C. Optic neuritis and myelitis. Ophthalmic Review

1914;34: 193-209.6 Beck GM. A case of diffuse myelitis associated with optic

neuritis. Brain 1927;50:687-703.7 Stansbury FC. Neuromyelitis optica (Devic's disease).

Presentation of five cases with pathological study andreview of the literature. Arch ophthalmol 1949;42:292-335;465-501.

8 Mandler RN, Davis LE, Jeffery DR, Kornfeld MK. Devic'sneuromyelitis optica: a clinicopathological study of 8patients. Ann Neurol 1993;34:162-8.

9 Motumara S, Tabira T, Kuroiwa Y. A clinical comparativestudy of multiple sclerosis and neuro-Behcet's syndrome.J Neurol Neurosurg Psychiatry 1980;43:210-3.

10 April RS, Vansonnenberg E. A case of neuromyelitis optica(Devic's syndrome) in systemic lupus erythematosus:clinicopathological report and review of the literature.Neurology 1976;26: 1066-70.

11 Kinney EL, Beroff RL, Rao NS, Lay MF. Devic's syn-drome and systemic lupus erythematosus. A case reportand review of the literature. Arch Neurol 1979;36:643-4.

12 Tola MR, Granieri E, Caniatti L, et al. Systemic lupus ery-thematosus presenting with neurological disorders. JNeurol 1992;239:61-4.

13 Goldman M, Herode A, Borenstein S, Zanen A. Optic neu-ritis, transverse myelitis and anti-DNA antibodies nineyears after thymectomy for myasthenia gravis. ArthRheum 1984;27:701-3.

14 Al-Deeb SM, Yaqub BA, Kjoja WO. Devic's neuromyelitisoptica and varicella. .NNeurol 1993;240:450-1.

15 Hughes RAC, Mair WGP. Acute necrotic myelopathy withpulmonary tuberculosis. Brain 1973;100:223-38.

16 Silber MH, Willcox PA, Iowen RM, Unger A.Neuromyelitis optica (Devic's syndrome) and pulmonarytuberculosis. Neurology 1990;40:934-8.

17 Mathews WB. In: Mathews WB, ed. McAlpine's multiplesclerosis. Edinburgh: Churchill Livingstone, 171-2.

18 Fukazawa T, Hamada T, Tashiro K, et al. Acute transversemyelitis in multiple sclerosis. _7 Neurol Sci 1990;100:217-22.

19 Ortiz de Zarate JC, Tamaroff L, Sica REP, Rodriguez JA.Neuromyelitis optica versus subacute necrotic myelitis:part II.Anatomical study of two cases. _7 Neurol NeuorsurgPsychiatry 1968;31:641-5.

20 Leonardi A, Arata L, Farinelli M, et al. Cerebrospinal fluidand neuropathological study in Devic's syndrome.Evidence of intrathecal immune activation. .7 Neurol Sci1987;82:281-90.

21 Francis DA, Brown A, Miller DH, Wiles CM, Bennett ED,Leigh N. MRI appearances of the CNS manifestations ofMycoplasma pneumonia: a report of two cases. _7 Neurol1988;235:441-3.

22 Kesselring J, Miller DH, Robb SA, et al. Acute dessimi-nated encephalomyelitis: MRI findings and the distinc-tion from multiple sclerosis. Brain 1990;113:291-302.

23 Kidd D, Thorpe JW, Thompson AJ, et al. Spinal cord MRIusing multi-array coils and fast spin echo. II. Findings inmultiple sclerosis. Neurology 1993;43:2632-7.

24 Willoughby EW, Grochowski E, Li DK, Oger J, KastrukoffLF, Paty DW. Serial magnetic resonance scanning inmultiple sclerosis: a second prospective study in relapsingpatients. Ann Neurol 1989;25:43-9.

25 Campi A, Filippi M, Comi E, Martinelli V, Baratti C,Rovaris M, Scotti G. Acute transverse myelopathy: spinaland cranial MR study with clinical follow up. AJNRAmINeuroradiol 1995;16: 115-23.

26 Compston DAS. Risk factors for multiple sclerosis: race orplace?I Neurol Neurosurg Psychiatry 1990;53:821-3.

27 Swingler RJ, Compston DAS. The distribution of multiplesclerosis in the United Kingdom. _7 Neurol NeurosurgPsychiatry 1986;49:1115-24.

28 Trojano M, Avolio C, Manzari C, Calo A, De Robertis F,Serio G, Livrea P. Multivariate analysis of predictive fac-tors of multiple sclerosis course with a validated method toassess clinical events. .7 Neurol Neurosurg Psychiatry 1995;58:300-6.

29 Paty DW, Oger JJF, Kastrukoff LF, et al. MRI in the diag-nosis of multiple sclerosis: prospective comparison ofclinical evaluation, EPs, oligoclonal banding and CT.Neurology 1988;38: 180-5.

30 Thompson AJ, Kermode AG, McManus DG, et al.Patterns of disease activity in multiple sclerosis: clinicaland magnetic resonance imaging study. BMJ 1990;300:631-4.

31 Miller DH, McDonald WI, Blumhardt LD, et al. Magneticresonance imaging in isolated non-compressive spinalcord syndromes. Ann Neurol 1987;22:714-23.

32 Tashiro K, Ito K, Mauro Y, et al. MR imaging of spinalcord in Devic's disease. _7 Comput Assist Tomogr 1987;11:516-7.

33 Fazekas F, Offenbacher H, Strasser-Fuchs S. MRI ofneuromyelitis optica: evidence for a distinct entity. .7Neurol Neurosurg Psychiatry 1994;59:1140-2.

387

on January 10, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.60.4.382 on 1 April 1996. D

ownloaded from