controversies in lupus: nervous system involvementbluestein et al'9 1976 lcassay/absorption lca 45 *...

Annals of the Rheumatic Diseases 1995; 54: 159-167

REVIEW

Controversies in lupus: nervous systeminvolvement

G AW Bruyn

Despite great improvements in the ability totreat systemic lupus erythematosus (SLE)nephritis, management of nervous systemmanifestations of patients with SLE remainsunsatisfactory. Nervous system involvementoccurs in about 50% of all patients at sometime during the course of their illness, with animpact on their lives which can vary frominsignificant to disabling. Indications that thenervous system is involved include the organicbrain syndrome (disorientation, forgetfulness,disturbances of attention), seizures, chorea,stroke, and frank episodes of psychosis.Controversy as to the best approach to patientswith central nervous system (CNS) SLEsurrounds the pathogenesis, diagnosticdifficulties, and treatment. This article focuseson these three topics.

epilepsy, coma, stroke, chorea-athetosis,neurocognitive dysfunction, cerebellar ataxia,headache, cranial neuropathy, and transversemyelitis5 --a variety which is insufficientlyaccounted for by direct clinicopathologicalcorrelation. This is the first big controversy inCNS lupus: is one unifying disease mechanismresponsible for this diversity of clinicalpresentations or do separate mechanisms playa role?Hypotheses advanced to explain the

spectrum of CNS manifestations associatedwith SLE may be divided into four categories:immune complex vasculitis; neurone reactiveautoantibodies; thrombosis associated withantiphospholipid antibodies; and cytokineenhanced autoimmunity.

PathogenesisThe pathogenesis of CNS SLE leads both tothe observed histopathological findings'-3 andto the clinical manifestations; at the same timeit forms the main target of treatment. In alandmark study, Johnson and Richardson'described the pathology of cerebral lupus.They found 80% of 24 postmortem brainspecimens to be microscopically abnormal;multifocal microinfarcts and increased peri-capillary microglial cells were the predominantabnormalities. Abnormalities of the brainvessels include a vasculopathy involving smallarteries (< 100 ,um), in addition to fibrinthrombi occluding the lumens of affectedvessels (fig 1). Features of the vasculopathyinclude hyalinisation of the meningeal, sub-cortical and cortical arterioles, in addition toperivascular lymphocytic infiltrates, endothelialproliferation, thrombosis, capillary wallthickening, and vasculitis (fig 2). Vasculitiswith inflammatory cells within the vessel wallwas found in only three of the 24 autopsyspecimens ofJohnson and Richardson. Similarfindings were reported in two subsequentclinicopathological studies-one performedbetween 1955 and 1977 on 57 patients,2 andthe most recent between 1972 and 1989.3Infrequently, the lesions may be restricted tothe cerebellum and mid-brain.4The association of the microvascular lesions

with microinfarcts suggested that occlusion ofsmall vessels is the basis for the damage to thenervous system. While the histopathologicalfeatures are straightforward, the clinicalspectrum of CNS involvement is broad.Neurological manifestations include psychosis,

IMMUNE COMPLEX DEPOSITIONIn contrast to the evidence available in lupusnephritis, evidence of a true vasculitis in brainsof lupus patients is not consistent: data forcomplement dependent inflammatory injury(consumption oftotal haemolytic complement,and of the third and fourth components) arenot consistent;9"1 high titres of DNAautoantibodies have not been found in thecerebrospinal fluid (CSF);11 true vasculitis isfound in only about 10% of patients withcerebral lupus;' and the spectrum of CNSmanifestations of lupus is unique for thisdisease and is not reproduced in any of thevasculitides.

Nevertheless, the lack of extensive histo-logical evidence for a true vasculitis involvingthe cerebral vessels does not preclude theimplication of immune complexes in thepathogenic process: pathology similar to thatfound in the lupus brain has been described inthe coronary arteries of a mouse model of SLE,in which immune complexes appear to causeocclusion and subsequent myocardialinfarction in the absence of an inflammatoryresponse.12 Immune complexes consisting ofDNA-anti-DNA complexes have beenidentified in the choroid plexus of patients withneuropsychiatric lupus by both immuno-fluorescence and electron microscopy. 3However, choroid tissue from patients withoutCNS involvement was not studied. Similarfindings were reported in a controlled study,but choroid tissues from rheumatoid arthritispatients and from SLE patients without CNSinvolvement also contained immunoglobulindeposition.'4 Given these findings, it might beconcluded that the highly vascular choroid

Department ofRheumatology,Medisch CentrumLeeuwarden,8934 AD Leeuwarden,The NetherlandsG AW BruynAccepted for publication21 September 1994

159

on June 16, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.54.3.159 on 1 M

arch 1995. Dow

nloaded from

http://ard.bmj.com/

Bruyn

V, a: tB 'v,

<S e ;: {t

.3.". |.i-

40

@~''

~ ~

FiurICe trl erou sstm bolem ndtoiS E vendingsombsi nd lheadsubpedyalceebaltisu, impa irmelnfrtiofn eurowngal led n. eI hlasrbeng

showns thahtpatients neurosyechiaictric

have dsrupntionheprmailtof theblood-banaflradariemay balsoreincreasedinrteaIg

synthesisatcompared wirulthn lupuse

withot fo eatrv e g nd l ad t

VitouceerlV etrs'

NEURONE REACTIVE ANTIBODIES

Animal studies have supported the concept ofneurone reactive antibodies directly accoun-ting for neuropsychiatric SLE: administrationof antibodies to brain constituents can induceseizures or behavioural changes.'7 18 Over the

past two decades, a number of studies havereported an array of neurone reactive autoanti-bodies 16 19 20 22-26 28 (table 1) found in bothserum and CSF of lupus patients. Ashistological examinations have failed to revealcell death as a possible mechanism, it isassumed that the autoantibodies exert theireffects by interfering with the cell's ability as aresponder in the neuronal network; how theydo this remains uncertain. Neurone reactiveautoantibodies can be divided into lympho-cytotoxic antibodies (LCA) cross reactive withbrain tissue, and neuronal antibodies which aretargeted directly to neuronal antigens-whichmay be neuronal membranes or intra-cytoplasmic constituents of neuronal cells.Controversy exists as to the definition oflymphocytotoxicity and, as a consequence, asto the prevalence of LCA, the antigenicspecificities shared by LCA and neuronalantibodies, the various reactivities of LCA, andthe clinical relevance of these reactivities.Lymphocytotoxic antibodies (LCA). Membranereactive autoantibodies (for example anti-bodies against erythrocytes, lymphocytes, andplatelets) are well recognised in SLE and areresponsible for the haematological featuresfrequently found in the disease.2' Cross reactivitybetween LCA and brain cells was first observedby Bluestein and Zvaifler'9 and subsequentlyconfirmed by other investigators."6 20 21 24 25 28LCA have been found to correlate with variousneuropsychiatric manifestations,'9 23 24 26-28but most consistently with cognitive dys-function and visuospatial deficits.16 28 How-ever, the specificity of LCA for brain tissue hasbeen questioned in some studies'1 16 22 inwhich the percentage of LCA was similarin SLE patients with and without CNSinvolvement." 16 These inconsistencies maybe explained not only by methodologicaldifferences, but also by the variability indefining neuropsychiatric SLE. Although it isunclear what causal role LCA play, mostevidence indicates that they are not producedin reaction to brain damage but may have aspecific role in the pathogenesis of cerebrallupus, particularly cognitive dysfunction andvisuospatial deficit.'6 28Neuronal membrane autoantibodies. Brainreactive antibodies to membrane antigens havebeen demonstrated in the serum of SLEpatients and have been found to correlate withCNS disease.29 Bluestein et al24 found that IgGantineuronal antibodies in CSF, but not inserum, correlated with diffuse CNS disease,but other investigators did not confirm thesefindings.26 Most studies used neuroblastomacell lines as the antigen source, but it isdoubtful if those primitive tumour antigens arealso present on normal brain cells.30 Antigenicvariability is great: one study found anautoantibody reactivity with the 50 kDaneuronal antigen associated with active CNSlupus in 19 of 20 patients tested.3' It is not clearif this 50 kDa neuronal membrane antigen issimilar to the 50-52 kDa reactivity in lupussera that contain LCA.28 Another study did notshow surface antigens;26 yet another found a97 kDa antigen present on neuronal cells

160


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from

http://ard.bmj.com/

Controversies in lupus: nervous system involvement

9. i. A'~~~~

4'.N~~~~~~~~~~~~~~4

~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 4I

'6* ..4~~~~~~~~~~~~~~~~~~~~~.U .. .4A i

A

-~~~~~~~~~~~~~~~~~~~~~~~~~~m

4......4..4.~~~~~~~~~AM

-4.~~

3=.

I.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l**44.4 4.4 .4 .4~~~~~~~~~PFiue2Dti ofi .Vsbeae mnster,acmlto4fmcohgs

asr'Yeswlin,ad nipetaiatin (Hemt 4ian4osn4Hoznt4arereens450um4 1'

reonie bysm4SEsra3.ecueofti~~~~~vrailt om netiaos aesugse

tam others, tacumuatio ofa it ti .r Ha mant ob in s p o siba

e ilS LEe

Intracytoplasmic targeted antibodies. Antibodiesto intracytoplasmic constituents of neuronalcells-including anti-ribosomal P 26 33-36antineurofilament antibodies,37 and anti-Ro(SS-A) or anti-La(SS-B) antibodies,38-have been found in serum and CSF of patientswith lupus. Since anti-Ro(SS-A) antibodiescan cause heart block, it is postulated that suchantibodies may profoundly interfere withcellular processes. They may enter neuronalcells by an as yet undefined access mechanism;alternatively, they may react with neuronalmembrane antigens expressing P protein orneurofilament epitopes. Anti-ribosomal Pautoantibodies also recognise epitopes onvarious human cells, for example humanhepatoma cells, human neuroblastoma cellsand, to a lesser extent, human fibroblasts.33The presence of antigen target on non-neuronal tissue may indicate a pattern ofsensitisation outside the nervous system;furthermore, autoantibodies can bind tosurface P antigen and affect the function ofcells. These autoantibodies have been found in90% of SLE patients with psychosis,26 34suggesting a high specificity. However, otherinvestigators have questioned the specificity, asthey found anti-P antibodies in 50% of SLEpatients without psychosis.35 For thoroughdiscussion of anti-P autoantibodies in CNSSLE, the reader is referred to an excellentrecent review.36 Antibodies to neurofilamentproteins, which form part of the cytoskeletonof neuronal cells, have been demonstrated tooccur in neuropsychiatric SLE but, again, arenot specific to this entity.37 They may alsooccur in other disorders such as Alzheimer'sdisease and Creutzfeldt-Jakob disease.

ANTIPHOSPHOLIPID ANTIBODIES

Antibodies to phospholipids comprise a diversegroup that includes the lupus anticoagulantand the anticardiolipin autoantibodies. It haslong been thought that antiphospholipid (aPL)antibodies were responsible for the non-inflammatory vasculopathy characterised bythickening of the walls of small arteries as aresult of intimal proliferation. The originaldescription of the lupus anticoagulant datesback to its observation in a group of lupuspatients who presented with stroke.39 Initial

Table 1 Neuronal antibodies in neuropsychiatnic systemic lupus erythematosus

Authors Year Assay Type ofNA CeUls kiled or absorption (%o) Clinical correlation antibody/CNS disease?NP-SLE Non-NP-SLE

Bluestein et al'9 1976 LC assay/absorption LCA 45 * 26 Yeswith human brain

Bresnihan et al 20 1977 LC assay/absorption LCA 55 * 30 Yeswith human brain

Winfield et al" 1977 LC assay LCA 50 NS 85 NoPussell et al22 1982 LC assay/absorption LCA 65 * 41 Yes

with human brainLong et al'6 1990 LC assay LCA 32 NS 23 Negative with overall NP-SLE;

positive with cognitive impairementDenburg et a]'4 1994 T lymphocyte line LCA NM NM Not with overall NP-SLE;

significantly with visuospatial deficitWilson et al23 1979 Neuroblastoma/MHA NA 75 45 IgG association with diffuse CNS diseaseBluestein et al'4 1981 Neuroblastoma NA 74 * 11 Association only for CSF antibodyGolombek et al'6 1986 Neuroblastoma NA 57 NS 50 No clinical correlationKelly et all' 1987 Neuroblastoma/MHA NA 17 NS 8 No clinical association

LC = Lymphocytotoxic; LCA = lymphocytotoxic antibodies; NA = neuronal antibodies; NM = not mentioned; NP = neuropsychiatric; MHA = mixedhaemabsorption assay.*Significant difference between the two groups; NS = non-significant difference.

161


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from

http://ard.bmj.com/

162

studies emphasised an association with focalneuropsychiatric SLE, such as stroke,amaurosis fugax,40 and transverse myelitis.4'Subsequent studies have found an associationwith more diffuse manifestations of cerebralSLE, such as seizures,42 chorea43 and multi-infarct dementia44 (but not with psychosis).The multiple features associated with aPL,

known as antiphospholipid antibody syndrome(APS) are thought to be the consequence ofsmall infarcts. Endothelial cell swelling,hyperplasia, mitotic figures, and a novelangiomatoid thrombovascular complex (notvasculitis) are the characteristic cerebral andrenal histopathological features of bothprimary and secondary APS.45 46 It has beenproposed that the endothelial swelling iscaused by the incorporation of fragments ofplatelet thrombi into the vascular endo-thelium46 a process hypothetically facilitatedby von Willebrand factor.47 Other plasmaproteins such as 32-glycoprotein I andprothrombin-Ca, which have a high affinity fornegatively charged phospholipids, may bind tothe lipid surface of endothelial cells, exposingnovel epitopes and perseverating the immuneresponse. This mechanism may activateplatelets, finally resulting in thrombosis.Repetitive episodes of thrombosis in thesesmall vessels might lead to incorporation ofplatelet fragments and lead to the thickeningand irregularity of small vessels.48 In long-standing neuropsychiatric SLE, small vesselabnormalities may lead to the microinfarctionsfound at postmortem.` 46 48

CYTOKINES

Cytokines (polypeptide products of activatedlymphocytes that participate in a variety ofcellular responses) may play a dual role inthe development or persistence of CNSinvolvement in SLE. Interferons mayincrease the severity of autoimmune diseasein animals,49 but may themselves beresponsible for psychosis, as has beendemonstrated in patients with hepatitis B.50Intrathecal synthesis of interferon alfa5' 52 andinterleukin-653 (a cytokine with B cell stimu-latory properties and interleukin- 1 likeproperties) has been reported in SLE patientswith CNS involvement. In the patients withpsychosis, interferon concentrations in CSFwere significantly greater than in SLE patientswith seizures alone.52 Despite these reportsoffering some circumstantial evidence thatcytokines may be involved in the pathogenesisof cerebral lupus, many questions remainconcerning the role of cytokines in lupus; it isstill not clear, for example, if cytokineconcentrations are increased during a diseaseflare. It is evident that the role of cytokines inlupus requires further study. 4

DiagnosisIn reviewing the extensive range ofimmunoserological, electrophysiological andneuroimaging techniques which have beeninvestigated as possible markers of CNS lupus,

it is apparent that there is still no single testwhich is diagnostic. Diagnosis of neuro-psychiatric SLE, therefore, is a matter ofclinical acumen. Selection of a test from thebroad array available should take into accountthat a positive case may, at best, be merelysupportive of the diagnosis, while a negativeresult does not necessarily preclude a diagnosisof cerebral lupus.Cerebrospinal fluid (CSF) examination. CSFabnormalities are commonly found but arenon-specific. Generally, both a modestpleocytosis and increased protein concen-trations are found. The foremost purpose ofthis procedure is to rule out CNS infection; itshould be borne in mind, however, that CNSpleocytosis may be absent in infection.Neuropsychometric testing (NPT). For mostclinicians, NPT has some obvious dis-advantages. It is not readily available, is time-consuming, and has no indication in an acutesituation as it requires a stable patient.Cognitive impairment, however, can beidentified and quantified in a standardised andrepeatable manner by NPT. In this way,Carbotte et al55 demonstrated that 66% of 62individuals from an unselected lupuspopulation had cognitive impairment. Notonly did more than 80% of patients withcurrent or past neuropsychiatric involvementhave abnormal neuropsychometric tests, butalso 42% of those with no prior CNSsymptoms were found to have cognitivedysfunction. Ginsburg et al56 reported on theprevalence of cognitive deficits in a group of 49randomly selected ambulatory patients withSLE and 40 with rheumatoid arthritis. Theyfound a significantly higher prevalence ofcognitive impairment amongst the SLEpatients. Subsequent studies57-5 reported asubstantially lower frequency of cognitivedysfunction (about 20%) in SLE patients,reflecting selection bias in the formerstudies.55 56Thus NPT appears to be not only a sensitive,

but also a convenient, inexpensive, and non-invasive diagnostic tool in the assessment ofcognitive dysfunction in lupus, especially inthat group who have vague symptoms and anabsence of clinical signs. Its main disadvantageis its susceptibility to learning effects, and theinfluence of various confounding factors suchas the influence of treatment, physical illness,and pschychological distress on cognitivefunction. In addition to aiding diagnosis,neuropsychiatric examination may help thelupus patient to understand better thosecognitive problems encountered at home, atwork, and in social situations; in this way, it canbe utilised to anticipate and address difficultiesin everyday functioning.Electroencephalography (EEG). EEG is of littlediagnostic value in neuropsychiatric lupus.The electroencephalographic findings arefrequently abnormal, showing diffuse waveslowing without any distinctive pattern, unlessfocal features are apparent clinically. Diag-nostic sensitivity has been reported variously asbetween 330/o and 850/o of SLE patients withCNS involvement; a figure for the diagnostic

Bruyn


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from

http://ard.bmj.com/


specificity is more difficult to define, as no largeprospective controlled studies have addressedthis question. The technique of quantativeEEG (QEEG) has been applied in aprospective study of 52 patients with SLE.60QEEG sensitivity appeared to be 87%, whereasthe specificity was 750/o. The technique provedto be particularly useful in the subgroup thatis the most difficult to assess, namely thosewith various forms of affective disorders anddepression in the absence of neurological signs.Since this technique is not available in themajority of hospitals, experience is limited.Angiography. Cerebral angiography is aninsensitive diagnostic technique in CNS lupus,except where symptoms or signs of majorvessel haemorrhage, thrombosis, or vasculitisdevelop.6' The procedure may be helpfulbefore operation when cerebral bleeding hasoccurred. As angiography is an invasivetechnique with low sensitivity, it is not usedroutinely in the assessment of neuropsychiatricSLE.Computed tomography (CT). The first CTscanning studies in cerebral lupus wereconducted in the late 1 970s and early1980s.62 64 CT is most useful in detecting grossmorphological abnormalities such as largeintracranial infarcts, intracranial haemorr-hages, and cerebral atrophy. It is unclear if thecerebral atrophy is caused by the disease itselfor by glucocorticosteroid therapy. As thepredominant neuropathology in neuro-psychiatric lupus is that of microinfarction, CTis frequently normal, even in the presence offocal syndromes such as cerebrovascularaccidents or focal seizures.65Magnetic resonance imaging (MRI). MRI ismore sensitive than CT in demonstratingparticular morphological abnormalities such asoedema and small infarcts.65-70 Threepredominant patterns of MRI abnormalitiesare recognised: irreversible high intensitylesions compatible with cerebral infarction,multiple hyperintense areas of the whitematter, probably representing microinfarctionsand which are associated with aPL,69 70 andfocal hyperintensity of the grey matter. It isnoteworthy that the grey matter abnormalitiesmay be reversible with clinical improve-ment,67 68 suggesting resolution of oedema. Inone study of 28 SLE patients with acute CNSevents, cranial MRI scans were obtained withina month from the onset of symptoms.68 In allthose patients with focal symptoms, MRI wasable to detect cerebral lesions; CT scanningdetected the lesions in only 38% of them. MRIshowed lesions in 1 00/o of episodes of seizures,but in patients demonstrating organic brainsyndromes or psychosis it had a sensitivity ofonly 20%.Compared with CT, MRI is much the more

sensitive method of detecting focal lesions inthe brain of symptomatic SLE patients.Limitations are the poor specificity of thelesions found on MRI-one cannot make thediagnosis of cerebral SLE from radiographicfindings alone. This may pose a problem indifferentiating some white matter lesions seenin SLE patients from those found in patients

with ischaemic cerebral disease or multiplesclerosis.Radionuclide scanning. After the report of90-100% abnormal technetium brain scans ina group of lupus patients with active CNSdisease,7' this technique subsequently becameobsolete until recently, when single photonemission computed tomography (SPECT)proved worthwhile in cerebral lupus.72 Ina study of 20 patients with SLE whowere clinically suspected to have 'active'neurological involvement, Nossent et al founda sensitivity of almost 90%, but a lowspecificity of 330/o.73 More recently Rubbertet al showed an abnormal cortical perfusion in90% of patients with mild symptoms of CNSdisease (cognitive deficits) compared with 10%of SLE patients without CNS symptoms.74These studies support the notion that, pendingvalidation, SPECT may become an objectivetool in differentiating mild CNS disease inlupus from functional conditions such asdepression.Positron emission tomography (PET). PETdetects reduced cerebral metabolism withthe use of radiolabelled deoxyglucose. PETscanning has been able to detect abnormalitiesof glucose uptake and a localised disturbanceof cerebral blood flow in active neuro-psychiatric lupus.75 76 Two patterns arerecognised: focal glucose hypometabolism,and a diffuse non-homogeneous distributionpattern of increased and decreased glucosemetabolism. In a study of 10 SLE patients withneurological symptoms and three patients withno neurological involvement, PET scanningshowed abnormalities in all 10 with activeinvolvement, but was normal in the three otherpatients; the pathological changes on PETcorresponded with clinical features.76 BecausePET scanning requires the proximity of acyclotron, its clinical availability will remainlimited, and it will therefore not easily find aplace in the clinical management of lupus.

ManagementThe management of cerebral lupus requiresthat the presence of cerebral lupus isestablished and its severity assessed.Sometimes forgotten is the fact that the mostfrequent cause of central nervous systemmanifestations in lupus is not the lupus itself,but infection. Infection of the CNS should bevigorously sought and treated in all cases. Inaddition, factors such as hypertension shouldbe under control. The ability unequivocally toestablish the presence and severity of cerebrallupus remains a mission yet unresolved by therheumatology community. Because the revisedARA criteria77 are largely inadequate indefining CNS SLE, investigators have concen-trated on standardising a classification andsystem of definition of neurological events.78However, until strict criteria for recognition ofCNS SLE are laid down, reports correlatingclinical findings with pathology, serologicalparameters, and radiology may be biased andunreliable, and it will not be possible to testtreatment modalities rigorously.

163


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from

http://ard.bmj.com/

Bruyn

It is not surprising, therefore, thatuncertainty and controversy dominate the fieldof management of cerebral lupus. Should anSLE patient presenting with stroke be treatedin a totally different way than a patient withorganic brain syndrome? Merely for practicalreasons, it might be sensible to classify CNSSLE into separate groups with variousmanifestations.6No controlled study of any treatment

strategy has so far been conducted inneuropsychiatric lupus. The principles oftreatment of cerebral lupus lupus are triple:anticoagulants, corticosteroids, and immuno-suppressives (table 2). Anticoagulants shouldbe given for manifestations associated withhypercoagulability. Steroids are administeredto reduce inflammation and the immuneresponse. Finally, immunosuppressives, forexample cyclophosphamide and azathioprine,suppress circulatory B and T lymphocytes, andautoantibody formation by B lineage cells inlymphoid tissues.Anticoagulation for whom? Since the elucidationof the role of antiphospholipid antibodies inthrombotic manifestations in SLE patients,39 41growing agreement on anticoagulation hasemerged steadily. Lupus patients simply withpresence of these antibodies are currently notconsidered to be at high risk of developingthrombotic events and do not receiveprophylactic anticoagulation. Patients withSLE and a history of bland stroke, however,have a 69% chance of a recurrent stroke.79Therefore, recurrent cerebral thromboticevents-whether associated with antiphos-pholipid antibodies or not-are now importantindications for initiating anticoagulationtherapy.79 80 In all cases of stroke, cardiogenicemboli should be considered as possible causesas they form a major cause of stroke in lupus.Although consensus has not yet been reachedand clinical experience is largely lacking, othernervous system manifestations associatedwith antiphospholipid antibodies (seizures,42chorea,43 migraine,80 or transverse myelitis4' 81)may become important treatment indicationsfor anticoagulant therapy in the very nearfuture. It is not known if patients with organicbrain syndrome may benefit from anti-coagulation; as their prognosis is generallyfavourable,59 administration of any medication,including anticoagulants, is not imperative.The best mode and duration of anti-

coagulation is yet unknown. Possible optionsinclude low dose aspirin (75 mg daily), war-farin, or subcutaneous heparin. In addition,hydroxychloroquine has been shown to possess

Table 2 Uncertainties of therapy in neuropsychiatric lupus erythematosus

Feature Anticoagulation Corticosteroids Cyclophosphamide

OBS + +Seizures ++ ++ ?Acute psychosis ++ ++Coma ++ ++Stroke (bland) ++ +Stroke (bloody)Chorea ++ ++Transverse myelitis + ++ ++Migraine + ?

OBS = Organic brain syndrome; ++ =worthwhile therapy; + =probably worthwhile therapy;? = uncertain effect; -= probably not effective or contraindicated.

anticoagulant properties. Antiplatelet coagu-lants such as aspirin may be ineffective inprevention of recurrent thrombosis in patientswith antiphospholipid antibodies, as retro-spective studies have shown that warfarin orcoumarin is more effective than aspirintherapy;82 83 data on low molecular weightheparin injections were too limited for anyconclusions to be drawn.82 Therefore, inpatients with major thrombosis such as stroke,warfarin is currently the preferred choice oftreatment; as an additional advantage, it iseffective against cardiogenic emboli. In orderto be effective, however, warfarin therapyshould be highly intensive (at an internationalnormalised ratio of between 2-5 and 4 0); it isobvious that such an anticoagulation level isnot without risk for major bleeding. Durationof therapy should be lifelong, as recurrence ofthrombosis may occur after cessation of theanticoagulation.83

Antimalarial drugs also have an anti-coagulant effect. Withdrawal of antimalarialtreatment was studied in patients with SLE ina randomised, placebo controlled clinicaltrial.84 Five of the patients (23%) takingplacebo and one of the patients (4%)continuing to take hydroxychloroquine hadsevere exacerbations of disease activity thatprompted their withdrawal from the study.Amongst the exacerbations in the placebogroup, one patient developed transversemyelitis requiring treatment with high dosecorticosteroids. Such observations may focusattention on the place of antimalarials in theorder of choice of therapy, which shouldperhaps be reconsidered in the future.Corticosteroids form a mainstay in the treatmentof SLE, although this dogma itself is largelyempirical. In all cases in which antineuronalantibodies may play a role (cognitive dys-function and psychosis) its use should beconsidered. Patients with vasculitis should alsobe treated with corticosteroids, but nervoussystem vasculitis occurs infrequently. Assteroids reduce oedema around infarcted areas,patients with acute bland stroke may improvewith steroids, but the administration of corti-costeroids for the prevention of stroke isjustified only when there is sufficient systemicactivity,79 as they do not predictably decreaseantiphospholipid antibody concentrations inplasma, nor do they protect against subsequentstrokes.6 79 High dose steroids are indicated incases of coma, seizures, psychosis, chorea, andtransverse myelitis, but even in these seriousconditions the effect of corticosteroid therapyis unpredictable, as controlled studies havenot been performed. The starting dose ofcorticosteroids in cerebral lupus is also notknown, but according to most authors isprobably high (. 1 mg/kg per day for four to sixweeks before gradual reduction). Intravenouspulse therapy has been extensively used inthe treatment of lupus since the originaldescription of its use in SLE nephritis.85 Therehave been several reports of the use of pulsesteroid therapy in the treatment of non-renalSLE, including neuropsychiatric lupus.8"89Eyanson et al reported two acutely ill patients

164


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from

http://ard.bmj.com/


who failed to respond to oral glucocorticoidtherapy, but who responded dramaticallyto intravenous methylprednisolone pulsetherapy.86 Dutton et al reported three womenwith acute retrobulbar optic neuritis whoresponded to intravenous methylprednisolonetherapy;87 the diagnosis of SLE, however, wasnot clearly established. More recently, newperspectives have arisen for the treatment ofthe rare complication, transverse myelitis, bypulse steroids followed by cyclophosphamidefor a mean of six months.88

Controversy attends the possible role ofcorticosteroids in the induction of psychosisand the facilitation of infection. Sergent et a189retrospectively examined the effect ofcorticosteroids in 28 patients with 52 episodesof nervous system manifestations and foundthat 14 patients had 15 episodes of functionalpsychosis, possibly attributable to treatment.Five of the nine deaths in that study wereattributable to infection. Since the patientsreceived mean prednisone doses of more than100 mg/day for prolonged periods, side effectsshould be anticipated. Glucocorticoids are nowprescribed in much smaller doses thanformerly used.7 In the study by O'Connor,90of 11 previously psychotic patients whosubsequently received steroids, only three hada recurrence of their psychoses. The eightpatients who had no further psychotic episodesreceived essentially the same amount ofcorticosteroids as they had had at the time oftheir psychoses. That paper clearly showed thatlupus patients are particularly prone to developpsychotic episodes independently of steroidtherapy, although the latter may play somepart.Cytotoxic agents have been extensively studiedin renal lupus during the past two decades.Treatment comprising oral azathioprine incombination with prednisone was better thanprednisone alone in the management of lupusnephritis. No data exist on this drug in themanagement of CNS lupus. Another agent,cyclophosphamide, has also proved to beeffective in lupus nephritis in combination withprednisone. In one open prospective trial, theeffect of monthly intravenous infusions ofcyclophosphamide to nine patients with lupusnephritis, cerebral lupus, or both, wasstudied;9' although CNS disease improved ina small number of patients, there wereinsufficient data to permit conclusions aboutthe treatment of the neurological disease.Another study was of refractory CNS lupus innine patients, six of whom had beenunresponsive to high dose steroid therapy.92Six patients had a complete and two a partialrecovery; side effects were common, however.From these studies and others,7 and fromunpublished personal observations, it seemsthat intravenous pulse cyclophosphamidetherapy may be the best therapy for lupuspatients with progressive deterioration in levelof consciousness,7 acute psychosis or coma9"(GRV Hughes, personal communication).Furthermore, in combination with intravenouspulse methylprednisolone it is the treatment ofchoice for transverse myelitis.88

Experimental therapies, including low dosemethotrexate, cyclosporin, plasmapheresis,total lymphoid irradiation, anti-idiotypicantibodies, and intravenous IgG are the subjectof too limited experience for it to be possibleto recommend them at the present time.

SummaryWhat have we learned about CNS lupus inrecent years? An enormous amount ofknowledge on pathophysiology of anti-phospholipid antibodies, in particular, hasbeen gathered. Although hard evidence of adirect pathogenetic role of these antibodies incerebral lupus is still lacking, it is generally feltthat the multiple microinfarctions found in thebrains of lupus patients are related to theirpresence. Better understanding of thepathogenesis of cerebral lupus will come fromthe study of experimental models, as it hasbeen possible to develop an antiphospholipidantibody syndrome in mice.93Because no specific laboratory test for CNS

lupus is yet available, diagnosing the conditionremains a challenge to every clinician. Tech-niques including neuropsychometric testing,quantitative EEG, and SPECT scans havetaught us more about cognitive dysfunctionand psychosis in patients with SLE. Thesecategories remain the most difficult to define.The concept of hypercoagulability in SLE

patients has diverted the direction of therapyfrom immunosuppression towards anti-coagulation. It is of utmost importance thatrandomised trials are commenced in orderto determine the optimal mode of anti-coagulation for various groups of lupuspatients. It will be necessary to conduct suchtrials under strict inclusion criteria, based onwell defined patient categories. Such an enter-prise will require international co-operation ofinvestigators.

The specimens for microscopy were kindly provided byProf F Eulderink, Department of Pathology, AcademischZiekenhuis Leiden.

1 Johnson R T, Richardson E P. The neurological manifes-tations of systemic lupus erythematosus. Medicine(Baltimore) 1968; 47: 337-69.

2 Ellis S G, Verity M A. Central nervous system involvementin systemic lupus erythematosus: a review of neuro-pathologic findings: 57 cases 1955-1977. Semin ArthritisRheum 1979; 8: 212-21.

3 Hanly J G, Walsh N M G, Sangalang V. Brain pathologyin systemic lupus erythematosus. Jf Rheumatol 1992; 19:732-41.

4 Smith R W, Ellison D W, Jenkins E A, Gallagher P J,Cawley M I D. Cerebellum and brainstem vasculopathyin systemic lupus erythematosus: two clinico-pathologicalcases. Ann Rheum Dis 1994; 53: 327-30.

5 Mitsias P, Levine S R. Large cerebral vessel occlusivedisease in systemic lupus erythematosus. Neurology 1994;44: 385-93.

6 Futrell N, Schultz L R, Millikan C. Central nervous systemdisease in patients with systemic lupus erythematosus.Neurology 1992; 42: 1649-57.

7 Sibley J T, Olszynski W P, Decoteau W E, Sundaram M B.The incidence and prognosis of central nervous systemdisease in systemic lupus erythematosus. J Rheumatol1992; 19: 47-52.

8 HochbergM C, Boyd R E, Ahearn J M, et al. Systemic lupuserythematosus: a review of clinico-laboratory features andimmunogenetic markers in 150 patients with emphasis ondemographic subsets. Medicine (Baltimore) 1985; 64:285-95.

9 Jongen P J H, Boerbooms A M T, Lamers K J B, Raes B C,Vierwinden G. Diffuse CNS involvement in systemiclupus erythematosus: Intrathecal synthesis of the 4thcomponent of complement. Neurology 1990; 40: 1593-6.

165


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from

http://ard.bmj.com/

Bruyn

10 Hadler N M, Gerwin R D, Frank M M, Whitaker J N,Baker M, Decker J L. The fourth component ofcomplement in the cerebrospinal fluid in SLE. ArthritisRheum 1973; 16: 507-21.

11 Winfield J B, Brunner C M, Koffler D. Serologic studies inpatients with systemic lupus erythematosus and centralnervous system dysfunction. Arthritis Rheum 1978; 21:289-94.

12 Hang L M, Izui S, Dixon F J. (NZBxBXSB)Fl hybrid: amodel of acute lupus and coronary vascular disease withmyocardial infarction. J Exp Med 1981; 36: 154-6.

13 Atkins C J, Kondon J J, Quismorio F P, Friou G J. Thechoroid plexus in systemic lupus erythematosus. AnnIntern Med 1972; 76: 65-72.

14 Boyer R S, Sun N C J, Verity A, Nies K M, Louie J S.Immunoperoxidase staining of the choroid plexus insystemic lupus erythematosus. J Rheumatol 1980; 7:645-50.

15 Harbeck R J, Hoffman A A, Hoffman S A, Shucard D W.Cerebrospinal fluid and the choroid plexus during acuteimmune complex disease. Clin Immunol Immunopath1979; 13: 413-25.

16 Long A A, Denburg S D, Carbotte R M, Singal D P,Denburg J A. Serum lymphocytotoxic antibodies andneurocognitive function in systemic lupus erythematosus.Ann Rheum Dis 1990; 49: 249-53.

17 Karpiak S E, Graf L, Rapport M M. Antiserum to braingangliosides produces recurrent epileptiform activity.Science 1976; 194: 735-7.

18 Karpiak S E, Serokosz M, RapportM M. Effects of antiserato S-100 protein and to synaptic membrane fraction inmaze performance and EEG. Brain Res 1976; 102:313-21.

19 Bluestein H G, ZvaiflerN J. Brain-reactive lymphocytotoxicantibodies in the serum of patients with systemic lupuserythematosus. J Clin Invest 1976; 57: 509-16.

20 Bresnihan B, Oliver M, Grigor G, Hughes G R V. Brainreactivity of lymphocytotoxic antibodies in systemic lupuserythematosus with and without cerebral involvement.Clin Exp Immunol 1977; 30: 333-7.

21 Bresnihan B, Oliver M, Williams B, Hughes G R V. Anantineuronal antibody cross-reacting with erythrocytesand lymphocytes in systemic lupus erythematosus.Arthritis Rheum 1979; 22: 313-20.

22 Pussell B A, Blyth F, Charlesworth J A. Failure to detectbrain reactivity of lymphocytotoxins in cerebral lupus.Clin Exp Immunol 1982; 47: 133-7.

23 Wilson H A, Winfield J B, Lahita R G, Koffier D.Association of IgG anti-brain antibodies with centralnervous dysfunction in systemic lupus erythematosus.Arthritis Rheum 1979; 22: 458-62.

24 Bluestein H G, Williams G W, Steinberg A D. Cerebro-spinal fluid antibodies to neuronal cells: association withneuropsychiatric manifestations of systemic lupuserythematosus. AmJ Med 1981; 70: 240-6.

25 Kelly M C, Denburg J A. Cerebrospinal fluid immuno-globulins and neuronal antibodies in neuropsychiatricsystemic lupus erythematosus and related conditions.JrRheumatol 1987; 14: 740-4.

26 Golombek S J, Graus F, Elkon K B. Autoantibodies in thecerebrospinal fluid of patients with systemic lupuserythematosus. Arthritis Rheum 1986; 29: 1090-7.

27 Hanly J G, WalshN M, Fisk J D, et al. Cognitive impairmentand autoantibodies in systemic lupus erythematosus. BrJRheumatol 1993; 32: 291-9.

28 Denburg S D, Behmann S A, Carbotte R M, Denburg J A.Lymphocyte antigens in neuropsychiatric systemic lupuserythematosus. Relationship of lymphocyte antibodyspecificities to clinical disease. Arthritis Rheum 1994; 37:369-75.

29 How A, Dent P B, Liao S-K, Denburg J A. Antineuronalantibodies in neuropsychiatric systemic lupuserythematosus. Arthritis Rheum 1985; 28: 789-95.

30 Hanly J G, Rajaraman S, Behmann S, Denburg J A. A novelneuronal antigen identified by sera from patients withsystemic lupus erythematosus. Arthritis Rheum 1988; 31:1492-9.

31 Hanson V G, Horowitz M, Rosenbluth D, Spiera H,Puszkin S. Systemic lupus erythematosus patients withcentral nervous system involvement show autoantibodiesto a 50-kD neuronal membrane protein. J Exp Med 1992;176: 565-73.

32 Hanly J G, Hong C. Antibodies to brain integral membraneproteins in systemic lupus erythematosus.3'ImmunolMeth1993; 161: 107-18.

33 Koren E, Reichlin M W, Koscec M, Fugate R, Reichlin M.Autoantibodies to the ribosomal P proteins react with aplasma membrane-related target on human cells. I ClinInvest 1992; 89: 1236-41.

34 Bonfa E, Golombek S J, Kaufinan L D, et al. Associationbetween lupus psychosis and anti-ribosomal P proteinantibodies. NEnglJMed 1987; 317: 265-71.

35 Van Dam A, Nossent H, De Jong J, et al. Diagnostic valueof antibodies against ribosomal phosphoproteins. A crosssectional and longitudinal study. J Rheumatol 1991; 18:1026-34.

36 Teh L, Isenberg D A. Antiribosomal P protein antibodiesin systemic lupus erythematosus. Arthritis Rheum 1994;37: 307-15.

37 Robbins M L, Komguth S E, Bell C L, et al. Anti-neurofilament antibody evaluation in neuropsychiatricsystemic lupus erythematosus. Arthritis Rheum 1988; 31:623-8.

38 Mevorach D, Raz E, Shalev 0, Steiner I, Ben-Chetrit E.Complete heart block and seizures in an adult with

systemic lupus erythematosus. Arthritis Rheum 1993; 36:259-62.

39 Boey M L, Colaco M B, Gharavi A E, Elkon K B,Loizou S, Hughes G R V. Thrombosis in systemic lupuserythematosus: striking association with the presence ofcirculating lupus anticoagulant. BMJ 1983; 287: 1021-3.

40 Digre K B, Durcan F J, Branch D W, Jacobson D M,Varner M W, Baringer J R. Amaurosis fugax associatedwith antiphospholipid antibodies. Ann Neurol 1989; 25:228-32.

41 Alarcon-Segovia D, Deleze M, Oria C, et al. Anti-phospholipid antibodies and the antiphospholipidsyndrome in systemic lupus erythematosus. A prospectiveanalysis of 500 consecutive cases. Medicine (Baltimore)1989; 68: 353-65.

42 Herranz M T, Rivier G, Khamashta M A, Blasrer K U,Hughes G R V. Association between antiphospholipidantibodies and epilepsy in patients with systemic lupuserythematosus. Arthritis Rheum 1994; 4: 568-7 1.

43 Asherson R A, Derksen RHW M, Harris E N, et al. Choreain systemic lupus erythematosus and lupus-like disease:association with antiphospholipid antibodies. SeminArthritis Rheum 1987; 16: 253-9.

44 Asherson R A, Mercy D, Philipis G, et al. Recurrent strokeand multi-infarct dementia in SLE associated withantiphospholipid antibodies. Ann Rheum Dis 1987; 46:605-11.

45 Westerman E M, Miles J M, Backonja M, Sandstrom W R.Neuropathologic findings in multiinfarct dementiaassociated with aCL. Arthritis Rheum 1992; 35: 1038-41.

46 Hughson M D, McCarty G A, Sholer C, Brumback M D.Thrombotic cerebral arteriopathy in patients with theantiphospholipid syndrome. Mod Pathol 1993; 6: 644-53.

47 Lindsey N J, Dawson R A, Henderson F I, Greaves M,Hughes G R V. Stimulation of von Willebrand factorantigen release by immunoglobulin from thrombosisprone patients with systemic lupus erythematosus and theanti-phospholipid syndrome. J Rheumatol 1993; 32:123-6.

48 Ellison D, Gatter K, Heryet A, Esiri M. Intramural plateletdeposition in cerebral vasculopathy of systemic lupuserythematosus. J Clin Pathol 1993; 46: 37-40.

49 Adam C, Thoua Y, Ronco P, Verroust P, Tovey M,Morel-Marover L. The effect of exogenous interferon:acceleration of autoimmune and renal disease in NZB/WFl mice. Clin Exp Immunol 1980- 40: 373-82.

50 McDonald E M, Mann A H, Thomas H C. Interferon asmediators of psychiatric morbidity: an investigation in atrial of recombinant alpha-interferon in hepatitis-Bcarriers. Lancet 1987; ii: 1175-7.

51 Lebon P, Lenoir G R, Fischer A, Lagrue A. Synthesis ofintrathecal interferon in systemic lupus erythematosuswith neurological complications. BMJ 1983; 287:1165-7.

52 Shiozawa S, Kuroki Y, Kim M, Hirohata S, Ogijno T.Interferon-alpha in lupus psychosis. Arthritis Rheum 1992;35: 417-22.

53 Hirohata S, Miyamoto T. Elevated levels of interleukin-6 incerebrospinal fluid from patients with systemic lupuserythematosus and central nervous system involvement.Arthritis Rheum 1990; 33: 644-9.

54 Gordon C, Emery P. Cytokines and the acute phaseresponse in SLE. Lupus 1993; 2: 345-7.

55 Carbotte R M, Denburg S D, Denburg J A. Prevalence ofcognitive impairment in systemic lupus erythematosus.J NervMentDis 1986; 174: 357-64.

56 Ginsburg K S, Wright E A, LarsonM G, et al. A controlledstudy of the prevalence of cognitive dysfunction inrandomly selected patients with systemic lupuserythematosus. Arthritis Rheum 1992; 35: 776-82.

57 Hanly J G, Fisk J D, Sherwood G, Jones E, Verrier Jones J,Eastwood B. Cognitive impairment in patients withsystemic lupus erythematosus. J7 Rheumatol 1992; 19:562-7.

58 Hay E M, Black D, Huddy A, et al. Psychiatric disorder andcognitive impairment in systemic lupus erythematosus.Arthritis Rheum 1992; 35: 411-6.

59 Hay E M, Huddy A, Black D, et al. A prospective study ofpsychiatric disorder and cognitive function in systemiclupus erythematosus. Ann Rheum Dis 1994; 53: 298-303.

60 Ritchlin C T, Chabot R J, Alper K, et al. Quantativeelectroencephalography. A new approach to the diagnosisof cerebral dysfunction in systemic lupus erythematosus.Arthritis Rheum 1992; 35: 1330-42.

61 Sanders E A J M, Hogenhuis L A H. Cerebral vasculitis aspresenting symptom of systemic lupus erythematosus.Acta Neurol Scand 1986; 74: 75-7.

62 Bilaniuk L T, Patel S, Zimmerman R A. Computedtomography of systemic lupus erythematosus. Radiology1977; 124: 119-21.

63 Gonzalez-Scarano F, Lisak L B, Bilaniuk L T,Zimmerman R A, Arkins P C, Zweiman B. Cranialcomputed tomography in the diagnosis of systemic lupuserythematosus. Ann Neurol 1979; 5: 158-65.

64 Carette S, Urowitz M B, Grosman H, St Louis E.Cranial computerized tomography in systemic lupuserythematosus. J'Rheumatol 1992; 9: 855-9.

65 Vermess M, Bernstein R M, Bydder G M, Steiner R E,Young I R, Hughes G R V. Nuclear magnetic resonance(NMR) imaging of the brain in systemic lupuserythematosus. 7 Comput Assist Tomogr 1983; 7: 461-7.

66 Sewell K L, Linch A, Aranow C B, Grayzel A I. Magneticresonance imaging versus computed tomographicscanning in neuropsychiatric SLE. Am J Med 1989; 86:625-6.

166


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from

http://ard.bmj.com/


67 Sibbitt W L, Sibbitt R R, Griffey R H, Eckel C,Bankhurst A D. Magnetic resonance and computedtomographic imaging in the evaluation of acuteneuropsychiatric disease in systemic lupus erythematosus.Ann Rheum Dis 1989; 48: 1014-22.

68 McCune W J, MacGuire A, Aisen A, Gebarski S.Identification ofbrain lesions in neuropsychiatric systemiclupus erythematosus by magnetic resonance imaging.Arthritis Rheum 1988; 31: 159-66.

69 Molad Y, Sidi Y, Gornish M, Lerner M, Pinkhas J,Weinberger A. Lupus anticoagulant: correlation withmagnetic resonance imaging of brain lesions. J Rheumatol1992; 19: 556-61.

70 Ishikawa S, Ohnishi K, Miyachi Y, Ishizaka H. Cerebrallesions in systemic lupus erythematosus by magneticresonance imaging. Relationship to anticardiolipin anti-body.JRheumatol 1994; 21: 87-90.

71 Bennahum D A, Messner R P, Shoop J D. Brain scanfindings in central nervous system involvement bysystemic lupus erythematosus. Ann Intern Med 1974; 81:763-5.

72 Kushner M J, Tobin M, Fazekas F, et al. Cerebral bloodflow variations in CNS lupus. Neurology 1990; 40:99-102.

73 Nossent J C, Hovestadt A, Schonfeld D H W, Swaak A J G.Single-photon-emission computed tomography of thebrain in the evaluation of cerebral lupus. Arthritis Rheum1991; 34: 1397-403.

74 Rubbert A, Marienhagen J, Pirner K, et al. Single-photon-emission computed tomography analysis of cerebral bloodflow in the evaluation of central nervous system involve-ment in patients with systemic lupus erythematosus.Arthritis Rheum 1993; 36: 1253-62.

75 Hiraiwa M, Nonaka C, Toshiali A, Masaaki I. PET insystemic lupus erythematosus: relation of cerebralvasculitis to PET findings. Am J Neuroradiol 1983; 4:541-3.

76 Stoppe G, Wildhagen K, Seidel J W, et al. Positron emissiontomography in neuropsychiatric lupus erythematosus.Neurology 1990; 40: 304-6.

77 Tan E M, Cohen A S, Fries J F, et al. The 1982 revisedcriteria for the classification of systemic lupus erythe-matosus. Arthritis Rheum 1982; 25: 1271-7.

78 Singer J, Denburg J A. Diagnostic criteria for neuro-psychiatric systemic lupus erythematosus: the results of aconsensus meeting. J Rheumatol 1990; 17: 1397-402.

79 Futrell N, Millikan C. Frequency, etiology, and preventionof stroke in patients with systemic lupus erythematosus.Stroke 1989; 20: 583-91.

80 Asherson R A, Khamashta M A, Gil A, et al. Cerebro-vascular disease and antiphospholipid antibodies insystemic lupus erythematosus, lupus-like disease, and theprimary antiphospholipid syndrome. AmJ Med 1989; 86:391-9.

81 Lavalle C, Pizarro S, Drenkard C, Sanchez-Guerrero J,Alarqon-Segovia D. Transverse myelitis: a manifestationof SLE associated with antiphospholipid antibodies.JRheumatol 1990; 17: 34-7.

82 Rosove M H, Brewer PM C. Antiphospholipid thrombosis:clinical course after the first thrombotic event in 70patients. Ann Intern Med 1992; 117: 303-8.

83 Derksen R H W M, De Groot P G, Kater L,Nieuwenhuis H K. Patients with antiphospholipid anti-bodies and venous thrombosis should receive long ternanticoagulant treatment. Ann Rheum Dis 1993; 52: 689-92.

84 Canadian Hydroxychloroquine Study Group. Arandomized study of the effect of withdrawing hydroxy-chloroquine sulfate in systemic lupus erythematosus.N EnglJMed 1991; 324: 150-4.

85 Cathcart E S, Idelson B A, Scheinberg M A, Couser W G.Beneficial effects of methylprednisolone pulse therapy indiffuse proliferative nephritis. Lancet 1976; i: 163-6.

86 Eyanson S, Passo M H, Aldo-Benson M A, Benson M D.Methylprednisolone pulse therapy for nonrenal lupuserythematosus. Ann Rheum Dis 1980; 39: 377-80.

87 Dutton J J, Burde R M, Klingele T G. Autoimmuneretrobulbic optic neuritis. Am J Opthalmol 1982; 94:11-4.

88 Barile L, Lavalle C. Transverse myelitis in systemic lupuserythematosus-the effect ofIV pulse methylprednisoloneand cyclophosphamide. J7 Rheumatol 1992; 19: 370-2.

89 Sergent J S, Lockshin M D, Klempner M S, Lipsky B A.Central nervous system disease in systemic lupuserythematosus. AmjMed 1975; 58: 644-55.

90 O'Connor J F. Psychoses associated with systemic lupuserythematosus. Ann Intern Med 1959; 51: 526-36.

91 McCune W J, Golbus J, Zeldes W, Bohlke P, Dunne R,Fox D A. Clinical and immunological effects of monthlyadministration of intravenous cyclophosphamide insevere systemic lupus erythematosus. N EnglJ Med 1988;318: 1423-31.

92 Boumpas D T, Yamada H, Patronas N J, Klippel J H,Balow J E. Pulse cyclophosphamide for severe neuro-psychiatric lupus. QJ7Med 1991; 296: 975-84.

93 Smith H R, Hansen C L, Rose R, Canoso R T.Autoimmune MRL-lpr/lpr are an animal model for thesecondary antiphospholipid syndrome. J7 Rheumatol 1990;17: 911-5.

167


http://ard.bmj.com

/A

nn Rheum


arch 1995. Dow

nloaded from
http://ard.bmj.com/

controversies in lupus: nervous system involvementbluestein et al'9 1976 lcassay/absorption lca 45 *...

Documents