We considered 3 hypotheses for this study. The shaded areas represent patients with optic neuritis and/or myelitis. NMO indicates neuromyelitis optica; SLE, systemic lupus erythematosus; and SS, Sjögren syndrome.
Distribution of neuromyelitis optica (NMO)–IgG titers in the patients' serum samples. Titers were measured as the reciprocal of end point doubling dilutions. Serum samples represented below the horizontal line were negative for NMO-IgG. The lowest positive value is 1 in 60; the highest, 1 in 61 440. Ab indicates antibody; NMOSDs, NMO spectrum disorders; NOS, non–organ specific; SLE, systemic lupus erythematosus; and SS, Sjögren syndrome.
Pittock SJ, Lennon VA, de Seze J, Vermersch P, Homburger HA, Wingerchuk DM, Lucchinetti CF, Zéphir H, Moder K, Weinshenker BG. Neuromyelitis Optica and Non–Organ-Specific Autoimmunity. Arch Neurol. 2008;65(1):78-83. doi:10.1001/archneurol.2007.17
Neuromyelitis optica (NMO) is often associated with other clinical or serological markers of non–organ-specific autoimmunity.
To evaluate the relationship between NMO spectrum disorders (NMOSDs), including NMO, longitudinally extensive transverse myelitis, and recurrent optic neuritis, and autoimmune disease. We concentrated on the association with systemic lupus erythematosus (SLE), Sjögren syndrome (SS), or serological evidence of these disorders, which commonly is a source of diagnostic confusion.
Retrospective blinded serological survey.
Mayo Clinic College of Medicine, Rochester, and Centre Hospitalier Régional Universitaire de Lille.
Group 1 included 153 US patients with NMOSDs (78 with NMO and 75 with longitudinally extensive transverse myelitis) and 33 control subjects with SS/SLE. Group 2 included 30 French patients with SS/SLE, 14 with NMOSDs (6 with NMO, 6 with longitudinally extensive transverse myelitis, and 2 with recurrent optic neuritis), 16 without NMOSDs, and 4 with NMO without SS/SLE.
For group 1, NMO-IgG was detected in 66.7%, antinuclear antibodies in 43.8%, and Sjögren syndrome A (SSA) antibodies in 15.7% of patients with NMO and longitudinally extensive transverse myelitis. Five NMO-IgG–seropositive patients with NMOSDs had coexisting SLE, SS, or both. Antinuclear antibodies and SSA antibodies were more frequent in NMO-IgG–seropositive patients than in NMO-IgG–seronegative patients (P = .001). For group 2, NMO-IgG was detected in 5 of 14 patients (35.7%) with NMOSDs and SS/SLE and in 2 of 4 patients (50.0%) with NMO without SS/SLE (P = .59). We detected NMO-IgG only in patients with NMOSDs and not in 49 controls with SS/SLE but without optic neuritis or myelitis from the 2 cohorts (P = .01).
Neuromyelitis optica spectrum disorders with seropositive findings for NMO-IgG occurring with SS/SLE or non–organ-specific autoantibodies is an indication of coexisting NMO rather than a vasculopathic or other complication of SS/SLE.
Neuromyelitis optica (NMO) is characterized by optic neuritis and longitudinally extensive transverse myelitis (LETM, defined by a lesion extending 3 vertebral segments or longer on T2-weighted magnetic resonance images of the spine during an acute episode of myelitis). The course of NMO is relapsing in most cases.1 The serum autoantibody NMO-IgG is a sensitive and specific marker for NMO.2 The association between NMO and NMO-IgG has recently been confirmed by an independent European group.3 Its antigen is aquaporin-4 (AQP4), the predominant water channel protein in the central nervous system.4 The autoantibody NMO-IgG is detectable also in patients with recurrent optic neuritis without myelitis and in a high proportion of patients with a single episode or recurrence of LETM without optic neuritis.2 Detection of NMO-IgG in patients with a first LETM event predicts that LETM will relapse or that optic neuritis will develop in more than 50% of cases within 12 months.5 Brain lesions may also occur in patients with NMO, with or without symptoms, and in some cases may be hard to distinguish from those occurring in multiple sclerosis, although some hypothalamic and brainstem lesions are relatively specific for NMO.6
Patients with NMO often have an accompanying autoimmune disease, most commonly, but not limited to, systemic lupus erythematosus (SLE), Sjögren syndrome (SS), or a related profile of non–organ-specific autoantibodies.7- 9 Patients in whom transverse myelitis is accompanied by non–organ-specific autoantibodies are customarily classified as having disease-associated acute transverse myelitis,10,11 whether or not symptoms and signs of SS/SLE are present. Our observations suggest an alternative hypothesis to the prevailing presumption of a direct contribution of SS/SLE to the pathogenesis, namely, coexistence of NMO with non–organ-specific autoimmune disorders. The resolution of this issue has important diagnostic and therapeutic consequences because of emerging new treatment strategies for NMO.12,13
Herein we evaluate the association of NMO-IgG and non–organ-specific autoantibodies in patients with SS/SLE and in those with an NMO spectrum disorder (NMOSD), including NMO and at-risk syndromes such as LETM or recurrent optic neuritis. The apparent specificity of NMO-IgG for NMO allowed us to evaluate the following 3 hypotheses (Figure 1): (1) NMOSDs and SS/SLE are nonoverlapping conditions complicated by optic neuritis and myelitis; (2) these conditions are independent but coexist in the same patient; and (3) the serological findings are nonspecific and NMO-IgG or non–organ-specific antibodies may be seen in either condition.
The institutional review boards of the Mayo Clinic College of Medicine, Rochester, and Centre Hospitalier Régional Universitaire de Lille approved this study.
Group 1 consisted of 153 consecutive US patients (Table 1) with a diagnosis of NMO1 or a syndrome recognized to have a high risk of conversion to NMO (including single and recurrent LETM). All patients had at least 1 episode of myelitis extending 3 or more vertebral segments on magnetic resonance images of the spine. These patients were identified through an NMO clinical biospecimens database1,2,4 containing demographic, clinical, imaging, and laboratory data from patients who presented with syndromes compatible with NMO and underwent testing for NMO-IgG. Data were entered by one of the study neurologists (S.J.P., D.M.W., or B.G.W.) who had performed the evaluation of the patient or reviewed the medical record. Coexisting autoimmune conditions were also recorded. Classification criteria for SLE and SS were assessed by a rheumatologist for all patients with those diagnoses.14,15 Of the 153 patients in this study, 130 (85.0%) underwent evaluation at the Mayo Clinic. We included in addition 33 Mayo Clinic control patients with SS/SLE who did not have symptoms of NMO. Controls were ascertained serologically through a review of the medical records of patients undergoing evaluation in the rheumatology department who were seropositive for extractable nuclear antigens (ENAs) (Sjögren syndrome A [SSA] or B [SSB] antibodies) or antinuclear antibodies (ANAs; with or without anti–double-stranded DNA) at the Clinical Immunology Laboratory of the Mayo Clinic. We included only those patients who fulfilled classification criteria for SLE or SS.16
Group 2 consisted of 34 French patients with NMO (Table 2), with or without associated SS/SLE (n = 18), and patients with SS/SLE without NMO symptoms (n = 16). These patients were ascertained retrospectively and received their diagnoses at the Center Hospitalier Régional Universitaire de Lille using international rheumatological criteria for the diagnosis of SS and SLE. Neuromyelitis optica was diagnosed according to the criteria of Wingerchuk et al,1 and all patients with NMO had had at least 1 episode of myelitis extending 3 or more vertebral segments on magnetic resonance images of the spine, which was the same criterion used for group 1. (A longitudinally extensive spinal cord lesion has a 98% sensitivity and 83% specificity for NMO; Mayo Clinic 2006 criteria for NMO require a longitudinally extensive lesion or seropositivity for NMO-IgG.)17
We detected NMO-IgG with the use of indirect immunofluorescence.2 Two independent assessors (S.J.P. and V.A.L.), who were blinded to neurological and rheumatological diagnoses for all patients, classified every serum sample as positive or negative for NMO-IgG (with 100% concordance). Estimates of titration end points varied by no more than 1 doubling dilution.
Assays for ANAs and ENAs were performed by means of an enzyme immunoassay (EIA) kit (BioRad Diagnostics, Hercules, California) using antigens purified from HEp-2 cell nuclei. A negative ANA result was defined as less than 1 EIA U, and a positive ENA result was defined as more than 20 EIA U. Serum samples positive for ENA on EIA screening results were tested further by means of the antigen-specific EIA for anti-SSA/Ro, anti-SSB/La, anti-Sm, anti-U1RNP, anti-Jo-1, and anti-Scl-70.
Groups 1 and 2 are presented and analyzed separately because they represent 2 distinct and unrelated study cohorts collected with different ascertainment criteria. The US samples were from a cohort of patients with NMO and suspected NMO; the French cohort was a smaller, more selective population of patients with coexisting symptoms of SS/SLE and NMO. The diagnoses and classification were made independently by the 2 clinical groups. The samples of the French cohort were tested on coded samples and analyzed in a completely blinded fashion.
We compared the frequency of NMO-IgG, ANA, and ENA detection in the serum samples of patients diagnosed as having NMO or at-risk syndromes (with or without SS/SLE) with the frequency that these antibodies were detected in the serum samples of patients with SS/SLE without NMO symptoms. The analysis of significance used χ2 tests, Fisher exact tests, and unpaired, 2-tailed t tests (α = .05) as appropriate.
Figure 2 illustrates the NMO-IgG status and titers for all patients. Coexisting non–organ-specific autoantibodies and autoimmune disorders are given for group 1 in Table 1 and for group 2 in Table 2.
One hundred fifty-three patients were classified as having an NMOSD. Seventy-eight of these patients (51.0%) fulfilled diagnostic criteria for NMO and 75 (49.0%) had LETM (recurrent in 44 patients). One hundred thirty (85.0%) underwent clinical evaluation at the Mayo Clinic. Controls had SS (14 patients) or SLE (19 patients) without symptoms or signs of NMO. Eight had a coexisting neurological disorder (non-NMO) that affected the brain or peripheral nerve.
Of the 153 patients with NMOSDs, NMO-IgG was detected in 102 (66.7%) (Table 1). It was detected less frequently in patients with a single episode of LETM (10 of 31 [32.3%]) than in patients meeting the full criteria for NMO (61 of 78 [78.2%]) or relapsing LETM (31 of 44 [70.5%]) (P = .001).
Antinuclear antibodies or ENAs were detected in 72 of 153 patients with NMOSDs (47.0%) (Table 1). The frequency of ANAs in these patients was 43.8% (67 of 153). The ANA frequency did not differ significantly in subgroups with NMO (41 of 78 [52.6%]), single LETM (8 of 31 [25.8%]), or relapsing LETM (18 of 44 [40.9%]).
The frequency of ENAs in patients with NMO or patients who were at risk for NMO was 15.7% (24 of 153). As was the case for ANAs, the ENA frequency did not differ significantly in subgroups of patients with NMO (13 of 78 [16.7%]), single LETM (3 of 31 [9.7%]), or relapsing LETM (8 of 44 [18.2%]).
Among controls who had SS or SLE without NMO symptoms, NMO-IgG was not detected. Among patients with NMO symptoms, ANA or ENA detection was more frequent in those who were seropositive for NMO-IgG (60 of 102 [58.8%]) than in those lacking NMO-IgG (12 of 51 [23.5%]; P = .001, χ2 analysis). Similar results for ANAs or ENAs were found for NMO-IgG–seropositive vs NMO-IgG–seronegative patients with NMO (39 of 61 [63.9%] vs 5 of 17 [29.4%]; P = .02), single LETM (5 of 10 [50.0%] vs 4 of 21 [19.0%]; P = .08), or relapsing LETM (15 of 31 [48.4%] vs 3 of 13 [23.1%]; P = .08).
Five patients (all NMO-IgG seropositive) among the 153 with NMOSDs fulfilled clinical classification criteria for SS/SLE (2 for SLE, 2 for SS, and 1 for both). All were seropositive for ANAs (median titer, 5.4 U [interquartile range, 3.3-8.7 U]; reference limit, <1.0 U), and 4 of 5 were ENA seropositive (median titer, 127.0 U [interquartile range, 81.0-187.2 U]; reference limit, <25.0 U). All were seronegative for anticardiolipin antibodies. The diagnosis of SS or SLE predated the symptoms of NMO in 2 patients by 7 and 8 years and postdated the symptoms in 3 patients by 1, 1, and 8 years.
Other autoimmune disorders documented in the 153 patients with NMOSDs included autoimmune thyroid disease in 26 (17.0%), ulcerative colitis in 4 (2.6%), idiopathic thrombocytopenic purpura in 2 (1.3%), rheumatoid arthritis in 2 (1.3%), myasthenia gravis in 2 (1.3%), Raynaud phenomenon in 1 (0.7%), polymyositis in 1 (0.7%), and celiac disease in 1 (0.7%).
Eighteen of the 34 patients in the French cohort were classified as having an NMOSD. Ten were diagnosed as having NMO, of whom 6 had coexisting SS (4 patients), SLE (1), or both (1), and 4 had neither. Eight patients, 6 with LETM and 2 with recurrent optic neuritis, were deemed to be at risk for NMO.
Controls included 5 patients with SLE without neurological involvement and 11 with SS (5 without neurological involvement, 4 with peripheral neuropathy, and 2 with strokelike episodes). None of the patients with NMO or who were at risk for NMO, with or without coexisting SS/SLE, had symptoms other than those related to their optic nerves or spinal cord.
Among the 18 patients with NMO or who were at risk for NMO, NMO-IgG was detected in 7 (38.9%) (Table 2). Two of the 4 patients with NMO (50.0%) without coexisting SS or SLE were seropositive for NMO-IgG, whereas 5 of the 14 patients (35.7%) who fulfilled clinical classification criteria for SS (12 patients), SLE (1), or both (1) were seropositive (3 who had SS, 1 who had SLE, and 1 who had both SS and SLE). Antinuclear antibodies were detected in 15 (83.3%), SSA antibodies in 8 (44.4%), and SSB antibodies in 1 (5.6%) of the 18 patients with NMO or who were at risk for NMO.
Of the 16 controls with SS/SLE and no NMO symptoms, NMO-IgG was not detected in any. All 16 had non–organ-specific autoantibodies. Thus, NMO-IgG was restricted to patients with NMOSDs compared with the controls with SS/SLE (P = .01). Its frequency was similar in patients with NMOSDs with and without SS/SLE (P = .59).
This study demonstrates the specificity of NMO-IgG for distinguishing patients with optic neuritis and myelitis as an NMOSD manifestation from patients with multisystem autoimmune disorders plus other neurological syndromes. Non–organ-specific autoantibodies (ANAs and ENAs [SSA and SSB antibodies]) are frequently encountered in patients with NMO and are consistently positive in patients with SS/SLE. By contrast, the lack of NMO-IgG in patients with SS/SLE, except for those with a coexisting history of LETM, optic neuritis, or both, is consistent with our proposal that NMO, SS, and SLE are overlapping disorders that coexist in some patients, just as NMO may coexist with autoimmune thyroiditis, type 1 diabetes mellitus, or myasthenia gravis. Patients who have an NMO or an at-risk syndrome, with or without ANAs or ENAs, generally do not fulfill classification criteria for the diagnosis of SLE or SS.
Serological findings in the US and French patient groups were similar and complementary despite independent diagnoses by 2 groups of physicians on different continents using standard criteria. In both cohorts, a high proportion of patients with NMO or who were at risk for NMO were seropositive for NMO-IgG, and patients without NMO symptoms were uniformly seronegative. The fact that patients who have SS or SLE without NMO were consistently seronegative for NMO-IgG in both patient cohorts indicates that this autoantibody is not a nonspecific accompaniment of SS or SLE. The higher frequency of non–organ-specific autoantibodies in patients with NMO who are NMO-IgG seropositive than in patients with NMO who are NMO-IgG seronegative may reflect a more intense autoimmune response in the former group.
The coexistence of SS or SLE with NMOSDs probably reflects the predisposition of patients with NMO to multiple autoimmune diseases, both organ specific and non–organ specific, as exemplified by the variety of other coexisting autoimmune diseases documented in our US NMOSD cohort. These observations argue against a direct effect of SLE or SS in causing optic neuritis or myelitis, at least in the considerable proportion of patients with SLE or SS who are seropositive for NMO-IgG and experience these syndromes. The immunopathological features of the NMO spinal cord lesion18 are well characterized, in contrast to the nervous system complications of SS and SLE, for which few pathological reports are available.19- 21
We do not believe that the seropositivity for NMO-IgG that we detected in these patients is an epiphenomenon of optic nerve and spinal cord inflammation. The finding of NMO-IgG is highly specific to NMO and does not occur in many other conditions that have been studied as control diseases with alternative causes for severe myelopathy (eg, documented viral myelitis, vitamin B12 deficiency, sarcoidosis, and tumor) or optic neuropathy (eg, ischemic or compressive optic neuritis).2,22 Typically, NMO-IgG is seen at the very first symptom of NMO and predicts relapse and development of definite NMO. The target of NMO-IgG is AQP4, which occurs at the exact site of the deposition of immunoglobulins and products of complement activation in human NMO spinal cord lesions.18 Furthermore, recent data reveal selective loss of AQP4 immunostaining in all NMO lesions, regardless of stage and despite preservation of other markers of astrocytes, different from the increased immunostaining seen in active multiple sclerosis lesions.23 A novel inflammatory lesion in spinal cord and brainstem regions characterized by loss of AQP4 but lacking demyelination and necrosis was also reported.24 This loss of AQP4 colocalized with eosinophilic and plasma cells and the NMO-typical vasculocentric immunoglobulin and complement activation product deposition. These data and the beneficial responses reported for plasma exchange and rituximab therapies in NMO support a pathogenic role for the AQP4-specific IgG marker autoantibody of NMO.4,13
Thus far there are, to our knowledge, no reported immunopathological analyses of spinal cord lesions in patients with NMOSDs in the context of SS or SLE. None of 9 autopsy cases of NMO reported by Lucchinetti et al18 had SLE or SS, although 3 (33%) had other autoimmune disorders (hypothyroidism, pernicious anemia, or thrombocytopenic purpura). Immunopathological analysis of optic nerve and spinal cord lesions in patients with coexisting NMOSDs and SS or SLE is needed to advance our understanding of the relationship of these disorders.
Correspondence: Brian G. Weinshenker, MD, Department of Neurology, Mayo Clinic College of Medicine, 200 First St SW, Rochester, MN 55905 (email@example.com).
Accepted for Publication: July 2, 2007.
Author Contributions:Study concept and design: Pittock, de Seze, Vermersch, and Weinshenker. Acquisition of data: Pittock, de Seze, Vermersch, Homburger, Zéphir, and Weinshenker. Analysis and interpretation of data: Pittock, Lennon, de Seze, Vermersch, Homburger, Wingerchuk, Lucchinetti, Moder, and Weinshenker. Drafting of the manuscript: Pittock, Zéphir, and Moder. Critical revision of the manuscript for important intellectual content: Pittock, Lennon, de Seze, Vermersch, Homburger, Wingerchuk, Lucchinetti, Zéphir, and Weinshenker. Statistical analysis: Pittock. Administrative, technical, and material support: Pittock and Homburger. Study supervision: Pittock, Lennon, de Seze, Vermersch, Lucchinetti, Moder, and Weinshenker.
Financial Disclosure: Dr Lennon is a named investor on a patent application filed by the Mayo Foundation for Medical Education and Research that relates to the NMO antigen and its application to the diagnosis of NMO. Drs Lennon, Lucchinetti, and Weinshenker have intellectual property related to the discovery of the NMO autoantigen. Dr Weinshenker served as a consultant to Genentech Inc for development of a clinical trial for NMO, which does not relate directly to the subject of this article.
Funding/Support: This study was supported by the Ralph Wilson Medical Foundation, the Olson Foundation, and the Mayo Foundation.
Additional Contributions: Denice Bredlow provided assistance in the manuscript preparation. Aaron Maixner, BS, Matt Hangge, BS, and Thomas Kryzer provided technical assistance.