Abstract
Objectives: To evaluate the spectrum of neuromyelitis optica (NMO), including characteristics of the index events (optic neuritis [ON]) and myelitis), neuroimaging, CSF, and serologic studies, and to evaluate the long-term course.
Methods: Review of 71 patients with NMO evaluated at the Mayo Clinic between 1950 and 1997.
Results: NMO was either monophasic or relapsing. Patients with a monophasic course (n = 23) usually presented with rapidly sequential index events (median 5 days) with moderate recovery. Most with a relapsing course (n = 48) had an extended interval between index events (median 166 days) followed within 3 years by clusters of severe relapses isolated to the optic nerves and spinal cord. Most relapsing patients developed severe disability in a stepwise manner, and one-third died because of respiratory failure. Features of NMO distinct from “typical” MS included >50 cells/mm3 in CSF (often polymorphonuclear), normal initial brain MRI, and lesions extending over three or more vertebral segments on spinal cord MRI.
Conclusions: Clinical, laboratory, and imaging features generally distinguish neuromyelitis optica from MS. Patients with relapsing optic neuritis and myelitis may have neuromyelitis optica rather than MS. Patients with a relapsing course of neuromyelitis optica have a poor prognosis and frequently develop respiratory failure during attacks of cervical myelitis.
Neuromyelitis optica (NMO) (Devic’s syndrome) was originally regarded as a monophasic syndrome consisting of acute, severe (“transverse”) myelitis and bilateral simultaneous or sequential optic neuritis (ON) occurring in close temporal succession and resulting in paraplegia and blindness.1-3 However, subsequent studies variably included patients with less severe clinical attacks, unilateral ON, symptoms separated by months or years, or recurrent exacerbations.4-9 The spectrum and natural history (frequency of attacks, mortality, and long-term morbidity) of NMO, particularly the relapsing form, remain undefined. Furthermore, whether NMO is a subtype of MS or a distinct disease entity remains controversial.5-9 We describe the demographic, clinical, CSF, and MRI features of 71 patients with NMO to address these issues.
Methods.
Case ascertainment.
We reviewed the medical records and MRI studies of all patients diagnosed with NMO, Devic’s syndrome, or Devic’s disease between 1950 and 1993 at the Mayo Clinic. We personally evaluated all patients with NMO seen at our MS Clinic or hospital between 1993 and 1997. Questionnaires were used to obtain long-term data for patients not followed to the time of the study. Retrospective and prospective data, including those obtained by questionnaire, were pooled for analysis.
Case definition.
To evaluate the importance of traditional diagnostic criteria, we considered two retrospective definitions: “strict criteria” (bilateral ON and myelitis, the index events, occurring within 2 years of one another without symptomatic disease outside of the optic nerve and spinal cord) and “not meeting strict criteria” (unilateral ON or development of a second index event over a period greater than 2 years).
Data collection.
We recorded clinical, CSF, and MRI data. An ordinal scale was devised to quantify optic nerve and spinal cord impairment (table 1).
Quantification of optic nerve and spinal cord impairment
We evaluated the results of CSF obtained within 4 weeks of the onset of a new relapse for white blood cell (WBC) counts and differential because only “acute” CSF studies showed significant pleocytosis. All CSF studies were used to evaluate the presence of oligoclonal banding and immunoglobulin abnormalities.
Abnormal brain MRIs were classified as either satisfying or not satisfying radiologic criteria for MS.10 Lesion length, cord swelling, gadolinium enhancement, and atrophy were recorded for spinal cord lesions.
Statistical methods.
Because the distribution of much of the quantitative data was skewed, medians and ranges are presented; Wilcoxon rank sum tests were used to assess differences between groups. Chi-square or Fisher’s exact tests were used for dichotomous variables. A log-rank test was used to compare survival between groups.
Results.
Patient ascertainment and classification.
We identified 93 patients diagnosed with NMO, 81 through medical records and 12 from our clinic. Twenty-two patients were excluded from analysis because they had not experienced one of the two index event types or the diagnosis was later changed (n = 20), or because of incomplete data (n = 2). Two patients were diagnosed postmortem; they had myelitis with subclinical optic nerve demyelination found at autopsy.
The remaining 71 patients were grouped using two criteria: the case definition (whether or not the strict criteria were met) and the subsequent clinical course (monophasic or relapsing). Patients with a relapsing course had further relapses after their index events, whereas those with a monophasic course did not (minimum 3 years follow-up). Figure 1 illustrates how the cohort was divided. Patients with monophasic disease failed to meet the strict criteria due to unilateral rather than bilateral ON. The patients in the relapsing group did not meet strict criteria because they had unilateral ON (n = 9) or more than a 2-year interval between index symptoms (n = 8). Sixty-one of 69 patients (88%) contacted returned follow-up questionnaires.
Figure 1. Patient ascertainment and classification. NMO = neuromyelitis optica, MPS = monophasic course, fulfilling strict diagnostic criteria, MPN = monophasic course, not fulfilling strict diagnostic criteria, RPS = relapsing course, fulfilling strict diagnostic criteria, RPN = relapsing course, not fulfilling strict diagnostic criteria.
Diagnostic criteria.
To address the validity of our strict definition, we compared the patients who met the strict criteria with those who did not for both the monophasic and relapsing groups. There were no important differences in demographics, clinical or paraclinical features, or outcome (data not shown). Therefore, we performed all subsequent comparisons between monophasic and relapsing groups regardless of the index event interval and number of optic nerves involved (i.e., whether they met the strict criteria).
Demographics.
Demographic characteristics are shown in table 2. The monophasic and relapsing groups differed with respect to sex ratio and age at onset. There were nonwhites (Asian = 3, African-American = 2, and Indian = 1) only in the relapsing group. Only relapsing patients (n = 4) reported a family history of MS. We have not identified a patient with a monophasic course since 1988.
Demographic features and index events
Antecedent and concomitant conditions.
Viral illnesses preceded NMO in a minority of patients in each group (see table 2). Two patients in the monophasic group received swine flu vaccination just before onset of NMO. Autoimmune diseases were present in one third of patients in the relapsing group, including hypothyroidism (n = 9), Sjögren’s syndrome (n = 4), pernicious anemia (n = 3), ulcerative colitis (n = 1), primary sclerosing cholangitis (n = 1), and idiopathic thrombocytopenic purpura (n = 1).
Index events.
Isolated unilateral ON or myelitis was the presenting feature in 43 of 48 patients (90%) in the relapsing group and 11 of 23 patients (48%) in the monophasic group (table 3). Simultaneous bilateral ON and myelitis (all within 24 hours) occurred in 7 of 23 patients (30%) in the monophasic group but in none of the relapsing group.
Neurologic impairment: Index events and last follow-up
The time between the first event and the point when all the index events (myelitis and bilateral ON; original NMO definition) had occurred was longer in the relapsing group than it was in the monophasic group (median 166 versus 5 days, range 2 to 730 versus 0 to 151 days; p = 0.0001).
Index events were more severe in the monophasic group (see table 3). In the monophasic group, 21 of 39 affected eyes (54%) for which quantitative information was available became completely blind (no light perception) compared with 20 of 72 (28%) in the relapsing group. Sixteen of 23 patients (70%) with a monophasic course were paraplegic at nadir compared with 15 of 48 (31%) in the relapsing group.
Most patients improved by at least one level on the visual scale (monophasic group 34 of 42 eyes [81%], relapsing group 68 of 87 eyes [78%]). Motor scores improved by at least one level in 18 of 23 (78%) patients in the monophasic group and 42 of 48 patients (88%) in the relapsing group. At best recovery (before further exacerbations in the relapsing group) motor and sphincter function were significantly worse in the monophasic group.
Lhermitte’s symptom (17/48, 35%), paroxysmal tonic spasms (17/48, 35%), and radicular pain (16/48, 33%) often accompanied acute myelitis in the relapsing group. Radicular pain rarely occurred in the monophasic group (2/23, 9%), and Lhermitte’s symptom and tonic spasms were not recorded in this group.
Characteristics of the relapsing group.
Relapses were confined to recurrent ON or myelitis in all except five patients, who experienced facial numbness (n = 2), vertigo (n = 2), and cerebellar tremor (n = 1). The median number of total relapses per patient was five (range 1 to 18); 47 of 48 patients (98%) had more than one relapse. Figure 2 demonstrates that patients with a relapsing course often have periods of quiescence interspersed with frequent attacks. The interval between the final index event and the first relapse (defining a relapsing course) occurred within 1 year in 55% of patients, within 3 years in 78%, and within 5 years in 90%.
Figure 2. Scatterplot of relapses in patients with a relapsing course grouped as to whether strict diagnostic criteria were fulfilled (RPS) or not fulfilled (RPN).
Respiratory failure and mortality.
Respiratory failure caused by acute cervical myelitis occurred 19 times in 16 relapsing patients (33%) and only twice in 2 monophasic patients (9%). Both monophasic patients recovered. Fifteen of the 16 relapsing patients (93%) with respiratory failure died; 3 had recovered from a first episode only to succumb to a second. The 5-year survival rate (figure 3) was 90% in the monophasic group and 68% in the relapsing group (p = 0.06). All deaths in the relapsing group were due to respiratory failure.
Figure 3. Survival analysis of patients with a monophasic (dashed line) and relapsing (solid line) course.
Long-term impairment.
The mean disease duration at last follow-up was 16.9 years in the monophasic group and 7.7 years in the relapsing group. At last available assessment, visual acuity, motor strength, and sensory function were worse in the relapsing group than they were in the monophasic group (see table 3). Sixty percent of relapsing patients (26/43) with complete examination information were functionally blind (acuity 20/200 or worse) in at least one eye compared with 22% of the monophasic patients (5/23) (p = 0.014, two-sided rank sum test). Permanent monoplegia or paraplegia occurred in 25 of 48 patients (52%) in the relapsing group and 7 of 23 patients (31%) in the monophasic group (p = 0.086). Eleven of 17 surviving monophasic patients (65%) could walk with no or unilateral assistance compared with 17 of 32 relapsing patients (53%).
Laboratory studies.
Antinuclear antibodies (titer >1:40) were present in 12 of 25 relapsing patients tested (48%) and 0 of 6 monophasic patients (p = 0.059). Antibodies to extractable nuclear antigen were present in four patients with Sjögren’s syndrome.
CSF.
Twenty of 23 monophasic patients (87%) and 47 of 48 relapsing patients (98%) had at least one CSF examination (table 4). Pleocytosis (>5 WBC/mm3) was present in 11 of 15 monophasic patients (73%) and 31 of 38 in the relapsing group (82%). Greater than 50 cells/mm3 were observed in 5 of 14 patients (36%) in the monophasic group and 13 of 38 patients in the relapsing group (34%). Neutrophils were present in at least one CSF study in 7 of 14 monophasic patients (50%) and 21 of 35 relapsing patients (60%). Only the CSF protein differed significantly between the monophasic and relapsing groups.
CSF characteristics of neuromyelitis optica
MRI: Brain.
Features typical or distinct to NMO are illustrated in figure 4. Brain MRI was performed in 28 patients—27 relapsing and 1 monophasic. The median interval from the first index event to the initial brain MRI was 7.5 months (range 0.5 to 194). Optic nerve enhancement was seen in 5 of 6 patients (83%) studied within 2 weeks of an episode of ON. The brain parenchyma was normal in 22 of 28 initial scans (79%); only 3 of 28 (11%) satisfied the criteria of Paty et al.10 for MS. Three patients had increased T2 signal in the medulla that was contiguous with a longitudinally extensive upper cervical cord lesion (see figure 4, C and D).
Figure 4. Selected MR images of the brain. (A) Diffuse enhancement of the optic chiasm on gadolinium-enhanced T1-weighted coronal images. (B) A few periventricular lesions are seen in one patient on a T2-weighted axial image. (C and D) Lesion extending rostrally from the cervical cord into the medulla on gadolinium-enhanced T1-weighted sagittal and on T2-weighted axial images.
Sixteen relapsing patients had more than one brain MRI; the initial scan was normal in 14. Follow-up MRIs showed that 7 remained normal (mean interval between first and last MRI, 11 months; range 1 to 39), 6 developed nonspecific abnormalities that did not meet the criteria of Paty et al.10 for MS (mean interval 23 months, range 3 to 85), and 1 patient “converted” to meet MS criteria 103 months later.
MRI: Spinal cord.
Twenty-three relapsing patients had a total of 50 spinal cord MRIs performed within 8 weeks of the onset of an episode of myelitis; 47 of 50 (94%) were abnormal. In 44 of 50 (88%), longitudinal, confluent lesions extended across 3 or more (often 6 to 10) vertebral segments (figure 5). Cord swelling (25/50 scans, 50%) and gadolinium enhancement (32/50 scans, 64%) were commonly observed. Five of 23 patients (22%) with follow-up cord imaging developed focal cord atrophy.
Figure 5. Selected MR images of the spinal cord. (A) Longitudinally extensive spinal cord swelling and hyperintensity on T2-weighted sagittal images accompanied by (B) linear enhancement on gadolinium-enhanced T1-weighted sagittal image. (C) Spinal cord atrophy is evident on this T1-weighted sagittal image.
Treatment.
Corticosteroids were associated with improvement in 55 of 69 acute attacks (80%) with available data. Five relapsing patients experienced corticosteroid dependence, with recurrent myelitis when the dosage was reduced or the medication discontinued. Plasma exchange was associated with improvement in four of six steroid-unresponsive patients. Attack frequency decreased after treatment with beta-interferon in three of nine patients (median duration 4.5 months) and with azathioprine or cyclophosphamide in three of nine patients (median duration 4 months).
Discussion.
Allbutt11 recognized the association of visual loss and spinal cord disease in 1870, and Erb12 provided the first thorough description of NMO in 1880. Devic’s1 review in 1894 of 16 cases plus one fatal case of his own led to his eponymic association with NMO. Most initial reports of NMO described a severe, monophasic disorder. However, in 1927 Beck5 reported a 15-year-old girl with four optic nerve or spinal cord attacks in 13 months, each with intervening periods of remission lasting several months. McAlpine13 reviewed 22 cases in 1938 and noted that 7 had one or more remissions but, nevertheless, experienced a fatal outcome. Although a relapsing form of NMO has been recognized, its distinction from the monophasic type has been unclear. This report clearly distinguishes the two subtypes of NMO and provides long-term data on clinical course and outcome.
Patients with unilateral ON are indistinguishable from those with bilateral ON. Similarly, the interval over which patients developed index events had no diagnostic significance; rather, it helped predict whether patients will follow a monophasic versus a relapsing course. Therefore, patients with NMO may have unilateral ON and occasionally an interval of greater than 2 years between the first two index episodes.
Characteristics associated with a relapsing course include female sex, older age at onset, longer time interval between index events, and the presence of systemic autoimmunity. Bilateral ON and myelitis occurring within 1 month usually predicted a monophasic course. Conversely, almost all relapsing patients presented with isolated ON or myelitis and then experienced other index events over a period of greater than 3 months.
Our data are consistent with widely held beliefs that the clinical manifestations of NMO are more severe than those of “typical” MS.14 Patients with MS often enter a progressive phase of the disease,15 whereas those with NMO worsen from sequelae of attacks, perhaps because the attacks are so severe that they preclude appreciation of progressive axonal deterioration. Monophasic patients experienced more impairment from attacks than did relapsing patients, but their long-term outcome was better. Most monophasic patients retained adequate visual function and were ambulatory at last assessment, whereas the majority of relapsing patients became blind in at least one eye and nonambulatory. Exacerbations of MS causing complete blindness or paraplegia with severe residual deficit are uncommon. Respiratory failure, the cause of death in one-third of the relapsing group, is virtually unknown in MS.
CSF and MRI findings also distinguish NMO from MS. Neutrophilic pleocytosis or >50 nucleated cells/mm3 in CSF was common in NMO but rare in classical MS.16,17 Oligoclonal bands, present in CSF in up to 90% of MS patients,18 were uncommon in NMO; others have also made this observation.8,9
Brain MRI is normal or nonspecific, as also recognized by others.8,9,19 Some patients developed white matter lesions on serial imaging, but rarely do these meet radiologic criteria for MS. In relapsing-remitting MS, approximately 5% of patients fail to meet standard MRI criteria, and only 1% are normal. As described by others in patients with NMO,8,9,20 we observed longitudinally extensive lesions in the spinal cord. In contrast, cord lesions in MS are rarely longer than a single vertebral segment.21,22
We used our results to design diagnostic criteria for NMO (table 5). Diagnostic criteria are necessary to evaluate prospectively whether certain high-risk presentations (e.g., ON, myelitis, and negative brain MRI) herald NMO. Early diagnosis will be critical if effective long-term treatment to prevent attacks becomes available because many patients develop severe, irreversible impairment early in the disease. Three absolute criteria (ON, acute myelitis, and no clinical expression of disease outside of the optic nerve and spinal cord) and six supportive criteria have discriminating value (see table 5). MRI and CSF criteria are used to support a diagnosis of NMO if present but do not eliminate the diagnosis if absent. These criteria, though arbitrary, were derived from our data. They require prospective validation.
Proposed diagnostic criteria for neuromyelitis optica
The prognosis of relapsing NMO is relatively poor compared with MS. Relapses consist almost exclusively of ON and myelitis. They occur early, in clusters, and at unpredictable intervals (see figure 2). The combination of prednisone and azathioprine reduced attack frequency in an uncontrolled series.23
The pathogenesis of NMO and its relationship to MS remains uncertain. The inflammatory necrosis detected on pathology may reflect the severity of the inflammatory response rather than its nature. Thickened and hyalinized blood vessels in the spinal cord and optic nerve have been reported in NMO,8,24 but the number of patients investigated is small. Whether the vascular lesions are an essential part of the pathology or a secondary response is unknown. Aside from its clinical course, other observations suggest that NMO is distinct from MS. The high prevalence of autoantibodies9,25 and numerous reports of inflammatory opticospinal disease associated with connective tissue disorders9,26-28 suggest a defect in B-cell autoimmunity. Unlike MS, there is a racial predilection for nonwhites.9,29 Japanese patients with “opticospinal” MS behave similarly to relapsing NMO patients30-32 and differ immunogenetically from Japanese with Western MS, consistently being HLA DR2 negative.33
We acknowledge the limitations of this retrospective study. Ascertainment and referral biases may influence the results. Some patients we might have classified as having NMO may have been diagnosed with MS and missed by our record search. Our initial NMO definition, though arbitrary, was based upon prior studies. However, we objectively assessed and discarded the notion that bilateral ON and myelitis occurring within a specific short interval are essential to the diagnosis. We believe that the spectrum of NMO is wider than previously recognized, and the condition is likely under-diagnosed. The diagnostic criteria derived from these data require prospective evaluation of “high-risk” patients, such as those with bilateral ON and a normal brain MRI, to determine their validity in discriminating patients with “typical” MS from those with the generally poorer prognosis of NMO.
Acknowledgments
Supported by the Mayo Clinic and Mayo Foundation.
Acknowledgment
Katherine Archer, Dawn Wick, and Tischa Agnessi assisted with data entry and statistical analysis. Laura Irlbeck assisted with manuscript preparation.
Footnotes
-
Presented in part at the 48th annual meeting of the American Academy of Neurology; San Francisco, CA; March 1996.
- Received September 15, 1998.
- Accepted April 24, 1999.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
