Article Text

This article has a correction. Please see:

Download PDFPDF
Neuromyelitis optica
  1. Anu Jacob1,
  2. Mike Boggild2
  1. 1Specialist Registrar, The Walton Centre for Neurology and Neurosurgery, Liverpool, UK
  2. 2Consultant Neurologist, The Walton Centre for Neurology and Neurosurgery, Liverpool, UK
  1. Correspondence to:
 Dr M Boggild, Consultant Neurologist, The Walton Centre for Neurology and Neurosurgery, Liverpool L9 7LJ, UK; mike.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Neuromyelitis optica also known as Devic’s disease is an uncommon, immune mediated demyelinating condition of the central nervous system affecting predominantly the spinal cord and optic nerves. Since 1894 when Eugene Devic summarised 17 known cases of optic neuritis and myelitis the relation between neuromyelitis optica and multiple sclerosis (MS) has been controversial. The overlapping clinical features and the propensity for patients with neuromyelitis optica to express a range of auto-antibodies can result in misdiagnosis, of both MS and other autoimmune disorders. However, clinical, radiological, and immunopathological studies suggest neuromyelitis optica is distinct from MS. The recent identification of an apparent disease specific antibody—termed NMO-IgG (against the aquaporin-4 water channel) implicates humoral immunity.2,3 This further differentiates the disorder from MS and suggests that treatment to prevent relapses should be aimed primarily at humoral B cell mediated immunity if a pathogenic role for NMO-IgG is confirmed.


Neuromyelitis optica is an uncommon disorder in Western populations; based on observed cases among a population of 3 million over 10 years in North West England we estimated a minimum incidence of 0.4/million/year and a prevalence of 4/million, representing only one in 200 patients with demyelinating disease in this population.4 This contrasts with a much higher incidence in populations of Asian, Afro-Caribbean, and South American descent implying underlying genetic mechanisms in the expression of demyelinating disease. In all populations there is a strong female predominance, of >3:1 compared with males.4,5 The mean age of onset is around 40, although cases have been reported in childhood.


The cardinal clinical features of the disorder are transverse myelitis, which is often longitudinally extensive, and optic neuritis. These two index events can occur simultaneously, in rapid succession, or they can be separated by many years. The optic neuritis can be unilateral or bilateral. Some patients have repeated episodes of optic neuritis before myelitis occurs and vice versa (the nomenclature of the disease at this stage is relapsing myelitis or relapsing optic neuritis). Most of those affected (>80%) go on to have repeated relapses (relapsing NMO) though a minority may have only the index events (monophasic disease). Relapses are generally more disabling than those in patients with typical MS. Whereas in MS disability develops largely in the progressive phase of the disease, in neuromyelitis optica disability is acquired as a consequence of relapses; progressive disability without relapses is rare in our experience.

In white populations most patients presenting with optic neuritis and myelitis are likely to have MS rather than neuromyelitis optica. Features which may help distinguish the disorder from MS clinically are the more severe myelitis, optic neuritis with poor recovery, and no involvement of other parts of the neuraxis. However the disease spectrum may be wider than currently accepted (see discussion below).

Eugène Devic (1858–1930). Image courtesy of Professor Christian C, Hôpital Neurologique Pierre Wertheimer, Lyon, France.

Eugène Devic was born in 1850 in La Cavalerie, a small village in Aveyron, in southern France. He trained at the medical college of Lyon and later worked on a variety of subjects, and was especially interested in typhoid fever and cardiovascular diseases. In the field of neurology, he wrote about infantile chorea, “polyneuritic psychosis”, mental disorders in typhoid fever, cerebral glioma, corpus callosum tumours, meningeal angiosarcoma, and post-hemiplegic contracture. In December 1892 he saw a 45 year old French woman for intractable headache and depression with “general weakness” at the Hôtel-Dieu Hospital of Lyon. On 27 January urinary retention appeared, followed by complete paraplegia and bilateral blindness. She died from bedsores on 4 March 1893. The case and pathological examination, which confirmed lesions in the spinal cord and optic nerves, was presented as a clinicopathological study at the First Congress of Internal Medicine in Lyon in 1894. Here Devic mentioned 16 other similar cases reported in Europe and the USA. These 17 cases were studied in detail in the doctoral thesis of Fernand Gault —”De la neuromyélite optique aiguë”— in the same year. Neuromyelitis optica was called “Devic’s disease” after Acchiote proposed this eponym in 1907. In his paper Devic named the disorder “neuromyélite optique” or “neuroptico-myélite”. The two cardinal questions he raised more than a century ago “Why such a peculiar localisation?” and “What is the intimate nature of the process?” still remain largely unanswered.1

Clinical, radiological, and immunopathological studies suggest neuromyelitis optica is distinct from MS

A variety of diagnostic criteria for the disorder have been formulated and are summarised elsewhere.6 None is perfect and it is likely that they will be revised in the light of emerging clinical and laboratory data. In general we would consider the diagnosis in the presence of:

  • longitudinally extensive myelitis (usually more than three vertebral segments)

  • optic neuritis

  • normal brain MRI, or if abnormal, atypical for MS—see below.

The prognosis also differs from MS. Early reports suggested a five year survival of 68%, death often resulting from severe spinal cord disease and respiratory compromise.5 More recent series suggest a better outcome, possibly reflecting better case ascertainment or treatment. However there is no doubt that disability is acquired earlier in neuromyelitis optica than MS; this and a mean relapse rate of around two per year make early diagnosis and therapy imperative.


The most characteristic radiological feature is a longitudinally extensive cord lesion, extending often over three or more spinal segments and expanding the cord (fig 1). This is usually hypointense on T1 and hyperintense on T2 MRI. Lesions are centrally situated within the cord and patchy contrast enhancement is often seen. Occasionally lesions can be identified in the optic nerves. Though classically NMO has been defined by lack of brain lesions or symptoms, it has increasingly been noted that up 60% of patients with otherwise typical relapsing NMO (many with positive NMO-IgG) can have lesions on brain MRI. These range from extension of high cervical cord lesions into the brain stem, diencephalic, brainstem or cerebral lesions “atypical” for MS and in minority MS-like lesions. However all these patients have long cord lesions which seems to be a specific feature, distinguishing these cases from MS.9

The most characteristic radiological feature is a longitudinally extensive cord lesion, extending often over three or more spinal segments and expanding the cord


Cerebrospinal fluid acutely may reveal a prominent pleocytosis of either lymphocytes or neutrophils and raised protein. In contrast to MS, there are usually no oligoclonal bands (in over 80%). Lennon and co-workers at the Mayo clinic recently reported the discovery of NMO-IgG, which may be the first “disease specific” antibody in CNS demyelinating disease.3 The antibody, identified initially from Western blots in patients screened for possible paraneoplastic antibodies, is reported to have a sensitivity of 73% and a specificity 91% for neuromyelitis optica and was also positive in a significant proportion of patients deemed to be at high risk of neuromyelitis optica (that is, patients with recurrent optic neuritis or myelitis). Despite the apparent association of NMO-IgG with neuromyelitis optica, independent confirmation of this finding and indeed evidence of pathogenesis is still awaited. The Mayo group has also recently reported that the target antigen for NMO-IgG appears to be the aquaporin-4 water channel, located in astrocytic foot processes at the blood-brain barrier.2 To date, testing for NMO-IgG is only available through the Mayo clinic (Rochester, USA; at a cost of $500/sample). If the role of the antibody is confirmed and the assay becomes readily available, the hope is clearly that the relation between neuromyelitis optica and “neuromyelitis optica spectrum” disorders (see below) will be clarified, and predictive testing for neuromyelitis optica in patients with isolated myelitis or severe optic neuritis may be possible.

A range of positive auto-antibodies (including ds-DNA) have been reported in up to 40% of patients.4 This can give rise to diagnostic difficulties. We would consider these to be epiphenomena occurring in the context of disordered humoral immunity. However there are reports of patients with unambiguous systemic lupus erythematosus, Sjögren’s syndrome, and mixed connective tissue disease who have developed neuromyelitis optica. Whether they have two separate autoimmune diseases, or neuromyelitis optica is a consequence of the primary disorder, is impossible to determine clinically. Testing for NMO-IgG and pathological examination in such cases may be able to clarify this.


Extensive necrosis, demyelination, and often cavitation across multiple spinal cord segments, involving grey and white matter with perivascular infiltrates, prominent macrophages, eosinophils, and vascular hyalinisation is typical. Deposition of complement in a ring pattern on the outer surface of blood vessels and in a rosette perivascular pattern has been elegantly demonstrated. Prominent perivascular IgG reactivity and IgM deposition in a rosette pattern implicate these as sites of immune mediated damage.8


The rarity of the condition inevitably limits the evidence for therapeutic interventions. As in most immune mediated disorders management consists of treatment of relapses, therapy for the underlying disease, symptom control, and rehabilitation. For the general neurologist managing a patient with neuromyelitis optica the approach to relapses and indeed underlying disease therapy is perhaps most comparable to that of the more common antibody mediated disorder myasthenia gravis.


High dose corticosteroids and supportive care remain the mainstays of management of relapse. In view of the severity of relapses and likely need for maintenance treatment our policy is to follow a course of intravenous methylprednisolone (1 gram daily for 3–5 days) with a gradual taper of oral prednisolone over several months, from an initial dose of 1 mg/kg/day. An alternate day maintenance dose of the order of 10–20 mg prednisolone is generally our target in patients with relapsing disease.

A minority of patients fail to respond to adequate steroid therapy or relapse rapidly and in such cases there is a role for therapeutic plasma exchange. We consider it early—within weeks of symptom onset. In the North American randomised trial of plasma exchange in severe demyelinating events, patients with neuromyelitis optica were overrepresented among the responders, with a 60% response rate (versus 6% overall for sham exchange). In this study plasma exchange was undertaken within three months of onset of relapse.


Most patients follow a relapsing course, often acquiring substantial disability within two or three relapses. Immunosuppression appears to reduce the relapse rate. The first report of successful treatment was a series of seven patients treated with prednisolone (tailing dose as above) and azathioprine (at 2.5–3 mg/kg). In our initial case series of 42 patients, relapse rates were reduced by over 80% in patients established on immunosuppressive therapy (most frequently azathioprine). This has therefore remained our initial therapy option in patients with relapsing, but reasonably stable, disease. For patients intolerant of azathioprine, mycophenylate mofetil is a reasonable alternative (but without any evidence base), with the possible advantage of more rapid onset of action.


  • Although neuromyelitis optica is uncommon it is a rapidly disabling yet treatable disorder.

  • Early recognition and diagnosis followed by prompt, carefully supervised immunosuppressive treatment in relapsing patients is paramount. A blind, quadriparetic, ventilator dependent person, from a treatable disorder, is a tragedy that one should strive to avoid at all costs.

  • Management appears distinct from that of MS and should probably be in the hands of clinicians with an interest in demyelinating disease.

In patients who have “breakthrough” disease on azathioprine, or who present with frequent severe relapses, more aggressive immunosuppression may be necessary. Recent small case series have reported on the successful use of rituximab (a B cell depleting monoclonal antibody) and mitoxantrone. We have used both of these agents in small numbers of patients without complications. Interferon beta, the mainstay of treatment in relapsing MS, does not appear to be effective.


Pain, stiffness, bladder, and bowel symptoms need to be tackled. Tonic spasms seem to be much commoner than in MS and in our experience usually respond to carbamazepine. Rehabilitation, physiotherapy, mobility, and visual aids are often needed. Some patients with high cervical cord lesions will need long term home ventilatory support. There is a UK based self-help group (telephone +44 (0)151 529 6100) and a patient friendly website (


There may be a number of related disorders, a full discussion of which is beyond the scope of this review. Patients with idiopathic relapsing myelitis, Asian “opticospinal” MS, and chronic relapsing inflammatory optic neuropathy may form part of the “neuromyelitis optica spectrum”. We have seen several patients presenting with severe myelitis without visual symptoms but delayed visual evoked responses, some of whom have subsequently developed optic neuritis. It seems likely that such patients have neuromyelitis optica although they fall outside current criteria for the disorder. Traditionally patients with clinical or radiological findings outside the optic nerves and spinal cord have also been excluded from the diagnosis of neuromyelitis optica. There are however recent reports of patients with clinically typical neuromyelitis optica who have developed brain lesions on MRI. Interestingly these patients have been shown to have positive NMO-IgG, and histopathology (of the brain lesions) similar to classical neuromyelitis optica, widening the spectrum still further. More studies with a validated disease marker in the near future should clarify these relationships. Genetic links between neuromyelitis optica and MS have been hypothesised and will perhaps open a window to the better understanding of both conditions.9

Figure 1

Sagittal T2 MRI appearances of longitudinally extensive cervical myelitis (arrow), usually centrally located within the cord.


We thank Ms Karen Reeves at the ABN and Neurologists around the UK who have referred patients to the UK NMO study. The UK NMO study is an ongoing registry of NMO and related diseases, recruited by the British Neurological Surveillance Unit (BNSU), through which to date over 70 cases have been identified and are being followed.

This article was reviewed by Professor Reinhard Hohlfeld, Munich, Germany.


Linked Articles

  • Errata
    BMJ Publishing Group Ltd

Other content recommended for you