Article Text

Download PDFPDF

Paediatric neuromyelitis optica: clinical, MRI of the brain and prognostic features
  1. Michael Absoud1,
  2. Ming J Lim1,
  3. Richard Appleton2,
  4. Anu Jacob3,
  5. Joanna Kitley4,
  6. M Isabel Leite4,
  7. Michael G Pike5,
  8. Angela Vincent4,
  9. Evangeline Wassmer6,
  10. Patrick Waters4,
  11. Mark Woodhall4,
  12. Cheryl Hemingway7,
  13. Jacqueline Palace4
  1. 1 Department of Children's Neurosciences, Evelina Children's Hospital at Guy's & St Thomas’ NHS Foundation Trust, Kings Health Partners Academic Health Science Centre, London, UK
  2. 2 The Roald Dahl EEG Unit, Pediatric Neurosciences Foundation, Alder Hey Children's Hospital, Liverpool, UK
  3. 3 Department of Neurology, The Walton Centre for Neurology and Neurosurgery, Liverpool, North Lincolnshire, UK
  4. 4 Nuffield Department of Clinical Neurosciences, Oxford University Hospitals National Health Service Trust, University of Oxford, Oxford, UK
  5. 5 Department of Neurology, Oxford Children's Hospital, Oxford University Hospitals National Health Service Trust, Oxford, UK
  6. 6 Department of Neurology, Birmingham Children's Hospital, Birmingham, UK
  7. 7 Department of Neurology, Great Ormond Street Hospital for Children, London, UK
  1. Correspondence to Dr Michael Absoud, Department of Children's Neurosciences, Staircase D South Wing, Evelina Children's Hospital at Guy's & St Thomas’ NHS Foundation Trust, London SE1 7EH,UK; michael.absoud{at}

Statistics from


Neuromyelitis optica (NMO) is a severe and rare inflammatory condition characterised by presence of optic neuritis (ON) and transverse myelitis (TM) and distinct from multiple sclerosis (MS).1 Permanent disability is more attack-related than in MS. Autoantibodies to the astrocytic water channel protein aquaporin-4 (AQP4), found in majority of cases, contribute to astrocytic and myelin damage.1 We describe the clinical, MRI, outcome and prognostic features of paediatric onset NMO in relation to AQP4-Ab status.


We conducted retrospective case ascertainment and case note review of paediatric onset (<17 years) cases from the UK national NMO service ( database (2009–2012), with ethical approval as part of the UK NMO study. All had testing for AQP4-Ab (cell-based assay). Inclusion criteria were as in Wingerchuk et al.2 Descriptive statistics, univariate associations and Kaplan-Meier lifetable analysis were used to explore differences between AQP4-Ab positive/negative cases. MRI scans were reviewed at multidisciplinary meetings. MRI brain characteristics were compared to those of a national cohort sample of all known AQP4-Ab negative first episode central nervous system-acquired demyelination cases and not satisfying Wingerchuk criteria but with abnormal MRI brain scans.3


Twenty-two cases were ascertained and 20 met inclusion criteria (table 1); 18 (90%) females and 12 (60%) AQP4-Ab positive. Both excluded cases were AQP4-Ab positive, one had monophasic unilateral ON (3 years follow-up) and one had two TM attacks (5 years follow-up). No demographic features distinguished AQP4-Ab positive from negative cases. The demyelination AQP4-Ab negative control cohort (n=29) at 1 year had: acute disseminated encephalomyelitis (ADEM) (4), ON (2), acute TM (11) and MS (13) diagnoses. At presentation, 8 (40%) NMO cases had unilateral ON; 4 (20%) bilateral ON; 3 (15%) TM; 3 (15%) simultaneous TM and ON; 2 (10%) ADEM. At last follow-up, 55% (11/20) had an episode involving the brain and 19/20 (95%) had longitudinally extensive TM (LETM). First relapse phenotypes were: ON (n=8), TM (n=6), TM and ON (n=1) or brain syndrome (n=3). There were no differences with regards first attack or relapse phenotype (see online supplementary eTable 1) in relation to AQP4-Ab status.

Table 1

NMO cohort description and phenotype characteristics compared to control group

Six cases (30%) had abnormal brain MRI at disease onset; 11/20 (55%) within the first year and 15/20 (10 AQP4-Ab positive, 5 AQP4-Ab negative) by last follow-up (table 1 and see online supplementary eFigure 1). There were no apparent differences with regards to MRI brain lesion location in children with NMO (n=15) compared with demyelination controls (n=29). NMO brain scan lesions compared with controls were large (>2 cm), acute lesions largely resolved on repeat imaging, and often showed T1 hypointense lesions.

Four relapsing AQP4-Ab negative cases (50%) with serum available tested for antimyelin oligodendrocyte glycoprotein (MOG) antibody (long assay) were negative. None of the seven AQP4-Ab were negative cases, but 2/10 AQP4-Ab positive cases tested had cerebrospinal oligoclonal bands.

Using life table analysis (see online supplementary eFigure 2), mean time to relapse=2.4 years in AQP4-Ab negative cases (95% CI 1.1 to 3.6 years) versus 0.76 years (95% CI 0.43 to 1.1 years) for positive cases (p=0.03 log-rank-test, Mantel-Cox). Most were relapsing (6/8 AQP4-Ab negative and 12/12 positive). Annualised relapse rate (ARR) for AQP4-Ab positive cases was 0.70/year versus 0.38/year for negative cases (z score=2.316; p=0.21).

In AQP4-Ab positive cases, 10/12 had visual acuity of <6/60 Snellen in at least one eye (0/8 negative) and 58% had bilateral severe visual impairment (<6/60 Snellen). Three AQP4-Ab negative cases but no positive cases were wheelchair dependent on follow-up. Disease modifying treatments used included oral prednisolone, azathioprine, rituximab, IVIG, mycophenolate mofetil, mitoxantrone and ofatumumab.


We describe a national historical paediatric onset NMO longitudinal cohort compared to a contemporary incident national cohort sample.3 As in other series, NMO predominantly affects females; all AQP4-Ab negative cases were females. At onset, majority had a presentation which included ON (75%) compared to TM (30%). Importantly, 2/20 cases presented with ADEM. Over half had a brain syndrome episode during the disease course and the majority (95%) had LETM. Of 67 NMO paediatric cases previously described (see online supplementary eTable 2), 52/65 (80%) tested positive for AQP4-Ab. In our previously reported national incident prospective cohort (n=125) with 1 year outcome three had relapsing NMO (2 AQP4-Ab positive) and one had unilateral ON (AQP4-Ab positive) with poor visual outcome.4 A paediatric NMO spectrum disorders cross-sectional study from one lab, where 58/88 consecutive NMO-IgG seropositive patients were followed up for median 1 year,5 showed 38 (66%) cases fulfilled NMO diagnostic criteria.2 Our NMO cases were ascertained from the UK clinical population via the national NMO service, and also by identifying AQP4-Ab tested cases from this service.

Brain lesions in paediatric NMO occur more commonly (>50%) than adult NMO (25%). In our cohort three-quarters had abnormal MRI brain scan, and 30% had abnormality at disease onset. In our cohort, the location of the MRI brain lesions did not differentiate NMO from other acquired demyelinating disorders or between AQP4-Ab positive/negative cases. However, we describe distinct NMO lesion nature in that they tend to be large (>2 cm); acute lesions resolving on repeat scan and frequently showed T1 hypointense lesions. Interestingly, NMO AQP4-Ab positive cases had corpus callosum changes (n=4/10) and were Barkhof space criteria positive (n=3/10), while none of the NMO AQP4-Ab negative cases had either of the features.

Four of relapsing AQP4-Ab negative cases in our study were negative for MOG-Ab. As such it is likely that there are yet undiscovered autoantigens in seronegative cases.

Most (90%) NMO cases were relapsing, while AQP4-Ab positive cases had a higher ARR and shorter time to relapse compared to negative ones. The ARR in our AQP4-Ab positive cases is similar to that found in other cohorts (see online supplementary eTable 2). In our mixed treatment onset population, in order to conduct a clinical trial to reduce ARR from 0.70 to 0.35 (negative binomial model; dispersion index=0.43; power=80%; α=0.05), would require sample size n=84 per arm.

Paediatric NMO has been reported to have better disability prognosis than adult NMO. In our cohort, three children were wheelchair dependant and half had visual impairment. AQP4-Ab positivity is associated with early recurrence and visual disability.

NMO may be difficult to identify at first presentation due to similarity of clinical phenotype to other acquired demyelinating syndromes. Identifying distinct MRI brain lesion features as described may aid to raise suspicion. In order to minimise risk of early visual failure, a low threshold for AQP4-Ab testing in any initial presentation of paediatric demyelinating disease is recommended.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

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.