Article Text

Download PDFPDF
Original research
Early predictors of disability of paediatric-onset AQP4-IgG-seropositive neuromyelitis optica spectrum disorders
  1. Valentina Camera1,2,3,
  2. Silvia Messina1,3,
  3. Kariem Tarek Elhadd4,
  4. Julia Sanpera-Iglesias5,
  5. Romina Mariano1,3,
  6. Yael Hacohen6,7,
  7. Ruth Dobson8,
  8. Stefano Meletti2,9,
  9. Evangeline Wassmer10,
  10. Ming J Lim5,11,
  11. Saif Huda4,
  12. Cheryl Hemingway7,
  13. Maria Isabel Leite1,3,
  14. Sithara Ramdas12,13,
  15. Jacqueline Palace1,3
  1. 1 Nuffield Department of Clinical Neuroscience, University of Oxford, Oxford, UK
  2. 2 Department of Biomedical, Metabolic and Neurosciences, University of Modena and Reggio Emilia, Modena, Italy
  3. 3 Department of Clinical Neurosciences, John Radcliffe Hospital, Oxford University Hospitals NHS Trust, Oxford, UK
  4. 4 The Walton Centre for Neurology and Neurosurgery, The Walton Centre NHS Foundation Trust, Liverpool, UK
  5. 5 Children’s Neurosciences, Evelina London Children's Hospital, London, UK
  6. 6 Department of Neuroinflammation, Queen Square MS Centre, University College London, London, UK
  7. 7 Department of Paediatric Neurology, Great Ormond Street Hospital for Children, London, UK
  8. 8 Preventive Neurology Unit, Wolfson Institute of Preventive Medicine, Queen Mary University of London, London, UK
  9. 9 Neurology Unit, Azienda Ospedaliero-Universitaria di Modena, Modena, Italy
  10. 10 Department of Paediatric Neurology, Birmingham Women’s and Children’s Hospitals NHS Foundation Trust, Birmingham, UK
  11. 11 Department of Women and Children’s Health, King's College London, London, UK
  12. 12 Department of Paediatric Neurology, Oxford Radcliffe Hospitals NHS Trust, Oxford, UK
  13. 13 Department of Paediatrics, University of Oxford, Oxford, UK
  1. Correspondence to Dr Jacqueline Palace, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK; jacqueline.palace{at}


Objective To describe onset clinical features predicting time to first relapse and time to long-term visual, motor and cognitive disabilities in paediatric-onset aquaporin-4 antibody (AQP4-IgG) neuromyelitis optica spectrum disorders (NMOSDs).

Methods In this retrospective UK multicentre cohort study, we recorded clinical data of paediatric-onset AQP4-IgG NMOSD. Univariate and exploratory multivariable Cox proportional hazard models were used to identify long-term predictors of permanent visual disability, Expanded Disability Status Scale (EDSS) score of 4 and cognitive impairment.

Results We included 49 paediatric-onset AQP4-IgG patients (38.8% white, 34.7% black, 20.4% Asians and 6.1% mixed), mean onset age of 12±4.1 years, and 87.7% were female. Multifocal onset presentation occurred in 26.5% of patients, and optic nerve (47%), area postrema/brainstem (48.9%) and encephalon (28.6%) were the most involved areas. Overall, 52.3% of children had their first relapse within 1 year from disease onset. Children with onset age <12 years were more likely to have an earlier first relapse (p=0.030), despite showing no difference in time to immunosuppression compared with those aged 12–18 years at onset. At the cohort median disease duration of 79 months, 34.3% had developed permanent visual disability, 20.7% EDSS score 4 and 25.8% cognitive impairment. Visual disability was associated with white race (p=0.032) and optic neuritis presentations (p=0.002). Cognitive impairment was predicted by cerebral syndrome presentations (p=0.048), particularly if resistant to steroids (p=0.034).

Conclusions Age at onset, race, onset symptoms and resistance to acute therapy at onset attack predict first relapse and long-term disabilities. The recognition of these predictors may help to power future paediatric clinical trials and to direct early therapeutic decisions in AQP4-IgG NMOSD.

  • paediatric neurology

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.

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 spectrum disorder (NMOSD) is an inflammatory demyelinating disease of the central nervous system (CNS) affecting predominantly the optic nerves and the spinal cord.1 Approximately 60%–80% have disease-specific aquaporin-4 antibodies (AQP4-IgGs) in the serum, which, due to the astrocytic location of the AQP4 water channels, lead to a primary autoimmune astrocytopathy2–4 with a characteristic relapsing course. Only about 5% of AQP4-IgG-seropositive NMOSDs have a paediatric onset.4 5 Data on long-term clinical outcomes of AQP4-IgG-seropositive paediatric patients are sparse due to the rarity of the disease,6–9 and further evidence is needed particularly because of the new therapies coming through.5 This is a long-term outcome study of a relatively large single-country paediatric-onset AQP4-IgG NMOSD cohort recruited from several UK neurology centres.


We retrospectively analysed data prospectively collected from databases and clinical notes (recorded from 1980 to July 2020) of patients with paediatric-onset (defined as <18 years) AQP4-IgG NMOSD (diagnosed according to the 2015 diagnostic criteria1) from six tertiary UK neuromyelitis optica centres: (1) John Radcliffe Hospital, Oxford, UK (adult and paediatric); (2) Great Ormond Street Hospital, London, UK (paediatric); (3) Walton Centre, Liverpool, UK (adult and paediatric); (4) Evelina London Children’s Hospital, UK (paediatric); (5) Birmingham Children’s Hospital, Birmingham, UK (paediatric); and (6) Royal London Hospital, London, UK (adult). The presence of serum AQP4-IgGs was identified in the Autoimmune Neurology Laboratory using a cell-based assay as described by Waters et al.10

All data were collected as standard clinical care at the respective centres (6 monthly neurological and ophthalmological follow-ups in remission, serum AQP4-IgG testing, diagnostic MRIs and cerebrospinal fluids -CSFs- where diagnostically useful) and subsequently shared according to local trust policy with the coordinating centre in Oxford. Information was collected on sex, age at onset, race, onset attack type and severity at nadir, relapses, acute and maintenance immunosuppressive therapy, time to first relapse, time to long-term visual disability (best eye worse than 6/36 for longer than 6 months), motor disability (walk ≤500 m unaided for longer than 6 months, Expanded Disability Status Scale (EDSS) score of 4) and cognitive impairment (defined by neuropsychological assessments or documented learning disability requiring extra support at school).

Neurocognitive testing was performed using a varied battery of age-appropriate standard tests (online supplemental table 1) of cognition dictated by clinical need as outlined11 by clinical psychologists/neuropsychologists working with tertiary adult and/or paediatric neuroscience centres. These tests typically include a measure of episodic memory, language, attention and executive functioning. Cognition was considered impaired if any test within these domains fell below the fifth percentile. In patients where formal cognitive assessment was not performed, cognitive impairment was pragmatically defined as documented learning disability requiring extra support at school, which was defined as requiring a minimum of a special educational needs coordinator. Box 1 details the variables collected for each patient and definitions of the clinical phenotypes, severity scoring, and visual, motor and cognitive disability outcomes.

Supplemental material

Box 1

List of the collected clinical variables and definitions

  • Sex: male and female.

  • Age at onset (continuous variable): <12 years; age at onset subgroups: 12–18 years.

  • Race: white, Asian, black and mixed.

  • Disease duration (months).

  • Onset syndromes

    • Monolateral or bilateral optic neuritis (ON).

    • Transverse myelitis (TM).

    • Area postrema syndrome and/or brainstem syndrome (BS).

    • Cerebral syndrome (CS) (acute disseminated encephalomyelitis (ADEM) syndrome, diencephalic syndromes or encephalopathies without typical MRI features of ADEM).

    • Multifocal syndromes.

  • Severity of the onset attack, defined as

    • Inability to walk unaided at TM attacks nadir.

    • Visual acuity 6/60 or worse in affected eye at ON attacks nadir.

    • Documented reduction of consciousness or vomiting for more than 7 days with weight loss at CS and BS attacks nadir.

  • Age (years) and time (months) from onset to aquaporin-4 antibodies detection.

  • Acute therapy of the onset attack: intravenous methylprednisolone, intravenous immunoglobulins, plasma exchange and no therapy.

  • Relapsing or monophasic course.

  • Time to first relapse (months).

  • Annualised relapse rate before and after the immunosuppressive therapy initiation.

  • Time from onset to immunosuppressive therapy initiation (months).

  • First and current long-term immunosuppressive therapy: azathioprine, mycophenolate mofetil and rituximab, methotrexate and cyclophosphamide.

  • Discontinuation of long-term immunosuppressive therapy due to relapse or intolerance.

  • Development of permanent visual disability after the onset attack or during the disease course,* defined as visual acuity in the best eye worse than 6/36 on the Snellen chart.

  • Time to permanent visual disability (months).

  • Development of permanent motor disability after the onset attack or during the disease course,* defined as a persisting Expanded Disability Status Scale (EDSS) score of 4, 6 or 8.

  • Time to permanent motor disability (EDSS score 4) (months).

  • Development of cognitive impairment after the onset attack or during the disease course,* defined by neuropsychological assessments and/or documented learning disability requiring extra support at school.

  • Time to cognitive impairment (months).

  • Death after the onset attack or during the disease course.

  • Time to death (months).

  • *Recorded at 6 months from onset attack or confirmed at 6 months after the last relapse.

Statistical analysis

Descriptive and group comparison analyses were conducted considering the low number of observations. Categorical variables were presented as absolute frequencies and percentages and continuous variables as mean±SD or median and range. Mean and median differences between the two groups were analysed using unpaired t-test or Mann-Whitney U test, respectively. Analysis of variance and Kruskal-Wallis tests were used to compare means and medians of more than two groups. Fisher’s exact test was applied to compare proportions among groups. We used Kaplan-Meier curves to depict time to first relapse, to visual disability, to EDSS score of 4 and to cognitive impairment (dependent variables) among groups, and groups differences were compared with log-rank test. Univariate cox proportional hazard model was performed to calculate survival rates of the aforementioned dependent variables. Independent continuous and categorical variables included sex, age at disease onset, onset clinical presentations, onset attack severity, acute therapy, second-line acute therapy and time to immunosuppression. An exploratory multivariable Cox proportional hazard model was performed using those predictors resulting in a p value of <0.10 in the univariate model. We evaluated the possible violation of the proportional hazard assumption with Schoenfeld residuals. Statistical analysis was performed using Stata V.14.0 software.


Study population

We collected data on 49 patients with paediatric-onset AQP4-IgG-seropositive disease with a median disease duration of 79 months (range 2–401). Female to male ratio was 7:1 and median current age was 21 years (range 7-54). In our cohort, 38.8% (n=19) of patients were white; 34.7% (n=17) were black; 20.4% (n=10) were Asian; and 6.1% (n=3) were mixed. Mean age at disease onset was 12±4.1 years.

Twelve patients were previously included in a retrospective international multicentre analysis, but additional ‘time to’ data, cognitive outcomes and acute attack treatment were obtained for this analysis.9

Table 1 shows the demographic and clinical descriptive features and group comparisons. Patients with mixed black ancestry were included in the black ancestry group as they showed similarities in demographic and clinical features.

Table 1

Demographic, onset and disease course clinical features among groups in AQP4-IgG NMOSD with paediatric onset

Demographic and onset clinical findings

White children had significant younger age at onset compared with the other races (p=0.008), while black children were older than other races (p=0.025) (table 1). Children frequently presented with area postrema and/or brainstem syndrome (BS) (48.9%, n=24), unilateral or bilateral optic neuritis (ON) (47%; n=23, 12 had bilateral ON and 11 unilateral ON), cerebral syndrome (CS) (28.6%, n=14) and transverse myelitis (TM) (24.5%; n=12, 9 had longitudinally extensive TM and 3 patients had unknown spinal cord lesion length as MRI was not available at the time of their onset). Multifocal presentation, defined as two or more of the aforementioned CNS syndromes, was seen in 26.5% (n=13) of patients. No difference in onset presentation was noted when analysing patients according to sex, age at onset or race. A severe onset attack was seen in overall 75.5% (n=37) of patients. Onset TM events were severe (unable to walk unaided at nadir) in 9 out of 12 patients (75%), and 2 patients (16.7%) were left with an EDSS 6 score of (unable to walk >100 m unaided) after the onset event. Onset ON events were severe (visual acuity 6/60 or worse in the affected eye) in 18 out of 23 patients (78.3%), and 6 patients (26%) were left with visual disability (visual acuity in best eye worse than 6/36 on Snellen chart) after the onset event.

Disease course

A relapsing course was observed in 83.7% (n=41) of children with a median time to first relapse of 12 months (IQR=4–26). Overall, 52.3% patients with a minimum disease duration of 1 year had a relapse within 1 year from onset, and 66.7% of patients with a minimum disease duration of 2 years had a relapse within 2 years from onset. Median time to first relapse was 26 months (IQR=17–37) in those starting the immunosuppressive therapy around the onset attack (n=14), while those who were not immunosuppressed at their first relapse (n=32) had a median time to first relapse of 5 months (IQR=3–16) (p=0.039). It is noteworthy that those with monophasic disease had a significantly shorter follow-up time compared with those with relapsing disease (table 1), but they also had a non-significant shorter time to immunosuppressive treatment (median time 4.5 months vs 11 months, p=0.065). Age at onset between 12 and 18 years was associated with approximately half the risk of relapse over time compared with those less than 12 years at onset (figure 1A and table 2), despite no difference in time to immunosuppression between those aged <12 years and those aged 12–18 years at onset as observed (table 1). No patients had progression of disability outside of relapses, and of the eight who did not relapse, none developed progressive disability.

Figure 1

Most significant Kaplan-Meier curves illustrating early predictors affecting the probability of remaining free from first relapse, visual disability, EDSS score of 4 and cognitive impairment. (A) Time to first relapse was significantly shorter in children aged <12 years at onset than those aged 12–18 years. (B) Time to permanent visual disability was shorter in those who had ON at onset than those who did not. (C) Time to permanent visual disability was shorter in the white than black and Asian, and this difference was mainly due to the onset attack outcomes. (D) Time to EDSS score of 4 was significantly longer in Asian than in black and white. (E) Time to cognitive impairment was significantly shorter in those who presented a CS at onset than those who did not. (F) Time to cognitive impairment was significantly shorter in those who were treated with second-line acute therapy (intravenous immunoglobulin or plasma exchange) than those who did not receive second-line acute therapy (no therapy or intravenous methylprednisolone). AT, acute-onset attack therapy; CS, cerebral syndrome; EDSS 4; Expanded Disability Status Scale Score of 4; ON, optic neuritis; P, log-rank test p-value

Table 2

Survival and univariate analysis for early predictors of time to first relapse and time to long-term disability outcomes

Acute and maintenance therapy

At first clinical presentation, 65.3% (n=32) of patients were treated with intravenous methylprednisolone (IVMP), and 20.4% (n=10) required a second-line acute therapy (plasma exchange (PLEX) and/or intravenous immunoglobulins (IVIG)). Children who required treatment beyond IVMP were most likely to be black (p=0.018) and to present with multifocal involvement of CNS (p=0.004). At first clinical presentation, 30.6% (n=16) of patients did not have acute therapy: mild sensory symptoms or an under-recognised area postrema syndrome (n=10), unilateral ON (n=5) and one patient presenting with a encephalitic illness. Excluding the initial prednisolone cover, we initiated the first non-steroid maintenance immunosuppressive therapy after a median time of 10.5 months (range 2–400). Asian children received non-steroid maintenance therapy after a significant longer median time from onset (p=0.030) and had a higher mean number of relapses (p=0.036) as compared with the white and black groups (table 1). In order of frequency, the most used immunosuppressants were azathioprine (AZA, 2–2.5 mg/kg/day) (61.2%), mycophenolate mofetil (MMF, 10–20 mg/kg two times per day up to a maximum of 1.5 g two times per day if adult weight) (14.3%) and rituximab (RTX, repeated every 6 months or when CD19 count rises; <50 kg 375 mg/m2, every week for four doses ;>50 kg two doses, 2 weeks apart) (12.2%) (table 1). At the last review, 16.7% (n=1) of patients had discontinued first-line RTX due to neutropenia (median disease duration 55 months, range 2–140). Discontinuations with AZA and MMF were 43.3% (n=13) (median disease duration 94 months, range 2–401) and 85.7% (n=6) (median disease duration 45 months, range 21–283), respectively. The percentages of discontinuation due to relapses were 53.8% (n=7) for AZA and 66.7% (n=4) for MMF. The percentage of discontinuation due to a lack of tolerability was 46.2% (n=6) for AZA and 33.3% (n=2) for MMF. There were no differences in medication use according to sex, age at disease onset and race (table 1). A shorter median time to initiating immunosuppressive therapy occurred in children presenting with TM (median 4 months, range 0–400) compared with those presenting with non-TM symptoms (median 11 months, range 0–216) (p=0.049).

Disability outcomes

The onset attack left 10 patients (20.4%) with at least one residual disability: 6 with visual disability (all bilateral ON related), 4 with EDSS score of 4 (3 TM related and 1 BS related) and 2 with cognitive impairment (both CS), with 2 patients having more than one disability. In this cohort with a median disease duration of 79 months (range 2–401), 36.7% (n=18) of patients had visual disability; 18.4% (n=9) reached EDSS score 4; 10.2% (n=5) reached EDSS score 6; and 2% (n=1) reached EDSS score 8. Of patients presenting with ON, 26.1% (n=6) were left with visual disability from onset. Of patients presenting with TM, 25% (n=3) were left with at least an EDSS score of 4 due to the onset attack. Cognitive impairment was present in 24.5% (n=12) at follow-up, and all cases were related to attacks involving the brain: 2 from the onset attack and 10 during later relapses (of whom 5 had CS at onset). Five required educational support, and 7 of the 12 had in addition a formal neuropsychometric assessment. Although the protocols varied, the the most common findings were impairment of processing speed and executive and attention functions. Of patients presenting with CS attacks, 14.3% (n=2) were left with cognitive impairment. During follow-up, two patients (4%) died. Neither died from a relapse; one died due to a choking event 1 year after discharge, having been left with bulbar problems from onset attack and was diagnosed postmortem. The second had cardiorespiratory arrest during the COVID-pandemic (COVID-19 PCR negative) after discharge at home. By the cohort median disease duration of 79 months, 34.3% (12/35) reached visual disability; 20.7% (6/29) reached an EDSS score 4; 14.8% (4/27) reached an EDSS score 6; 4.2% (1/24) reached an EDSS score 8; 25.8% (8/31) reached cognitive impairment; and 4.2% (1/24) died.

Early predictors of long-term disability

Table 2 summarises log-rank and univariate HRs relative to early predictors of the first relapse, visual disability, EDSS score 4 and cognitive impairment. Figure 1 shows Kaplan-Meier curves estimating the cumulative probability of remaining free from the aforementioned outcomes in relation to the most significant early clinical–demographical features. Multivariate Cox proportional hazard model results for each outcome are provided in table 3.

Table 3

Multivariable Cox hazard model for early predictors of time to visual and cognitive disability

Visual disability

ON phenotype was strongly associated with permanent visual disability (HR=7.8, p=0.002, 95% CI 2.2 to 27.7), even when excluded from the onset attack (HR=4.34, p=0.041, 95% CI 1.06 to 17.8) (figure 1B). None of the five patients who had second-line acute therapy in addition to IVMP for onset ON (black=3, white=1 and Asian=1) developed permanent visual disability vs 46.2% (6/13) of those ON treated with IVMP only (white=5, Asian=1 and black=0).

White children were approximately four times more likely to develop permanent visual disability compared with black children (HR=4.1, p=0.032, 95% CI 1.13 to 14.8) (figure 1C). Older children (12–18 years) at onset tended to have a non-significant longer time to visual disability than younger children (<12 years) (HR=0.44, p=0.093, 95% CI 0.17 to 0.15) (table 2). Longer time to immunosuppression was associated with a lower risk of developing visual disability (HR=0.98, p=0.026, 95% CI 0.97 to 0.99). However, this association was lost if patients who develop visual disability from onset attack were excluded (HR=0.98, p=0.067, 95% CI 0.98 to 1.00), possibly due to earlier introduction of immunosuppression.

Finally, from the exploratory multivariable Cox hazard model, white race and an onset which includes ON were both associated with an increased risk of visual disability (white race: HR=12.10, p=0.022, 95% CI 1.91 to 76.61) (ON: HR=6.70, p=0.016, 95% CI 1.43 to 31.34) (table 3).

Expanded Disability Status Scale 4

There was a trend towards non-Asians reaching an EDSS score of 4 considerably earlier than Asians (p=0.054) (figure 1D). Those with a TM onset attack did not reach EDSS score of 4 earlier than other primary attack sites (table 2).

Cognitive impairment

CS presentations were associated with a greater risk of cognitive impairment than other presentations (HR=3.22, p=0.048, 95% CI 1.02 to 10.2) (figure 1E), and this association remained after exclusion of the CS onset attacks (HR=5.52, p=0.017, 95% CI 1.36 to 22.5). Patients resistant to IVMP and treated with second-line acute therapy were more likely to develop early cognitive impairment (HR=3.99, p=0.034, 95% CI 1.11 to 14.4) (figure 1F). However, when these two factors were included in an exploratory multivariate hazard model, the significance was lost (table 3).


Key results and interpretation

This study reports the longest follow-up of paediatric-onset AQP4-IgG-seropositive NMOSDs to date (79 months, range 2–401) and represents patients within a single country with similar environmental and treatment protocols. Some patients were diagnosed clinically at onset and confirmed when the AQP4-IgG result became available, the longest after 32 years.

Children presenting under 12 years relapsed earlier than those aged 12–18 years. The non-white predominance that has been reported in adults was also observed in this study (2011 England and Wales census for 18–24 years old12), and predominance of black race was greater in those aged 12–18 years as compared with those under 12 years of age. Black children were also more likely to be refractory to IVMP and require treatment with acute second-line therapies. Prognostic differences were also seen: earlier time to (1) motor disability in non-Asians, (2) visual disability in white race and (3) cognitive impairment with IVMP resistant onset CS. Children with ON onset attacks and those with ON who did not escalate to second-line acute therapies also reached visual disability earlier.

Existing studies on long-term outcomes of paediatric-onset NMOSD were limited to mixed cohorts of seropositive and seronegative patients,8 had shorter disease duration,7 9 or had more heterogeneity across different countries and healthcare systems.8 However, we noted comparable demographic onset features (median age at onset, female to male ratio and black predominance) in a single-country (USA), 1 year of disease duration study on paediatric-onset AQP4-IgG disease.7 We reported an elevated incidence of area postrema syndrome as onset presentation compared with other paediatric AQP4-IgG NMOSD studies.7 9 This may relate to ‘hindsight’ diagnoses and improved awareness of the area postrema clinical presentation. RTX was superior to AZA and MMF as a first-line therapy, with less discontinuations (none of the six patients followed up for a median of 4.5 years had relapses). This is in keeping with a recent multicentre study that reported no relapses in children treated with first-line RTX followed up for a median time of 2 years.9 Despite methodological differences, similar studies found a higher probability of developing visual disability than motor disability in patients with paediatric-onset AQP4-IgG.4 7–9 The only study investigating predictors of disability in children found a similar percentage of patients with cognitive impairment (25.4%) although with a shorter median disease duration of 4 years, and a similar relationship between long-term visual disability development and ON presentations to our study.9

The use of survival analysis adjusts for varied disease durations among the different subgroups categorised by age, sex, race, onset presentations and therapy and permitted to report the risk of outcome at different timelines. To our knowledge, this method has not been used before to look at early predictors of disability in children and allowed us to compare our findings to a three-centre study on predominantly adult-onset patients (Oxford and Liverpool, UK, and Sendai, Japan),13 which used the same disability outcomes definitions.

Compared with Kitley et al,13 black race predominance was greater in our paediatric-onset cohort (34.7% in children vs 20.3% in predominantly adults). We noted this was particularly marked in those with onset 12–18 years. Although racial differences in age of puberty are reported with earlier age in black than white children, this seems unlikely to explain the rate of predominance of white children in younger ages and black children in older ages.14–16 Unfortunately, the date of the pubertal age was not available in our cohort. The proportion of patients presenting with CS, BS and multifocal syndromes was greater in children than adults, while the proportion of those presenting with TM or with ON was lower and similar, respectively, in our paediatric-onset cohorts than in predominantly adult-onset cohort. Children receiving second-line acute therapy due to IVMP resistance were mostly black and with multifocal presentations. As a consequence of such severity, clinicians administered immunosuppressive therapy earlier and after lower annualised relapse rates to black than to the other races, possibly resulting in a lower disability burden than the white race. Despite good diagnostic assays, children still had early relapses, especially in the younger age group (median time to first relapse of 5 months), often before immunosuppressive therapy was started (median time of 10.5 months). Time to first relapse in the older children group (17 months) was similar to the predominantly adult-onset cohort13 (14 months); however, that older study noted longer delay to immunosuppression than in our older children and therefore even the 12-18-year-old age group may have more active disease than adults.

The survival analysis on long-term disability outcomes confirmed the age effect on visual and motor disability.8 13 Despite similar median disease duration, children had double the risk of visual disability compared with the predominantly adult cohort (36.7% of children vs 18% of predominantly adults) but were three times less likely to reach EDSS score 6 (10.2% of children vs 34% of predominantly adults) and were 10 times less likely to reach EDSS score 8 (2% of children vs 23% of predominantly adults). Four per cent of children vs 9% of predominantly adults died over the similar disease durations.

We found that, as in adults and in another paediatric study,9 13 ON onset predicted long-term visual disability, and we found interesting that its acute treatment was a determinant factor. Paediatric patients with ON at onset had an higher proportion of severe onset attacks at nadir and had a greater residual visual disability rate after onset attack than predominantly adult-onset patients.13 Our paediatric patients with TM at onset had a higher proportion of severe onset attacks, but interestingly, a lower proportion was left with residual motor disability compared with the adult-onset cohort.13 This effect may highlight a possible increased susceptibility of the optic nerve and a reduced susceptibility of the spinal cord to AQP4-IgG-related inflammation in children compared with adults.17 Moreover, while ON presentation was predictive of shorter time visual disability, even excluding the onset attack influence, TM presentation did not predict long-term motor disability. However, the early use of immunosuppressive therapy in patients with TM over those without TM may have contributed. Patients with ON onset refractory to IVMP who were not treated with second-line acute therapy were more likely to develop early visual disability. Several studies have demonstrated the safety and efficacy of treating refractory ON with PLEX, even in children.18–20 Hence, we recommend the use of second-line acute therapy in any case that is refractory to IVMP to prevent residual blindness and, although the number was too small to study the time interval, it is likely the sooner acute therapy is given, the better the outcome.21

We found that white race was predictive of long-term visual disability. Although in the prevalent adult study13 black race was more likely to develop visual disability, in our paediatric onset cohort, white race had the highest risk, especially at onset attack and even adjusting for age at onset. This may reflect the inclusion of very young patients (<12 years) in our study. We found race to be predictive also of motor disability (EDSS score 4), as Asian children reached EDSS score 4 later than the other ethnic groups, despite having longer disease duration time, higher number of relapses and longer time to immunosuppression. These data are consistent with findings on a racial influence on motor disability.13

We found that cognitive impairment developed in around a quarter of children, and it was particularly frequent in CS presentations refractory to IVMP, even when the influence of the onset attack was removed. However, these predictors lost their significance in the multivariable Cox hazard model analysis, possibly because these predictors were interdependent. A larger prospective study may be able to establish whether these or other predictors are present. The proportion of cognitive impairment in AQP4-IgG NMOSD adults was estimated to be 29%–57%22; however, routine cognitive testing in AQP4-IgG-seropositive patients is not consistently done and is difficult to perform in very young children, thus a direct comparison is not possible. The prolonged effect of AQP4-IgG related brain inflammation might cause irreparable damage on the cortex and its connections,23 possibly altering or delaying the development of cognitive abilities in children. Cognition should be actively assessed in children and help offered from an early stage during rehabilitation and, because the risk of cognitive damage related to future attacks is greater in this group, more aggressive immunosuppression may be indicated.


This study was limited by the low number of observations due to the rarity of the disease, and results were not corrected for multiple comparisons. A prospective study analysing a bigger sample of children with this disease might be useful to perform a multivariate analysis with multiple comparison correction. Noting similar observations across different studies adds weight to our findings, and new findings require validation in future studies. Although 12 patients were included in a previous report,9 our study had several additional features: it was a single-country study, which removes heterogeneity of management regimes; with a longer disease duration (median 79 months vs 48 months), time to event was used instead of outcome at the end of follow-up; detailed cognitive outcomes, details of the onset attack severity and its treatment were included. Due to the long follow-up, patients presenting long ago might have been treated differently from those presenting currently. Future paediatric predictive studies should include genetic, environmental factors (ie, passive smoking, vitamin D/sun exposure and dietary habits) serum, immunological and CSF biomarkers including Neurofilaments light chain (NFL) and glial fibrillary acidic protein (GFAP).24 25


Paediatric-onset AQP4-IgG NMOSD is a disabling disease affecting predominantly the black children followed by Asian and then white children in the UK. Children have a more severe onset and shorter time to first relapse than adults and show a good response to first-line RTX compared with AZA and MMF. AQP4-IgG-seropositive children appear more susceptible to optic nerve damage and less to spinal cord damage than older population studies, which may reflect structural differences affecting capacity to repair. Second-line acute therapy may reduce visual disability in ON attacks. White race and ON presentation predict future visual disability. Non-Asian ethnicities are more likely to develop an EDSS score of 4, and TM at onset is not predictive of motor disability as this often occurred during subsequent TM attacks. A steroid-refractory onset attack involving cerebral hemispheres predisposes to long-term cognitive impairment development. With improved predictive data in paediatric AQP4-IgG NMOSD, clinical trial powering can be more accurately calculated, and individualised monitoring, support and treatment regimens can be offered with the aim of improving outcomes in children.

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Ethics approval

The study procedure was approved by the Oxford Research Ethics Committee C (ref 10/H0606/56). All patients signed the informed consent.


We thank the European Charcot Foundation for its support to this project.



  • Twitter @silvia06, @drruthdobson,

  • Contributors VC, SMes, SMel, MJL, CH, SH, SR and JP: conception and design of the work; VC, SMes, KTE, JS-I, RM, YH, RD, EW, MIL and SR: acquisition and analysis of data for the work; VC, SMes, KTE, JS-I, RM, YH, RD, EW, SMel, MJL, CH, SH, MIL, SR and JP: drafting the work or revising it critically for important intellectual content and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Funding European Charcot Foundation research fellowship funded Dr. Valentina Camera. No award/grant number.

  • Competing interests VC reports funding from the European Charcot Foundation supporting for the present manuscript. No award/grant number. SM reports grants from Biogen, Novartis, Bayer, Merck, Almirall and Roche; RD reports grants from Barts Charity, MS Society of Great Britain, Horne Family Charitable Trust and Celgene; grants and consulting fees from Merck and Biogen, consulting fees from Roche, Teva, Sanofi Genzyne and Janssen; MJL reports consulting fees from Advisory Board Octapharma and from Advisory Board Novartis; non-financial support from Advisory Board CSL Behring; grants from NIHR, GOSH Charity and Action Medical Research; CH reports consulting fees from Novartis, Biogen, UCB and Roche; JP reports grants and consulting fees from Merck Serono, Novartis, Biogen Idec, Teva, Abide and Bayer Schering; grants from MS Society and Guthie Jackson Foundation; payment for expert testimony and support for attending meetings and/or travel from Chugai Pharma and Bayer Schering, Alexion, Genzyme, MedImmune, EuroImmun, MedDay and ARGENX.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.