Introduction
Over the past 5 years, interest in and knowledge about pediatric multiple sclerosis (MS), including its treatment, pathogenesis, demographics, natural history, and MRI and laboratory features, have increased considerably. In this article, we review the currently available literature on this disorder, including the most recent advances in research. In general, studies of pediatric MS in North America have focused specifically on the population under the age of 18 years, although many other studies include only those under the age of 16 years. We will review studies of MS with onset under 18 years of age, including those studies using an earlier cut-off point.
Definitions
According to consensus definitions from the International Pediatric MS Study Group (IPMSSG),1 pediatric MS can be diagnosed after two clinical episodes of CNS demyelination that are separated by at least 30 days. No lower age limit is specified. According to these definitions, the Barkhof adult brain MRI criteria can be used to meet the requirement for lesion dissemination in space. Three of the following four features should be demonstrated: first, nine or more white matter lesions or one gadolinium-enhancing lesion; second, three or more periventricular lesions; third, one juxtacortical lesion; and fourth, an infratentorial lesion. These adult MRI criteria have not, however, been validated in children. The combination of an abnormal cerebrospinal fluid (CSF) test and two lesions on MRI, of which one must be in the brain, can also meet the dissemination in space criteria. The CSF must show either at least two oligoclonal bands (OCBs) or an elevated IgG index.
MRI might also be used to satisfy criteria for dissemination in time following the initial clinical event, even in the absence of a new clinical demyelinating event. New T2-bright or gadolinium-enhancing foci must develop 3 months or more after the initial clinical event.
Importantly, an episode consistent with the clinical features of acute disseminated encephalomyelitis (ADEM) cannot at present be considered to be the first event of MS, although our experience over the past 4 years is challenging this idea, as detailed below (see Figure 1 for a diagnostic algorithm).1 Limited diagnostic criteria are available to clarify the clinical distinction between ADEM and a first attack of MS, emphasizing the need for accurate biomarkers that can distinguish monophasic from recurrent demyelinating conditions.
Figure 1 | Diagnostic algorithm of pediatric onset demyelinating disorders.
Abbreviations: ADEM, acute disseminating encephalomyelitis; CIS, clinically isolated syndrome; CRION, chronic relapsing inflammatory optic neuropathy; NMO, neuromyelitis optica; RRMS, relapsing–remitting multiple sclerosis.
Epidemiology
Incidence and prevalence
The worldwide prevalence of pediatric MS is unknown, but data are available from individual countries or MS centers. A large series of 17,934 adult MS cases enrolled from 13 participating centers in France and Belgium identified 394 individuals (2.2%) who gave a history of MS onset at the age of 16 years or under.2 Other studies using large data sets, including 4,399 cases from a US center in Boston,3 3,375 cases from Italy,4 and 3,223 cases from Canada,5 have found prevalence rates of reported MS onset in childhood ranging from 3.1–4.4% of all MS cases. Age cut-offs in these studies were 18 years in the Boston study and 16 years in the Italian and Canadian studies.
The worldwide incidence of pediatric MS is also unknown, although a study from Canada of initial demyelinating events occurring before the age of 18 years, including first events of MS, neuromyelitis optica, ADEM, complete transverse myelitis and recurrent optic neuritis, reported an incidence of 0.9 per 100,000 individuals.6 Taken together, pediatric MS cases represent a small but important subset of the MS population. Whether the incidence of MS in children has increased in the past few decades, as has been reported in adults, is unclear.
Demographic features
The female:male ratio of pediatric MS varies by age. Below the age of 6 years, the ratio of girls to boys is 0.8:1. This ratio increases to 1.6:1 between the ages of 6 and 10 years, and to 2.1:1 for children over the age of 10 years, compared with a ratio of 
3:1 in adults.7
MS in adults is most common in non-Hispanic white individuals, but several studies have pointed out marked racial and ethnic variability in the pediatric population in North America. North American data published on race and ethnicity in pediatric MS reflects US government definitions, which divide race into five categories (American Indian or Alaska Native, Asian, black or African American, Native Hawaiian or other Pacific Islander, and white), and ethnicity into two categories (Hispanic or Latino, and not Hispanic or Latino).8 In a pediatric MS center in Boston, a higher proportion of blacks or African Americans was recorded in the pediatric-onset MS group than in the adult-onset MS group (7.4% versus 4.3%).3 Pediatric MS in African Americans seems to have a more severe presentation early on in the disease course compared with cases in white individuals.9 Among a sample seen at the Pediatric MS Center on Long Island in New York, higher proportions of individuals of African American race, Hispanic ethnicity, and first-generation American origin were observed in pediatric than in adult MS cases. The proportion of pediatric MS African American or Hispanic cases was also greater than in the pediatric census for the same geographical area.10 A review of pediatric and adult patients followed up in a Canadian center showed that the pediatric cases were more likely to report Caribbean, Asian or Middle Eastern ancestry and less likely to report European ancestry than the adult cases.11 A similar phenomenon has been observed among American children with MS relative to adults.10 The reasons behind this greater diversity in ethnicity, race and ancestry in pediatric MS remains unclear, but it could represent a combination of genetic and environmental influences, as well as changing regional demographic factors affecting that age range in North America (for example, increased influx of immigrants of non-European ancestry across North America in recent years).
Etiology
Pathophysiology and immunology
Limited data are available regarding the underlying immunopathophysiology of pediatric MS, as no comprehensive studies have been performed. In addition, no systematic studies have compared the immunopathophysiology of pediatric MS with that of adult MS. A study published in 2008 of a large cohort of children with CNS inflammatory demyelination, type 1 diabetes or CNS injury demonstrated that children with these conditions exhibited heightened peripheral T-cell responses to a wide array of self antigens.12 Children with autoimmune diseases or CNS injury also exhibited abnormal T-cell responses against multiple cow milk proteins.12 A smaller study evaluating T-cell responses to myelin basic protein (MBP) and myelin oligodendrocyte glycoprotein (MOG) epitopes in adult and pediatric MS found similar responses to specific peptides—predominantly MBP83–102, MBP139–153, MBP146–162, MOG1–26, MOG38–60 and MOG63–87—in both groups.13 Interestingly, responses to fetal MBP were minimal, and again were similar in both groups. A tetramer approach, which enables detection of autoantibodies against folded proteins, revealed that up to 20% of children with ADEM, but none with pediatric MS, demonstrated elevations of anti-MOG antibodies.14
CSF studies have demonstrated that children under 11 years of age exhibit a distinct cellular profile in relation to their adolescent counterparts (D. Chabas et al., unpublished work). Compared with children 11 years or older, younger children with their first MS event were more likely to lack OCBs or an elevated IgG index, and had a higher percentage of neutrophils in their CSF, suggesting prominent activation of the innate immune response, as opposed to the typical activation of the adaptive response seen in older patients.
Studies examining markers of axonal damage in the CSF found minimal changes in most children with MS; however, a subgroup with prominent clinical symptoms at the time of CSF examination exhibited elevated levels of tau protein.15 The implications of this finding are unclear.
Environmental risk factors
Environmental factors have been shown to have a pivotal role in adult MS susceptibility. Relevant factors include the patient's latitude of residence, exposure to viruses—in particular, Epstein–Barr virus (EBV)—smoking, and vitamin D status.
Determining the importance of viral exposure in adult MS has proved difficult, given the lag time between the exposure and disease onset. Furthermore, by adulthood, most individuals have encountered most common viruses, but only a minuscule fraction of the population will have developed MS. The pediatric population provides a unique opportunity to study the role of viruses in MS, given the close temporal relationship between the infection and MS onset, and the fact that children are less likely than adults to have been exposed to a wide range of viruses. Banwell et al. compared 137 children with definite MS in the Americas and Europe with age-matched controls.16 No difference was observed between the two groups with regard to seropositivity to cytomegalovirus, herpes simplex virus type 1, varicella zoster virus (VZV) and parvovirus B19, but EBV seropositivity was associated with an almost threefold increased likelihood of MS. Similarly, Pohl et al. studied 147 patients with pediatric MS and paired controls and found that seropositivity for EBV was more prevalent among the patients than the controls (99% versus 72%, P = 0.001).17
Data from the US Pediatric MS Network confirm the substantial association between EBV and increased pediatric MS susceptibility.18 This association seems to be stronger for EBV nuclear antigen 1 (EBNA1) than for the EBV viral capsid antigen, and remains so after adjusting for age and race.18 Further supporting the specific role of EBNA1 in pediatric MS susceptibility, higher EBNA1 titers are found in positive pediatric MS patients than in EBNA1-positive controls. Moreover, the humoral response to EBNA1 in patients with pediatric MS targets at least two regions of EBNA1 that do not seem to be targeted by antibodies in age-matched controls.19
Mikaeloff et al. evaluated 137 children with MS and 1,061 controls for clinical episodes of chicken pox.20 77% of the MS population had a history of chicken pox, compared with 85% of the control population (odds ratio 0.58, 95% CI 0.36–0.92), suggesting a possible protective effect of VZV.
Besides common viral exposures, concern has been raised over the use of vaccinations—most recently, hepatitis B vaccine—and the subsequent development of MS. Mikaeloff et al. found that there was no increased risk of developing a first episode of childhood MS up to 3 years post vaccination in the French population.21 These authors did, however, show a trend for the hepatitis B vaccine to increase the risk of MS in the longer term. These findings need to be confirmed in larger cohorts. A second study by the same authors found no increase in the relapse rate after a first episode of CNS inflammatory demyelination in childhood when the patients were subsequently vaccinated against hepatitis B or tetanus.22 Furthermore, the risk of conversion to MS did not seem to increase after vaccination, although the study was insufficiently powered to detect a small increase in risk.
The same group evaluated the risk of childhood-onset MS in relation to exposure to passive smoking.23 The investigators compared 129 cases of pediatric MS in France with 1,038 controls matched for age, sex and place of residence. The risk of a first episode of MS in individuals exposed to parental smoking was found to be over twice that in individuals whose parents did not smoke, and was even higher for those with prolonged exposure of 10 years or more.
In adults, low levels of vitamin D have been linked to MS susceptibility. This relationship has not yet been confirmed in children, although levels of 25-hydroxyvitamin D3 are lower than the normal range in the vast majority of pediatric patients seen at pediatric MS clinics.24
Whether the association of EBV infection or exposure to smoking with increased pediatric MS susceptibility overlaps with socioeconomic status, or whether these associations are independent, or even synergistic, is currently unclear. Further studies are required to explore the precise mechanisms leading from specific environmental exposures to disease onset.
Diagnosis
Differential diagnosis
White matter abnormalities on MRI in childhood have a broad range of differential diagnoses. Inflammatory, infectious, metabolic and neurodegenerative disorders can all present with such abnormalities.25 Careful history taking and physical examination are necessary, but not sufficient, to distinguish between demyelinating disorders, acute infectious or vascular processes, and chronic degenerative or metabolic processes (Box 1). Later in this article, we review imaging, laboratory and clinical features that help to establish the diagnosis of MS in children.
In many cases, monophasic demyelinating conditions are difficult to distinguish from a first demyelinating episode of pediatric MS, especially in younger patients. ADEM has been defined as a first clinical event, affecting the CNS, of inflammatory or demyelinating etiology with an acute or subacute onset and involving a polysymptomatic presentation. According to proposed consensus definitions published in 2007, encephalopathy—defined by the IPMSSG as behavioral change, such as confusion or excessive irritability, or alteration in consciousness, such as coma or lethargy—is required to reach a diagnosis of ADEM.1 Studies applying the IPMSSG definitions in retrospective cohorts suggest that encephalopathy is useful in distinguishing clinically isolated syndrome (CIS—the first clinical MS event) from ADEM.26, 27 Some patients initially presenting with an ADEM phenotype, however, experience recurrent demyelinating events without encephalopathy, sometimes with accumulation of T2-bright lesions in the brain, and are later diagnosed with MS or neuromyelitis optica.28 Increasingly, data suggest that encephalopathy can be associated with a first episode of MS, especially in younger patients.16, 29, 30, 31 Rather than being disease specific, therefore, the presence of encephalopathy could be related to immaturity of the brain or the immune system in younger patients.
No reliable predictive clinical, biological or MRI markers are currently available to identify a first episode of MS in children who present with an ADEM-like initial demyelinating event. Patients with an ADEM-like presentation whose MRI is suggestive of MS at onset do, however, have an increased risk of developing a second episode.32
Clinical presentation
Over 25 retrospective and prospective studies involving more than 10 patients have described the clinical presentation of pediatric MS, amounting to a total 1,573 patients, 316 of whom were prospectively evaluated.4, 5, 26, 29, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54 Both retrospective and prospective MS studies report similar ages of onset and female:male ratios, but considerable variability exists in the frequency of initial MS symptoms. In a combined analysis of 4 prospective studies, 28% of patients presented with cerebellar findings (range 12–52%), which is significantly higher than the 11% of patients evaluated in 21 retrospective studies (P <0.0001). Similarly, sensory deficits were identified in 27% of patients in the prospective studies but in only 15% of patients included in the retrospective studies (P <0.0001). These differences could reflect recall bias in the retrospective studies, although other neurological signs that would be expected to be recalled with higher frequency (for example, motor deficits), or underreported (for example, optic neuritis) in younger children, were reported with equal frequency in retrospective and prospective studies (27% with motor deficits; >20% with optic neuritis). Brainstem deficits were also identified at similar rates (19% of patients in retrospective studies; 22% in prospective studies). Monosymptomatic presentation is typical in adult MS, but polysymptomatic presentation is more common in pediatric MS. Over one-third (35%) of retrospectively analyzed patients and over half (53%) of prospectively analyzed pediatric patients (P <0.0001) exhibited multiple neurological deficits at presentation. Systematic recording of neurological deficits from the onset of disease could account for the increased frequency of polysymptomatic presentation reported in prospective studies. Most studies did not include sufficient detail to assess whether the symptoms were attributable to a single lesion (for example, a brainstem lesion) or to multiple active lesions.
Mental status has not been consistently described in either retrospective or prospective studies, and is sometimes not discussed at all. Encephalopathy was reported in a subset of studies (8 of 21 retrospective studies, accounting for 46% of the retrospectively analyzed patients, and 2 of 4 prospective studies, accounting for 60% of the prospectively analyzed patients). Criteria for altered mental status or encephalopathy, if described, were not used consistently across studies. Overall, 16% (29 of 180) of prospectively analyzed patients were described as having encephalopathy—a significantly higher proportion than the 8.7% (62 of 712) recorded in the subset of retrospectively analyzed patients (P <0.0001). These differences could be attributable to recall bias, or to heightened awareness of evaluating mental status during demyelinating episodes.
Diagnostic testing
Diagnostic testing for pediatric MS includes serum and CSF evaluation, visual evoked potentials (VEPs), and MRI.
Cerebrospinal fluid analysis
As noted above, the CSF profile in childhood-onset MS can vary by age. The IgG index has been found to be elevated in 68% of adolescents (
11 years of age) with MS, but in only 35% of younger children (<11 years of age; D. Chabas et al., unpublished work). Conversely, younger children have more neutrophils in the CSF than older children. These features tend to depend on age rather than disease duration. The distinct CSF IgG and cellular profile in younger children tends to vanish on repeat CSF analysis (mean 19 months after initial analysis), suggesting a transient immunological phenomenon associated with disease onset. Interestingly, absence of neutrophils in the CSF at disease onset is predictive of an earlier second neurological episode.
One study has reported OCBs to be present in the CSF of up to 92% of children with MS,55 but the presence or absence of this feature could depend on the individual laboratory, the disease duration and/or the age of the patient. Another study also found OCBs to be less frequent in younger children (43%) than in adolescents (63%; D. Chabas et al., unpublished work). OCBs are found in up to 29% of cases of ADEM.29, 30, 55, 56 Within the French KIDMUS (Kids with Multiple Sclerosis) cohort, 94% (69 of 72) of children who were positive for OCBs went on to develop MS.48 Only 40% of patients with established MS in this study had OCBs, however, suggesting that this feature has low sensitivity but high specificity for the development of MS when present at disease onset. These results must be interpreted with caution, as information regarding timing of lumbar puncture and OCB detection was not provided in this paper, nor is it clear how many children included in the study underwent analysis of OCBs and/or other CSF markers.
Visual testing
Approximately one-third of children who go on to have a diagnosis of MS experience optic neuritis as an initial presenting symptom,32, 45 and an even higher proportion might experience subclinical abnormalities of the visual pathway.57 The limited capacity of standard Snellen charts to measure subtle visual dysfunction is well documented in the adult MS population.58 Low-contrast letter acuity charts (Sloan charts) have been shown to provide a sensitive and reliable assessment of visual acuity in the pediatric MS population.59
Other tests, such as VEPs or pattern-reversal VEPs, have been shown to be of diagnostic utility in childhood MS. In one study, almost half of patients with this condition showed increased visual latency that revealed a second focus of demyelination before a second clinical attack.57
Optical coherence tomography (OCT), which is often used in patients with glaucoma, is now being applied to pediatric patients with MS. This procedure uses near-infrared light to quantify the thickness of the retinal nerve fiber layer, which contains only nonmyelinated axons. OCT has been shown to provide a sensitive evaluation of the thickness of this layer, which correlates with optic atrophy, in children with MS.59 Taken together, VEP, OCT and Sloan chart testing can provide objective evidence of previous inflammatory insult to the optic nerve in the pediatric MS population. These approaches could help both in establishing a diagnosis of MS and in monitoring disease progression.
MRI scanning
In the past, limited data have suggested that pediatric MS patients might not meet the MRI criteria for adult MS.60 In a small study, pediatric patients with MS were reported to have fewer T2-bright foci and more-frequent tumefactive MS lesions on brain MRI than was reported in adults with MS (Figure 2).31 More-recent data collected at disease onset, however, have shown that children with MS could have a higher lesion burden on their initial brain MRI scan than adults, especially in the brainstem and cerebellum.61 This finding should raise concern, as both higher lesion burden and brainstem and cerebellar involvement have been reported to be associated with a worse outcome in adults.
Figure 2 | Brain MRI scans of young patients with multiple sclerosis.
a | 7 year-old child. b | 16 year-old adolescent. Note the large, poorly defined lesions in the scan of the younger child.
The fact that younger children with MS (age <11 years) can present with atypical MRI features is intriguing, and could complicate diagnosis and possibly delay the initiation of preventative therapies.31 Brain lesions in younger children are typically large with poorly defined borders, and are frequently confluent at disease onset (Figure 2). Such T2-bright foci in younger children can vanish on repeat scans, unlike those seen in teenagers or adults. Disease processes in the developing brain, including the immune response, might, therefore, be different from those in older patients.31
Several studies have evaluated the utility of MRI in the diagnosis of MS. One study suggested that the presence of well-defined periventricular and corpus callosal lesions is a specific but insensitive predictor of MS after a first attack in childhood (KIDMUS criteria).32 Another study has confirmed these general findings, and has found that the presence of any of the Barkhof criteria for MS, as well as lesions that are perpendicular to the corpus callosum and are small or well defined, could be effective in differentiating children with MS from those with a monophasic disease at the time of disease presentation.62 A small study using some of the same patients reported that a combination of any two of the following criteria can distinguish a first attack of pediatric MS from ADEM:63 presence of T1 black holes; presence of two or more periventricular lesions; and absence of diffuse bilateral lesions. Importantly, this study failed to analyze the presence of gadolinium-enhancing foci, which can also appear dark on T1-weighted imaging. Advanced imaging techniques have not yet been evaluated for the assessment of future risk of MS in children, and the use of standard MRI techniques to diagnose MS early in young patients must be refined.
Pathology
Pathology can lead to a definitive MS diagnosis, but biopsies are rarely performed in children owing to concerns regarding potential morbidity from the procedure. Biopsies are usually performed on patients with atypical clinical or radiological presentations. No studies have yet described pathological differences between adult and pediatric MS.
Disease progression
Prognostic factors
Besides MRI features, no clear prognostic factors at the time of presentation determine whether a child with an acute demyelinating event will go on to develop MS. Clinical studies have been hampered in part by inconsistency of definitions used across publications, and the small numbers of individuals studied at any one site. In the available studies, the risk of developing MS after ADEM has been reported as 0% in a study from Argentina,30 9.5% in a study from San Diego,64 and 18–29% in studies from France.32, 48 Variability in the criteria used to define ADEM and pediatric MS might have contributed to the wide range of published incidence figures.
The French KIDMUS study group examined pediatric patients with an initial demyelinating event, including CIS-like and ADEM-like events, and showed that overall 57% developed a second attack during a mean follow-up period of 5.4 years.48 86% of patients with initial optic neuritis developed MS, whereas 50% of those with an initial brainstem syndrome developed MS. Overall, the main positive predictive factors for the development of MS were age at onset 10 years and optic nerve involvement. A relatively low risk of developing MS was found in patients with mental status change at presentation, suggesting that the presence of encephalopathy could be a negative predictive factor. Of patients with an initial diagnosis of ADEM, 29% developed MS. In a subsequent publication by the same group, after the diagnosis of ADEM had been redefined to include 'change in mental status' as a qualifying criterion, 18% of children were found to develop MS, as defined by the development of a second event during follow-up.65
Clinical course
The initial clinical course in most patients with childhood-onset MS is relapsing–remitting, with remission of neurological symptoms being followed by relapse rather than progressive neurological disability in 85.7–100% of cases.5, 34, 66 After several days or weeks, patients recover partially or fully from their exacerbations and remain clinically stable until they develop their next MS relapse. During the relapsing–remitting phase, disability can occur as a result of incomplete recovery from relapses. Some patients later develop a more insidious progression of disability with or without superimposed exacerbations. This phase of the disease is called the secondary progressive phase.
Relapses
The annualized relapse rate in pediatric-onset MS has been estimated to be between 0.38 and 0.87 for the whole relapsing–remitting period in the few studies that have examined patients with mean disease duration of 10 years or more.5, 37, 43 A prospective study of patients with MS seen at a large MS center showed that patients with an onset before the age of 18 years had a higher relapse rate during the first few years of their disease than adults seen at the same institution.67
The quality of recovery after subsequent relapses during the relapsing–remitting phase has been reported to be good, at least in the very early stages. This observation raises the question of whether children have less irreversible neuronal injury during their clinical relapses or/and have a better ability to repair than adults.
Evolution to the secondary progressive phase
The proportion of pediatric-onset MS patients reported as reaching the secondary progressive phase is highly variable from one study to the next, largely because of variability in the duration of follow-up.5, 45, 66, 68 The median time between onset of the disease (first neurological episode) and conversion to secondary progression, therefore, is more informative. The estimated median time from onset of MS to the secondary progressive phase has been reported to be 16–28 years in pediatric-onset MS, compared with 7–19 years for patients with adult-onset MS. Patients with childhood-onset MS tend, however, to be younger when they reach this stage (median age 31–41 years compared with 37.5–52 years in adult-onset MS). In summary, patients with childhood-onset MS convert to secondary progression on average 10 years later than patients with adult onset, but they are, on average, 10 years younger when they reach this phase of the disease.
Time to irreversible disability
Substantial interindividual variability exists in terms of rate of disability progression in MS, regardless of the age at onset.4, 34, 68 The estimated median time from MS onset to use of a unilateral device to ambulate is reported to be 28–29 years in pediatric MS, compared with 18 years for the adult group.5, 66 Despite the fact that it takes approximately 10 years longer for patients with childhood-onset MS to reach irreversible limitation of ambulation than patients with adult-onset MS, patients with childhood-onset MS reach these disability scores at a younger age (often around the time that they start a family and are young professionals). This observation contradicts the generally accepted idea of a more-favorable prognosis in the younger age group. The efficacy of disease-modifying therapies (DMTs) for MS with respect to delaying the development of the secondary progressive phase has yet to be studied.
Cognitive and psychosocial outcomes
The idea that patients with adult-onset MS experience cognitive defects, including those of complex attention, efficiency of information processing, executive function, processing speed and long-term memory, is widely accepted.69 Studies on early-onset MS have confirmed this association, with longitudinal data suggesting the presence of cognitive impairment in over half of patients with this condition after 10 years of follow-up.70, 71 Preliminary data from two North American centers (Stony Brook, NY, USA72 and Toronto, Canada73), describing outcomes in 37 and 10 patients, respectively, suggest marked cognitive impairment in a large proportion of patients with pediatric MS. In the larger study, impairments in attention, memory and confrontation naming were found in over one-third of patients. These impairments showed strong correlations with Expanded Disability Status Scale (EDSS) scores, number of relapses and total disease duration, and were found after a relatively short disease duration (19 months on average). Depression was noted to be present in half of the cases.
Similarly, the 10 children evaluated by the Toronto group were found to have deficits in executive function, processing speed and working memory, although no correlation with EDSS scores was found. The EDSS scores seen in this group were relatively low, however, with no child having a score worse than 1.5.73 These studies were both limited in that they lacked controls, were cross-sectional and so did not provide longitudinal data, and used relatively brief follow-up periods. In addition, testing was not performed systematically after a given disease duration (for example, 1 year). A small longitudinal study of 12 patients has provided preliminary data suggesting progressive cognitive decline,74 but again this was a small study that lacked controls.
A larger, multi-institutional, cross-sectional, Italian study evaluating 66 children with relapsing–remitting MS (RRMS) and 57 healthy controls was published in 2008.75 The group used a protocol that included IQ testing, as well as tests of verbal memory, visual memory, attention, executive functioning, expressive language, and receptive language. After an average disease duration of 3 years, almost one-third of patients met the criteria for significant cognitive impairment. Neither disease duration nor number of relapses was found to be of relevance when children with MS and cognitive impairment were compared with those who showed no such impairment. Strikingly, almost three-quarters of children with MS in this study reported fatigue, although in contrast to the earlier American study only 6% reported depression. Over half of the children reported that MS had negative effects on their everyday life and schooling.
The functional impact of cognitive impairment in children with MS is not clearly documented in these studies, but one might postulate that these impairments would translate to an increased need for school-based interventions and accommodations. Certainly, cognitive decline, disability level and a progressive course have all been found to influence social and work-related function in adults with early-onset MS.70 Importantly, several of the studies noted above showed variably increased rates of depression and fatigue, as well as negative effects on school and everyday life. These findings underline the need for multidisciplinary care, with particular attention being paid to school-related accommodations, as well as the emotional well-being of this patient population.
One study has reported on health-related quality of life in children with CIS or early MS.76 Although mean disease duration in this cohort of 51 patients was short (2.3 
1.8 years) and disability levels were low (median EDSS score 1.5), patients nevertheless had marked reductions in their health-related quality of life scores, emphasizing the dramatic consequences for the lives of children with this disease.
In a longitudinal, multicenter study from Italy, 70% of patients with pediatric MS showed worsening of cognitive deficits over 2 years.77 Most of the patients were on DMTs for MS, however, and otherwise remained relatively stable from a clinical perspective. Further longitudinal studies including controls should help us to elucidate whether progressive deterioration in cognitive function correlates with disease duration, and furthermore whether such deterioration corresponds to decline in social functioning. Finally, as a clear correlation between disease duration and cognitive impairment was not found in these preliminary studies, evaluation of neuropsychological outcomes in relation to MRI measures of atrophy could provide further insight into the factors that affect functional outcome.
Disease-modifying therapies
Four first-line disease-modifying therapies—glatiramer acetate, intramuscular and subcutaneous interferon (IFN)-
1a, and subcutaneous IFN-1b—have been approved for treatment of RRMS in the adult population, but data regarding these therapies in children are limited. At present, most treatment decisions in children are based in part on treatment studies performed in adults. The relatively small number of patients with pediatric MS presents practical barriers to performing double-blind, randomized controlled trials on first-line DMTs in this population, and no such trials have yet evaluated the efficacy of agents approved for adult MS or new agents. Use of DMTs also varies from country to country because of issues including—but not limited to—variations in insurance coverage and rate of patient follow-up.
Interferon 
The clinical benefit of IFN- therapy in RRMS is thought to be mediated via several mechanisms, including the inhibition of proinflammatory cytokines, induction of anti-inflammatory mediators, reduction of lymphocyte migration, and inhibition of autoreactive T-cells.78, 79 In adult MS, studies have demonstrated an 30% reduction in exacerbation rate compared with placebo for periods of 2–3 years.80
Two retrospective studies have evaluated the effects of IFN- on relapse rates in pediatric MS. An open-label study published by an Italian cooperative group reported a reduction in annualized relapse rate from 1.9 to 0.4 in 52 children on IFN-1a.81 Limitations of this study include a lack of untreated controls, its retrospective and nonrandomized nature (introducing the possibility of patient selection bias), and the relatively small number of patients included. Furthermore, the standard deviations of the initial and follow-up annualized relapse rates showed overlap (1.9 1.1 and 0.4 0.5, respectively), indicating that the actual effect of the medication could be marginal.
Some of these issues were addressed in a recent study published by the Kid Sclérose en Plaques (KIDSEP) group.82 This retrospective, nonrandomized, comparative cohort study evaluated the timing of the first attack after initiation of IFN- in 24 patients out of a cohort of 197 patients with pediatric MS. The risk of a first attack was reduced in children on IFN- at 1 year (hazard ratio [HR] 0.31, 95% CI 0.13–0.72) and 2 years (HR 0.40 (95% CI 0.2–0.83) compared with children not given IFN-. After 4 years, the HR was 0.57 (95% CI 0.30–1.10). This result suggests that although the relapse rate might decrease initially on IFN- therapy, the benefit is less clear after 4 years. The study might, however, have been underpowered to show an effect after this time. No significant differences in EDSS scores were found between the treated and untreated groups. Effects on neurocognitive performance were not explored in this study.
Retrospective case series suggest that IFN-1a and IFN-1b are safe and well tolerated in children at the same doses as those used in adults.83, 84, 85, 86 Reported adverse effects include flu-like symptoms (35–65%), leucopenia (8–27%), thrombopenia (16%), anemia (12%), and transient elevation in transaminases (21–33%).83, 84, 86, 87 Injection-site reactions (>66%), abscesses (6%) and injection-site necrosis (6%) can occur in children taking the subcutaneous formulation.86
An optimal dosing schedule for IFN- therapy has not been established in pediatric MS. Retrospective studies have, however, described titration schedules that follow adult protocols; that is, gradual titration to 30 mcg once weekly for intramuscular IFN-1a, and 22 mcg three times weekly or 44 mcg three times weekly for subcutaneous IFN-1a.81, 82, 83, 84, 85, 86, 87 Children over the age of 10 years seem to be able to tolerate full doses of IFN-, although tolerance might be reduced in the younger population.84 In the US Pediatric MS Network series of children with MS receiving therapy (mean follow-up 3.5 years, n = 264), of the children started on intramuscular IFN-1a therapy, 42 of 97 (42%) required change to another therapy, 24 owing to breakthrough disease and 18 owing to adverse effects and/or compliance problems.88 Of those who were initially started on subcutaneous IFN-1a therapy, 21 of 74 (28%) required a change to another therapy, 10 owing to breakthrough disease and 11 owing to adverse effects and/or compliance problems. Of the 31 patients started on subcutaneous IFN-1b, 15 of 31 (48%) required a change to another therapy, 10 owing to breakthrough disease and 5 owing to adverse effects and/or compliance problems.
Little information is available regarding neutralizing antibodies to IFN- in the pediatric MS population, although one small study has suggested that positive neutralizing antibodies might be less commonly seen in pediatric MS than in the adult population (E. A. Yeh et al., unpublished work).
Glatiramer acetate
Glatiramer acetate is the acetate salt of a mixture of synthetic polypeptides composed of L-alanine, L-glutamic acid, L-lysine and L-tyrosine. This drug is designed to mimic human MBP, and is postulated to induce a myelin-specific response mediated by suppressor T lymphocytes and to inhibit specific effector T lymphocytes, as well as affecting the function of antigen-presenting cells.89 Glatiramer acetate was found to reduce the number of relapses by 29% in adults with RRMS over a period of 2 years.90 One small retrospective study describing the use of glatiramer acetate in seven children with MS suggested that the medication is well tolerated.91 In the US Pediatric MS Network series of 56 children with MS who were initially started on this medication, 12 of 58 (21%) required change to another therapy—9 owing to breakthrough disease and 3 owing to adverse effects—over a mean follow-up period of 3.5 years.88
Treatment failure
First-line treatment failure is a concern in both adult and pediatric MS. The currently available first-line DMTs are accepted to be only partially effective, resulting in a reduction in relapse rate of 30% in the adult population.92 Treatment failure can also arise because of intolerable adverse effects or the presence of an unacceptable level of breakthrough disease (either in terms of severity or frequency), continued presence of gadolinium-enhancing lesions on MRI, or progression of disability and/or disease despite adherence to medication. The biological mechanisms underlying poor response to therapy have not been elucidated, but could involve heterogeneity of disease processes between individuals owing to genetic, immunological or environmental variability.
Compliance is a key issue in the treatment of children with MS, as daily to weekly injections of medication are required, and are frequently associated with adverse effects. 15% of children followed at the six centers participating in the US Network of Pediatric MS Centers of Excellence changed therapies owing to compliance issues.88
Definition of treatment failure is a challenging issue, and has been the subject of considerable debate among clinicians treating adult-onset and pediatric-onset MS. Consensus criteria for the adult MS population, which were proposed by Cohen et al. in 2004,93 include the presence of more than one relapse per year, no decrease in relapse rate, incomplete recovery from relapses and/or accumulation of disability, new brainstem and/or spinal cord lesions on MRI, polyregional disease, and worsening motor and/or cognitive impairment. In general, these criteria are reserved for patients who have been on therapy for at least 6 months.
Consensus definitions of breakthrough disease in the pediatric MS population are not available, and the use of second-line therapies has not been well studied in this population. At present, many practitioners adhere to the guidelines that are used for the adult population. Under these guidelines, approximately one-quarter of children with MS experience breakthrough disease, prompting a switch to a second-line therapy an average of 1.5 years after starting a first-line therapy.88
Second-line agents
The use of second-line agents, such as cyclophosphamide, mitoxantrone, mycophenolate mofetil, daclizumab, rituximab or natalizumab, in pediatric MS has been described in retrospective case series and reports with limited follow-up (E. A. Yeh et al., unpublished work).88, 94, 95, 96, 97 Owing to the retrospective, open-label nature of the studies, no firm conclusions regarding the efficacy and safety of these agents can be drawn. A retrospective study of cyclophosphamide use in 17 children with MS suggested a temporary reduction in relapses and disease progression. Use of this drug was, however, associated with secondary bladder cancer in one case, and with amenorrhea and infertility in several patients.97 Further studies evaluating the short-term and long-term safety as well as the efficacy of these agents are needed. For ethical reasons, randomized, placebo-controlled trials of second-line agents are unlikely to be performed in pediatric patients.
Conclusions and future directions
Pediatric MS represents a relatively rare but important entity, as it provides unique insights into disease processes related to MS. Studies have suggested substantial variability in presenting symptoms and laboratory and imaging features between children with prepubertal disease and those with postpubertal disease. Given the distinct features that characterize onset before puberty, the possible contribution of hormonal influences and maturation of the immune system and CNS on pediatric MS deserve further study.
No DMTs have been approved by the FDA for the treatment of children with MS, although the currently available first-line therapies for adults seem to be safe and well tolerated in this population. Further studies are required to assess the safety and efficacy of second-line treatments in children with MS.
Validation of the studies reviewed in this article, as well as studies to define additional risk factors, clinical features and biomarkers, are needed to further improve our capacity for early recognition of acquired CNS demyelinating diseases in children. Newer imaging modalities, such as diffusion tensor imaging, magnetization transfer ratio and volumetric analysis, are likely to have roles in the future elucidation of biological processes involved in disease pathogenesis in pediatric MS.
Review criteria
Articles on which this Review was based were found via several approaches. PubMed was queried using search terms including "pediatric", "juvenile", "adolescent", "childhood", "MS", "multiple sclerosis", "treatment", "therapy", "pathology", "etiology", "outcome", "neuropsychological", "MRI", "demyelination" and "demyelinating disease". The scope was limited to papers published between 1966 and the present day. The 'related articles' function of PubMed was used to broaden the search. English-language references that contained information relevant to the Review were included. Non-English-language papers were used if no equivalent English-language papers were found, and if the papers made a notable observation with regard to the pediatric multiple sclerosis field. Bibliographies were searched for additional references that were missed via the PubMed searches. Abstracts were cited only if information regarding an important issue relevant to clinical practice was unavailable from published papers.
