Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Multiple Sclerosis

Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

PRIMER

Corrected: Author Correction

Multiple sclerosis
Massimo Filippi1,2*, Amit Bar-​Or3, Fredrik Piehl4,5,6, Paolo Preziosa1,2, Alessandra Solari7,
Sandra Vukusic8 and Maria A. Rocca1,2
Abstract | Multiple sclerosis (MS) is the most common chronic inflammatory, demyelinating and
neurodegenerative disease of the central nervous system in young adults. This disorder is a
heterogeneous, multifactorial, immune-​mediated disease that is influenced by both genetic and
environmental factors. In most patients, reversible episodes of neurological dysfunction lasting
several days or weeks characterize the initial stages of the disease (that is, clinically isolated
syndrome and relapsing–remitting MS). Over time, irreversible clinical and cognitive deficits
develop. A minority of patients have a progressive disease course from the onset. The pathological
hallmark of MS is the formation of demyelinating lesions in the brain and spinal cord, which can be
associated with neuro-​axonal damage. Focal lesions are thought to be caused by the infiltration
of immune cells, including T cells, B cells and myeloid cells, into the central nervous system
parenchyma, with associated injury. MS is associated with a substantial burden on society owing
to the high cost of the available treatments and poorer employment prospects and job retention
for patients and their caregivers.

Multiple sclerosis (MS) is a chronic, inflammatory, Diagnosis is based on the demonstration of the dissem­
demyelinating and neurodegenerative disease of the ination of demyelinating lesions to different regions of
central nervous system (CNS). MS is a heterogen­ the CNS (dissemination in space (DIS)) and over time
eous, multifactorial, immune-​mediated disease that is (dissemination in time (DIT)), which can be demon­
caused by complex gene–environment interactions. strated using clinical evaluation or paraclinical tools
The pathological hallmark of MS is the accumula­ once MS-​mimicking disorders have been excluded.
tion of demyelinating lesions that occur in the white MRI has a high sensitivity for detecting disease-​related
matter and the grey matter of the brain and spinal abnormalities, including the presence of demyelinat­
cord. The clinical manifestations and course of MS are ing lesions and, accordingly, the use of this imaging
hetero­geneous; in most patients, reversible episodes of modality has substantially changed the diagnosis of
neurological deficits (known as relapses) that usually MS. Additionally, MRI is helpful for monitoring dis­
last for days or weeks characterize the initial phases ease activity and the response to disease-​modifying
of the disease (that is, clinically isolated syndrome treatments (DMTs). Combined with improved under­
(CIS) and relapsing–remitting MS (RRMS); Fig. 1). standing of the immunological and neurobiological
Over time, the development of permanent neurologi­ disease processes underlying MS, improvements in
cal deficits and the progression of clinical disability diagnosis have led to the development of many new
become prominent (known as secondary progressive treatments that can substantially reduce disease activ­
MS (SPMS); Fig. 1). A minority of patients have a pro­ ity in many patients and delay, at least partially, the
gressive disease course from onset, which is referred progression of MS.
to as primary progressive MS (PPMS); Fig. 1). Each In this Primer, we review current knowledge on
subtype of MS can be classified as active or not active the epidemiology and pathophysiology of MS and
on the basis of clinical assessment of relapse occur­ describe the clinical presentations and the classification
rence or lesion activity detected using MRI1; more­ of clinical phenotypes. The current diagnostic tools
over, patients with PPMS or SPMS, can be classified and their prognostic value are discussed, in addition
according to whether disability has progressed over a to how treatment of the disease has evolved. Finally,
given time1,2. key outstanding questions in the field are considered,
MS typically affects young adults, with an onset including the identification of features specific to the
between 20 years and 40 years of age and has a higher pathological substrates of MS, the development of bio­
*e-​mail: filippi.massimo@hsr.it prevalence in women, although some patients experi­ markers sensitive to disease-​related changes, the opti­
https://doi.org/10.1038/ ence their initial demyelinating event during child­ mization of treatment at an individual patient level and
s41572-018-0041-4 hood or adolescence, typically with an RRMS form3,4. the assessment of the impact of comorbidities.

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 1

0123456789();
Primer

a systematic review estimated an overall incidence of


Author addresses
3.6 per 100,000 person-​years in women and 2.0 per
1
Neuroimaging Research Unit, Institute of Experimental Neurology, Division of 100,000 person-​years in men10 and demonstrated an
Neuroscience, San Raffaele Scientific Institute, Vita-​Salute San Raffaele University, increased female to male ratio over time from an esti­
Milan, Italy. mated 1.4 in 1955 to 2.3 in 2000 (ref.10). The increased
2
Department of Neurology, Institute of Experimental Neurology, Division of
female preponderance of MS suggests a possible role of
Neuroscience, San Raffaele Scientific Institute, Vita-​Salute San Raffaele University,
Milan, Italy.
environmental risk factors that mainly affect women
3
Department of Neurology and Center for Neuroinflammation and Experimental (for example, occupation, increased cigarette smoking,
Therapeutics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, obesity, birth control and childbirth)9,11.
PA, USA. MS symptoms are the main direct cause of death
4
Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden. in >50% of patients with MS, although infections and
5
Department of Neurology, Karolinska University Hospital, Stockholm, Sweden. suicide are substantially increased compared with the
6
Neuroimmunology Unit, Center for Molecular Medicine, Karolinska University Hospital, general population12. The life expectancy of patients is
Karolinska Institute, Stockholm, Sweden. reduced by 7–14 years, but this decreased life expectancy
7
Unit of Neuroepidemiology, Fondazione IRCCS Istituto Neurologico Carlo Besta, is less evident in recent estimates12. Excess standardized
Milan, Italy.
mortality values are higher in men than women, in
8
Service de Neurologie, Sclérose en Plaques, Pathologies de la Myéline et
Neuro-inflammation, Fondation Eugène Devic EDMUS Contre la Sclérose en Plaques,
patients with PPMS than those with RRMS and in those
Hôpital Neurologique Pierre Wertheimer, Hospices Civils de Lyon, Lyon, France. with higher disability13–16.

Risk factors
Epidemiology The causes of MS are still unknown, although this disease
MS is one of the most widely studied neurological dis­ is known to result from interplay of genetic susceptibility
eases in terms of epidemiology and it is the primary and environmental risk factors.
cause of non-​traumatic disability in young adults. RRMS
typically has an onset between 20 years and 35 years of Lifestyle and environmental factors. Many environmen­
age, whereas PPMS typically begins at ~40 years of age, tal factors can contribute to the risk of MS and might
although up to 10% of patients experience their initial be present and therefore increase disease risk during
demyelinating event during childhood or adolescence4. a particular time frame. Substantial evidence supports a
Approximately 2.3 million people have MS5 worldwide, period of susceptibility to environmental risk factors for
and this disease is associated with a high societal eco­ MS during adolescence17 (Table 1), although exposure to
nomic burden, which has increased over time. The some factors might be relevant during other phases of
economic burden of MS was estimated as ~14.6 billion life (such as low vitamin D level during pregnancy)18.
euros in 2010 within Europe and 4.3 billion dollars in the Identifying the role of lifestyle or environmental risk fac­
United States in 2013 (refs6,7). tors of MS is difficult and large prospective studies are,
The prevalence of MS varies between countries with few exceptions, rare. The most well-​established risk
(Fig. 2). MS is mainly found in individuals of European factors are Epstein–Barr virus (EBV) infection in ado­
descent and is rare in Asian, black, Native Americans lescence and early adulthood, tobacco exposure through
and Māori individuals8. Prevalence estimates range from active or passive smoking, a lack of sun exposure,
2 per 100,000 individuals in Asia to ~1 per 1,000 indi­ low vitamin D levels and obesity during adolescence
viduals in Western countries, although a prevalence of (Table 1). Other, less-​established risk factors include
1 per 400 individuals has been reported in some countries night work, excessive alcohol or caffeine consumption
with a high latitude9. Indeed, in many studies, a higher and history of infectious mononucleosis17.
latitude correlates with increased prevalence and inci­ Owing to the immune-​mediated pathogenesis of
dence of MS, mainly in Europe and North America9,10. MS, infectious diseases have been suggested as possi­
Genetic factors, in particular the distribution of the ble triggers for disease onset. Of the different pathogens
HLA-​DRB1 haplotype, might account in part for the lati­ investigated, EBV infection is the most consistently and
tudinal gradient, but environmental risk factors that vary robustly associated17,19. To this end, it is noteworthy that
with latitude might also be involved. Of these factors, up to 100% of patients with MS are seropositive for EBV
low vitamin D levels owing to a lack of sun exposure is according to epidemiological studies20. The mechanism
the most likely candidate (see Risk factors, below). by which EBV infection increases the risk of MS is not
The prevalence of MS has increased since the 1950s, clear, but molecular mimicry leading to the generation
especially in women9; this finding might represent a true of cross-​reactive T cells and antibodies has been pro­
increase in disease burden but might also be attributed posed17,19. Despite data supporting an increased risk
to improved access to medical facilities, better diag­ of MS with EBV infection, a direct causal relationship
nostic accuracy and increased life expectancy owing to remains difficult to establish.
improved management. However, these reasons cannot Smoking has been consistently demonstrated as a
explain the female preponderance. The female to male risk factor for MS and has an odds ratio (OR) of ~1.6
ratio of MS, which seems to decrease with higher lati­ (ref.17). The risk of MS and smoking is dose-​dependent: a
tude, has increased to ~3:1 in the 2010s from a 2:1 ratio higher amount of smoking and cumulative smoking are
in the 1950s, despite no difference in the incidence of both associated with increased risk. Passive exposure to
disease in males and females in some regions (such smoking has also been associated with increased risk of
as Norway, the United States and Italy)9,11. In 2008, MS17. Moreover, smoking has also been linked to faster

2 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

disability progression and to higher risk of conversion as polymorphisms in genes encoding human leukocyte
from RRMS to SPMS21. A direct toxic effect of some antigen (HLA), to confer a higher risk of MS. As some
smoke components (promoting lung irritation) and of the environmental factors are modifiable, preventive
an indirect systemic effect (mediated by the peribron­ strategies might be possible in the future; this strategy
chial lymphatic tissue) have been proposed to explain might also be relevant after disease onset, as some of
the association. these risk factors have an influence on disease course
Sun exposure, particularly exposure to ultraviolet-​B and prognosis (see below).
radiation, is the major determinant of vitamin D levels
and tends to decrease with increasing latitudes. Thus, Genetic factors
vitamin D levels have been proposed to underlie the ‘lati­ The prevalence of familial MS is ~13% for all MS pheno­
tude effect’ in MS prevalence. Several studies suggest an types23. The risk of recurrence within families increases
association between low vitamin D levels and increased with the percentage of genetic sharing24; for example,
risk of MS and an increased disease activity (in terms of the age-​adjusted risk in monozygotic twins is 35%, as
clinical relapses and MRI activity)17,22, suggesting a pro­ compared with 6% in dizygotic twins and 3% in sib­
tective effect of normal vitamin D levels throughout the lings24. The heritability of MS is polygenic and involves
disease course. Although the mechanisms of action of polymorphisms in several genes, each of which is asso­
vitamin D are not fully clear, some data suggest that the ciated with a small increase in disease risk. Among these,
active form of vitamin D (1,25-dihydroxycholecalciferol) polymorphisms in HLA class I and HLA class II genes
has a role in the modulation of immune function17,22. convey the highest risk of MS17,24.
Interestingly, some of these risk factors, notably Genome-​w ide association studies have identi­
EBV infection, obesity during adolescence and smok­ fied >200 genetic risk variants for MS; each variant has
ing, can interact with genetic risk factors for MS, such a small effect on risk of disease, and different combi­
nations of these variants likely contribute to genetic
susceptibility in different patients25. Most of these vari­
Pre-symptomatic Relapsing–remitting
CIS Secondary progressive ants encode molecules involved in the immune system
(such as the HLA genes on chromosome 6, including
Inflammatory
relapses HLA-​DRB1*15:01 polymorphisms, and polymorphisms
Clinical in IL2 and IL7R) and are associated with a higher risk
disability of other systemic immune disorders. Polymorphisms
(RRMS and in genes involved in T cell activation and proliferation
SPMS)
Clinical
(such as IL2 and IL7R) are a major feature of the disease,
Clinical disability

disability together with polymorphisms in other components of


(PPMS) adaptive and innate immunity (such as genes that modu­
late tumour necrosis factor (TNF))26–28. Risk genes of MS
Clinical
threshold do not overlap with those of other neurodegenerative
diseases, whereas mutations in only a few genes that
have clear functions in the nervous system have been
associated with an increased risk of MS (for example,
MANBA and GALC). As previously mentioned, some
polymorphisms, particularly those in HLA genes, might
interact with environmental risk factors (Table 1). For
Time example, the HLA-​DRB1*15:01 allele, which conveys
Fig. 1 | Clinical course of MS. The National Multiple Sclerosis Society Advisory an increased risk of MS, but not the protective HLA-​
Committee on Clinical Trials in multiple sclerosis (MS)316 defined four clinical A*02 allele, confers a significantly higher risk of MS in
courses of MS: relapsing–remitting MS (RRMS), secondary progressive MS (SPMS), smokers (OR 13.5)29, in individuals with EBV infec­
primary progressive MS (PPMS) and progressive relapsing MS (PRMS)316. RRMS accounts tion (OR 16.0)30 and in those with adolescent obesity
for ~85% of patients and is characterized by the occurrence of relapses at irregular
(OR 16.2)31. Moreover, polymorphisms in genes involved
intervals with complete or incomplete neurological recovery133,317,318; the average
relapse frequency is ~1.1 per year early in the disease course but seems to decrease with in vitamin D metabolism (such as GC and CYP24A1)32
advancing disease, increasing neurological dysfunction and age319. Most patients are associated with an increased risk of MS17. Further
with RRMS will develop SPMS, which is characterized by progressive, irreversible efforts are required to elucidate how environmental risk
disability that occurs independently of the presence of relapses316. Conversion to SPMS factors interact with MS susceptibility genes to contrib­
occurs in ~2–3% of patients per year316. Approximately 10–15% of patients present with ute to early disease mechanisms in the immune system
PPMS, which is characterized by disease progression from the onset, resulting in gradual, and the CNS.
progressive and permanent neurological deficits for >1 year without relapses159,316.
PRMS is rare and is characterized by progressive disease from the onset, with acute Mechanisms/pathophysiology
relapses (with or without full clinical recovery) and periods of continuing progression Pathology
between relapses316. A revision of these phenotypes has been proposed1 and includes
The pathological hallmark of all MS phenotypes is
clinically isolated syndrome (CIS)1 to denote those patients whose first clinical
presentation has characteristics of inflammatory demyelination that could be MS but focal plaques (also known as lesions), which are areas
who do not fulfil its diagnostic criteria1. Within each subtype, disease can be classified as of demyelination that are typically located around post-​
active or not active, which are defined by the occurrence of relapses or lesions detected capillary venules and are characterized by breakdown
using MRI. Another important modifier of the progressive stages is the inclusion of of the blood–brain barrier (BBB). The mechanisms of
whether disability has progressed over a given time period. BBB breakdown are incompletely understood but seem

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 3

0123456789();
Primer

Patients with MS
(per 100,000 individuals)
0.00–5.00
5.01–20.00
20.01–60.00
60.01–100.00
>100.00
Data not provided

Fig. 2 | Worldwide prevalence of MS. The prevalence of multiple sclerosis (MS) varies between countries. In general,
the prevalence of MS is higher in countries of higher latitude and in Western countries. The MS International Federation’s
Atlas of MS, 2013.

to involve direct effects of pro-​inflammatory cytokines are less frequent in patients with PPMS and SPMS owing
and chemokines (such as TNF, IL-1β and IL-6) prod­ to a reduced frequency of inflammatory events in these
uced by resident cells and endothelial cells, as well as patients. PPMS and SPMS are mainly characterized by
indirect cytokine-​dependent and chemokine-​dependent inactive lesions. Inactive lesions are sharply circum­
leukocyte-​mediated injury33,34. The dysregulation of the scribed, hypocellular and have well-​defined demyelin­
BBB increases the trans-​endothelial migration of acti­ ation, reduced axonal density, reactive astrocyte gliosis,
vated leukocytes, including macrophages, T cells and variable microglial activation only in the periplaque
B cells, into the CNS, which leads to further inflammation white matter (without macrophages) and a lower den­
and demyelination, followed by oligodendrocyte loss, sity of lymphocytes than active lesions42–46. However,
reactive gliosis and neuro-​axonal degeneration35,36. inflammatory mechanisms still have a role in PPMS
Plaques occur in both white matter and grey matter and SPMS44,45,47–49; indeed, active or mixed (inactive and
and are typically found throughout the CNS, including active) lesions represented up to 57% of all lesions in
in the brain, optic nerve and spinal cord37–39. Although patients with progressive MS in one study, and active
the anatomical location of white matter lesions is asso­ lesions correlated with a more-​severe disease course49.
ciated with specific clinical manifestations of MS, the Other forms of plaques include chronic active plaques
total volume of these lesions is only moderately cor­ and slow expanding lesions. Chronic active plaques are
related with overall clinical disability and cognitive more frequent in patients with MS with a longer dis­
impairment40,41 owing to the involvement of other patho­ ease duration and in SPMS and are characterized by
physiological mechanisms, such as the occurrence of macrophages at the edge of the lesion, with fewer macro­
grey matter lesions and normal-​appearing brain tissue phages in the lesion centre (Fig. 3). Slow expanding
damage, which affect both grey matter and white matter. lesions, which are typically found in patients with SPMS,
are characterized by an inactive centre with demyelin­
White matter lesions ation, activated microglia at the lesion edge and few
The earliest phases of MS (CIS and RRMS) are typically macrophages containing myelin debris, but transected
characterized by active demyelinating lesions. These axons are also observed, suggesting a very slow rate of
lesions have heavy lymphocyte infiltration (mainly CD8+ ongoing demyelination and axonal damage42,44–46.
T cells and CD20+ B cells, with fewer CD4+ T cells), acti­
vated microglia (particularly at the lesion edge and con­ Normal-​appearing white matter. In addition to the
taining myelin debris), macrophages (containing myelin focal lesions typically observed in patients with MS,
debris) and large, reactive (sometimes multinucleated) macroscopically normal white matter (that is, normal-​
astrocytes42,43. By contrast, active demyelinating plaques appearing white matter (NAWM)) often shows signs

4 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

Table 1 | Lifestyle and environmental risk factors for MS


Risk factor Odds HLA gene Combined Effect during Immune Level of
ratio interaction odds ratioa adolescence system implied evidence
Smoking ~1.6 Yes 14 No Yes +++
EBV infection (seropositivity) ~3.6 Yes ~15 Yes Yes +++
Vitamin D level <50 mM ~1.4 No NA Probable Yes +++
Adolescent obesityb ~2.0 Yes ~15 Yes Yes +++
CMV infection (seropositivity) 0.7 No NA Unknown Yes ++   
Night work ~1.7 No NA Yes Yes ++
Low sun exposure ~2.0 No NA Probable Yes ++
Infectious mononucleosis ~2.0 Yes 7 Yes Yes ++
Passive smoking ~1.3 Yes 6 No Yes +
Organic solvent exposure ~1.5 Unknown Unknown Unknown Unknown +
Oral tobacco or nicotine 0.5 No NA Unknown Yes +
consumption
Alcohol ~0.6 No NA Unknown Yes +
Coffee ~0.7 No NA Unknown Yes +
+, non-​replicated observations that require further study; ++, case–control observations that have been replicated and/or
supported by independent methods; +++, high level of evidence from large prospective studies or a case–control observation that
is supported by Mendelian randomization studies; CMV, cytomegalovirus; EBV, Epstein–Barr virus; HL A, human leukocyte antigen;
MS, multiple sclerosis; NA, not applicable. aCombined odds ratio for the non-​genetic factor and HL A allele. bAdolescent obesity
defined as body mass index >27 at 20 years of age). Adapted from ref.17, Springer Nature Limited.

of diffuse inflammation and neuro-​axonal damage45,50. formation is supposed to be promoted by pro-​inflam­


Abnormalities of NAWM have been observed in patients matory media­tors released from the meninges or
with RRMS but are more severe in those with progres­ present in the cerebrospinal fluid60. Compared with
sive disease and include decreased fibre density owing white matter lesions, cortical lesions typically display
to axonal degeneration and demyelination, small round less BBB breakdown, less oedema, a lower degree of
cell infiltration (mainly lymphocytes), macrophage inflammation (characterized by fewer infiltrating acti­
infiltration, widespread microglia activation and glio­ vated micro­glia and macro­phages61) and more efficient
sis50. NAWM was previously considered secondary to myelin repair occurring after demyelination, suggest­
the axonal damage within focal lesions, although these ing that different mechanisms determine lesion for­
diffuse changes poorly correlate with the number, size, mation in the white matter and the grey matter62,63.
location and destructiveness of focal white matter lesions Cortical lesions are associated with variable degrees of
in the brain50 and spinal cord, suggesting that they might transected neurites, neuronal apoptosis and loss of neu­
occur independently51. rons, neuro-​axons and glial cells, together with a sub­
stantial loss of synapses61,64,65 (Fig. 3). Decreased synaptic
Grey matter lesions. Extensive cortical demyelina­ density has also been described in the normal-​appearing
tion is observed in the forebrain50,52 and cerebellum53 cortex in patients with MS without cortical lesions, sug­
in patients with MS, occurs from the earliest phases gesting that synaptic loss might be in part independent
of the disease (that is, also in patients with radiolog­ from focal demyelination in the cortex65.
ically isolated syndrome54 (see below) and CIS55) and According to their location within the grey matter,
is more widespread in patients with PPMS and SPMS, four different types of cortical lesions have been identi­
in extreme cases of which >60% of the cortex can be fied in patients with MS61,66: type I lesions are located at
affected. Lesions can also occur within deep grey mat­ the cortico-​subcortical border and affect both the grey
ter nuclei56,57 and in the grey matter of the spinal cord, matter and the white matter; type II lesions are small
in which grey matter demye­lination is more extensive perivenous intracortical lesions that do not affect white
and widespread than in the white matter37,39. Although matter or the pial surface of the brain; type III lesions
the mechanisms underlying the differences in the extent extend inward from the subpial layers of the cortex
of demyelination in the grey matter and white matter (subpial lesions); and type IV lesions extend through
have not been clarified, it could be due to differences the whole width of the cortex but without passing the
in the mechanisms of promoting demyelination and border between the cortex and the white matter. Type III
the presence of pro-​demyelinating soluble factors in the cortical lesions are the most frequent in patients with
cerebrospinal fluid37,39. MS and are characterized by subpial areas of demye­
Cortical lesions are predominantly found in corti­ lination, which involve the cortical ribbon of several
cal sulci and in deep invaginations of the brain surface gyri and are often related to meningeal inflammatory
and are often topographically related to inflamma­ infiltrates58,59 usually not extending beyond layers 3 and
tory infiltrates in the meninges58,59. Moreover, their 4 of the cortex.

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 5

0123456789();
Primer

a b c d e

f g h i

Fig. 3 | Post-​mortem histopathological findings in MS. a | A tissue block from the superior frontal gyrus (SFG) showing
normal-​appearing grey and white matter (upper rectangle) and a macroscopically visible mixed grey and white matter
lesion (lower rectangle) was obtained from a 70-year-​old patient with secondary progressive multiple sclerosis (MS) with a
disease duration of 33 years who died owing to euthanasia, with a post-​mortem delay of 6 hours. b–e | Stained sections of
normal-​appearing grey and white matter are shown. f–i | Stainined sections of a mixed grey and white matter lesion are
shown. Proteolipid protein labelling (parts b,f) to quantify myelin confirmed the presence of a mixed grey and white
matter lesion (part f). In the same lesion, Bielschowsky (part g) and NeuN (part h) staining revealed axonal injury and
neuronal shrinkage and loss, respectively. Compared with normal-​appearing grey and white matter (part e), sections stained
for ionized calcium binding adaptor molecule 1, a marker of microglia, showed a higher density of microglia in the rim of the
lesion (part i), mainly in the white matter edge. The patient donor gave written informed consent for the use of his tissue and
medical records for research purposes, and he was registered at the Netherlands Brain Bank, Amsterdam, Netherlands.

Remyelination and degeneration. Remyelination can astrocytes72. Together with peripheral immune cells,
occur in MS62,63,67, has been suggested as a mechanism CNS-​resident cells secrete a range of inflammatory
of clinical recovery after a relapse and could represent mediators that can recruit inflammatory cells into the
a target for future therapies68. Remyelination gives rise CNS, lead to neuronal demyelination and induce inflam­
to the so-​called shadow plaques that are characterized mation within the CNS parenchyma. In addition, both
by global or patchy remyelination, a sharp demarcation peripheral and CNS-​compartmentalized inflammatory
from the surrounding NAWM and axons with thin mye­ mechanisms are involved in MS pathophysiology. In
lin sheaths and shortened internodes62,63,69,70. The extent particular, CNS-​resident cells that sense homeostatic
of remyelination is very heterogeneous, although it disturbances, mainly microglia and astrocytes, can also
is generally limited and restricted to the lesion border or is produce neurotoxic inflammatory mediators (such as
patchy, and has been demonstrated in ~40–50% of white cytokines, chemokines and reactive oxygen species)
matter lesions and in up to 90% of grey matter lesions, that can promote and sustain neuro-​axonal damage and
although different values have been reported in some neurodegeneration in MS47 (Fig. 4). Despite the notion
studies62,63,67. The variability in remyelination depends on that CNS-​compartmentalized inflammation likely con­
several factors, including patients’ age, disease duration, tributes to CNS injury, it is poorly targeted by currently
lesion location, the presence of oligodendrocyte progeni­ available treatments73 and needs further study48,74.
tor cells and axonal integrity48; substantial remyelination
is frequently observed during the earlier phases of MS T cell involvement. The historical view of MS, on the basis
and in younger individuals, whereas it is more sparse or of studies of patients and studies using the most com­
absent in PPMS and SPMS71. monly used animal model of MS (that is, experimental
Of the neuropathological findings in MS, neuro-​axonal autoimmune encephalomyelitis (EAE)), is that relapses
loss is of particular interest, as it corresponds to neuro­ are principally mediated by aberrantly activated and/or
degeneration. In MS, neurodegeneration occurs from the insufficiently regulated pro-​inflammatory CNS-​specific
earliest phases of disease and might contribute to irre­ effector T cells, including CD4+ T cells and CD8+ T cells,
versible clinical disability45. Whether the degree of axonal that traffic to the CNS parenchyma and cause perivascu­
loss correlates with the severity of MS is unknown and lar demyelination, glial cell activation and neuro-​axonal
requires further study. Different mechanisms occurring injury47,75. One potential cause of aberrant effector T cell
at different stages of MS might drive n ­ eurodegeneration activation is an insufficiency in the function of regu­
as a primary and/or secondary phenomenon45,48. latory T (Treg) cells and resistance of CNS-​specific effector
T cells to Treg cell-​mediated regulation76,77. Indeed, several
Immune pathophysiology abnormalities in circulating Treg cells have been observed
Our understanding of the underlying immunopatho­ and implicated in MS, including decreased expression
physiology of MS has evolved. The traditional view of forkhead box protein P3 (FOXP3) by Treg cells and/or
of T cell-​mediated MS relapses has been altered to deficient regulatory capacity of FOXP3-expressing
include the involvement of key bidirectional inter­ CD25hiCD127low natural Treg cells (which arise in the thy­
actions between several immune cell types, includ­ mus and are a separate lineage to induced Treg cells)78–80.
ing T cells, B cells and myeloid cells in the periphery, In addition, decreased numbers or deficient regulatory
and resident cells of the CNS such as microglia and responses have also been suggested for CD46-expressing

6 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

induced type 1 regulatory cells, CD39-expressing thought to contribute to effector T cell trafficking from
Treg cells, IFNγ-​expressing Treg cells and follicular Treg cells lymphoid structures and blood to the CNS in the EAE
in blood in patients with MS, which could promote model and possibly in patients with MS99,100. In addition,
­aberrant ­effector T cell function81–83. junctional adhesion molecule-​like (JAML) has a role in
The most widely implicated pro-​inflammatory effec­ the migration of CD8+ T cells and monocytes across the
tor T cells are IL-17-expressing CD4+ T cells (known as brain endothelium, whereas MUC18 is used by CD8+
T helper 17 cells (TH17 cells)) and CD8+ T cells that might T cells and CD4+ T cells to access the CNS, and nin­
be increased in the periphery and in the CNS in patients jurin 1 is selectively implicated in the CNS migration of
with MS. These cells are speculated to contribute to myeloid cells99,101. Also of note, in addition to the post-​
direct injury of oligodendrocytes and neurons (although capillary venule BBB endothelial cells (which are the site
the exact mechanisms that direct injury have not been of the classical perivascular MS lesions), immune cells
defined) and to indirect tissue injury through the acti­ might enter the CNS via the subarachnoid space and the
vation of other cells, such as macrophages84–87. Other blood–CSF barrier. Identifying the molecules that are
effector T cell subsets with a role in MS include IFNγ-​ involved in the trafficking of subsets of immune cells
secreting CD4+ T cells (TH1 cells) and granulocyte– across distinct CNS barriers might guide development of
macrophage colony-​stimulating factor (GM-​CSF)- more selective therapeutic targeting. The earliest molec­
expressing CD4+ and CD8+ T cells; the role of GM-​CSF ular mechanisms underlying new inflammatory lesion
is not fully defined in MS, but GM-​CSF has been shown formation in MS might involve abnormalities in these
to activate myeloid cells and CD8+ mucosal-​associated barriers, which enable immune cell infiltration from
invariant T (MAIT) cells in the EAE model88–90. the periphery102–104.
The aberrant T cell activation in MS requires anti­ The biology underlying remission in MS is not well
gen presentation to T cells by antigen-​presenting cells understood, but it is unlikely to merely represent a pas­
(APCs) such as B cells and myeloid cells (macrophages, sive decline in the pro-​inflammatory effector cell activity
dendritic cells and microglia) in the periphery and the and is likely to involve mechanisms that downregulate
CNS, although the responsible antigens have not been immune responses, such as Treg cell activity76. In addi­
routinely identified91 (Box 1). Myelin-​related antigens are tion, remission is likely to involve activation-​induced
suspected to be involved, although there is no consen­ cell death wherein activated pro-​inflammatory cells
sus, and some studies have suggested antigens on the might have upregulated surface molecules that make
neuronal or glial cell surface. Important bidirectional them more susceptible to killing by other immune
interactions between T cells and myeloid cells that can cells105. Indeed, several studies have suggested that
shape their effector responses (both pro-​inflammatory apoptosis of immune cells (such as myelin-​reactive
or anti-​inflammatory responses) have long since T cells) could exert positive effects by switching off
been recognized75,92. Pro-​inflammatory APCs such as CNS inflammation106.
B cells and myeloid cells can drive TH1 cell and TH17 cell
responses, which might have a role in immune cell inter­ B cell involvement. A role for B cells in the development
actions and the trafficking that underlies relapses in MS. of MS relapses has emerged on the basis of impressive
To this end, circulating myeloid cells in patients with results of selective B cell-​targeting therapies (such as anti-​
MS have an overly pro-​inflammatory profile, includ­ CD20 antibodies) in MS107. The role of a small subset
ing the expression of the microRNA miR-155 and pro-​ of CD20-expressing T cells (which are also depleted with
inflammatory cytokines such as TNF, IL-12, IL-6, IL-23 anti-​CD20 therapy) remains of interest, although this
and IL-1β, which are involved in TH1 cell and TH17 cell subset has not been ascribed a particular pathogenetic
differentiation93–95. function in MS108.
How aberrantly activated immune cells access Healthy individuals typically have low levels of
the CNS in MS is of ongoing interest and therapeutic antibodies in the CNS (the normal ratio is ~1:300
importance. Despite the fact that the CNS was con­ of CNS to periphery); patients with MS have an abnor­
sidered immune privileged, with the BBB thought to mally increased production of antibodies within the
restrict entry of cells and macromolecules from the CNS, which can be detected, for example, as increased
circulation, as previously mentioned, BBB breakdown immunoglobulin synthesis rates and the presence of
has been observed in patients with MS, which is specu­ cerebrospinal fluid-​restricted oligoclonal bands (OCBs).
lated to facilitate the migration of pro-​inflammatory This finding was the basis for anti-​B cell therapies in
cells into the CNS parenchyma. In addition, a lymphatic MS, although interestingly, the reduction in relapse
drainage system has been demonstrated in the CNS96. rate with anti-​CD20 therapy was associated with little
The immune system can interact continuously with the or no change to the cerebrospinal fluid immunoglob­
CNS as part of normal immune surveillance and, in MS, ulin profile in patients109,110, suggesting an antibody-​
bidirectional trafficking likely takes place during the independent role of B cells in MS relapses. These
course of disease97,98. After activation in the periphery, antibody-​independent functions are likely to be the
immune cells upregulate cell surface molecules such as contribution of B cells to cascades of cellular immune
chemokine receptors and adhesion molecules, which interaction in the periphery and/or their ability to attract
enables efficient tissue infiltration, including to the CNS. and activate T cells and myeloid cells in the CNS72.
Indeed, chemokine receptors, such as CC-​chemokine Indeed, B cells from patients with MS have an abnor­
receptor 6 (CCR6), CCR2 and CCR5, and cell surface mal propensity to produce pro-​inflammatory cytokines
glycoprotein MUC18 (also known as MCAM) are (including IL-6, GM-​CSF, TNF and lymphotoxin-α

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 7

0123456789();
Primer

Meningeal SAS Astrocyte Early disease


vessel Pial basement membrane
Glia limitans CNS parenchyma
Soluble mediators
recruit immune cells Soluble mediators
promote inflammation
CD8+ at distal sites
T cell
Perivascular DC
or macrophage
Perivascular immune Oligodendrocyte
T cell cell accumulation
reactivation Myelin
sheath
Antibody
Vessel Neuron
CD4+ activation Degraded
T cell CD8+ TH1 cell IFNγ myelin
MAIT protein
IL-17
cell Demyelination
TH17 cell
Monocyte GM-CSF

Cerebrospinal fluid
Phagocytosis
Complement
proteins Ependyma
B cell
Choroid plexus
Microglial Capillary
Pericyte cell
Plasma cell Choroid plexus macrophage

T cell reactivation by choroid


plexus and meningeal APCs Late disease

Meningeal tertiary
lymphoid-like structures
promote glia limitans damage
and astrocyte dysfunction

Follicular DC
Oligodendrocyte
progenitor

Clonally CCL2 Metabolic


expanded GM-CSF stress
B cells
Osteopontin
NO
ROS Ionic
Glutamate RNS Energy
accumulation imbalance deficiency

Neuro-axonal and
oligodendrocyte
damage and death

Neurodegenerative
processes promote
further damage
at distal sites

(LTα)) and are deficient in regulatory cytokines such profile of B cells from patients with MS can induce
as IL-10 (refs111–116). One subset of pro-​inflammatory aberrant TH1 cell and TH17 cell responses through TNF
B cells, CD27+ GM-​CSF-expressing memory B cells, and IL-6 and can induce pro-​inflammatory myeloid cell
which produce high levels of TNF and IL-6 but do not responses (principally through GM-​CSF), which could
express IL-10, is found in increased numbers in the contribute to the cellular immune cascades involved
circulation of patients with MS and has an exaggerated in relapses111–116. In line with this finding, anti-​CD20
response profile112. The abnormal cytokine response B cell-​depleting therapy reduces the pro-​inflammatory

8 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

◀ Fig. 4 | Immune system dysregulation within the central nervous system in early against myelin antigens, such as myelin basic protein
and late MS. Immune cells are believed to enter the central nervous system (CNS) in (MBP) or myelin-​oligodendrocyte glycoprotein (MOG)
multiple sclerosis (MS) through the blood vessels of the blood–brain barrier (BBB), the and the inward rectifying potassium channel Kir4.1 (also
subarachnoid space (SAS) and the choroid plexus (dashed arrows). In MS relapses, which known as KCNJ10)119–121, have not led to the same patho­
are more prominent in the early phases of disease, underlying mechanisms involve the
genetic implications for specific CNS-​directed antibodies
infiltration of cells of the innate and adaptive immune systems, such as CD4+ and CD8+
T cells, B cells and myeloid cells, into the CNS parenchyma with perivascular distribution
as those observed in other conditions, such as anti-​
around post-​capillary venules of the BBB. These immune cells, together with resident aquaporin 4 antibodies in neuromyelitis optica spectrum
activated microglia and astrocytes, are thought to contribute to oligodendrocyte injury, disorders (NMOSDs) and anti-​N-methyl-​d-aspartate
demyelination and neuro-​axonal injury through cell contact-​dependent mechanisms (NMDA) antibodies in NMDA encephalitis. In addition,
and the secretion of soluble factors. In later stages of the disease, the episodic infiltration the presence of circulating anti-​MOG antibodies in a
of immune cells into the CNS is diminished. Mechanisms contributing to ongoing tissue subset of patients with CNS inflammatory demyelinat­
injury (and the clinical manifestations of progressive disease) are thought to include ing disease, including NMOSD122, has been associated
neurodegeneration, in terms of neuro-​axonal, astrocyte and oligodendrocyte damage, with clinical and imaging features that are not typical
owing to acute or chronic oxidative stress promoted by innate and adaptive immune cell of MS123,124 (even if they have been also described in
activation, mitochondrial dysfunction, extracellular free iron accumulation, loss of myelin
up to 5% of patients with MS), which are mainly charac­
trophic support, hypoxia, altered glutamate homeostasis and a pro-​inflammatory
environment, with possible involvement of cytotoxic factors and complement activation. terized by severe brainstem and spinal cord involvement,
Chronic inflammation is potentially mediated by ongoing CNS-​compartmentalized a severe disease course with high relapse rates and failure
inflammation involving meningeal immune cell infiltrates (for example, B cells) that can in response to several DMTs125.
form lymphoid-​like structures and by CNS-​resident innate cells (for example, microglia).
For example, CC-​chemokine ligand 2 (CCL2) and granulocyte–macrophage colony-​ Progressive MS. In addition to cascades of the cellular
stimulating factor (GM-​CSF) produced by astrocytes can promote microglia recruitment immune interactions in the periphery that contribute to
and activation, and astrocytes can limit remyelination by preventing the differentiation MS relapses, ongoing inflammation in the CNS might
of oligodendrocyte progenitor cells into mature oligodendrocytes. APC, antigen-​ contribute to the propagation of injury in patients with
presenting cell; DC, dendritic cell; MAIT, mucosal-​associated invariant T; NO, nitric oxide; PPMS and SPMS (Fig. 4). In particular, inflammation
RNS, reactive nitrogen species; ROS, reactive oxygen species; TH1, T helper 1; TH17,
may differ in individuals with progressive MS compared
T helper 17. Adapted from ref.47, Springer Nature Limited.
with RRMS and is characterized by a lower frequency of
inflammatory relapses (waves of infiltration of activated
responses of TH1 cells and TH17 cells and reduces mye­ immune cells into the CNS in a perivascular distrib­
loid cell pro-​inflammatory responses in the periphery of ution). Additionally, a CNS-​compartmentalized inflam­
patients with MS111,112. By contrast, the (largely naive) mation is evident, involving, for example, CD8+ T cells
B cells that re-​emerge after discontinuation of anti-​CD20 and plasma cells that survive and persist in the CNS
treatment111,112,114 have reduced secretion of GM-​CSF, or surrounding meninges and possibly also involving
IL-6 and TNF but increased IL-10 secretion; whether microglia and astrocyte inflammatory responses43,45,47,48.
these cells have an immune-​regulatory effect in a subset CD8+ T cells might be quiescent memory cells that
of patients that potentially contributes to the durabil­ promote further tissue damage when exposed to and
ity of the treatment effect and whether the treatment activated by their target antigen43. The different inflam­
effect lasts until the re-​emergence of pro-​inflammatory matory mechanisms in PPMS and SPMS might contrib­
memory B cells are of interest. ute to the lack of efficacy of DMTs, which typically have
MS relapses might also be driven by alterations systemic anti-​inflammatory activity45,48.
in the balance between pro-​inflammatory and anti-​ Ongoing questions relate to how relapse biology is
inflammatory B cells. This is supported by the obser­ involved in the initiation and maintenance of CNS-​
vation that, aside from anti-​CD20 therapies, all other compartmentalized inflammation, which, at least at
approved therapies for MS affect memory B cell some point in the disease process, is maintained in the
responses (reviewed previously72). In addition, the absence of obvious relapses. The subpial demyelinating
finding that atacicept (a recombinant fusion protein that cortical injury that is present from the earliest phase
inhibits B cells) exacerbated MS relapses in clinical trials of the disease and is more widespread in patients with
lends further support to this hypothesis117. Atacicept progressive MS reportedly involves a graded pattern of
leads to selective loss of several subsets of B cells (includ­ neuronal loss and microglial activation45,126,127, which
ing plasmablasts and plasma cells) but spares memory could be consistent with a ‘surface-​in’ process, such as
B cells, which might result in a more pro-​inflammatory that mediated by one or more toxic substances in the cere­
B cell profile, therefore, aggravating disease. brospinal fluid. In this regard, the extent of meningeal
The antibody-​independent functions of B cells do not inflammation is associated with the extent of subpial cor­
preclude a role for antibodies in MS pathophysiology. tical injury126 and with higher levels of pro-​inflammatory
However, antibody levels in the CNS do not substantially cytokines such as IFNγ, TNFα, LTα and IL-6 in the cere­
change following anti-​CD20 treatment118, suggesting brospinal fluid of patients60. The potential for meningeal
that antibodies are unlikely to be critically involved in immune cells to contribute to CNS injury has also been
triggering relapses. It is possible that antibodies could noted above in the context of cytotoxic CD8+ T cells that
persist in the CNS for a long period of time after treat­ may enter the CNS through the meninges to respond
ment; however, if the antibodies were relevant, the effects to local (potentially EBV-​infected) B cells60,128 (Box 1).
of treatment would not be quick or substantial while the In addition, B cells from patients with MS can secrete
antibodies do not change. Studies of circulating antibod­ unidentified factors that are toxic to oligodendrocytes and
ies in patients with MS, including antibodies directed neurons in vitro129,130. The CNS inflammation in patients

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 9

0123456789();
Primer

might, in turn, foster B cell persistence and propagation to brainstem and/or cerebellar syndromes) or the cere­
of CNS-​compartmentalized inflammation131. Future bral hemispheres (cerebral hemispheric syndrome;
research will aim to elucidate whether and how bidire­c­ Fig. 5; Table 2). During the disease course of RRMS,
tional interactions between meningeal immune cells further clinical episodes can occur (known as relapses);
and underlying brain cells contribute to the propaga­ these episodes last for ≥24 hours and occur in the
tion of non-​relapsing inflammation and progressive absence of fever, infection or clinical features of enceph­
injury to CNS structures adjacent to the cerebrospinal alopathy (for example, altered consciousness or epilep­
fluid and how such processes could interact with and/or tic seizures)132. Symptoms of a clinical attack typically
respond to the degenerative mechanisms described show an acute or sub-​acute onset, worsen over days or
above (reviewed previously45). weeks, reach a peak severity within 2–3 weeks and remit
Despite occurring at early disease stages, neuro-​ to a variable degree, ranging from minimal resolution to
axonal degeneration is common in progressive disease. complete recovery normally 2–4 weeks after reaching
The mechanisms of neuro-​axonal degeneration include maximum deficit133.
neuronal apoptosis owing to acute or chronic oxida­ Optic neuritis is the first neurological episode in
tive stress promoted by innate and adaptive immune ~25% of patients and is associated with a conver­
cell activation, mitochondrial dysfunction and extra­ sion to clinically definite MS in 34–75% of patients
cellular free iron accumulation, loss of myelin trophic between 10 years and 15 years after clinical onset134–136.
support, hypoxia, altered glutamate homeostasis and a Approximately 70% of patients with MS have optic neu­
pro-​inflammatory environment, with possible cytotoxic ritis during the course of the disease134–136. Optic neuritis
factors and complement activation45,48. is characterized by a partial or total visual loss in one eye
with a central scotoma (a blind spot in the visual field),
Diagnosis, screening and prevention dyschromatopsia (deficiency of colour vision) and pain
Clinical presentation within the orbit that is worsened by eye movement134–136
The clinical presentation of MS is heterogeneous and (Table 2). During fundus oculi examination using ophthal­
depends on the location of demyelinating lesions within moscopy, the optic nerve head appears normal if inflam­
the CNS. Although no clinical findings are unique to mation is limited to the retrobulbar portion of the
MS, some are highly characteristic of the disease. nerve, but approximately one-​third of patients might
Typically, the onset of MS is characterized by an ini­ have inflammation of the optic disc (papillitis) and disc
tial clinical attack (defined as CIS) in ~85% of patients, oedema owing to anterior optic neuritis. Patients with­
which consists of an unpredictable episode of neuro­ out visual complaints with suspected MS should be eval­
logical dysfunction owing to demyelinating lesions in uated for more subtle manifestations of optic neuritis,
the optic nerve (leading to optic neuritis), spinal cord such as an afferent pupillary defect or abnormalities at
(leading to myelitis), brainstem or cerebellum (leading paraclinical tests (for example, visual evoked potentials,
optical coherence tomography (OCT) or MRI).
Sensory symptoms are the first clinical manifesta­
Box 1 | Autoantigens in MS
tion in up to 43% of patients with MS and are mainly
The antigenic targets of the aberrant immune cell activation in multiple sclerosis (MS) caused by myelitis or brainstem syndromes137. Sensory
remain incompletely defined. Historically, the focus of investigation was on myelin symptoms include paresthesia (commonly described
proteins that are commonly used to induce autoimmune encephalomyelitis (EAE) in as numbness, tingling, pins-​and-needles feeling, tight­
experimental models, such as myelin basic protein (MBP), proteolipid protein and ness, coldness and/or swelling of the limbs or trunk),
myelin-​oligodendrocyte glycoprotein75,92. Indeed, several studies in patients support a Lhermitte sign138 (a transient symptom described as
role for myelin-​reactive T cells in MS owing to the increased frequency, stability and/or
an electric shock radiating down the spine or into the
pro-​inflammatory response profiles of these cells in patients compared with
controls92,299. However, most healthy individuals also have T cells (and B cells) that are
limbs with flexion of the neck), impairment of vibra­
reactive to the same myelin antigens as patients with MS, therefore, the mere presence tion and joint position sensation, and reduced pain and
of such autoreactive cells is insufficient to induce disease. Non-​myelin antigens might light touch perception. These symptoms can temporar­
be relevant in early MS pathogenesis, such as axo-​glial apparatus molecules that have ily worsen with increased body temperature (known as
been implicated in paediatric-​onset MS300. T cell activation by an infectious agent that Uhthoff phenomenon).
has similarities with central nervous system (CNS) antigens (known as molecular Motor manifestations are the initial symptoms in
mimicry) has been postulated as a mechanism for triggering MS and MS relapses. 30–40% of patients and affect almost all patients during
In particular, a strong epidemiological association has been demonstrated between the course of the disease139. Motor symptoms are charac­
Epstein–Barr virus (EBV) infection and risk of MS in the earliest phases of MS, close to its terized by pyramidal signs (such as Babinski sign, more-​
biological onset301,302; EBV shares a molecular sequence with MBP, and aberrant CD4+
pronounced reflexes and clonus), paresis and spasticity.
and CD8+ T cell responses to EBV have been reported in patients with MS303,304.
Moreover, EBV can transform and activate B cells in vivo, and it is plausible that, in
Brainstem and cerebellar symptoms are present in up to
patients with MS, EBV contributes to pro-​inflammatory B cell activation in the 70% of patients with MS139, which include impairment in
periphery and, in turn, mediates aberrant activation of CNS-​reactive T cells that are ocular movements (such as nystagmus (involuntary eye
involved in MS relapses. In addition, some studies have demonstrated EBV-​infected movement), oscillopsia (a visual phenomenon in which
B cells and plasma cells in patients with MS, which are located adjacent to CD8+ T cells items in the visual field seem to move) and diplopia
expressing cytotoxic molecules, such as perforin and granzyme128,305. The process by (double vision)), ataxia and gait imbalance, dysmetria
which immune cell activation to additional CNS antigens might be triggered as a (poor coordination) and decomposition of complex
consequence of CNS injury and exposure of additional antigenic targets has been movements, slurred speech and dysphagia (difficulty
referred to as epitope spreading. This process is well demonstrated in EAE, with limited swallowing). The extent of sphincter and sexual dysfunc­
studies suggesting this might also occur in patients with MS92,306.
tion often parallels the degree of motor impairment in the

10 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

a b c

d e

Fig. 5 | Radiological examples of demyelinating events in MS. 3T MRI sequences from five patients with clinically
isolated syndrome (CIS) suggestive of multiple sclerosis (MS), within 5 days from clinical onset, are shown. Focal lesions
(arrows) can be observed in: the right optic nerve in a patient with acute optic neuritis (part a); the left pons and the right
middle cerebellar peduncle in a patient with diplopia (part b); the cerebellar hemispheres in a patient with vertigo (part c);
the cervical spinal cord in a patient with paresthesia and Lhermitte sign (part d); and the left cerebral hemisphere in a
patient with right sensorimotor hemisyndrome (part e).

lower extremities, and the dysfunction usually becomes Affective disturbance occurs in up to two-​thirds of
permanent late in the disease course, affecting 34–99% patients, of which depression is the most common mani­
of patients140. The most common symptom of bladder festation146. Pain is reported in up to 43% of patients and
dysfunction is urinary urgency, but hesitancy, frequency can include trigeminal neuralgia, dysesthetic pain, back
and urge incontinence can also occur140. Constipation is pain, visceral pain and painful tonic spasms147. Typically,
more common than faecal incontinence, and men with the prevalence and the severity of all clinical manifesta­
MS often have erectile dysfunction and impotence. tions previously described are higher in patients with
Other symptoms include cognitive impairment, PPMS and SPMS than in those with RRMS.
fatigue and affective disturbance. Overall, 40–70% of Several qualitative and semi-​quantitative scales have
patients with MS have cognitive impairment, which can been proposed to grade the clinical manifestations
start in the earliest phases of the disease41. Cognitive of MS. Of these, the Expanded Disability Status Scale
deficits can predict conversion to clinically definite MS (EDSS)148 is the most widely accepted measure of clin­
in patients with CIS141, are more frequent and more-​ ical disability. The EDSS is a scale that ranges from 0
pronounced in chronic progressive MS, worsen over (a completely normal neurological examination) to 10
time and affect patients’ daily life activities41. Common (death owing to MS) and provides 8 subscale measure­
cognitive symptoms include impairment in information ments (functional system scores) that include the main
processing speed, episodic memory, attention, efficiency functional domains affected by MS, including pyra­
of information processing and executive function41. Up midal, cerebellar, brainstem, sensory, bowel and bladder,
to 95% of patients experience fatigue142. Several mech­ visual, mental and other domains.
anisms have been suggested to promote the occurrence
of fatigue in patients with MS. Fatigue can be associated Diagnostic criteria
with relapses and can persist after the attack has sub­ The diagnosis of MS is primarily based on clinical cri­
sided, but it can also be a feature of daily life and can teria; in most patients, the occurrence of two or more
be present for years. Several strings of evidence support clinically distinct episodes of CNS dysfunction with
the hypothesis of a central origin of MS-​related fatigue at least partial resolution is sufficient for diagnosis of
owing to a dysfunction of cortico-​subcortical circuits, RRMS. Although the diagnosis can be made on the
mainly involving structural damage in fronto-​parietal basis of clinical criteria alone, MRI can support, sup­
regions and the basal ganglia143. Sleep disorders (for plement or replace some clinical criteria owing to the
example, insomnia, obstructive sleep apnoea and restless sensitivity and specificity of this imaging modality in
legs syndrome) are found in up to 54% of patients with demonstrating demyelinating lesions, as well as DIS and
MS144 and might also promote fatigue145. DIT149 (Box 2).

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 11

0123456789();
Primer

Table 2 | Typical and atypical clinical presentations of MS


Presentation Typical or Onset Involvement Signs or symptoms Recovery
atypical
presentation
Optic neuritis Typical Sub-​acute Unilateral • Afferent pupillary defect Gradual
to chronic • Central visual blurring or scotoma recovery within
(hours to • Reduced visual acuity 2–4 weeks after
days) • Dyschromatopsia (colour blindness) reaching peak
• Normal optic disc or optic disc swelling severity
• Mild unilateral orbital pain that is worsened by eye
movements
Atypical Acute Bilateral • Peripheral visual loss Progressive
(seconds • Altitudinal visual loss worsening or no
to minutes) • Retinal haemorrhages or exudates recovery
• Severe optic disc swelling
• No light perception
• No or severe orbital pain
• Photophobia
Brainstem Typical Sub-​acute Unilateral and localized • Unilateral or bilateral internuclear ophthalmoplegia Gradual
and/or and/or • Multidirectional nystagmus recovery
cerebellar chronic • Sixth cranial nerve palsy starting within
syndromes (hours to • Ataxia or gait imbalance 2–4 weeks after
days) • Vertigo reaching peak
• Facial numbness or sensory loss severity
• Dysmetria and decomposition of complex
movements
• Dysarthria and slurred speech
• Dysphagia
• Hearing loss
• Nausea
Atypical Acute Alternating syndromes • Vascular territory signs Progressive
(seconds • Isolated trigeminal neuralgia worsening or no
to minutes) • Fluctuating ocular or bulbar weakness recovery
• Fever
• Meningism
Myelitis Typical Sub-​acute • Incomplete transverse • Sensory involvement: paresthesias (numbness, Gradual
and/or myelitis tingling, pins-​and-needles feeling, tightness, recovery
chronic • Asymmetric involvement coldness and/or swelling of the limbs or trunk), starting within
(hours to Lhermitte sign, impairment of vibration and joint 2–4 weeks after
days) position sense, decreased pain and light touch reaching peak
perception and Uhthoff phenomenon severity
• Motor deficits: pyramidal signs (Babinski sign,
bright reflexes and clonus), spastic paresis and/or
weakness (asymmetric) and spasticity
• Sphincter dysfunction: urinary urgency, hesitancy,
urge incontinence, constipation and faecal
incontinence
• Sexual dysfunction: erectile dysfunction and
impotence
Atypical Acute • Complete transverse • Progressive and symmetrical spastic paraparesis Progressive
(seconds myelitis • Progressive sensory ataxia (posterior column worsening or no
to minutes) • Complete Brown-​ involvement) recovery
Séquard syndrome • Sharp level to all sensory modalities
• Cauda equina syndrome • Segmental loss of pain and temperature sensation
• Anterior spinal artery • Areflexia and/or spinal shock
territory lesion • Acute urinary retention
• Localized or radicular • Severe pain
spinal pain
Cerebral Typical Sub-​acute Unilateral • Hemisyndrome (corticospinal tract involvement): Gradual
hemispheric and/or hemiparesis and hemisensory deficits recovery
syndromes chronic • Campimetric deficits (optic radiation involvement) starting within
(hours to 2–4 weeks after
days) reaching peak
severity
Atypical Acute Bilateral • Encephalopathy Progressive
(seconds • Epilepsy worsening or no
to minutes) • Cortical blindness recovery
• Headache
• Intracranial hypertension
MS, multiple sclerosis.

12 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

The original Schumacher diagnostic criteria required culminating in the latest revision of the MS diagno­
evidence based on clinical examination alone for DIS stic criteria (that is, the 2017 revision of the McDonald
and DIT150. Although these criteria were developed criteria)152 (Box 2; Fig. 6).
before the introduction of MRI, they are still used as
a reference tool for basic definitions of DIS, DIT and MRI. MRI has a high sensitivity for detecting macro­
relapse. These criteria were subsequently modi­fied to scopic abnormalities in the brain and spinal cord in
include laboratory diagnostic studies such as assess­ patients with MS. Abnormal MRI owing to the pres­
ment of cerebrospinal fluid, evoked potentials and ence of focal lesions is observed in almost all patients
neuroimaging151. Cerebrospinal fluid findings sup­ with MS and in most patients with CIS. MS lesions have
porting a diagnosis of MS include a normal or mildly typical MRI signal and location characteristics, which
raised white cell count and protein levels, increased aid in the diagnosis. Lesions usually appear as multi­
immunoglobulin G (IgG) index and the presence of focal, ovoid areas of increased signal on T2-weighted
cerebrospinal fluid-​specific IgG OCBs. Cerebrospinal images, with lesions commonly located in periventricu­
fluid-​specific OCBs were included in the latest revi­ lar, juxtacortical and infratentorial regions of the brain
sion of the MS diagnostic criteria152 on the basis of their and in the spinal cord (Figs 5,6). The administration
high prevalence in patients with MS (up to 88%)153 of gadolinium-​based contrast agents and the acquisition of
and owing to their role in predicting evolution to clin­ post-​c ontrast T1-weighted images enable active
ically definite MS (see below)154,155. Evoked potentials, lesions to be distinguished from inactive lesions; signal
including sensory (visual, somatosensory and brain­ enhancement, which underlies active lesions, occurs
stem auditory) and motor evoked potentials, assess owing to increased BBB permeability and corresponds
functionally relevant pathways and can identify clini­ to areas with ongoing inflammation. Lesions that persis­
cally silent lesions in the CNS, which might be missed tently appear hypointense on post-​contrast T1-weighted
during standard routine clinical examination156. MRI images (so-​called black-​holes) are associated with more
was formally included in the diagnostic algorithm of severe tissue damage than that seen with lesions that
patients with CIS and suspected MS in 2001 (ref.157). do not appear dark on such images. This hypointen­
The original criteria have been revised for clarification sity is suggestive of demyelination and axonal loss160.
and to simplify their use in the clinical setting55,149,158,159, Recommendations aimed at optimizing and standard­
izing the use of MRI of the CNS in clinical practice have
been given161.
Box 2 | The 2017 revised criteria for the diagnosis of MS In the diagnostic criteria for MS, MRI is used to con­
Relapsing–remitting MS firm DIS or DIT for RRMS, and it has been included in
• At least two clinical relapses and objective clinical evidence on neurological the diagnostic criteria for PPMS (Box 2; Fig. 6). The latest
examination of at least two lesions with distinct anatomical location, or at least two revision of the MS diagnostic criteria152 included the
clinical relapses and objective clinical evidence of one lesion and clear-​cut historical count of symptomatic lesions for the definition of DIS
evidence of a prior relapse involving a lesion in a distinct anatomic location and DIT, which enables the simplification of the appli­
• At least two clinical relapses and objective clinical evidence of one lesion; in addition, cation of MRI criteria without losing their accuracy55
DIS should be demonstrated by either a second clinical relapse implicating a different (Box 2). The inclusion of cortical lesions as part of the
CNS site or using MRIa diagnostic criteria is also relevant, as these lesions are
• One clinical relapse and objective clinical evidence of two or more lesions; specific for MS, although improvement in their detec­
in addition, DIT should be demonstrated by a second clinical relapse, or using MRIb or tion is still necessary. At present, ~18% of cortical lesions
demonstration of cerebrospinal fluid-​specific OCBs confirmed by pathological studies can be detected using
• One clinical relapse and objective clinical evidence of one lesion; in addition, DIS MRI, most of which are type I lesions, whereas type III
should be demonstrated by a second clinical relapse implicating a different CNS site lesions (subpial) are difficult to detect even with advanced
or using MRI, whereas DIT should be demonstrated by a second clinical relapse, or MRI techniques.
using MRI or demonstration of cerebrospinal fluid-​specific OCBs
The growing application of MRI has substantially
Primary progressive MS increased the identification of asymptomatic individuals
A disease course characterized by progression from onset, 1 year of disability with brain MRI abnormalities suggestive of MS, which is
progression (retrospectively or prospectively determined) independent of clinical referred to as radiologically isolated syndrome (RIS)162.
relapse and two of the following criteria:
Up to 34% of patients with RIS have a clinical attack
• One or more T2-hyperintense lesions in at least one area in the brain characteristic of within 5 years. Male sex, younger age and the presence
MS (periventricular, cortical and/or juxtacortical or infratentorial)
of spinal cord lesions increase the risk of having a first
• Two or more T2-hyperintense lesions in the spinal cord with no distinction between clinical event163. More studies are necessary to further
symptomatic or asymptomatic lesions
define RIS and the differential diagnosis of this dis­
• Demonstration of cerebrospinal fluid-​specific OCBs order and to develop recommendations to monitor and
CNS, central nervous system; DIS, dissemination in space; DIT, dissemination in time; MS, eventually treat these patients.
multiple sclerosis; OCB, oligoclonal band. aOne or more T2-hyperintense lesions in at least two Aside from use in diagnosis, MRI has also gained a
of four areas of the CNS (periventricular, juxtacortical (by combining cortical or juxtacortical
fundamental role in monitoring treatment efficacy (use
lesions), infratentorial and spinal cord lesions), with the removal of the distinction between
symptomatic and asymptomatic lesions. bSimultaneous presence of gadolinium-​enhancing in monitoring inflammation and/or neurodegener­
and non-​enhancing lesions at any time, with the removal of the distinction between ation) and in the early recognition of treatment-​related
symptomatic and asymptomatic lesions, or a new T2-hyperintense and/or gadolinium-​ adverse effects (for example, progressive multifocal
enhancing lesion on follow-​up MRI with reference to a baseline scan irrespective of the timing leukoencephalopathy (PML) and other opportunistic
of the baseline MRI.
infections)164.

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 13

0123456789();
Primer

a b c d e

f g h i j k

Serum Cerebrospinal
fluid

Fig. 6 | 2017 McDonald Criteria for demonstration of DIS and DIT in a patient with CIS suggestive of MS.
a–e | Dissemi­nation in space (DIS) can be demonstrated by one or more T2-hyperintense lesions in two or more of
four typical areas of the central nervous system (arrows), with the removal of the distinction between symptomatic
and asymptomatic lesions. Periventricular lesions are shown in part a. Cortical or juxtacortical lesions to define
juxtacortical involvement are shown in parts b and c. Infratentorial lesions are shown in part d. A spinal cord lesion is
shown in part e. f–j | Dissemination in time (DIT) can be demonstrated by a simultaneous presence of gadolinium-​
enhancing and non-​enhancing lesions (parts f,g) at any time and with the removal of the distinction between
symptomatic and asymptomatic lesions, a new T2-hyperintense and/or gadolinium-​enhancing lesion on follow-​up
MRI, with reference to a baseline scan (parts h–j), irrespective of the timing of the baseline MRI, or the presence of
cerebrospinal fluid-​specific oligoclonal bands (OCBs), which are not visible in the serum (part k). Several white matter
lesions are visible on the fluid-​attenuated inversion recovery (FL AIR) sequence (part f); one showed gadolinium-​
enhancement (arrowhead) on the post-​contrast sequence (part g), whereas the majority were non-​enhancing (dots).
Compared with the baseline FL AIR sequence (part h), a new T2-hyperintense and gadolinium-​enhancing lesion
(circled) is visible on follow-​up FL AIR (part i) and on post-​contrast sequences (part j). CIS, clinically isolated syndrome;
MS, multiple sclerosis.

Biomarkers and prognostic factors MS that alert clinicians to reconsider the differential
Several biomarkers and prognostic factors associated diagnosis in more detail. These red flags include lon­
with conversion from CIS to MS and with disability pro­ gitudinal extensive transverse myelitis extending over
gression in patients with CIS and early RRMS have been three spinal cord segments (suggesting a diagnosis of
identified, including environmental, genetic, clinical, NMOSD) and the presence of bleeds or infarcts, which
laboratory and MRI features132 (Table 3). Patients with CIS might suggest a cerebrovascular aetiology168,169.
and brain lesions at MRI (including patients with one
lesion only) have a >80% chance of developing MS within Prevention
20 years165,166. Although predicting the long-​term clinical As previously discussed, MS is a complex disease that
outcome of patients with MS, including the severity of is caused by the possible interactions between genetic
disability, is more difficult than predicting conversion to and environmental factors; some of these factors are
MS in patients with CIS, several risk factors have been modifiable and, therefore, represent a promising basis
identified (Table 3). A progressive disease from the onset for MS prevention170. Up to 80% of the population that
and a faster rate of disability accumulation in the first is at risk of MS (particularly in Western countries)
2–5 years are predictors of poor outcomes in PPMS132,167. have sub­optimal low serum vitamin D levels; avail­able
data suggest that the risk of MS is reduced if serum
Differential diagnosis 25-hydroxy vitamin D levels are >100 nmol per litre.
The range of diseases that mimic the clinical manifesta­ Accordingly, vitamin D supplementation might reduce
tions and MRI features of MS is wide (Box 3), and there­ MS incidence and promote potential benefits for the pre­
fore, careful exclusion of other neurological dis­orders vention of other systemic diseases, like osteoporosis170. In
is essential during the diagnostic work up for MS. The line with this, studies are ongoing to evaluate the effect of
identification of clinical and paraclinical features that vitamin D supplementation alone or in addition to DMTs
are not suggestive of MS might reduce the chance of a to reduce MS-​related disease activity (that is, occur­
false positive diagnosis. On this basis, MRI red flags have rence of relapses and formation of new white matter
been identified in individuals with clini­cally suspected lesions at MRI) and disability progression171.

14 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

In addition, smoking is associated with an increased and to promote more healthy lifestyles (through, for
risk of MS and poorer prognosis, therefore, promot­ example, reduction of food intake and increased phys­
ing smoking cessation might be a straightforward ical activity) are strongly suggested for people at risk
intervention to reduce MS incidence, disability pro­ of developing MS (especially adolescents and young
gression and the risk of comorbidities170. Overweight adults)170. Finally, on the basis of the evidence on the
and obesity are associated with a twofold to threefold role of EBV in MS, research aiming to develop specific
increased risk of MS and have a negative effect on treatments (including antiviral thera­pies, prophylactic
clinical and MRI disease-​related outcomes and with and therapeutic EBV vaccines and viral neutralization
several comorbidities; to this end, educational inter­ antibodies), which could be used for MS prevention,
ventions aiming to reduce body mass index (BMI) is ongoing170.

Table 3 | Features of CIS and early MS predicting conversion to definite MS and disability progression
Factor Association with Association Refs
conversion to MS with disability
progression
Environmental and lifestyle factors
Smoking Higher risk Poor prognosis 21,320

Low vitamin D levels Higher risk Unknown 155,321

EBV infection Higher risk Unknown 304

Obesity (particularly in childhood and adolescence) Higher risk Poor prognosis 322,323

Genetic factors
HLA-​DRB1*1501 Higher risk Unknown 324,325

Clinical factors
Non-​white ethnicity Higher risk Poor prognosis 326,327

Female sex Higher risk Good prognosis 328

Male sex Lower risk Poor prognosis 329

Older age Lower risk Poor prognosis 330,331

Younger age Higher risk Good prognosis 326,329

Onset with optic neuritis or somatosensory disturbances Lower risk Good prognosis 332,333

Onset affecting efferent pathways (for example, motor) Higher risk Poor prognosis 332

Monofocal onseta Lower risk Good prognosis 334,335

Multifocal onset b
Higher risk Poor prognosis 334,335

Cognitive impairment Higher risk Poor prognosis 141

Higher relapse rate in the first 2–5 years from disease onset NA Poor prognosis 167,329,336

Incomplete recovery after a relapse 329

Higher disability accumulation in the first 2–5 years from disease 167,329,336

onset
Shorter time to conversion to SPMS 329

Laboratoristic factors
Presence of cerebrospinal OCBs Higher risk Poor prognosis 153,155,333

High level of neurofilament light subunit Higher risk Poor prognosis 251,254,337

Neuroradiological factors
Higher number and volume of T2-hyperintense lesions Higher risk Poor prognosis 155,165,166,333

Lesions in infratentorial regions Higher risk Poor prognosis 338,339

Spinal cord lesions Higher risk Poor prognosis 340–342

Presence of gadolinium-​enhancing lesions Higher risk No data 155

New T2 lesions formation in the first 5 years Higher risk Poor prognosis 165,166

Optical coherence tomography factors


Lower ganglion cell and inner plexiform layer thickness Higher risk No data 266

CIS, clinically isolated syndrome; EBV, Epstein–Barr virus; MS, multiple sclerosis; NA, not applicable; OCBs, oligoclonal bands;
SPMS, secondary progressive multiple sclerosis. aInvolvement of a single functional system (for example, motor, somatosensory ,
visual, cerebellar or brainstem). bTwo or more functional systems involved simultaneously.

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 15

0123456789();
Primer

Box 3 | Common and relevant differential diagnoses of MS options for two decades mainly because of their excellent
safety profiles but also owing to their lower cost than
Infectious diseases168,169: newer drugs. However, although these therapies have
• Meningitis very low risk of severe adverse drug reactions, they
• Encephalitis (including progressive multifocal leukoencephalopathy) have only moderate clinical effectiveness and often poor
• Lyme neuroborreliosis tolerability owing to injection-​related adverse effects,
• Intracerebral abscess such as flu-​like symptoms with IFNβ and injection site
inflammation with both IFNβ and glatiramer acetate,
Genetic disorders307,308:
which frequently prompt treatment switches172 (Table 4).
• Leukodystrophies (for example, adrenoleukodystrophies) The increasing number of approved DMTs improves the
• Leber hereditary optic neuropathy possibility of tailoring therapy to individual patient
• Fabry disease needs with regard to efficacy, safety aspects and patients’
Metabolic disorders168,309: preferences. Although caution should be applied when
• Vitamin B12 deficiency comparing across studies owing to the heterogeneity
• Copper deficiency of patient cohorts and the lack of reliable comparative
studies among treatments, approximate measures of
Vascular disorders169,310: clinical efficacy can be deduced from some parameters
• Cerebral small-​vessel disease typically evaluated in randomized clinical trials (RCTs),
• Cerebral autosomal dominant arteriopathy with subcortical infarcts and including the relative reduction in annual relapse fre­
leukoencephalopathy (CADASIL) quency and the number of patients needed to treat (the
• Susac syndrome number of patients you need to treat to prevent one
• Primary angiitis of the central nervous system additional bad outcome, typically a relapse or disability
• Dural arteriovenous fistula progression)173,174 (Table 4).
Systemic immune-​mediated diseases169: The dominant current treatment strategy for RRMS,
called escalation therapy, is endorsed by the European175
• Systemic lupus erythematosus
and American176 guidelines and is, in part, dictated by
• Behçet disease
the label of different DMTs and public or insurance-​
• Sarcoidosis based regulations. The basis of escalation therapy is to
• Sjögren syndrome start with a safe but moderately effective DMT, typically
Non-​MS idiopathic inflammatory demyelinating diseases311–314: IFNβ, glatiramer acetate, teriflunomide or dimethyl
• Neuromyelitis optica spectrum disorders fumarate, and switch to another first-​line DMT in
• Myelin-​oligodendrocyte glycoprotein (MOG) encephalomyelitis patients with intolerable adverse effects or to a more
• Acute disseminated encephalomyelitis (especially in paediatric patients) effective DMT (second-​line or third-​line therapies) in
those with new relapses or MRI lesions. In patients with
• Chronic lymphocytic inflammation with pontine perivascular enhancement
responsive to steroids severe disease who do not respond to traditional DMTs,
autologous haematopoietic stem cell transplantation
Variants of multiple sclerosis (MS)168,315: might be effective177. However, owing to the availabil­
• Balo concentric sclerosis ity of more effective DMTs, such treatment failures are
• Schilder disease increasingly rare, and in general, <1% of patients with
• Marburg MS RRMS are candidates for transplantation177,178.
Headache Another treatment strategy, known as induction ther­
• Migraine169 apy, has been introduced on the basis of the availabil­
ity of more-​effective drugs and of the evolving concept
of treating patients earlier with more-​effective drugs
Management (such as alemtuzumab or ocrelizumab) to prevent the
The treatment of MS can be divided into DMTs that accumulation of irreversible CNS damage and clinical
are used to reduce inflammatory disease activity and disability. Induction therapy refers to a strong immuno­
its long-​term clinical consequences, the management of intervention that is started soon after a confirmed diag­
MS relapses and symptomatic treatments used for short-​ nosis in a patient with negative prognostic factors (that
term amelioration of MS symptoms such as fatigue, pain is, a higher disease activity (severe and frequent relapses
and spasticity. Several DMTs are available for the treat­ and a higher number of lesions at MRI) and an accu­
ment of RRMS, whereas only one DMT is approved for mulation of disability) (Table 3). This approach enables
PPMS. Additional DMTs are now in clinical trials a rapid reduction in disease-​associated inflammation by
for RRMS, PPMS and SPMS, and intense efforts are removing T cells, B cells and myeloid cells and possibly
being made to identify novel therapeutic targets. shifting towards a more tolerogenic condition owing to
a reset of the immune system, which can be followed
Current standard of care by use of less-​aggressive therapies as maintenance if
Disease-​modifying treatments for RRMS. A DMT needed179. One or more cycles of induction therapy
should be prescribed as soon as a patient has been diag­ can be performed, followed by a possible de-​escalation
nosed with RRMS or CIS to reduce the risk of disease therapy. However, when treatment with a high-​efficacy
progression. Injectable DMTs, such as IFNβ or glati­ drug is stopped, a careful evaluation of further treatment
ramer acetate, have been the main first-​line treatment selection is needed; particularly for therapies that do not

16 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

exert a substantial immune reset (such as natalizumab or Disease-​modifying treatments for progressive disease.
fingolimod), another high-​efficacy treatment should be DMTs that are used for the treatment for RRMS cannot
started to prevent potential disease reactivation175. prevent disease worsening in patients with PPMS and
However, despite the benefits of induction therapy, SPMS173. Mitoxantrone, a cytostatic drug, was approved
an important reason to restrict the use of more-​effective by the US FDA for SPMS in 2000, but its use is limited by
DMTs, such as second-​line or third-​line treatments, is cardiotoxic and mutagenic adverse effects185. Fingolimod
their risk profile. Collectively, older injectable DMTs and natalizumab have been tested in patients with
have the lowest risk of more-​serious adverse effects, PPMS186 or SPMS187, respectively, but did not demon­
although these therapies have more-​frequent but less-​ strate superiority over placebo. More promising data
serious adverse effects that affect tolerability179. By emerged from a placebo-​controlled study with rituxi­
contrast, more-​recent DMTs typically have a better tol­ mab in PPMS, in which the risk of disability progres­
erability but are associated with increased risks of severe sion was reduced in younger patients with signs of active
adverse effects, especially infections179 (Table 4). These inflammation on the baseline MRI scan188. A beneficial
effects include respiratory and urinary tract infections, effect of anti-​CD20 agents in PPMS was substantiated
herpesvirus reactivation and PML. In particular, PML, in a larger study with ocrelizumab, which significantly
which is an opportunistic infection affecting the brain reduced the risk of disability progression compared
owing to John Cunningham virus, has been described with placebo and led to the first approval of a DMT for
in patients who received natalizumab, dimethyl fumar­ PPMS189. Notably, the patients in this study were young
ate and fingolimod treatment. For natalizumab, the risk (mean 44.6 years of age), and 27% of patients had signs
of PML can be determined by testing patients for John of active inflammation on baseline MRI, suggesting that
Cunningham virus antibodies. In addition, newer DMTs earlier phases of the disease might be more responsive to
have substantial immunosuppressive activity, which has treatment. Collectively, data indicate that DMTs that act
been suggested to increase risk of long-​term malignancy, mainly on the adaptive immune system have a reduced
although the precise risk of this is still uncertain180. efficacy in progressive disease compared with in RRMS
Alemtuzumab treatment has been associated with auto­ but that treatment with anti-​CD20 DMTs such as ocre­
immune diseases181–183; of these, thyroid disease is the lizumab or rituximab should be considered, especially
most common autoimmune adverse effect associated in patients with shorter disease duration and/or signs of
with alemtuzumab and is found in up to one-​third of active inflammation on MRI.
patients, although more-​rare, serious autoimmune reac­
tions, including immune thrombocytopenic purpura, Relapses. The most established treatment for the acute
Goodpasture disease, neutropenia, haemolytic anaemia, management of MS relapses is high-​dose corticoster­
agranulocytosis and acquired haemophilia, have also oids. These drugs are associated with a faster func­
been observed. In addition, before starting a DMT, the tional recovery and protect against the occurrence of
teratogenic risks associated with the treatment must be more severe deficits in the first weeks after treatment
considered in women planning pregnancy. Of the treat­ but have unclear long-​term benefits. Current protocols
ments currently available, glatiramer acetate is the only typically include 3–5 days of intravenous methylpredni­
DMT considered safe to use during pregnancy. solone with or without oral tapering with prednisone.
One of the limitations of data derived from RCTs is Intramuscular administration of dexamethasone and
that trials include a selected group of mostly treatment-​ oral administration of high-​dose methylprednisolone
naive patients who do not have substantial comorbidi­ have an equivalent efficacy to intravenous administra­
ties and relevant DMT comparators, which means that tion. Relapses that do not respond to corticosteroids
we have an imprecise evidence base for tailoring DMT can be treated with plasma exchange (3–5 courses) or
strategies in these patient groups, therefore, careful moni­ intravenous immunoglobulin.
toring of patients with MS is constantly needed184. Also,
carrying out an individualized risk–benefit assessment Symptomatic treatments. Several different pharmaco­
for MS therapies is important, as the relative benefit of logical agents are used to treat the symptoms of MS, such
individual DMTs differs between patients and all DMTs as impaired walking capability, spasticity, pain, loss of
are associated with adverse effects of varying severities bladder and bowel control and neuropsychiatric symp­
(Table 4). For example, younger age and a more-​active toms190,191. However, for most therapies, the evidence
disease (in terms of relapse frequency and MRI activity) base for clinical efficacy in patients with MS is weak.
are associated with increased benefit regarding treatment-​ Only two symptomatic treatments have under­
associated long-​term protection of neurological func­ gone more extensive testing in MS: nabiximols for the
tions. As the risk–benefit ratio of therapy can change over treatment of spasticity and dalfampridine for walking
time, studies addressing the contexts in which DMTs can ability. Nabiximols can ameliorate spasticity in patients
safely be discontinued are also needed, even if current with MS192,193, and empirical evidence supports the use
guidelines suggest continuing a DMT if a patient is stable of baclofen, dantrolene, tizanidine and botulinum toxin
and shows no safety or tolerability issues175. Furthermore, A injections for the treatment of spasticity in restricted
the possible effect of combined therapies (the simultane­ muscle groups. Dalfampridine is a voltage-​dependent
ous treatment with two or more DMTs) should be further potassium channel blocker that improves the trans­
explored even though, at present, there is no convinc­ mission of nerve signals in demyelinated axons and
ing data suggesting that combinations of DMTs would improves the walking ability of people with MS194,195.
provide an added benefit in terms of efficacy. In addition to medical treatments, walking aids such as

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 17

0123456789();
Primer

Table 4 | Relapse rates and the main adverse effects of DMTs in relapsing–remitting MS
DMT Type Reduction of Adverse effects Refs
annualized
relapse ratea
First line
Glatiramer s.c. mixture of synthetic 30% • Injection site reactions (erythema, inflammation, 343

acetate polypeptides induration or pain at injection site)


• Flushing
• Chest tightness or pain
• Palpitations
• Anxiety
• Trouble breathing
IFNβ1a s.c. recombinant protein 32% • Injection site reactions (erythema, inflammation, 344

induration or pain at injection site)


IFNβ1a i.m. recombinant protein 32% • Flu-​like symptoms
345

IFNβ1b s.c. recombinant protein 34% • Leukopenia (neutropenia or lymphopenia) 346

• Thrombocytopenia
Pegylated s.c. pegylated 35% • Anaemia
347

IFNβ1a recombinant protein • Infections


• Thyroid dysfunction (hypothyroidism or
hyperthyroidism)
• Liver damage (transaminase increase)
• Fatigue
• Mood disturbances (depressive symptoms)
Teriflunomide Oral pyrimidine 34% • Headache 348,349

synthesis inhibitor • Diarrhoea


• Hair thinning or loss
• Liver damage (transaminase increase)
• Increased blood pressure
• Paresthesia
• Leukopenia (neutropenia or lymphopenia)
• Infections
Dimethyl Oral NRF2 agonist 49% • Flushing 350,351

fumarate • Liver damage (transaminase increase)


• Gastrointestinal disturbances (abdominal pain,
nausea and vomiting)
• Leukopenia (mainly lymphopenia)
• Infections
• PML
• Allergic reactions
Second line
Fingolimod Oral S1P inhibitor 54% • Reduced heart rate 352,353

• Increased blood pressure


• Leukopenia (mainly lymphopenia)
• Infections
• Liver damage (transaminase increase)
• Macular oedema
• PML
• Skin cancer (basal and Merkel cell carcinoma)
and melanoma
Daclizumab i.v. monoclonal 44%b • Liver damage (transaminase increase) 354

(withdrawn) anti-​CD25 antibody • Gastrointestinal disturbances (abdominal pain,


nausea and vomiting)
• Allergic reactions
• Infections
• Immune-​mediated encephalitis
Alemtuzumab i.v. monoclonal 52% • Infusion-​related reactions 355,356

anti-​CD52 antibody • Leukopenia (mainly lymphopenia)


• Infections
• Autoimmune reactions (immune
thrombocytopenia, immune thyroiditis and
immune glomerulonephritis)
• Cancers (thyroid cancer, melanoma and
lymphoproliferative disorders)
Cladribine Oral purine analogue 58% • Leukopenia (neutropenia or lymphopenia) 357

• Infections
• Rash
• Alopecia
• Cancers

18 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

Table 4 (cont.) | Relapse rates and the main adverse effects of DMTs in relapsing–remitting M­S­
DMT Type Reduction of Adverse effects Refs
annualized
relapse ratea
Second line (cont.)
Ocrelizumab i.v. monoclonal 45%b • Infusion-​related reactions 107

anti-CD20 antibody • Leukopenia (mainly lymphopenia)


• Decreased blood immunoglobulin
• Infections
• Cancers
Natalizumab i.v. monoclonal 69% • Infusion-​related reactions 358,359

anti-​VLA4 antibody • Allergic reactions


• Infections
• Progressive multifocal leukoencephalopathy
DMT, disease modifying treatment; i.m., intramuscular; i.v., intravenous; MS, multiple sclerosis; NRF2, nuclear factor erythroid 2-related
factor 2; PML, progressive multifocal leukoencephalopathy; S1P, sphingosine-1-phosphate; s.c., subcutaneous; VLA4, very late
antigen 4. aCompared with placebo unless otherwise stated. bCompared with IFNβ1a.

orthoses, crutches or walkers are important to improve might exert positive effects on the quality of life (QOL)
ambulation capacity. Traditional or electrical wheel of patients with MS206. Anxiety and depression, as well as
chairs and other mobility devices constitute important suicide, rates are elevated in MS, but studies addressing
means to preserve independence of movement among the efficacy of pharmaceutical and non-​pharmaceutical
patients with more-​advanced disease. interventions specifically in MS are rare, therefore,
Damage to sensory nerve tracts in MS leads to chronic treatment guidelines largely rely on data from non-​MS
neuropathic pain conditions for which gabapentinoids populations and include the use of selective serotonin
(such as gabapentin and pregabalin), tricyclic anti­ or noradrenergic reuptake inhibitors and/or cognitive
depressants and serotonin noradrenaline reuptake inhib­i­ behavioural therapy207,208. As a whole, offering patients
tors are first-​line treatments. Opioids such as tramadol or with MS comprehensive rehabilitation programmes to
codeine are second-​line treatments for moderate to severe address the wide range of MS-​associated symptoms, with
pain. In some countries, cannabinoids in the form of the aim of alleviating burdensome symptoms through
medical marijuana or synthetic drugs are recommended increasing the levels of physical activity, is important209.
as a possible third-​line option196. The management of
lower urinary tract symptoms consists of oral antimus­ Biomarkers for drug response
carinic drugs, administered alone or in combination MRI is the only tool that can reliably assess disease
with intermittent self-​catheterization, and the use of activity in MS164. In general, MRI can identify the
botulinum toxin A bladder instillations, neuromodula­ degree of inflammation (demonstrated by the quanti­
tion, indwelling catheters and surgery in patients with fication of contrast-​enhancing lesions and new T2
more-​severe symptoms197. lesions forming over time) and the degree of neuro­
Despite the high prevalence and clinical relevance degeneration (demonstrated by atrophy of the brain and
of cognitive impairment in patients with MS, effective spinal cord). Newly appearing T2-hyperintense MRI
treatments options are still lacking. The effects of symp­ lesions are a valid surrogate marker of treatment effi­
tomatic therapies such as modafinil and donepezil are cacy in phase II DMT studies and correlate with relapse
inconsistent; however, some DMTs (such as IFNβ, fingo­ frequency210. To this end, regular brain MRI scans are
limod and natalizumab) in combination with cognitive recommended in patients with MS to verify treatment
rehabilitation might improve or at least stabilize cogni­ effects owing to the frequent subclinical disease activity.
tive performances198. Fatigue and psychiatric comor­ However, some studies have suggested the deposition of
bidities are important contributors to loss of working gadolinium-​based contrast agents in the basal ganglia
ability and social participation of patients with MS199,200. and in the dentate nucleus of patients who underwent
The off-​label prescription of alertness-​improving drugs serial MRI acquisitions211. Although significant clinical
such as modafinil and amphetamine is common even consequences of these deposits have not been demon­
though evidence supporting the efficacy of these thera­ strated, further studies are required to better understand
pies in MS is poor or absent201,202. However, several the potential long-​term biological and clinical effects of
smaller studies suggest that novel approaches to treat gadolinium administration211.
fatigue, including alfacalcidol (a vitamin D analogue)203, In patients with MS, the development of long-​term
physical exercise202, cognitive behavioural therapy202,204, disability correlates with brain atrophy measures212,
deep transcranial magnetic stimulation205 and fatigue although these measurements have not been widely
management courses202, give some clinical benefit. introduced into the clinical routine for several rea­
Considering the high prevalence of sleep disorders sons. For example, the quantification of atrophy is still
in patients with MS, the treatment of an underlying challenging because high-​quality MRI sequences are
sleep disorder (continuous positive airway pressure for necessary, several factors (for example, BMI, genetic
obstructive sleep apnoea, drug therapy and cognitive factors, alcohol consumption, and so on) can influence
behavioural therapy for insomnia and drug therapy for the measurement of atrophy, particularly longitudinally,
restless leg syndrome) significantly reduced fatigue and and its application for single patient monitoring still

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 19

0123456789();
Primer

needs to be validated213. Although the time to detect the classical test theory and the Rasch model. The latter is a
effects of MS treatments on brain atrophy is longer than prescriptive psychometric approach which rests on the
that for the accumulation of lesions, one meta-​analysis assumption that a scale has the basic property of uni­
demonstrated that the treatment effect on brain atrophy dimensionality and meets the fundamental measurement
correlated with the effect on disability progression214. In requirements, enabling arithmetic operations such as
addition to radiographic imaging, soluble biomarkers addition and subtraction232. Two MS-​specific unidimen­
that are present in different body fluids have been pro­ sional QOL inventories were produced in the United
posed as markers of treatment effects in MS, but so far, Kingdom using the Rasch model: the Leeds MS QOL
none have been validated215. (LMSQOL) and the Patient-​Reported Outcome Indices
for MS (PRIMUS) QOL233,234.
Treatments currently under development For other patient-​reported outcome measures, lin­
Several novel drugs that have similar mechanisms of guistically validated versions are needed to appropri­
action to existing DMTs are in late-​stage clinical trials, ately compare data from different populations and to use
such as the sphingosine-1-phosphate (S1P) inhibitors QOL as an outcome measure in international trials228,229.
siponimod, ponesimod and ozanimod, as well as the The only example of an MS-​specific QOL instrument
anti-​CD20 monoclonal antibody ofatumumab216,217. concurrently developed in 14 languages is the MS
The development of antigen-​specific therapies is a major International QOL Questionnaire (MusiQOL)235. Other
future goal and would alleviate the problems, such as inventories were developed in US English230,231 or UK
infections, associated with long-​term immunosuppres­ English234 and linguistically validated in other languages.
sion. In fact, small trials reporting beneficial effects Another QOL scale property that is key for use as an
suggest that this may become feasible218,219. Progress has outcome measure in clinical trials is responsiveness (the
also been made in identifying therapeutic agents with ability of an instrument to measure a meaningful clin­
potential neuroprotective or remyelinating effects, in ical change)228,229, which remains insufficiently assessed
part by re-​purposing older drugs used in other con­ in the MS population236–240.
texts. For example, phenytoin (an anti-​epileptic drug) Although the measurement and reporting of MS
has shown promising neuroprotective effects in pre­ QOL data have evolved considerably, translation to
serving nerve fibres in acute optic neuritis220, and the clinical practice has been limited. Questionnaire length
anti-​allergy drug clemastine increases optic nerve con­ and complexity of score calculation and interpretation
duction in MS-​associated optic neuropathy221. Similarly, are recognized as major barriers to the use of QOL and
the cholesterol-​lowering drug simvastatin reduced the other patient-​reported inventories in clinical practice227.
rate of brain atrophy in SPMS222. Newly developed drugs Questionnaire length can be particularly challenging for
aimed at increasing remyelination, such as the monoclonal patients with fatigue, which is common in patients with
antibody opicinumab, are also in clinical testing223,224. MS. The 29-item MS QOL (MSQOL-29) was recently
Recently, different dietary-​based treatment strat­ produced from the MSQOL-54 using Rasch analysis239.
egies have been proposed in patients with MS225. Among A further step is the development of computerized
them, the oral supplementation of biotin (vitamin B7) adaptive testing, which would present questions that
was shown to improve or at least stabilize disability in are most relevant for an individual patient, reducing the
patients with PPMS or SPMS compared with placebo, questionnaire length and exposure to items that are not
and an RCT in SPMS is ongoing225. relevant or appropriate241. A multidimensional comput­
erized adaptive version of the MusiQOL has recently
Quality of life been devised242.
Symptoms of MS can negatively affect patient daily func­
tioning, relationships, work and leisure activities and Outlook
ultimately lead to reduced QOL226,227. Patient-​reported Key outstanding questions in MS include the following:
outcome measures are increasingly used to overcome the identification of paraclinical features (for example,
the limitations of clinician-​reported measures (such laboratory, neuroradiological or neurophysiological) that
as the EDSS) in fully capturing the patient’s experience of are more specific to the pathological substrates of MS,
the disease. Patient-​reported outcomes typically address which might increase diagnostic specificity and reduce
symptoms, functioning, satisfaction with care, treatment the risk of misdiagnosis; the development of biomarkers
adherence and perceived value of treatment. In addition that are more sensitive to disease-​related changes (such
to being patient reported, QOL measures are multi­ as inflammatory activity or disease progression); the
dimensional: at a minimum, they assess the consequences optimization of treatment at an individual patient level;
of the disease and treatments on the physical, psycho­ and an assessment of the effect of comorbidities.
logical and social domains of patients’ lives. One major
driver of the development of QOL measures for MS was Improving specificity in the diagnosis
the recognition by the regulatory agencies of the impor­ An awareness of an increased risk of misdiagnosis243
tance of this construct as an outcome measure for clinical owing to an oversimplification of the diagnostic criteria
trials, provided that QOL inventories are targeted to the and their incorrect application was one of the motivating
specific patient population228,229. The oldest MS-​specific factors for the 2017 revision of the McDonald criteria152.
QOL instruments are the 54-item MS QOL (MSQOL-54) The identification of distinctive MRI features of MS
and the Functional Assessment of MS (FAMS)230,231. might help to reduce the risk; for example, the consistent
The FAMS was produced using a combination of the identification of lesions around small vessels in MS and

20 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();
Primer

accordingly the MRI detection of a vein centrally inside plexiform layer and the retinal nerve fibre layer (RNFL)
white matter lesions have been proposed as biomarkers of have been the two most frequently studied layers of the
inflammatory demyelination. Indeed, the proportion retina in patients with MS. In patients with acute optic neu­
of lesions with a central vein is higher in MS244 than in ritis, dynamic modifications of RNFL thickness, charac­ter­
other conditions, including NMOSD245, CNS inflamma­ ized by an initial increase owing to acute inflammatory
tory vasculopathies246 and vascular diseases247. In addi­ oedema followed by atrophy within the subsequent
tion, the detection of a central vein in three white matter 3–6 months, have been detected, therefore, the occur­
lesions has good specificity, sensitivity and inter-​rater rence of substantial and irreversible damage is detectable
reliability for the differential diagnosis of MS in patients only after 6 months267. Meta-​analyses of TD-​OCT and
with MS with or without comorbidities, patients with SD-​O CT studies have demonstrated RNFL thinning
migraine and others with a misdiagnosis of MS248. Future to be most marked in the temporal quadrant, which is
work is required to assess the predictive value of the cen­ the receiving region for the macular fibres268,269. In addi­
tral vein sign for the development of MS in individuals tion to the RNFL, technological advances in OCT have
with CIS and to define a standardized imaging protocol allowed the segmentation of other retinal layers, such as
and to identify criteria for the clinical implementation the ganglion cell layer (GCL), which has been shown to
of central vein assessment244. have a different response to optic neuritis attacks, as this
layer has a faster onset of thinning without any acute-​
Novel biomarkers phase oedema. For this reason, macular GCL thickness
Several novel biomarkers are currently being investigated has been proposed as a superior early indicator of neu­
in an attempt to improve MS diagnosis and monitoring. ral changes following optic neuritis270. RNFL thinning
Among them, neurofilaments, OCT and measures of has been shown in all MS phenotypes and can start in
grey matter damage are receiving the most attention. patients with CIS271. RNFL is more severe in patients
with SPMS272 than in those with RRMS independent of
Neurofilaments. Neurofilaments are major compo­ a history of previous optic neuritis268,273. Progression
nents of the axonal cytoskeleton, consisting of light of RNFL and GCL thinning correlates with visual
(NfL), intermediate and heavy chains that are released deficits and worsening disability274. In addition, cor­
from damaged neurons and axons in neurological dis­ relations between OCT parameters and atrophy of the
orders249; these neurofilament chains can be quantified in whole brain275 and grey matter as detected using MRI
the blood and cerebrospinal fluid as a marker for neuro-​ have been observed276. As discussed previously274, evi­
axonal damage in MS215,250. Several studies have demon­ dence for the utility of OCT in monitoring therapeutic
strated that levels of NfL in cerebrospinal fluid are higher response in MS is accumulating.
in individuals with CIS who convert to MS than in indi­ The role of OCT has now expanded beyond the detec­
viduals who do not convert250,251 and also in patients with tion of damage owing to optic neuritis269. Assessment
RIS who develop a first clinical attack than in those who of the afferent visual pathway has been proposed as a
do not252. In addition, higher levels of NfL are associ­ model of acute demyelinating events of the optic nerve
ated with greater disability253, more frequent relapses254, and of the chronic effects of MS. The latter derives from
higher numbers of T2-hyperintense and gadolinium-​ the particular structure of retinal ganglion cells, which
enhancing lesions on MRI255,256 and more severe brain lack myelination in the retina. As a consequence, their
atrophy256,257. In patients with RRMS, higher levels of thinning may reflect a cascade from optic neuritis or an
NfL predicted more severe disability and evolution to overall process of neurodegeneration277.
SPMS after 14 years of follow-​up258, whereas in patients
with PPMS or SPMS, levels of NfL predicted the annual Grey matter damage. Improvements in MRI tech­
EDSS increase259. Levels of NfL in cerebrospinal fluid or niques have enabled the measurement of grey matter
serum could be used to monitor treatment effects, as pathology in vivo, including focal lesions, tissue loss
demonstrated by the reduction of NfL levels seen after and neuronal abnormalities. Even though correlative
treatment start with natalizumab260 or fingolimod261,262, MRI pathologi­cal studies have demonstrated that only
although further studies are still necessary before using a small part of grey matter lesions is detected by cur­
this quantification in the clinical setting. rent technologies278, this seems to be sufficient to pro­
vide clinically relevant information. Focal lesions in
Optical coherence tomography. OCT can generate the cortex are a distinctive feature in patients with MS,
high-​resolution images of the retina in a non-​invasive, and the presence of at least one cortical lesion (Fig. 6)
rapid and reproducible manner and in a multicentre identifies patients with CIS at higher risk of develop­
setting263; specific consensus guidelines for performing ing MS55. Cortical lesions are highly specific for MS,
and reporting OCT results have also been defined264,265. as they have not been detected in other neurological
Unlike time-​domain OCT (TD-​OCT), spectral-​domain disorders that can mimic MS so far, such as NMOSD279
OCT (SD-​OCT) enables visualization of individual reti­ or migraine280. The presence of both cortical lesions and
nal layers and, accordingly, the advent of SD-​OCT has grey matter atrophy are more pronounced in patients
marked a substantial leap forward in image resolution with progressive MS and correlates with more severe
and acquisition speed. disability281. In addition, the quantification of cortical
In MS, OCT might show asymptomatic optic nerve lesions and grey matter atrophy enables the long-​term
involvement in patients with CIS and might predict prognostication of worsening of disability and cognitive
conversion to MS266; the combined ganglion cell-​inner dysfunction212,282.

Nature Reviews | Disease Primers | Article citation ID: (2018) 4:43 21

0123456789();
Primer

Evaluation of grey matter damage might also be neurodegeneration, cognitive impairment, fatigue and
useful for assessing the effects of specific treatments283. QOL, are not considered291,292. As brain atrophy occurs
However, the standardization of acquisition and analysis faster in patients with MS than in healthy individuals
procedures for the assessment of grey matter involve­ and this atrophy is clinically relevant213, it has been sug­
ment, particularly in a longitudinal setting, is lacking. gested that atrophy evaluation (NEDA-4) should also be
Further studies with large sample sizes and a longitudi­ included in the monitoring of patients with MS293. The
nal setting are necessary to define reliable approaches integration of neuropsychological and blood or spinal
for quantification of grey matter damage and to establish fluid NfL levels, or patient-​reported outcomes and QOL
normative values for grey matter volume changes in both measures, has also been proposed287. However, before
healthy individuals and patients with MS282. including any of these measures in clinical practice, it is
crucial to define the cut-​off values that separate normal
Optimizing treatment decisions at an individual level physiological changes from pathological changes213,287.
The greater complexity of MS treatment scenarios
owing to the greater availability of DMTs has increased Tackling comorbidities
uncertainty in daily treatment decisions both in terms The comorbidities of patients with MS have only
of the initial type of treatment and when to switch recently begun to receive attention294,295. As is the case
between different treatments. In real-​world practice, it for many chronic diseases, physical and mental comor­
is often difficult to apply criteria derived from RCTs, bidities as well as adverse lifestyle or health factors,
which are typically run in standardized and controlled such as smoking and obesity, are common and can
situations. Also, RCTs typically have a short duration, affect the course of disease by modulating biological
and the long-​term effectiveness and safety of a given pathways that promote inflammation and immune
treatment cannot be assessed. The growing number of responses, influencing disability, the diagnostic delay
observational real-​world studies is providing insights between symptom onset and diagnosis, cognition, mor­
into predictors of MS treatment response, the compar­ tality and QOL295. Additionally, the presence of such
ative effectiveness of DMTs and long-​term treatment comorbidities is likely to affect treatment decisions and
effectiveness, therefore assisting in decision-​making treatment response.
for individuals in daily clinical practice284. For example, A perceived knowledge gap in the estimates of the
an analysis of data entered in the MSBase, a global MS incidence and effects of such comorbidities is apparent,
cohort study including thousands of patients with and recommendations have been formulated by experts
MS285, has enabled the identification of demographic, to address this gap296. The prevalence of physical comor­
clinical and paraclinical information that is helpful in bidities in patients with MS, including diabetes mellitus,
predicting individual response to DMTs at the time of hypertension, hyperlipidaemia, ischaemic heart disease,
their commencement. fibromyalgia and irritable bowel syndrome, increases
Regarding the monitoring of disease activity after with age, whereas a similar increase is not seen for psy­
starting a given treatment, it is now accepted that dis­ chiatric comorbidities (for example, depression and anx­
e­ase activity should be judged using combined models iety)295. Patients with comorbidities are usually excluded
based on the integration of clinical and paraclinical from clinical trials297, thereby limiting understanding of
information. No evidence of disease activity (NEDA), the treatment effects (and adverse effects) in patients
a combined measure including MRI data (such as no with MS who have concomitant comorbidities. A retro­
new T2-hyperintense or gadolinium-​enhancing lesions) spective population study found a reduced use of DMTs
and clinical measures (the absence of relapses or dis­ with increasing number of comorbidities298.
ability progression), has been proposed for a more com­ Notably, some of the measures that are usually
prehensive evaluation of treatment effects286. NEDA applied to monitor MS can be modified by the presence
has been assessed in clinical trials and observational of comorbidities. For example, this is the case for the
studies287,288, which have suggested that, even though a extent of T2-hyperintense lesions, whose number and
substantial proportion of patients maintained NEDA burden are influenced by hypertension and vascular dis­
status during the first 1–2 years of treatment, only a few ease, or brain atrophy, which is accelerated by smoking.
patients sustained NEDA status over several years with Consequently, incorporating the assessment, preven­
newer DMTs289,290. tion and management of comorbidities into the care of
The current definition of NEDA (NEDA-3) is patients with MS is required213.
strongly weighted towards inflammatory activity,
whereas other relevant aspects of the disease, such as Published online xx xx xxxx

1. Lublin, F. D. et al. Defining the clinical course of 3. Greer, J. M. & McCombe, P. A. Role of gender in 6. Gustavsson, A. et al. Cost of disorders of the brain
multiple sclerosis: the 2013 revisions. Neurology 83, multiple sclerosis: clinical effects and potential in Europe 2010. Eur. Neuropsychopharmacol. 21,
278–286 (2014). molecular mechanisms. J. Neuroimmunol. 234, 7–18 718–779 (2011).
This is a proposal for a redefinition of MS clinical (2011). 7. Chen, A. Y., Chonghasawat, A. O. & Leadholm, K. L.
courses on the basis of the inclusion of disease 4. Yeshokumar, A. K., Narula, S. & Banwell, B. Pediatric Multiple sclerosis: frequency, cost, and economic
activity (considering clinical relapse rate and multiple sclerosis. Curr. Opin. Neurol. 30, 216–221 burden in the United States. J. Clin. Neurosci. 45,
imaging findings) and disease progression. (2017). 180–186 (2017).
2. Krieger, S. C., Cook, K., De Nino, S. & Fletcher, M. 5. Multiple Sclerosis International Federation. 8. Rosati, G. The prevalence of multiple sclerosis in the
The topographical model of multiple sclerosis: Atlas of MS 2013: mapping multiple sclerosis world: an update. Neurol. Sci. 22, 117–139 (2001).
a dynamic visualization of disease course. around the world. MSIF.org https://www.msif.org/ 9. Koch-​Henriksen, N. & Sorensen, P. S. The changing
Neurol. Neuroimmunol. Neuroinflamm. 3, e279 wp-​content/uploads/2014/09/Atlas-​of-MS.pdf demographic pattern of multiple sclerosis
(2016). (2013). epidemiology. Lancet Neurol. 9, 520–532 (2010).

22 | Article citation ID: (2018) 4:43 www.nature.com/nrdp

0123456789();

You might also like