925369 review-article20202020 MSJ0010.1177/1352458520925369Multiple Sclerosis JournalE Iacobaeus, G Arrambide MULTIPLE SCLEROSIS MSJ JOURNAL Meeting Review Aggressive multiple sclerosis (1): Towards a Multiple Sclerosis Journal 2020, Vol. 26(9) 1031­–1044 definition of the phenotype DOI: 10.1177/ https://doi.org/10.1177/1352458520925369 1352458520925369 https://doi.org/10.1177/1352458520925369 © The Author(s), 2020. Ellen Iacobaeus, Georgina Arrambide , Maria Pia Amato, Tobias Derfuss , Sandra Vukusic , Bernhard Hemmer , Mar Tintore Article reuse guidelines: sagepub.com/journals- and Lou Brundin; for the 2018 ECTRIMS Focused Workshop Group*. permissions Abstract: While the major phenotypes of multiple sclerosis (MS) and relapsing–remitting, primary and Correspondence to: L Brundin secondary progressive MS have been well characterized, a subgroup of patients with an active, aggres- Department of Clinical Neuroscience, Karolinska sive disease course and rapid disability accumulation remains difficult to define and there is no consensus Institute, Bioclinicum, about their management and treatment. The current lack of an accepted definition and treatment guide- Karolinska University Hospital, Solnavägen 31, lines for aggressive MS triggered a 2018 focused workshop of the European Committee for Treatment Stockholm 171 64, Sweden. and Research in Multiple Sclerosis (ECTRIMS) on aggressive MS. The aim of the workshop was to dis- lou.brundin@ki.se Ellen Iacobaeus cuss approaches on how to describe and define the disease phenotype and its treatments. Unfortunately, Lou Brundin it was not possible to come to consensus on a definition because of unavailable data correlating severe Department of Clinical Neuroscience, Division disease with imaging and molecular biomarkers. However, the workshop highlighted the need for future of Neurology, Karolinska Institute and Karolinska research needed to define this disease subtype while also focusing on its treatment and management. University Hospital, Here, we review previous attempts to define aggressive MS and present characteristics that might, with Stockholm, Sweden additional research, eventually help characterize it. A companion paper summarizes data regarding treat- Georgina Arrambide Mar Tintore ment and management. Servei de Neurologia- Neuroimmunologia. Centre d’Esclerosi Múltiple de Catalunya, (Cemcat), Vall Keywords: Aggressive, highly active, multiple sclerosis, disability, relapsing/remitting, observational d’Hebron Institut de Recerca, studies Hospital Universitari Vall d’Hebron, Universitat Autònoma de Barcelona, Date received: 13 November 2019; revised: 6 April 2020; accepted: 16 April 2020. Barcelona, Spain Maria Pia Amato Department NeuroFarBa, Introduction treatments. This strategy is debatable in patients with University of Florence, Florence, Italy/IRCCS Multiple sclerosis (MS) is a serious, lifelong disa- a more aggressive form of disease, in which the suc- Fondazione Don Carlo bling disease with unpredictable outcomes. The clini- cessful therapeutic window of opportunity may be Gnocchi, Florence, Italy cal presentation and phenotype of MS are variable narrower than for those with less aggressive disease.3 Tobias Derfuss Departments of Neurology between patients and over time. It can encompass Patients with aggressive MS may be better treated if and Biomedicine, University various degrees of severity. A minority of patients given rapid access to therapies considered to be more Hospital Basel, University of Basel, Basel, Switzerland exhibit either a benign course with little disability effective at slowing or preventing disability accrual. Sandra Vukusic accrual over time or an ‘aggressive’ course with fre- This approach would decrease the probabilities of Service de neurologie, quent, severe relapses, incomplete recovery and rap- breakthrough disease with first-line DMTs. Therefore, Sclérose en plaques, Pathologies de la myéline idly accumulating and permanent disability.1 early identification of subjects with aggressive dis- et neuro-inflammation, and ease at onset or with disease that becomes more Centre de Référence des Maladies Inflammatoires The efficacy of available disease-modifying therapies aggressive over time is essential to both aid treatment Rares du Cerveau et (DMTs) for relapsing multiple sclerosis (rMS) and recommendations and decisions in the clinic. In addi- de la Moelle, Hôpital Neurologique Pierre progressive (primary progressive MS (PPMS) and tion, having a unified definition of aggressive MS Wertheimer, Hospices Civils secondary progressive MS (SPMS)) have been insuf- could be used as a tool to enrich enrolment and strat- de Lyon, Lyon/Bron, France; Centre des Neurosciences de ficiently studied in aggressive disease courses.2 For ify subjects for prospective clinical trials for such dis- Lyon, Observatoire Français some available DMTs, regulatory approval and insur- ease phenotype.4 In the current report, attempts to de la Sclérose en Plaques, INSERM 1028 et CNRS ance coverage mandate that patients must be consid- define aggressive/active MS are reviewed. The group UMR5292, Lyon, France; ered not responsive (based on ongoing clinical or was unable to come to consensus about a new, more Université Claude Bernard Lyon 1, Faculté de médecine radiological disease activity) or intolerant to first-line data-driven definition. However, we provide a Lyon Est, Lyon, France therapies before receiving access to more effective detailed description of clinical and paraclinical journals.sagepub.com/home/msj 1031 Multiple Sclerosis Journal 26(9) Bernhard Hemmer Department of Neurology, factors that are relevant to substantiate a definition, The term ‘ever malignant’ was used to define MS Klinikum rechts der Isar, such as imaging, neuropathological and immunologi- patients who reached an Expanded Disability Status School of Medicine, Technical University of cal findings, as well as biomarker and genetic factors Scale (EDSS) of 6.0 within 5 years, which corre- Munich, Munich, Germany; that should be studied to determine possible associa- sponded to 12.1% of a cohort of 478 MS patients.6 Munich Cluster for Systems Neurology (SyNergy), tions with the clinical phenotype of aggressive MS. ‘Ever malignant’ cases had significantly more Munich, Germany relapses, more motor symptoms and a higher fre- * Participants to the quency of a progressive onset.6 In a larger study, 4%– ECTRIMS-Focused Workshop on Aggressive MS Materials and methods 14% of subjects fulfilled the criteria for ‘aggressive are listed in Supplemental To explore aggressive MS, its definition and treat- MS’.7,8 Here, aggressive MS constituted one of the Table 1. ment, the European Committee for Treatment and three forms: form 1 included patients who reached an Research in Multiple Sclerosis (ECTRIMS) held a EDSS of 6 within 5 years, form 2 included those with focused workshop on aggressive MS in March 2018 EDSS ⩾6 at the age of 40 years and form 3 included in Brussels, Belgium, including 50 European, North those who entered SPMS within 3 years after rMS and South American, Asian and Australian neurolo- disease onset.7,8 Additional criteria for aggressive gists and other professionals with interest in the onset MS have been described from retrospective understudied subgroup of MS patients with aggres- assessment as (1) two or more relapses in the year sive disease. Presentations on the clinical, paraclinical after onset and two or more gadolinium-enhancing and biological characteristics of aggressive MS were (Gd+) lesions on brain magnetic resonance imaging followed by discussions about current treatment strat- (MRI) or (2) one relapse within a year after onset if it egies and recommendations for future research that results in sustained baseline EDSS score of 3 along will allow improved definitions of aggressive MS and with two or more Gd+ lesions.9 Based on these crite- more effective and tailored treatment of affected ria, 58 (7.3%) of 783 MS patients were considered to patients. To support information presented at the have aggressive onset MS.9 workshop and data which appeared after, while this manuscript was being prepared, we conducted litera- Other definitions of aggressive MS have been proposed ture searches in the English language using PubMed, when considering treatment options. One of the rare applying the following search terms: ‘multiple sclero- clinical trials aimed at highly active disease, testing sis’ OR ‘MS’ AND ‘aggressive’ and ‘multiple sclero- mitoxantrone induction therapy prior to interferon sis’ OR ‘MS’ AND ‘highly active’. (IFN) β1b, defined aggressive MS as ⩾2 relapses or an EDSS increase ⩾2 points in the 12 preceding months, ⩾1 Gd-enhancing lesion and baseline EDSS between Results 2.5 and 5.0.11 Clinicians faced with identifying patients with breakthrough disease eligible for treatment with Past and recent efforts to define aggressive/highly autologous haematopoetic stem cell transplantation active/malignant MS (aHSCT) have characterized such individuals as hav- Previous efforts to characterize severe or aggressive ing ‘highly active MS’, defined as subjects with ‘fail- MS focused on the group of patients with highly ure of at least one and up to three DMTs evidenced by active disease (frequent and severe relapses, rapid ongoing or increased clinical and MRI activity’.12 A worsening) and high inflammatory and neurodegen- recent review on treatment algorithms for aggressive erative activity, generally those with poor prognosis MS similarly added ‘unresponsiveness to first and/or and outcomes over relatively short periods of time second line treatment’, as a defining feature to be con- (Table 1). These have been variously categorized as sidered along with clinical and MRI parameters.10 those with ‘malignant MS’,5 ‘aggressive MS’10 or ‘highly active MS’.15 ‘Malignant MS’ was defined in In spite of these reports with various suggestions for 1996 as a ‘disease with a rapid progressive course, the definition of severe or aggressive MS, there is no leading to significant disability in multiple neurologic consensus about which, if any, of these definitions is systems or death in a relatively short time after dis- most appropriate or most useful in clinical practice or ease onset’.5 However, the authors of this statement in a research setting. All are hampered by the need for did not specify how to quantify any of these criteria, either a retrospective assessment of the disease course thus leaving the definition somewhat vague. In the or a prolonged prospective evaluation until fulfilling following years, they advised that a ‘malignant’ (as the proposed criteria and thus hinder the ability to make well as a ‘benign’) label may be misleading, vary over rapid and efficient treatment decisions. Efforts to iden- time and only be determined retrospectively, conclud- tify early risk factors for disease severity are ongoing in ing that such labels should be used with caution.1 inception cohorts and registries. Two studies, reported 1032 journals.sagepub.com/home/msj E Iacobaeus, G Arrambide et al. Table 1. Historical and contemporary attempts at defining aggressive/highly active multiple sclerosis. Definition Citation Author Malignant MS “A disease with a rapid progressive course, leading to significant disability Lublin et al.5 in multiple neurologic systems or death in a relatively short time after disease onset” Ever malignant “Patients that reached an EDSS of 6.0 within 5 years from onset.” Gholipor et al.6 MS Aggressive MS “Patients that reached an EDSS of 6.0 within 5 years from onset” (Form 1). Menon S et al.7 “Patients that reached an EDSS ⩾ 6.0 at the age of 40 years” (Form 2) Menon S et al.8 “Patients who entered SPMS phase within 3 years after rMS onset” (Form 3) Aggressive onset “MS patients with (a) ⩾2 relapses in the year after onset and ⩾2 Gd+ Kaunzner et al.9 MS lesions on brain MRI scan or (b) one relapse within 1 year after onset if it results in sustained baseline EDSS score of 3.0 along with ⩾2 Gd+ lesions”. Aggressive MS “rMS with one or more of the following features: (a) EDSS score of 4.0 Rush et al.10 within 5 years of onset. (b) Multiple (⩾2) relapses with incomplete resolution in the past year. (c) ⩾2 MRI scans showing new or enlarging T2 lesions or Gd+ lesions despite treatment. (d) No response to therapy with one or more DMTs for up to 1 year”. Aggressive ‘⩾2 relapses or an EDSS increase ⩾2 points in the 12 preceding months, Edan G et al.11 relapsing– ⩾1 Gd-enhancing lesion and baseline EDSS between 2.5 and 5.0’. remitting MS (ARMS) Highly active MS ‘Failure of conventional treatment and ⩾1 severe relapses and/or incomplete Saccardi et al.12 recovery from clinically significant relapses and ⩾1 Gd+ lesion of diameter ⩾3 mm or accumulation of ⩾0.3 T2 lesions/month in two consecutive MRI 6–12 months apart’. Aggressive MS ‘Reaching an EDSS ⩾6.0 within 10 years of disease onset’ Tintore et al.13 Aggressive MS ‘Reaching an EDSS ⩾6.0 within 10 years of disease onset’ Malpas et al.14 MS: multiple sclerosis; EDSS: Expanded Disability Status Scale; SPMS: secondary progressive multiple sclerosis; rMS: relapsing multiple sclerosis; MRI: magnetic resonance imaging; Gd+: gadolinium-enhancing lesions; DMT: disease-modifying treatment. since the conclusion of the workshop, defined aggres- years,16,27,28 short inter-attack interval21,16 and early sive MS as reaching an EDSS ⩾6.0 within 10 years of accrual of disability with superimposed attacks.16 disease onset. In both cohorts, approximately 5.0–6.0% Severe MS relapses have been more frequently of patients fulfilled this definition.13,14 Their findings reported among young persons with MS65 although are described in the clinical and MRI subsections in the older age at MS onset is associated with higher risk of following. disability worsening.24,17 After the Workshop, Malpas et al defined aggressive MS as reaching an EDSS ⩾6.0 within 10 years of disease onset. Indicators of an Characteristics of aggressive MS aggressive disease course included age >35 years at Clinical features. Rapid progression of disability is a symptom onset, EDSS ⩾3.0 in the first year and pres- commonly cited characteristic of aggressive MS. ence of pyramidal signs in the first year of disease Clinical manifestations that have been associated with evolution. Using the MSBase registry, patients with an accelerated worsening of disability in MS are sum- all three features had a 32.0% risk of fulfilling the marized in Table 2A, including (a) severe relapse(s) proposed aggressive MS definition.14 Some of these leading to an increase by one EDSS point or greater findings, such as older age and motor symptoms at than or equal to two points in any functional system disease onset, are in accordance with previous (FS) and/or the need for hospitalization and/or steroid attempts to define aggressive MS in retrospective therapy, (b) multifocal attacks,18,19 (c) incomplete analyses of prospectively acquired data.6,7 remission20,21 and (d) attacks affecting motor, cogni- tive, sphincter or cerebellar systems.14,18,22–26 Addi- MRI findings. MRI findings have played a central role tional clinical factors are frequent relapses in the first in diagnosis, individual evaluation in clinical practice journals.sagepub.com/home/msj 1033 Multiple Sclerosis Journal 26(9) Table 2. Potential parameters associated with more aggressive disease course (any one or a combination of these characteristics may signal aggressive disease and all are deserving of further exploration and if possible, validation in the context of assessing severity of disease and poor prognosis). A. Clinical features and demographics Male sex16,17 Older age at onset (>35 or >40 years)14,17,18 Severe relapses   ⩾1 point increase in EDSS or ⩾2 points in any functional system    Need of hospitalization and/or steroid therapy   Multifocal18,19   Incomplete remission20,21    Affecting motor, cerebellar, cognition, or sphincter functions18,22–26    Presence of pyramidal signs in the first year of disease evolution14 Frequent relapses in the first 5 years27,28 Short interval between relapses16,21 Early accrual of disability with superimposed attacks16 EDSS ⩾3.0 in the first year of disease evolution14 B. MRI findings High T2 lesion burden13,29,30 ⩾20 T2 lesions at the time of disease onset13 ⩾2 Gd+ lesions10,31 ⩾2 Gd+ lesions at the time of disease onset13 T1 black holes29,32 Infratentorial lesions30,33 Spinal cord lesion and/or atrophy30,34,35 Early discernible atrophy36 Cortical and deep grey matter atrophy32,37,38 Smouldering lesions39 C. Neuropathological features Transections of axons during inflammatory demyelination40 Chronic demyelination causing axonal degeneration41 Cortical demyelination with neuritic transection and neuronal death42 Synaptic spine loss43 D. Immunological features in the blood Shift in the balance between Tregs and proinflammatory T-cells44–47 Lower percentage of CD4+ Foxp3 Tregs, memory CD4 and CD8 cells and CD56high NK Cells44,46,47 Increased Th17 cell responses48 Low TCR diversity49 E. Biomarker correlates Detection of CSF IgM OCB50,51 Elevated IgG Index48 Elevated levels of NfL in CSF or serum52–55 CSF CXCL1356,57 CSF CHI3L158 CSF MMP959 F. Genetic markers HLA-DRB1*15:0160 PD-161 MGAT562 APOE-epsilon 463 BDNF64 MS: multiple sclerosis; EDSS: Expanded Disability Status Scale; Gd+: gadolinium enhancing lesions; TCR: T-cell receptor; CSF: cerebrospinal fluid; OCB: oligoclonal bands; NfL: neurofilament light chain; CHI3L1: chitinase 3-like 1 protein; MMP9: metalloma- trix protein 9. and as a research tool. It can be used to assess disease The presence of a high T2 lesion load, >2 Gd + lesions, severity and predict poor prognosis. Several imaging lesion topography, T1 black holes and early discern- biomarkers have been studied, although not necessar- ible atrophy are well recognized early predictors for ily focusing on aggressive MS (Table 2B, Figure 1). disease progression.13,29–34,68–70 A study of over 1034 journals.sagepub.com/home/msj E Iacobaeus, G Arrambide et al. Table 3. A selection of biomarkers in CNS and periphery and their association with prognosis. Biomarker Authors n Duration of MRI data available Major findings follow-up (months) Intrathecal IgG Gasperi et al. 673 48 Yes Intrathecal IgG synthesis (MS)48 is associated with worse prognosis Lipid-specific IgM- Margraner et al. 24 48 Yes Lipid-specific IgM-OCB are OCB (CIS)50 associated with worse MRI measures Lipid-specific IgM- Trangarajh et al. 81 >120 NA Lipid-specific IgM-OCB are OCB (MS)51 associated with worse long- term clinical progression CSF-NfL Salzer et al. 95 168 NA Initial CSF-NfL levels (MS)52 relate to long-term clinical disability CSF-NfL Hakansson et al. 41 48 Yes CSF-NfL levels are (MS)53 predictor of aNEDA-3 status after 4 years s-NfL Chitnis et al. 122 120 Yes Early s-NfL levels correlate (MS)54 with T2LV, BPF and fatigue but not EDSS at 10 years s-NfL Barro C et al. 259 78 Yes s-NfL levels correlate with (MS)55 relapse, brain/spinal cord atrophy and clinical severity CSF-CXCL12 and Krumbholz et al. 52 NA No MRI CSF-CXCL12 and 13 are CSF-CXCL13 (MS)66 Neuropathological elevated in MS CSF and in findings MS lesions CSF-CXCL13 Brettschneider 91 48 Yes CSF-CXCL13 levels predict et al. (CIS and conversion to MS MS)67 CSF-CXCL13 Khademi et al. 466 60 Yes CSF-CXCL13 levels predict (CIS/MS)54 MS, relapse rate and new lesions CSF-CXCL13 Puthenparampil 40 6-12 Yes CSF-CXCL13 is associated et al. (MS)55 with cortical thinning and high CSF leukocyte count CSF-CH13L1 Canto et al. 813 120 Yes CSF-CHI3L1 levels predict (CIS)67 Neuropathological conversion to MS and findings increase in disability n: number of patients; MS: multiple sclerosis; CIS: clinical isolated syndrome characteristic of MS; MRI: magnetic resonance imaging; NA: not applicable; OCB: oligoclonal bands; CSF: cerebrospinal fluid; s-NFL: serum neurofilament light chain; T2LV: T2 hyperin- tense lesion volume; BPF: brain parenchymal fraction; CXCL: C-X-C motif chemokine ligand; CH13L1: chitinase-3-like protein. aPatients with no relapses, no brain MRI activity (no new or enlarging T2 lesions or gadolinium-enhancing lesions) and no sustained disability worsening (EDSS progression) during follow-up were classified as showing no evidence of disease activity-3 (NEDA-3). 1000 patients with clinically isolated syndromes (CIS) Gd+ lesions on the brain MRI performed at the time of characteristic of MS showed that a high brain T2 the first attack. When both MRI conditions were ful- lesion load at baseline was the most robust predictive filled, the risk increased to almost 20.0%.13 The devel- factor for future disability accrual.68 Tintore et al.13 opment of >1 Gd+ lesion and new T2 lesions was used a Barcelona-based inception cohort and, like Mal- associated with poorer prognosis at follow-up.71 In pas et al.,14 defined aggressive MS as reaching an EDSS addition, the positive correlation between MRI event- ⩾6.0 within 10 years of disease onset. Early radiologi- free survival and progression-free survival after cal biomarkers associated with this outcome were the aHSCT in MS further supports the predictive value of presence of ⩾20 lesions on T2-weighted images or ⩾2 new MRI lesions for disease progression.72,73 journals.sagepub.com/home/msj 1035 Multiple Sclerosis Journal 26(9) Figure 1. Typical characteristics of aggressive MS on conventional MRI: (a)–(c) T2-FLAIR transverse sequences showing multiple nodular, confluent lesions predominantly affecting the periventricular and callosal topographies (see arrows) in a 33-year-old male patient with a first demyelinating attack. (d)–(f) T1 transverse sequences in the same patient after gadolinium administration show multiple nodular and ring-enhancing lesions. (g) (T2-FLAIR transverse sequence) and (h) (T2 sagittal sequence of the cervical spinal cord) show multiple infratentorial and spinal cord lesions in a 24-year-old female patient with relapsing–remitting multiple sclerosis, several of which showed contrast enhancement despite treatment with (i) and (j) fingolimod. The presence of ⩾1 spinal cord lesion, symptomatic meta-analysis also showed that spinal cord atrophy or asymptomatic, has been reported as a marker of significantly correlated with clinical disability.35 In a early and long-term disability accrual.30,34 A study of non-spinal CIS patients, spinal cord factors 1036 journals.sagepub.com/home/msj E Iacobaeus, G Arrambide et al. such as lesion number, change in lesion number and macrophage infiltration within active lesions corre- change in upper cervical cord cross-sectional area lates with axonal transection supports the role of were independently associated with reaching an inflammatory mechanisms in axonal loss.40 Axonal EDSS ⩾3.0 at 5 years. Inclusion of brain lesion load loss within chronic demyelinating lesions is thought and atrophy modestly increased the predictive power to be a consequence of loss of myelin trophic support, of the model,70 suggesting spinal cord red flags may rendering denuded axons vulnerable to the chronic be more relevant than brain red flags to predict disa- inflammatory environment with CD8+ cells, microg- bility, at least as measured by the EDSS. In another lia and oxidative stress factors.41 study of MS patients, the results suggested that the presence of spinal cord lesions may indicate a higher Additional factors contributing to axonal loss occur in risk of 6- and 11-year EDSS progression than infraten- the normal appearing white matter (NAWM) involv- torial lesions.74 In addition, the presence of ⩾2 Gd+ ing secondary Wallerian degeneration of injured lesions and ⩾1 spinal cord lesions at baseline and 1 axons in adjacent white matter plaques.79 Cortical year and ⩾1 new spinal cord lesions at 3 years of dis- lesions have been associated with prominent neuronal ease evolution were associated with conversion to apoptosis in addition to axonal and dendritic transec- SPMS at 15 years.30 Furthermore, the occurrence of tion, and the presence of such lesions has been sug- cortical lesions and cortical atrophy has been reported gested as a potential biomarker for risk of early as risk factors for early neurological deficit in MS.37 disability development.80 When assessing tissue loss, the accumulation of hypoin- Taken together, extensive widespread neuroaxonal tense lesions ( ‘black holes’) on T1 sequences corre- damage and loss in MS may be viewed as the neuro- lated with progression rate.29 As for early brain atrophy, pathologic counterpart to clinically aggressive MS, a study in CIS patients showed that a percentage brain but this needs to be validated in clinicopathological volume change (PBVC) decrease below −0.817% in the correlated studies. In clinical practice, ongoing axonal first year of disease evolution was an independent pre- loss may be the underlying pathology in patients with dictor of a shorter time to a second attack,75 which could sustained progression despite treatment with one or be an indicator of aggressive MS. Evidence of a positive more DMTs. correlation between the disability level and atrophy of the thalamus and basal ganglia has been supported in Peripheral immune correlates of highly active several studies.76,77 A significant association between MS. Alterations in the peripheral immune system deep grey matter brain atrophy rate and time to disabil- associated with highly active MS have been poorly ity progression was detected in a cohort of more than studied. However, the observation of sustained sup- 1200 MS patients.36 Interestingly, another study showed pression of active neuroinflammation following that the presence of isolated thalamic atrophy without aHSCT in patients with aggressive/highly active MS whole brain atrophy distinguished patients with a higher has provided insights into immune mechanisms that risk of fast disability progression.38 might be associated with a highly inflammatory clini- cal phenotype in MS81 (Table 2D). A significant shift More recently, the presence of smouldering lesions has in the balance between regulatory and proinflamma- been suggested as a new MRI marker of poor outcome tory cells in the circulation has been detected in in MS patients since it reflects persistent inflammation, patients with active MS that underwent successful delayed tissue repair and neurodegeneration.39 aHSCT.44–47 Proportionally increased levels of CD4 + Foxp3 + regulatory T-cells (Tregs), memory Neuropathological aspects of highly active MS. A CD4 + and CD8 + cells, CD56high natural killer (NK) major contributor to the development of permanent cells and diminished Th17 cell responses have been disability in MS is axonal loss, which is an ongoing observed during peripheral immune reconstitution process from disease onset.78 This suggests that there after aHSCT in MS.44–47 In addition, successful out- should be a strong correlation between rapid disabil- comes after aHSCT in MS patients with ‘a poor prog- ity accumulation and a high rate of axonal damage nosis’ correlated with more diverse T-cell receptor and loss (Table 2). Key mechanisms resulting in neu- (TCR) repertoires of CD4 + cells suggesting that low ronal dysfunction include (1) transection of axons by TCR diversity may be a risk factor for a highly inflam- an inflammatory attack, (2) chronic demyelination matory disease course.49 Altogether, these findings causing axonal degeneration, (3) cortical demyelin- indicate a possible association between dysregulated ation causing neuritic transection and neuronal death subgroups of T cells and NK cells with highly active and (4) synaptic spine loss.40,42,43 The observation that MS. Such data, not specifically gathered to define journals.sagepub.com/home/msj 1037 Multiple Sclerosis Journal 26(9) aggressive MS, deserve further exploration and vali- CXCL13 levels were associated with cortical atrophy dation as potential immune markers for the disease. in MS suggesting its value as a marker to predict dis- ability progression.57 CSF and blood biomarkers with prognostic values (Table 2E and Table 3). Intrathecal IgG synthesis at Since presence of cortical lesions may be considered MS onset was associated with increased risk of dis- as a risk factor for early MS disability development,32 ability worsening within 4 years although intrathecal there have been attempts to develop surrogate non- IgG synthesis is present in 60% of MS patients inde- imaging biomarkers for assessment of cortical dam- pendent of disease severity or prognosis.48 Further- age. Increased levels of a specific panel of more, the presence of IgM OCB correlated with proinflammatory cytokines and chemokines includ- higher disease activity, EDSS progression and brain ing TNF, sTNFR1, IFN γ, CXCL12, CXCL13, IL6, atrophy in rMS.50,51 IL8, IL10, BAFF, APRIL, LIGHT, TWEAK, MMP-2, pentraxin3 and sCD163 distinguished patients with Evaluation of neurofilament light chain (NfL) levels high cortical lesion load at diagnosis compared to in both CSF (cNfL) and serum (sNfL) has shown patients with low cortical lesion load.83 strong promise as a marker for disease activity and disability correlated with neuroaxonal damage in MS Other potential CSF biomarkers that reflect inflam- and has been implemented as part of routine clinical matory activity and may have prognostic value assessment in some countries. To date, however, a include proteolytic enzyme matrix metalloprotein- correlation of NfL levels with disease severity has ase-9 (MMP9), osteopontin and chitinase 3-like 1 not been studied. Levels of cNfL were associated protein (CHI3L1). MMP9 is considered to have a cen- with both new T2 lesions and brain volume loss dur- tral role in MS pathogenesis with functions relating to ing follow-up; cNfL at MS diagnosis correlated with CNS tissue destruction. The finding that progressive Multiple Sclerosis Severity Scale (MSSS) estimates MS with inflammatory activity had increased levels during a follow-up period of 8–20 years.52,53 In a of CSF MMP9 suggests its potential as a marker to related finding, sNfL levels correlated with risk of distinguish progressive patients that could benefit future relapses,53 disability worsening and brain T2 from anti-inflammatory therapies.59 CHI3L1 corre- lesion load.54,55 Increased sNfL was also found to be lated with disease severity in several inflammatory a negative predictor for MS, correlating with brain diseases and was an independent risk factor for devel- atrophy and disability development in early MS.54,55 opment of EDSS 3 in a multivariate study on more Some studies have attempted to establish predictive than 800 European CIS patients.58 Importantly, none cut-off values of NfL levels. Patients with cNfL lev- of these biomarkers have yet been established as reli- els >386 ng/L had a higher risk of conversion to able markers to predict aggressive MS and should SPMS than patients with lower levels,52 although this currently only be viewed as indicators with potential cut-off has not been fully validated. Another study prognostic values and as a fertile area for research and assessed stratified cNfl and sNfL measurements in validation. rMS patients during the first years of participation in a clinical trial on IFN-β versus placebo. Upper tertile Genetic markers of MS disease activity. MS risk year 2 cNfl and year 3 sNfL levels were predictive of genes have been investigated for their influence on reaching an EDSS ⩾6.0 at year 8 from inclusion MS disease outcome but the results have been con- compared to lower tertile levels. Similarly, upper ter- flicting63,84 (Table 2F). The most dominant MS risk tile year 4 sNfL concentration was predictive of gene, HLA-DRB1*15:01, correlated with develop- reaching an EDSS score ⩾6.0 at 15 years, with NfL ment of increased number of T2 MRI lesions and loss concentrations being complementary to 2-year brain of brain volume in MS.60 A German study demon- parenchymal fraction change in predicting long-term strated that PD-1 polymorphisms were associated outcomes.82 with a progressive disease course, which was consid- ered to be a consequence of impaired PD-1-mediated Evaluation of C-X-C motif chemokine ligand 13 inhibition of T cell activation.61 A genome-wide asso- (CXCL13), also called B-cell attracting chemokine 1 ciation study detected variants of the glycosylation (BCA-1), has recently been used to assess inflamma- enzyme MGAT5 that influenced the severity of MS62 tory activity in MS. The chemokine CXCL13 is con- and carriers of the APOE-epsilon 4 allele had a more sidered to reflect B-cell activity in the CNS, and severe disease progression.63 increased CSF CXCL13 levels were found to corre- late with conversion from CIS to clinically definite Non-immune-related genes such as those involved in MS and with relapse rates.56,66,67 Furthermore, CSF neuroprotective or regenerative mechanisms have 1038 journals.sagepub.com/home/msj E Iacobaeus, G Arrambide et al. emerged as possible risk factors for MS disease out- EDSS ⩾6.0 within 10 years of disease onset. Clinical come heterogeneity.85,86 Genetically determined risk factors of individuals who might be considered to axonal response to inflammation and environmental have aggressive MS are age >35 years at symptom factors might be associated with clinical outcome and onset, EDSS ⩾3.0 in the first year and presence of disability progression in MS. An intriguing example pyramidal signs in the first year of disease evolution.14 was the detection that allelic variants of the neuropro- Older age and motor symptoms at disease onset were tective ciliary neurotrophic factor (CNTF) were asso- also identified as risk factors in previous attempts to ciated with disease onset, course and severity of define aggressive MS.6,7 MRI risk factors include the experimental autoimmune encephalomyelitis (EAE) presence of ⩾20 T2 lesions or ⩾2 Gd+ lesions on the in mice.87 Furthermore, polymorphisms of the brain- brain MRI performed at disease onset.13 When pre- derived neurotrophic factor (BDNF) have in some sent, these factors may serve as ‘red flags’ to define a studies been shown to impact on grey matter tissue patient as a candidate for early, more aggressive thera- damage in MS.64 However, current data on gene vari- peutic intervention. This suggestion must be consid- ants associated with aggressive MS are very sparse ered a working definition that needs to be solidified and controversial; additional genetic studies may through future research, requiring validation in other determine whether the clinical cohort of aggressive cohorts to confirm or refute their short- and long-term MS can be genetically distinguished from other dis- value in the clinical practice. Conversely, we also have ease phenotypes. to acknowledge that several parameters potentially associated with aggressive MS (Table 2) may be use- ful only for research purposes to understand this phe- Achievements, gaps in knowledge and future notype (i.e. neuroaxonal damage). In this sense, it will perspectives be important to identify those with translational Clinical, paraclinical and laboratory assessments all potential. have the potential to be used in categorizing MS patients with a prognosis of poor outcomes over a As noted in the manuscript, the list of current ‘red short period of time, although most have not been flags’ for aggressive disease is far from complete. evaluated with this goal in mind. Disease severity and Notably missing are immunological assays, advanced activity are essential parameters when considering the imaging, pathological, genetic, and many potential right treatment for patients.88 While guidelines exist body fluid biomarkers mentioned above. Identifying for treatment of rMS, PPMS and SPMS, specific their differential contribution of relapse activity and guidance for how to approach treatment for severe, disease progression is difficult, as disability may occur rapidly aggressive MS are lacking.88 To develop such due to the consequences of inflammation and/or pure guidelines, and to better manage those patients with neurodegenerative mechanisms. Rather, it is important aggressive disease, it is important to clearly identify to study the degree to which inflammation and neuro- subjects with initial or evolving signs of aggressive degeneration factors (and their possible interactions) disease course as early as possible and offer them a are involved in the etiopathogenesis of aggressive MS. highly efficient treatment as soon as possible after This may aid in defining not only aggressive MS in disease onset to mitigate breakthrough disease and relapsing forms but also in progressive phenotypes in progression of permanent disability. which available evidence is even scarcer. Based on the discussion held during the ECTRIMS Importantly, whereas recent studies are based on find- workshop on aggressive MS and on recent results ings at disease onset or during the first year of the from retrospective studies with prospectively col- disease,13,14 potential risk factors should be assessed lected data, we believe that several identified clinical throughout the disease evolution. MS may become and imaging features, possibly along with elevated highly active after years of stability in some patients, CSF and/or serum NfL levels, might be considered as and persons with a high functional reserve may take the most promising characteristics of highly active longer to develop disability.89 MS (Table 4). In addition, as the proportion of patients with charac- A limitation of this proposal is that several of these teristics of aggressive MS is low, collaborative studies features were not studied in the context of aggressive based on large observational registries and focused MS or weighted against other suspected risk factors. cohorts are necessary. These studies enable prognosti- The identification of several risk factors, however, cation that is both generalizable and comprehensive. depended on establishing a working definition of In addition, robust analytical approaches of the most aggressive MS which, in two studies, was reaching an relevant risk factors, like Bayesian analyses, are a journals.sagepub.com/home/msj 1039 Multiple Sclerosis Journal 26(9) Table 4. ‘Red flags’ that may indicate highly active MS. With additional research and validation, one or more of the following characteristics, along with physician judgement, may be considered a red flag for poor short- and long-term prognosis and thus could imply initiation of a highly effective therapy. Clinical • At onset   -  Older age6,7   - Motor symptoms6,20 • During early follow-up    - Poor recovery from the first two relapses20    - High relapse frequency (⩾3) within the first 2 years after MS onset27    - Any signs of breakthrough disease within 2 years of treatment with a highly active DMT72,73 • Combined factors Age >35 years at symptom onset, EDSS ⩾3.0 in the first year, and presence of pyramidal signs in the first year of disease evolution (32.0% risk)14 Imaging • At onset   - ⩾2 Gd+ lesions10    - High T2 lesion load68,71    - Presence of spinal cord > infratentorial lesions33,34,74 • During early follow-up   - ⩾2 new Gd+ lesions30,71   - ⩾1 new spinal cord lesions30    - Atrophy measures, possibly spinal cord > brain; PBVC decrease below -0.817% in the first year37,70,75,76 Combined factors ⩾20 lesions on T2-weighted images and ⩾2 Gd+ lesions at disease onset (18.9% risk)13 Biomarkers • At the time of diagnosisa   - NfL CSF levels > 386 ng/L52 • During follow-up    - Upper tertile NfL levels in serum or CSF82 MS: multiple sclerosis; DMT: disease modifying treatment; EDSS: Expanded Disability Status Scale; Gd +: gadolinium enhancing; PBVC: percentage brain volume change; NfL: neurofilament light chain; CSF: cerebrospinal fluid. aIncludes relapsing and progressive forms of MS. requisite of a shift from studies of associations at the Declaration of Conflicting Interests group level towards individual prognosis. Afterwards, The author(s) declared the following potential conflicts replication of novel associations in independent vali- of interest with respect to the research, authorship and/ dation cohorts is essential. or publication of this article: E.I. has honoraria for advisory boards or lecturing for Sanofi Genzyme and While objective markers for MS disease severity are Merck outside the submitted work. G.A. has received desirable, one cannot ignore the value of clinical compensation for consulting services or participation judgement, independent of the presence or absence of in advisory boards from Sanofi, Merck and Roche; markers, especially in the context of patients being research support from Novartis; travel expenses for sci- followed over time by the same physician. entific meetings from Novartis, Roche and Stendhal; and speaking honoraria from Sanofi and Merck. M.P.A. Acknowledgements has received research grants and personal fees as a The authors wish to thank ECTRIMS for financial speaker and member of advisory boards by Biogen, and administrative support of the focused workshop Sanofi Genzyme, Merck Serono, Novartis, Roche and that served as the basis for this paper and all work- Teva, outside the submitted work. T.D. reports grants shop attendees (Supplemental Table 2) for their par- from Novartis and Biogen. Dr Derfuss’ institution ticipation. All attendees had the opportunity to review received financial support from Novartis, Biogen, and comment upon an advanced draft of the manu- Roche, Merck Serono, Celgene, Sanofi Genzyme, script prior to submission. We also wish to thank Dr MedDay, GeNeuro, Mitsubishi Pharma and Actelion Stephen C. Reingold for assistance with manuscript for his activities as steering committee or advisory writing and editing, supported by ECTRIMS. board member or consultant. S.V. has received grants, 1040 journals.sagepub.com/home/msj E Iacobaeus, G Arrambide et al. personal fees and non-financial support from Biogen, 3. Giovannoni G, Butzkueven H, Dhib-Jalbut S, et al. Celgene, Geneuro, Genzyme, Medday, Merck Serono, Brain health: Time matters in multiple sclerosis. Mult Novartis, Roche, Sanofi and Teva, outside the submit- Scler Relat Disord 2016; 9(Suppl. 1): S5–S48. ted work. B.H. has served on scientific advisory boards 4. Fisher E, Lee JC, Nakamura K, et al. Gray matter for Novartis and as DMSC member for AllergyCare atrophy in multiple sclerosis: A longitudinal study. and TG therapeutics; he or his institution have received Ann Neurol 2008; 64(3): 255–265. speaker honoraria from Desitin; holds part of two pat- 5. Lublin FD and Reingold SC. Defining the ents; one for the detection of antibodies against KIR4.1 clinical course of multiple sclerosis: Results of in a subpopulation of MS patients and one for genetic an international survey. Neurology 1996; 46(4): determinants of neutralizing antibodies to interferon. 907–911. M.T. has received compensation for consulting ser- 6. Gholipour T, Healy B, Baruch NF, et al. vices and speaking honoraria from Almirall, Bayer Demographic and clinical characteristics of Schering Pharma, Biogen-Idec, Genzyme, Merck malignant multiple sclerosis. Neurology 2011; 76(23): Serono, Novartis, Roche, Sanofi-Aventis and Teva 1996–2001. Pharmaceuticals. M.T. is co-editor of Multiple Sclerosis Journal ETC. L.B. has received honoraria for advisory 7. Menon S, Shirani A, Zhao Y, et al. Characterising boards for Biogen, Sanofi-Genzyme, Novartis and aggressive multiple sclerosis. J Neurol Neurosurg Psychiatry 2013; 84(11): 1192–1198. Teva and has received lecturing fees from Biogen, Novartis, Teva and Sanofi-Genzyme outside the sub- 8. Menon S, Zhu F, Shirani A, et al. Disability mitted work. progression in aggressive multiple sclerosis. Mult Scler 2017; 23(3): 456–463. Funding 9. Kaunzner UW, Kumar G, Askin G, et al. A study of The author(s) disclosed receipt of the following finan- patients with aggressive multiple sclerosis at cial support for the research, authorship and/or publi- disease onset. Neuropsychiatr Dis Treat 2016; 12: cation of this article: The workshop on which the 1907–1912. manuscript is based was supported in its entirety by 10. Rush CA, MacLean HJ and Freedman MS. the European Committee on Treatment and Research Aggressive multiple sclerosis: Proposed definition in Multiple Sclerosis (ECTRIMS). and treatment algorithm. Nat Rev Neurol 2015; 11(7): 379–389. ORCID iDs 11. Edan G, Comi G, Le Page E, et al. Mitoxantrone prior Georgina Arrambide https://orcid.org/0000-0002 to interferon beta-1b in aggressive relapsing multiple -2657-5510 sclerosis: A 3-year randomised trial. J Neurol Tobias Derfuss https://orcid.org/0000-0001-8431 Neurosurg Psychiatry 2011; 82(12): 1344–1350. -8769 Sandra Vukusic https://orcid.org/0000-0001-7337 12. Saccardi R, Freedman MS, Sormani MP, et al. -7122 A prospective, randomized, controlled trial of autologous haematopoietic stem cell transplantation Bernhard Hemmer https://orcid.org/0000-0001 for aggressive multiple sclerosis: A position paper. -5985-6784 Mult Scler 2012; 18(6): 825–834. Mar Tintore https://orcid.org/0000-0001-9999-5359 13. Tintore M, Arrambide G, Otero-Romero S, et al. The Supplemental material long-term outcomes of CIS patients in the Barcelona Supplemental material for this article is available inception cohort: Looking back to recognize aggressive MS. Mult Scler. Epub ahead of print 15 online. October 2019. DOI: 10.1177/1352458519877810. 14. Malpas CB, Manouchehrinia A, Sharmin S, et al. Aggressive form of multiple sclerosis can be predicted early after disease onset. Mult Scler 2019; References 25: 605–607. 1. Lublin FD, Reingold SC, Cohen JA, et al. Defining the clinical course of multiple sclerosis: The 15. Diaz C, Zarco LA and Rivera DM. Highly active 2013 revisions. Neurology 2014; 83(3): multiple sclerosis: An update. Mult Scler Relat 278–286. Disord 2019; 30: 215–224. 2. Tintore M, Vidal-Jordana A and Sastre-Garriga J. 16. Weinshenker BG, Rice GP, Noseworthy JH, et al. The Treatment of multiple sclerosis: Success from bench natural history of multiple sclerosis: A geographically to bedside. Nat Rev Neurol 2019; 15(1): 53–58. based study. 3. Multivariate analysis of predictive journals.sagepub.com/home/msj 1041 Multiple Sclerosis Journal 26(9) factors and models of outcome. Brain 1991; 29. Truyen L, van Waesberghe JH, van Walderveen MA, 114(Pt. 2): 1045–1056. et al. Accumulation of hypointense lesions ("black holes") on T1 spin-echo MRI correlates with disease 17. Manouchehrinia A, Zhu F, Piani-Meier D, et al. progression in multiple sclerosis. Neurology 1996; Predicting risk of secondary progression in multiple 47(6): 1469–1476. sclerosis: A nomogram. Mult Scler 2019; 25(8): 1102–1112. 30. Brownlee WJ, Altmann DR, Prados F, et al. Early imaging predictors of long-term outcomes in 18. Bergamaschi R, Berzuini C, Romani A, et al. relapse-onset multiple sclerosis. Brain 2019; 142(8): Predicting secondary progression in relapsing- 2276–2287. remitting multiple sclerosis: A Bayesian analysis. J Neurol Sci 2001; 189(1–2): 13–21. 31. Rio J, Rovira A, Tintore M, et al. Disability progression markers over 6-12 years in interferon- 19. Amato MP, Ponziani G, Bartolozzi ML, et al. A beta-treated multiple sclerosis patients. Mult Scler prospective study on the natural history of multiple 2018; 24(3): 322–330. sclerosis: Clues to the conduct and interpretation of clinical trials. J Neurol Sci 1999; 168(2): 96–106. 32. Fisniku LK, Chard DT, Jackson JS, et al. Gray matter atrophy is related to long-term disability in multiple 20. Scott TF and Schramke CJ. Poor recovery after the sclerosis. Ann Neurol 2008; 64(3): 247–254. first two attacks of multiple sclerosis is associated with poor outcome five years later. J Neurol Sci 2010; 33. Tintore M, Rovira A, Arrambide G, et al. Brainstem 292(1–2): 52–56. lesions in clinically isolated syndromes. Neurology 2010; 75(21): 1933–1938. 21. Trojano M, Avolio C, Manzari C, et al. Multivariate analysis of predictive factors of multiple sclerosis 34. Arrambide G, Rovira A, Sastre-Garriga J, et al. Spinal course with a validated method to assess clinical cord lesions: A modest contributor to diagnosis events. J Neurol Neurosurg Psychiatry 1995; 58(3): in clinically isolated syndromes but a relevant 300–306. prognostic factor. Mult Scler 2018; 24(3): 301–312. 22. Zarei M, Chandran S, Compston A, et al. Cognitive 35. Casserly C, Seyman EE, Alcaide-Leon P, et al. Spinal presentation of multiple sclerosis: Evidence for a cord atrophy in multiple sclerosis: A systematic cortical variant. J Neurol Neurosurg Psychiatry 2003; review and meta-analysis. J Neuroimaging 2018; 74(7): 872–877. 28(6): 556–586. 23. Langer-Gould A, Popat RA, Huang SM, et al. 36. Eshaghi A, Prados F, Brownlee WJ, et al. Deep gray Clinical and demographic predictors of long-term matter volume loss drives disability worsening in disability in patients with relapsing-remitting multiple multiple sclerosis. Ann Neurol 2018; 83(2): 210–222. sclerosis: A systematic review. Arch Neurol 2006; 37. Scalfari A, Romualdi C, Nicholas RS, et al. The 63(12): 1686–1691. cortical damage, early relapses, and onset of the 24. Jokubaitis VG, Spelman T, Kalincik T, et al. progressive phase in multiple sclerosis. Neurology Predictors of disability worsening in clinically 2018; 90(24): e2107–e2118. isolated syndrome. Ann Clin Transl Neurol 2015; 38. Hanninen K, Viitala M, Paavilainen T, et al. Thalamic 2(5): 479–491. atrophy without whole brain atrophy is associated 25. Stewart T, Spelman T, Havrdova E, et al. with absence of 2-year NEDA in multiple sclerosis. Contribution of different relapse phenotypes to Front Neurol 2019; 10: 459. disability in multiple sclerosis. Mult Scler 2017; 39. Elliott C, Wolinsky JS, Hauser SL, et al. Slowly 23(2): 266–276. expanding/evolving lesions as a magnetic resonance 26. Zipoli V, Goretti B, Hakiki B, et al. Cognitive imaging marker of chronic active multiple sclerosis impairment predicts conversion to multiple sclerosis lesions. Mult Scler 2019; 25(14): 1915–1925. in clinically isolated syndromes. Mult Scler 2010; 40. Trapp BD, Peterson J, Ransohoff RM, et al. Axonal 16(1): 62–67. transection in the lesions of multiple sclerosis. N Engl J Med 1998; 338(5): 278–285. 27. Scalfari A, Neuhaus A, Degenhardt A, et al. The natural history of multiple sclerosis: A geographically 41. Bitsch A, Schuchardt J, Bunkowski S, et al. Acute based study 10: Relapses and long-term disability. axonal injury in multiple sclerosis. Brain 2000; Brain 2010; 133(Pt. 7): 1914–1929. 123(Pt. 6): 1174–1183. 28. Confavreux C, Vukusic S and Adeleine P. Early 42. Peterson JW, Bo L, Mork S, et al. Transected clinical predictors and progression of irreversible neurites, apoptotic neurons, and reduced disability in multiple sclerosis: An amnesic process. inflammation in cortical multiple sclerosis lesions. Brain 2003; 126(Pt. 4): 770–782. Ann Neurol 2001; 50(3): 389–400. 1042 journals.sagepub.com/home/msj E Iacobaeus, G Arrambide et al. 43. Jurgens T, Jafari M, Kreutzfeldt M, et al. 56. Khademi M, Kockum I, Andersson ML, et al. Reconstruction of single cortical projection neurons Cerebrospinal fluid CXCL13 in multiple sclerosis: A reveals primary spine loss in multiple sclerosis. Brain suggestive prognostic marker for the disease course. 2016; 139(Pt. 1): 39–46. Mult Scler 2011; 17(3): 335–343. 44. Burman J, Fransson M, Totterman TH, et al. 57. Puthenparampil M, Federle L, Miante S, et al. BAFF T-cell responses after haematopoietic stem cell Index and CXCL13 levels in the cerebrospinal fluid transplantation for aggressive relapsing-remitting associate respectively with intrathecal IgG synthesis multiple sclerosis. Immunology 2013; 140(2): and cortical atrophy in multiple sclerosis at clinical 211–219. onset. J Neuroinflammation 2017; 14(1): 11. 45. Darlington PJ, Touil T, Doucet JS, et al. Diminished 58. Canto E, Tintore M, Villar LM, et al. Chitinase Th17 (not Th1) responses underlie multiple sclerosis 3-like 1: Prognostic biomarker in clinically isolated disease abrogation after hematopoietic stem cell syndromes. Brain 2015; 138(Pt. 4): 918–931. transplantation. Ann Neurol 2013; 73(3): 341–354. 59. Opdenakker G, Nelissen I and Van Damme J. 46. Karnell FG, Lin D, Motley S, et al. Reconstitution of Functional roles and therapeutic targeting of immune cell populations in multiple sclerosis patients gelatinase B and chemokines in multiple sclerosis. after autologous stem cell transplantation. Clin Exp Lancet Neurol 2003; 2(12): 747–756. Immunol 2017; 189(3): 268–278. 60. Okuda DT, Srinivasan R, Oksenberg JR, et al. 47. Darlington PJ, Stopnicki B, Touil T, et al. Natural Genotype-Phenotype correlations in multiple killer cells regulate Th17 cells after autologous sclerosis: HLA genes influence disease severity hematopoietic stem cell transplantation for relapsing inferred by 1HMR spectroscopy and MRI measures. remitting multiple sclerosis. Front Immunol 2018; 9: Brain 2009; 132(Pt. 1): 250–259. 834. 61. Kroner A, Mehling M, Hemmer B, et al. A PD-1 48. Gasperi C, Salmen A, Antony G, et al. Association polymorphism is associated with disease progression of intrathecal immunoglobulin G synthesis with in multiple sclerosis. Ann Neurol 2005; 58(1): 50–57. disability worsening in multiple sclerosis. JAMA 62. Brynedal B, Wojcik J, Esposito F, et al. MGAT5 Neurol 2019; 76(7): 841–849. alters the severity of multiple sclerosis. J 49. Muraro PA, Robins H, Malhotra S, et al. T Neuroimmunol 2010; 220(1–2): 120–124. cell repertoire following autologous stem cell 63. Chapman J, Vinokurov S, Achiron A, et al. transplantation for multiple sclerosis. J Clin Invest APOE genotype is a major predictor of long-term 2014; 124(3): 1168–1172. progression of disability in MS. Neurology 2001; 50. Magraner MJ, Bosca I, Simo-Castello M, et al. Brain 56(3): 312–316. atrophy and lesion load are related to CSF lipid- 64. Nociti V, Santoro M, Quaranta D, et al. BDNF rs6265 specific IgM oligoclonal bands in clinically isolated polymorphism methylation in multiple sclerosis: A syndromes. Neuroradiology 2012; 54(1): 5–12. possible marker of disease progression. PLoS ONE 51. Thangarajh M, Gomez-Rial J, Hedstrom AK, et al. 2018; 13(10): e0206140. Lipid-specific immunoglobulin M in CSF predicts 65. Waldman A, Ghezzi A, Bar-Or A, et al. Multiple adverse long-term outcome in multiple sclerosis. Mult sclerosis in children: An update on clinical diagnosis, Scler 2008; 14(9): 1208–1213. therapeutic strategies, and research. Lancet Neurol 52. Salzer J, Svenningsson A and Sundstrom P. 2014; 13(9): 936–948. Neurofilament light as a prognostic marker in 66. Krumbholz M, Theil D, Cepok S, et al. Chemokines multiple sclerosis. Mult Scler 2010; 16(3): 287–292. in multiple sclerosis: CXCL12 and CXCL13 up-regulation is differentially linked to CNS immune 53. Hakansson I, Tisell A, Cassel P, et al. Neurofilament cell recruitment. Brain 2006; 129(Pt. 1): 200–211. levels, disease activity and brain volume during follow-up in multiple sclerosis. J Neuroinflammation 67. Brettschneider J, Czerwoniak A, Senel M, et al. 2018; 15(1): 209. The chemokine CXCL13 is a prognostic marker in clinically isolated syndrome (CIS). PLoS ONE 2010; 54. Chitnis T, Gonzalez C, Healy BC, et al. 5(8): e11986. Neurofilament light chain serum levels correlate with 10-year MRI outcomes in multiple sclerosis. Ann Clin 68. Tintore M, Rovira A, Rio J, et al. Defining high, Transl Neurol 2018; 5(12): 1478–1491. medium and low impact prognostic factors for developing multiple sclerosis. Brain 2015; 138(Pt. 7): 55. Barro C, Benkert P, Disanto G, et al. Serum 1863–1874. neurofilament as a predictor of disease worsening and brain and spinal cord atrophy in multiple sclerosis. 69. Radue EW, Barkhof F, Kappos L, et al. Correlation Brain 2018; 141(8): 2382–2391. between brain volume loss and clinical and MRI journals.sagepub.com/home/msj 1043 Multiple Sclerosis Journal 26(9) outcomes in multiple sclerosis. Neurology 2015; 80. Kutzelnigg A, Lucchinetti CF, Stadelmann C, et al. 84(8): 784–793. Cortical demyelination and diffuse white matter injury in multiple sclerosis. Brain 2005; 128(Pt. 11): 70. Brownlee WJ, Altmann DR Alves Da, Mota P, et al. 2705–2712. Association of asymptomatic spinal cord lesions and atrophy with disability 5 years after a clinically 81. Atkins HL, Bowman M, Allan D, et al. isolated syndrome. Mult Scler 2017; 23(5): 665–674. Immunoablation and autologous haemopoietic stem- cell transplantation for aggressive multiple sclerosis: 71. Rudick RA, Lee JC, Simon J, et al. Significance of T2 A multicentre single-group phase 2 trial. Lancet 2016; lesions in multiple sclerosis: A 13-year longitudinal 388(10044): 576–585. study. Ann Neurol 2006; 60(2): 236–242. 82. Kuhle J, Plavina T, Barro C, et al. Neurofilament 72. Burman J, Iacobaeus E, Svenningsson A, et al. light levels are associated with long-term outcomes in Autologous haematopoietic stem cell transplantation multiple sclerosis. Mult Scler 2019; 2019: 1–9. for aggressive multiple sclerosis: The Swedish experience. J Neurol Neurosurg Psychiatry 2014; 83. Magliozzi R, Howell OW, Nicholas R, et al. 85(10): 1116–1121. Inflammatory intrathecal profiles and cortical damage in multiple sclerosis. Ann Neurol 2018; 83(4): 73. Burt RK, Balabanov R, Han X, et al. Association 739–755. of nonmyeloablative hematopoietic stem cell transplantation with neurological disability in patients 84. Schreiber K, Otura AB, Ryder LP, et al. Disease with relapsing-remitting multiple sclerosis. JAMA severity in Danish multiple sclerosis patients 2015; 313(3): 275–284. evaluated by MRI and three genetic markers (HLA- DRB1*1501, CCR5 deletion mutation, apolipoprotein 74. Dekker I, Sombekke MH, Balk LJ, et al. Infratentorial E). Mult Scler 2002; 8(4): 295–298. and spinal cord lesions: Cumulative predictors of long-term disability. Mult Scler. Epub ahead of print 85. Alcina A, Ramagopalan SV, Fernandez O, et al. 2 August 2019. DOI: 10.1177/1352458519864933. Hexose-6-phosphate dehydrogenase: A new risk gene 75. Perez-Miralles F, Sastre-Garriga J, Tintore M, et al. for multiple sclerosis. Eur J Hum Genet 2010; 18(5): Clinical impact of early brain atrophy in clinically 618–620. isolated syndromes. Mult Scler 2013; 19(14): 86. Ponath G, Lincoln MR, Levine-Ritterman M, et al. 1878–1886. Enhanced astrocyte responses are driven by a genetic 76. Jacobsen C, Hagemeier J, Myhr KM, et al. Brain risk allele associated with multiple sclerosis. Nat atrophy and disability progression in multiple Commun 2018; 9(1): 5337. sclerosis patients: A 10-year follow-up study. J 87. Hoffmann V, Pohlau D, Przuntek H, et al. A null Neurol Neurosurg Psychiatry 2014; 85(10): mutation within the ciliary neurotrophic factor 1109–1115. (CNTF)-gene: Implications for susceptibility and 77. Tao G, Datta S, He R, et al. Deep gray matter atrophy disease severity in patients with multiple sclerosis. in multiple sclerosis: A tensor based morphometry. J Genes Immun 2002; 3(1): 53–55. Neurol Sci 2009; 282(1–2): 39–46. 88. Montalban X, Gold R, Thompson AJ, et al. 78. De Stefano N, Narayanan S, Francis GS, et al. ECTRIMS/EAN Guideline on the pharmacological Evidence of axonal damage in the early stages of treatment of people with multiple sclerosis. Mult multiple sclerosis and its relevance to disability. Arch Scler 2018; 24(2): 96–120. Neurol 2001; 58(1): 65–70. Visit SAGE journals online 89. Krieger SC, Cook K, De Nino S, et al. The journals.sagepub.com/ 79. Bjartmar C, Kinkel RP, Kidd G, et al. Axonal loss in topographical model of multiple sclerosis: A dynamic home/msj normal-appearing white matter in a patient with acute visualization of disease course. Neurol Neuroimmunol SAGE journals MS. Neurology 2001; 57(7): 1248–1252. Neuroinflamm 2016; 3(5): e279. 1044 journals.sagepub.com/home/msj