1

Année 2012 Thèse n°1952

THÈSE pour le

DOCTORAT DE L’UNIVERSITÉ BORDEAUX 2

Ecole doctorale Sciences de la Vie et de la Santé Mention: Sciences, Technologie, Santé

Option: Neurosciences

Présentée et soutenue publiquement Le 7 Novembre 2012

Par Aurélie RUET

Née le 5 Août 1979 à Villefranche sur Saône

Prédiction du diagnostic et du pronostic aux stades précoces de la Sclérose en Plaques

Thèse dirigée par Mr le Professeur Bruno BROCHET

Membres du Jury Rapporteurs et examinateurs: Mme le Professeur Iris-Katharina PENNER Mr le Professeur Jean PELLETIER Examinateurs: Mme le Professeur Maria-Pia AMATO Mr le Professeur Pierre CLAVELOU Mme le Docteur Christine LEBRUN-FRENAY

2

PUBLICATIONS

Peer-reviewed publications

- Original articles accepted

Ruet A , Deloire M, Hamel D, Ouallet JC, Petry K, Brochet B. Cognitive impairment, health related-related quality of life and vocational status in the early stages of multiple sclerosis. J Neurol 2012 [Epub ahead of print].

Ruet A , Deloire MS, Ouallet JC, Molinier S, Brochet B. Predictive factors for multiple sclerosis in patients with clinically isolated spinal cord syndrome. Mult Scler 2011; 17: 312-8.

Deloire MSA, Ruet A , Hamel D, Bonnet M, Dousset V, Brochet B. MRI predictors of cognitive outcome in early multiple sclerosis. Neurology 2011; 76: 1161-7

Collongues N, Marignier R, Zéphir H, Blanc F, Vukusic S, Outteryck O, Fleury M, Ruet A , Borgel F, Thouvenot E, Moreau T, Defer G, Derache N, Pelletier J, Audoin B, Debouverie M, Labauge P, Gout O, Camu W, Brassat D, Brochet B, Vermersch P, Confavreux C and de Seze J. High-risk syndrome for neuromyelitis optica: a descriptive and comparative study. Mult Scler 2011; 17: 720-4.

Collongues N, Marignier R, Zéphir H, Papeix C, Fontaine B, Blanc F, Rodriguez D, FleuryM, Vukusic S, Pelletier J, Audoin B, Thouvenot E, Camu W, Barroso B, Ruet A , Brochet B, Vermersch P, Confavreux C, de Seze J. Long-term follow-up of neuromyelitis optica with a pediatric onset. Neurology 2010; 5: 1084-8.

Deloire M, Ruet A , Hamel D, Bonnet M, Brochet B. Early cognitive impairment in multiple sclerosis predicts disability outcome several years later. Mult Scler 2010; 16: 581-587.

Collongues N, Marignier R, Zéphir H, Papeix C, Blanc F, Ritleng C, Tchikviladzé M, Outteryck O, Vukusic S, Fleury M, Fontaine B, Brassat D, Clanet M, Milh M, Pelletier J, Audoin B, Ruet A , Lebrun-Frenay C, Thouvenot E, Camu W, Debouverie M, Créange A, Moreau T, Labauge P, Castelnovo G, Edan G, Le Page E, Defer G, Barroso B, Heinzlef O, Gout O, Rodriguez D, Wiertlewski S, Laplaud D, Borgel F, Tourniaire P, Grimaud J, Brochet B, Vermersch P, Confavreux C, de Seze J. Neuromyelitis optica in France: A multicenter study of 125 patients. Neurology 2010; 74: 736-42.

- Original articles submitted

Ruet A , Deloire M, Charré-Morin J, Hamel D, Brochet B. Cognitive impairment differs between primary progressive and relapsing-remitting multiple sclerosis. Submitted in Neurology 2012, in revision.

Ruet A , Deloire MSA, Charré-Morin J, Hamel D, Brochet B. Validity and predictive value of a new computerised cognitive test for the detection of information processing speed impairment in multiple sclerosis. Submitted in Mult Scler 2012, in revision.

3

- Original articles in preparation

Ruet A , Arrambide G, Brochet B, Auger C, Simon E, Rovira A, Montalban X, Tintoré M. Early predictors of multiple sclerosis after a clinically isolated syndrome.

Other publications

Ruet A , Brochet B. Syndrome cliniquement isolé: prise en charge thérapeutique. Dossier: Sclérose en plaques: les traitements- 3ème partie. Neurologies 2012, Volume 15-N°148, p 3-8.

Brochet B, Ruet A . Les traitements de fond de seconde intention dans la sclérose en plaques rémittente-récurrente. Revue Neurologique Pratique Neurologique-FMC 2012; 3:91-99.

Brochet B, Ruet A . Les douleurs au cours de la sclérose en plaques. Acte des 24èmes

Entretiens annuels de la Fondation Garches-Handicap et douleur. 2011. Editions GMSanté.

Brochet B., Ruet A . Apport de l’IRM à la prédiction de l’incapacité dans la sclérose en Plaques. La lettre du Neurologue Suppl au vol XV-n°3 p3-5.

Ruet A . Les atteintes de la mémoire dans la sclérose en plaques dès les stades précoces de la maladie. I-Troubles cognitifs et Sclérose en plaques-Neurologies 2010; Volume Hors Série N°126-Cahier 2.

Ruet A. Les myélites aigües. Abstract Neurologies 2009.

Ruet A , Brochet B. Quelle est la place de la cognition dans la sclérose en plaques? La lettre du Neurologue 2009; Suppl au Vol XIII; N°11.

Ruet A . Congrès ECTRIMS, ACTRIMS et LACTRIMS, Montréal Sept 2008. SEP et Neurosciences 2008.

4

Oral communications

Myélites partielles, myélites transverses. Club francophone de la sclérose en plaques 2012, Bischoffsheim.

Screening for cognitive impairment in multiple sclerosis patients using a new computerised test. 1rst International meeting on cognition in multiple sclerosis (IMSCOGS) 2012, Bordeaux.

L’atteinte cognitive est-elle comparable entre les formes RR et PP de SEP? Réunion Cognition du Club Francophone de la Sclérose en plaques 2012, Bordeaux.

Spinal cord syndromes: predictive factors for multiple sclerosis. French and Spanish meeting 2012, Palma.

MRI at the beginning: Radiologically isolated syndromes. Rev Neurol (Paris) 2010; 166 Suppl 2: A1-A202.

Acute myelopathies in adults: 135 cases. French Neurological Society. Rev Neurol (Paris) 2008; 164: 775.

Posters

Ruet A , Deloire M, Hamel D, Ouallet JC, Brochet B. Prediction of vocational status by cognitive impairment in early MS patients: a 7-year longitudinal study. First international multiple sclerosis cognition society congress 2012, Bordeaux.

Ruet A , Charré-Morin J, Brochet B, Deloire M. Cognitive deficit in relapsing-remitting and primary progressive multiple sclerosis. First international multiple sclerosis cognition society congress 2012, Bordeaux.

Ruet A , Charré-Morin J, Brochet B, Deloire M. Cognitive impairment in multiple sclerosis: comparison between primary progressive and relapsing-remitting patients. 12èmes Journées scientifiques de l’école doctorale 2012.

Ruet A , Charré-Morin J, Brochet B, Deloire M. Cognitive impairment in relapsing-remitting and primary-progressive multiple sclerosis patients: comparison with matched controls. Mult Scler 2011; 17 Suppl 10: S53-S276.

Ruet A , Charré-Morin J, Brochet B, Hamel D, Deloire M. How to detect cognitive impairment in relapsing-remitting and primary-progressive multiple sclerosis. Mult Scler 2011; 17 Suppl 10: S277-S505.

Ruet A , Hamel D, Deloire M, Brochet B. Physical and cognitive predictors of employment status in relapsing-remitting multiple sclerosis patients: a 7-year longitudinal study. American Academy of Neurology 2011, Honolulu.

Ruet A , Hamel D, Deloire M, Brochet B. Physical and cognitive predictors of employment status in relapsing-remitting multiple sclerosis patients: a 7-year longitudinal study. 11èmes Journées scientifiques de l’école doctorale 2011, Arcachon.

5

Hamel D, Ruet A , Deloire M, Ouallet JC, Brochet B. Quality of life and cognition in early multiple sclerosis. Mult Scler 2010; 16 Suppl 10. S7- S352.

Collongues N, Marignier R, Zephir H, et al. Neuromyelitis optica in paediatric populations. Mult Scler 2009; 15 Suppl 9: S31-S150.

Collongues N, Marignier R, Zephir H, et al. Optic neuritis or myelitis associated with NMO-IgG antibodies: high risk syndrome or false-positive? Mult Scler 2009; 15 Suppl 9: S151-269.

Ruet A , Molinier S, Deloire M, et al. Predictive features of conversion to multiple sclerosis in clinically isolated syndromes affecting the spinal cord. Mult Scler 2008; 14 Suppl 1; S29-S294.

Ruet A , Molinier S, Deloire M, et al. MRI and CSF criteria for multiple sclerosis in patients with clinically isolated syndromes affecting the spinal cord. Mult Scler 2008; 14 Suppl 1; S29-S294.

Ruet A et al. T regulators cells in relapsing-remitting multiple sclerosis patients compared to matched controls. French Neurological Society. Rev Neurol (Paris) 2008.

Ruet A , Ferrer X. Alpha-dystroglycanopathies. Rev Neurol (Paris) 2008; 164 Suppl 2; A1-A215.

Ruet A , Jeannin S, Debruxelles S, Brochet B. Relapsing myelitis revealing cœliac disease. Rev Neurol (Paris) 2007; 163 Suppl 1; 7-273.

Posters in preparation

Ruet A , Arrambide G, Brochet B, Auger C, Simon E, Rovira A, Montalban X, Tintoré M. Predictive factors for MS in patients with clinically isolated syndromes .

Ruet A , Deloire M, Ouallet JC, Brochet B. Prediction of vocational status 10 years after the MS diagnosis by early information processing speed assessment .

Ruet A , Deloire M, Ouallet JC, Dousset V, Brochet B. Early diffuse brain tissue injury predicts disability ten years later in RRMS patients .

6

Chapitre 1:

INTRODUCTION

7

Introduction

La sclérose en plaques (SEP) est une maladie chronique, auto-immune, inflammatoire, et neurodégénérative. Il s’agit de la pathologie chronique la plus répandue parmi les maladies neurologiques invalidantes de l’adulte jeune en France et concerne 94,7 pour 100.000 (94,3 à 95,1) habitants. Les taux de prévalence chez les femmes et les hommes sont respectivement de 130,5 (129,8 à 131,2) et 54,8 (54,4 à 55,3) (Fromont et al., 2010). La SEP est une pathologie comportant de multiples facettes et peut toucher toutes les parties du système nerveux central (SNC). Les personnes atteintes de SEP peuvent présenter un large éventail de symptômes, tels qu’un déficit moteur, sensoriel, visuel, cérébelleux, vésico-sphinctérien, et sexuel. Ces symptômes peuvent survenir simultanément ou indépendamment dans le temps. Les principaux défis à relever dans la gestion des patients qui présentent des symptômes évocateurs de SEP sont de déterminer si un diagnostic de SEP peut être établi et d'évaluer la gravité de la maladie. Ces défis seront abordés dans la première, puis dans la deuxième partie de cette thèse.

Une difficulté réside dans l’hétérogénéité présente au cours de cette maladie. L’évaluation de l’activité clinique (poussée, invalidité) est insuffisante pour prédire l’évolution à long terme de la SEP. Parmi les paramètres cliniques proposés pour caractériser le handicap de la maladie, les déficits cognitifs sont apparus comme un candidat intéressant. Les troubles cognitifs sont fréquents chez les patients ayant une SEP, mais ils sont souvent sous-estimés à la fois par les patients et par les cliniciens. La relation entre les troubles cognitifs dans la SEP et l’atteinte diffuse cérébrale nous a incités à étudier l’atteinte cognitive comme un marqueur pronostique dans la SEP. La détection de ces déficits cognitifs dans la SEP sera abordée dans la dernière partie de cette thèse.

1. Prédiction du diagnostic de la SEP après un synd rome clinique isolé

Après avoir exclu les diagnostics différentiels, la démonstration de la dissémination des lésions dans l'espace et dans le temps est la base du diagnostic de SEP. Environ 85% des patients ayant une SEP débutent leur maladie par la survenue de poussées dans le cadre d’une SEP dénommée SEP rémittente récurrente (SEP-RR). Le premier événement clinique démyélinisant est appelé syndrome clinique isolé (SCI). Un diagnostic de SCI suggère la possibilité d’une SEP. Un SCI se produit généralement chez les jeunes adultes et peut affecter de manière isolée (SCI de type monofocal) ou simultanément (SCI de type polyfocal) les nerfs optiques, le tronc cérébral, la moelle épinière, ou, moins fréquemment, les structures supratentorielles du cerveau (Miller et al., 2012). Cependant, les premiers épisodes qui ne sont jamais suivis par une autre poussée sont cliniquement indiscernables des épisodes qui sont la première manifestation d’une SEP-RR active.

8

1.1. Démarche pour le diagnostic de la SEP

Le neurologue qui prend en charge un patient présentant un SCI évocateur de SEP doit suivre plusieurs étapes avant de pouvoir faire le diagnostic de SEP. Il doit répondre aux questions suivantes:

- Est-ce le SCI est typique d'une SEP? L'histoire clinique et l'examen neurologique doivent être compatibles avec un SCI évocateur de SEP. Après une réunion de consensus, il a été proposé une classification en 3 catégories des caractéristiques cliniques des SCI en fonction de leurs probabilités à être associés à un diagnostic de SEP: SCI généralement observés dans la SEP; SCI sont moins courants, mais possibles, dans la SEP, et SCI atypiques qui ne sont pas censés êtres rencontrés dans la SEP (Miller et al., 2008).

- Les symptômes sont-il compatibles avec une maladi e démyélinisante inflammatoire? Le neurologue doit exclure les syndromes évocateurs d’une pathologie non-démyélinisante à partir des données cliniques et paracliniques (telles que l'imagerie par résonance magnétique (IRM)) et les tests de laboratoire (analyses du sérum et du liquide céphalorachidien (LCR)). Des drapeaux rouges en faveur de diagnostics différentiels spécifiques ont été proposés dans un récent consensus (Miller et al., 2008).

- Existe-t-il des lésions multiples dans le système nerveux central? Un diagnostic de SEP peut être établi à partir des données cliniques uniquement si le patient a présenté au moins 2 épisodes cliniques typiques et évocateurs et des signes en lien avec au moins 2 lésions distinctes du SNC (Poser et al., 1983;. McDonald et al., 2001; Polman et al., 2005; Polman et al., 2011). Après avoir exclu d'autres diagnostics, et s'il n'y a pas de meilleures explications pour l'événement clinique, le diagnostic de SEP est fondé sur l'association de la dissémination des lésions dans l'espace (DIS) et dans le temps (DIT). Les données paracliniques peuvent aider à identifier la DIS au niveau de multiples régions affectées dans le SNC et en démontrant la DIT à travers différentes périodes.

1.2. Les critères diagnostiques de la SEP

Il n'existe pas de test spécifique et unique pour établir un diagnostic de SEP. La seule façon de prouver le diagnostic de la SEP est l'examen histopathologique des tissus à partir de plusieurs sites touchés au sein du SNC.

Les critères diagnostiques de la SEP peuvent être remplis uniquement en fonction des données de l'histoire clinique et de l'examen neurologique. L'interprétation correcte des symptômes et des signes neurologiques est une condition fondamentale pour un diagnostic non erroné de SEP.

Selon les critères de Schumacher, une sclérose en plaques cliniquement définie (SEP-CD) est définie par la survenue de signes objectifs de dysfonctionnement du système nerveux central compatibles avec une SEP en lien avec des lésions à au moins 2 sites (Schumacher et al., 1965). Deux épisodes ou plus de 24 heures au moins doivent être séparés par au moins 1 mois. Le diagnostic est posé par un neurologue si les signes ne peuvent pas être expliqués par d'autres maladies.

9

Les critères diagnostiques ont évolué au fil du temps et ont été révolutionnés par l'introduction de l'IRM en 1981 (Young et al., 1981). Les critères de Poser sont largement utilisés dans la pratique clinique, les études en épidémiologie et les essais thérapeutiques (Poser et al., 1983). Ces critères sont souvent utilisés comme étant la référence pour le diagnostic de la SEP. Selon ces critères, une SEP-CD est définie par la survenue de 2 épisodes cliniques et des preuves cliniques d'au moins 2 lésions distinctes au niveau du SNC; ou 2 attaques et des preuves cliniques d'une et des preuves paracliniques d'une autre lésion distincte. Un diagnostic de SEP peut être cliniquement probable si un patient a présenté 2 attaques et des signes cliniques d'une lésion, ou 1 attaque et des preuves cliniques de 2 lésions séparées; ou 1 attaque, des signes cliniques d'1, et la preuve paraclinique d'une autre lésion distincte. Les examens d'électrophysiologie, l'examen du LCR, et l'imagerie sont utilisés pour compléter les preuves pour le diagnostic de SEP dans les cas où les critères cliniques ne sont pas suffisants (absence de deuxième événement clinique ou de second site du SNC concerné). Ainsi, un diagnostic de SEP probable ou certaine peut être porté en fonction des données des divers examens.

L'IRM est un outil utile pour démontrer la dissémination spatiale et temporelle des lésions après un SCI avant la survenue d'un deuxième événement clinique. Plusieurs critères IRM ont été proposés (tableau 1), et les critères de DIS et DIT ont évolué au fil du temps (tableau 2) (McDonald et al., 2001; Polman et al., 2005; 2011). Un comité international a publié des critères diagnostic avec une importance croissante aux données de l'IRM en 2001 (McDonald et al., 2001). Le diagnostic de la SEP selon les critères de McDonald peut être faite chez les patients après un SCI si les données d'IRM et les résultats de laboratoire fournissent des preuves pour les DIS et DIT, ce qui appelé "diagnostic de SEP McDonald». Différentes versions des critères de McDonald ont été proposées (tableaux 2-3) et permettent l'établissement d'un diagnostic précoce de la SEP avant l'apparition d'un deuxième événement clinique dans certains cas. L'application des critères de McDonald après un SCI typique peut réduire l'incertitude des patients, l'anxiété et le temps d'attente entre le début du SCI et l'annonce du diagnostic de SEP (Heesen et al., 2003). Les traitements de fond ont été testés dans les essais pivots chez des patients inclus après un SCI typique avec des anomalies sur les IRM du cerveau (Jacobs et al., 2000; Kappos et al., 2006; Comi et al., 2009; Comi et al., 2012). Les agences américaines et européennes du médicament ont fourni les autorisations de prescription de ces traitements après un SCI en cas de diagnostic d'une «SEP McDonald».

1.3. Utilisation de l'IRM et du LCR après un SCI

L'incorporation des données de l'IRM dans les critères diagnostiques souligne la valeur indéniable de l'IRM dans cette pathologie. Comme l'IRM peut révéler de nombreuses lésions cliniquement silencieuses, l'inclusion des résultats de l'IRM dans les critères de McDonald augmente la sensibilité de la détection de la SEP.

Les critères diagnostiques de la SEP ont été obtenus à partir d'études d'IRM recherchant l'apparition d'une deuxième poussée après un SCI typique. Barkhof et collaborateurs ont comparé différents critères d'IRM utilisés pour prédire une SEP-CD dans une cohorte de 74 patients ayant eu un SCI (Barkhof et al., 1997). Dans cette étude, les examens cliniques et les résultats du LCR n'étaient pas pris en considération pour l'analyse

10

de la prédiction du diagnostic de SEP-CD. Les critères de Barkhof et collaborateurs ont ensuite été modifiés par Tintoré et collaborateurs qui ont proposé la présence de ≥3 des 4 critères de Barkhof avec la possibilité de substituer une lésion prenant le gadolinium par la présence de ≥9 lésions en T2 (tableau 1) (Tintoré et al., 2000). Plus récemment, Swanton et collaborateurs ont étudié une cohorte de patients ayant un SCI pour prédire le diagnostic de SEP-CD (Swanton et al., 2006). Ils ont proposé de nouveaux critères pour la SEP en simplifiant les critères de Barkhof-Tintoré. Leur objectif était de démontrer la DIS sur la base d une seule IRM en ne considérant que les lésions visibles en T2 et leurs topographies spécifiques. Sur la base des critères de Swanton, la DIS (Swanton et al., 2006) peut être démontrée par au moins 1 lésion en T2 dans au moins 2 des 4 sites considérés comme caractéristiques de la SEP (juxtacorticale, périventriculaire, sous-tentorielle, et dans la moelle épinière), les lésions dans les régions symptomatiques sont exclues chez les patients présentant un syndrome de la moelle épinière ou du tronc cérébral. Dans ces critères, ni l injection de gadolinium ni l examen du LCR ne sont requis pour démontrer une DIS. Les critères de Swanton et collaborateurs ont été étudiés dans une étude multicentrique rétrospective importante pour évaluer leur performance dans la prédiction de SEP-CD après un SCI typique (Swanton et al., 2007). Les nouveaux critères proposés étaient plus sensibles et moins spécifiques que les critères de DIS de McDonald 2005 (≥ 3 critères de Barkhof-Tintoré) dans la prédiction de SEP-CD mais sans compromettre leur exactitude. Ces critères de DIS permettent d estimer la probabilité d'avoir une SEP après un SCI typique et sont plus à considérer comme des facteurs de risque de la SEP plutôt que des critères diagnostiques.

Bien que la survenue d’un deuxième événement clinique conforte le diagnostic positif de la SEP, la tendance actuelle est d'établir le diagnostic de la SEP dès que possible pour commencer précocement un traitement de fond dans le but d’éviter ou du moins de limiter l’installation d’un handicap irréversible. Le neurologue doit garder à l'esprit que les résultats de laboratoire et l'IRM sont des outils complémentaires utiles pour évaluer le diagnostic de la SEP, mais ces résultats ne remplacent pas l'expérience pratique du clinicien. Kurtzke disait: «La SEP est ce qu’un bon clinicien pourrait appeler SEP" (1983).

Dans un cas typique de SCI, l'examen du LCR est actullement discuté. Néanmoins, l'utilisation des données du LCR permet d’aider à écarter les diagnostics alternatifs tels que les infections (par exemple, la neuroborréliose ou la neurosyphilis). De plus, la détection de bandes oligoclonales (BOC) surnuméraires ou différentes de celles retrouvées dans le sérum offre un soutien supplémentaire pour le diagnostic de la SEP, même si leur présence n'est pas spécifique à la SEP. Un LCR peut être considéré comme «positif» s’il existe un index d’immunoglobulines G (IgG) élevé ou lors de la présence de BOC supplémentaires dans le LCR différentes de celles retrouvées dans le sérum (Freedman et al., 2005). Ainsi, un LCR positif reflète la nature auto-immune de la maladie. Dans les critères de Poser, les données du LCR peuvent être utilisées pour le diagnostic de la SEP. Dans les critères de McDonald de 2001 et de 2005, un LCR positif peut aussi être utilisé pour démontrer le critère de DIS en l’absence de critères de Barkhof visibles à l’IRM (≥2 lésions visibles à l’IRM et compatibles avec un diagnostic de SEP sont nécessaires en plus du LCR positif). Dans les critères révisés de McDonald 2010, les données du LCR ne sont pas incluses dans les critères de DIS.

11

En conclusion, un diagnostic de SEP nécessite la démonstration de la dissémination des lésions dans l'espace et dans le temps après un premier événement clinique évocateur de la SEP et l’absence de meilleures explications. Ces exigences peuvent être remplies par les données cliniques, seules ou avec l’aide des données de l'IRM et de l'examen du LCR. Les conclusions issues de l’IRM et, dans une moindre mesure, celles de l’étude du LCR ont été étudiées pour leur capacité à prédire un diagnostic de SEP-CD après un SCI typique. Néanmoins, aucune étude à ce jour n'a cherché à savoir si ces facteurs peuvent indépendamment prédire le développement d’une SEP-CD après un SCI typique sans sélection à priori des éléments cliniques et paracliniques (examen du LCR et de l'IRM). Les facteurs prédictifs précédents issus des différentes versions des critères de McDonald, tels que les critères dits de DIS, sont des combinaisons de facteurs identifiés par l'étude seule de l’IRM (Barkhof et al., 1997) ou après proposition par un groupe d’experts (Swanton et al., 2006). Bien que ces critères soient sensibles pour le diagnostic de la SEP, il est important de déterminer sans sélection a priori des paramètres à la fois cliniques, d’imagerie par IRM et issus de l’étude du LCR identifiés à partir d'une cohorte de patients ayant présenté un SCI typique.

2. Prédiction du pronostic de la sclérose en plaque s rémittente-récurrente

Chez les patients ayant un diagnostic établi de SEP-RR, le principal défi est d'évaluer la gravité de la maladie. Une des difficultés tient compte de l'hétérogénéité de l'activité et de l'évolution de la maladie. L'échelle «Expanded Disability Status Scale» (EDSS) est utilisée comme la référence pour évaluer le score d'invalidité des patients atteints de SEP depuis de nombreuses années (Kurtzke, 1983). Un score fonctionnel composite appelé «Multiple Sclerosis Functional Composite» (MSFC) est également utilisé dans de nombreux essais cliniques pour estimer la mesure de la progression de la maladie (Fisher et al., 1999; Rudick et al., 2001). Le critère de jugement utilisé dans les études est souvent basé à partir de l’utilisation des niveaux de l'échelle EDSS à atteindre ou de la «Disability Status Scale» (DSS), qui est une version simplifiée de l'échelle EDSS.

2.1. Etat actuel des connaissances

Il manque de facteurs prédictifs cliniques de l’invalidité à long terme détectables lors des premiers stades de la SEP. Des études épidémiologiques récentes suggèrent que le taux de progression de l'invalidité dans les premières années de la maladie est un facteur prédictif d'invalidité à long terme. La prédiction d'une telle progression à un stade plus précoce serait utile pour la gestion des malades.

Le substrat anatomique du handicap accumulé dans la SEP est surtout le reflet du cumul de la perte axonale, qui est un facteur déterminant des lésions et de l’atrophie cérébrale. Les troubles cognitifs ont été observés au début de la SEP, ont été associés à des marqueurs d’IRM reflétant l’atteinte cérébrale diffuse, même aux stades précoces de la maladie. Par conséquent, l’atteinte cognitive pourrait être un candidat intéressant comme marqueur pronostique détectable aux stades précoces de la SEP.

12

2.2. Atteinte cognitive comme marqueur pronostique au début de la SEP

Il a été suggéré que les mécanismes biologiques conduisant aux manifestations cliniques aiguës diffèrent de ceux qui responsables de l’invalidité irréversible à long terme (Bjartmar et al., 2001; Trapp et Nef, 2008). L’accumulation de la perte axonale est de plus en plus documentés, même au début de la maladie, à l’origine d’une invalidité permanente (Trapp et al., 1998). Ainsi, il est important de déterminer quels marqueurs pourraient être associés à des lésions axonales diffuses au niveau du cerveau. Nous avons supposé que les déficits cognitifs étaient de bons candidats cliniques, puisqu’il a été montré que l’atteinte cognitive était associée aux marqueurs d’IRM reflétant plus l’atteinte diffuse cérébrale que les lésions focales dans une cohorte de patients ayant une SEP-RR récemment diagnostiquée (Deloire et al., 2005).

2.2.1. Les troubles cognitifs dans la SP

2.2.1.1. Fréquence des troubles cognitifs dans la S EP

Charcot a rapporté la présence de troubles cognitifs dans la SEP il y a plus d'un siècle dans les «Leçons sur les maladies du système nerveux central» (1877). Mais, les déficits cognitifs ont longtemps été sous-estimés par les cliniciens et les chercheurs. Ces symptômes non moteurs ont été considérés comme faisant partie des stades les plus avancés de la SEP. Depuis 30 ans, ce sujet a suscité un intérêt croissant, et de nombreuses preuves sont en faveur de la présence fréquente des troubles cognitifs dans la SEP (revue par Langdon et al., 2011; Chiaravalloti et al., 2008; Brochet, dans Amato, 2011), même dès les premiers stades de la maladie. Les estimations de la fréquence de ces troubles cognitifs chez les patients ayant une SEP varient en fonction de la composition de l'échantillon, de la définition utilisée pour les rechercher, et la nature des tests neuropsychologiques (NP) inclus dans la batterie. Ainsi, les taux de fréquence varient entre 43 à 70% des premiers stades aux stades les plus avancés de la maladie (Rao et al., 1991a; Brochet, dans Amato 2011). Certaines études réalisées en centres hospitalo-universitaire ont rapporté une fréquence du dysfonctionnement cognitif entre 54 et 65% dans la SEP (Peyser et al., 1980; Lyon-Caen et al., 1986; Truelle et al., 1987). En revanche, Rao et al. ont rapporté un taux de fréquence des troubles cognitifs à 43% à partir d’une large batterie NP administrée dans un échantillon hétérogène de patients ayant une SEP (Rao et al., 1991a).

Les déficits cognitifs ont été rapportés à tous les stades et dans tous types de la maladie, y compris au décours du SCI, lors de la SEP-RR, SEP-SP, SEP-PP, et même au cours de la SEP dite bénigne (BSEP). Dans une cohorte locale de 44 patients atteints de SEP rémittente au début de la maladie, les troubles cognitifs ont été détectés dans 45% des cas dans les 6 mois après le diagnostic de la SEP (atteinte cognitive définie par ≥ 2 tests NP anormaux en dessous du cinquième percentile par rapport aux valeurs des sujets sains contrôles) (Deloire et al., 2005). Dans une étude multicentrique italienne de 550 patients atteints de SEP-RR, les troubles cognitifs ont été détectés chez 34,9% des patients (avec la même définition que ci-dessus), et chez 19,5% des patients avec une définition plus stricte (atteinte cognitive à ≥3 tests NP en-dessous du cinquième percentile des contrôles) (Patti et al., 2009). Les troubles cognitifs sont généralement moins fréquents (25 à 30%) et plus

13

focaux (Lyon-Caen et al., 1986;. Feinstein et al., 1992;. Callanan et al., 1989; Pelosi et al., 1997;. Achiron et Barak, 2003; Feuillet et al., 2007; Zipoli et al, 2010) chez les patients au stade de SCI évocateur de la SEP que chez les patients ayant un diagnostic posé de SEP-RR. Ces études suggèrent que les déficits cognitifs sont fragmentés lors des stades précoces de la maladie, et que le dysfonctionnement affecte principalement la vitesse de traitement de l'information (VTI), la mémoire de travail, l'attention, et la fluence verbale. Fait intéressant, il a été rapporté que la proportion de patients atteints de troubles cognitifs augmente considérablement dans les années suivant le SCI (Reuter et al., 2011), passant de 29% au stade de SCI à 54% 5 ans plus tard. Plus récemment, il a été démontré que les sujets atteints de syndromes radiologiques isolés (RIS) évocateurs de la SEP peuvent aussi présenter des troubles cognitifs, à ce stade préclinique de la maladie (Hakiki et al., 2008; Lebrun et al., 2010; Amato et al., 2012). Le profil de dysfonctionnement cognitif dans les 2 cohortes de sujets ayant un RIS était semblable à celui qui est communément observé chez les patients atteints de SEP-RR, et les troubles cognitifs ont été détectés chez un tiers des sujets (de 27,6% (Amato et al., 2012) à 30,8% (Lebrun et al., 2010)). En revanche, à des stades avancés de la maladie, mais avec un niveau d’incapacité physique considéré léger à l'aide de l'échelle EDSS, les troubles cognitifs ont été détectés chez 45% des cas à partir d'un groupe de 163 patients ayant une BSEP avec un score EDSS ≤3,0 après ≥ 15 ans d’évolution clinique de la maladie (Amato et al., 2006). En fait, la proportion réelle des patients BSEP pourrait être surestimée par l'absence d'évaluation des troubles cognitifs. Ces résultats soulignent la nécessité de considérer le fonctionnement cognitif chez les patients ayant une SEP pour évaluer la gravité de la maladie. Notamment, il a été proposé une modification de la définition de la BSEP avec inclusion de l'évaluation cognitive (Rovaris et al., 2009).

Contrairement à la SEP-RR, peu d'informations sont disponibles concernant le dysfonctionnement cognitif chez les patients ayant une forme de SEP-PP. La fréquence réelle et la nature de l’atteinte cognitive chez les patients ayant une SEP-PP ne sont pas déterminées. Ceci reflète principalement les défauts méthodologiques des études antérieures, avec l’utilisation d’échantillons hétérogènes de patients ayant une SEP en termes de formes cliniques ou de durée de la maladie, de groupes de sujets contrôles inadéquats, et des différences d'âge et de niveau d'éducation connus entre les patients ayant une SEP-RR et ceux ayant une SEP-PP. Par ailleurs, les relations entre les troubles cognitifs et les variables cliniques telles que la durée de la maladie, le niveau d'invalidité et le profil d’évolution de la maladie sont contradictoires (Langdon et al., 2011).

2.2.1.2. Type de déficits cognitifs dans la SEP

Différents domaines cognitifs peuvent être affectés chez les patients ayant une SEP: la VTI, l'attention, la mémoire épisodique et de travail, les fonctions exécutives, le langage (Brochet, dans Amato, 2011; Chiaravalloti et al., 2008.). La répartition des troubles cognitifs a été rapporté dans une méta-analyse incluant 57 études axées sur la cognition de 3891 patients atteints de SEP-RR selon la méthode de la taille d'effet (Prakash et al., 2008). Notamment, il a été établi que les patients ayant une SEP sont plus susceptibles d’avoir une atteinte aux tests de VTI comme le «Symbol-Digit-Modalities Test» (SDMT) et le «Paced Auditory-Serial-Addition-test» (PASAT), et que l’atteinte de la VTI est un déficit central dans la SEP (DeLuca et al., 2004; Forn et al., 2008).

14

2.2.2. Détection des déficits cognitifs dans la SEP

L'administration d'une large batterie NP effectuée par un neuropsychologue qualifié est la référence pour la détection des déficits cognitifs dans la SEP. La batterie NP la plus couramment utilisée dans la SEP est la celle qui a été proposée par Rao (1990), comprenant des tests de VTI, de l'attention, de la mémoire épisodique verbale et visuo-spatiale, et de fluence verbale. Une batterie NP de 90 minutes appelée batterie minimale pour l’évaluation des fonctions cognitives dans la SEP (MACFIMS) a émergé d’un récent consensus. Cette batterie est composée de 7 tests NP, couvrant les domaines cognitifs communément atteints dans la SEP (VTI/ attention, mémoire de travail, mémoire épisodique, fonctions exécutives, et langage) (Benedict et al., 2002.). Les facteurs de confusion comme la fatigue et la dépression peuvent influencer les performances cognitives, et doivent être pris en compte pour l'évaluation cognitive. Les plaintes cognitives ne reflètent pas les performances aux tests NP, mais sont plus associées aux symptômes dépressifs (Deloire et al., 2006;. Benedict et al., 2003; 2004a). Benedict et al. ont étudié la valeur d'un auto-questionnaire (MSNQ) rempli par les patients et leurs aidants (Benedict et al., 2003). En fait, l'estimation des capacités cognitives des patients par l'entourage était plus fiable que les autoévaluations des patients eux-mêmes (Benedict et al., 2003; 2004a).

Cependant, cette évaluation NP complète prend du temps, et n'est pas applicable dans la pratique clinique quotidienne. Le défi est de déterminer les tests NP pertinents à utiliser pour la détection du dysfonctionnement cognitif dans la SEP, et sélectionner les patients qui nécessitent une évaluation complémentaire par un expert. Le SDMT a été proposé comme un bon candidat pour la détection des troubles cognitifs par rapport aux autres tests NP dans un échantillon de patients atteints de SEP-RR au début (Deloire et al., 2006), et dans un échantillon mixte de patients atteints de SEP-RR et de SEP-SP (Parmenter et al., 2007). Ce test NP de substitution symboles/chiffres effectué en 90 secondes fait partie à la fois de la batterie BRB-N (Rao, 1990) et de la MACFIMS (Benedict et al., 2006a). Notamment, de nombreuses preuves existent en faveur de la bonne fiabilité de ce test (Benedict et al., 2008; Benedict et al., 2012a). Récemment, le SDMT a ainsi été proposé pour l'évaluation cognitive minimale dans la SEP pour estimer la VTI (BICAMS) (Langdon et al., 2012).

En plus de l'atteinte de la VTI, les troubles de la mémoire sont fréquents dans la SEP. Ainsi, en plus du SDMT, il a été recommandé d'inclure le California Verbal Learning Test-Second Edition (CVLT-II) (Delis et al., 2000) pour évaluer la mémoire épisodique verbale, et le BVMT-R pour l'étude rapide de la mémoire épisodique visuospatiale (Benedict, 1997) L'application de cette brève évaluation cognitive (BICAMS) doit être prise en compte dans les études de recherche et dans la pratique clinique des patients ayant une SEP.

2.2.3. Cognition et IRM

Le substrat pathologique des troubles cognitifs dans la SEP n'est toujours pas élucidé, et pourrait refléter les dommages à l'intérieur et à l'extérieur des lésions visibles au niveau de la substance blanche (SB) associés à l’atteinte de la substance grise (SG).

15

3. OBJECTIFS de cette thèse

Dans la première partie de cette thèse, l'objectif principal est d'identifier les marqueurs prédictifs du diagnostic de la SEP chez des patients ayant présenté un SCI typique. Après un SCI, même si le risque de SEP est non nul, tous les patients n’auront pas un diagnostic de SEP posé au cours du temps. Nous proposons d'étudier les données cliniques, du LCR et celles de l’imagerie par IRM au décours d’un SCI typique et évocateur de SEP afin de pouvoir identifier des marqueurs prédictifs précoces du diagnostic de la SEP.

Deux étapes seront abordées:

1. Notre premier objectif était de déterminer des marqueurs biologiques, cliniques et d’imagerie par IRM permettant de prédire le diagnostic de la SEP après un SCI dans un échantillon homogène de patients ayant présenté une myélite aigüe partielle (Article 1 ).

2. La capacité des 3 facteurs identifiés à partir de la cohorte des SCI médullaires (âge ≤ 40 ans, présence de BOC surnuméraires dans le LCR, et ≥3 lésions périventriculaires) pour prédire le diagnostic de la SEP a été évaluée dans un large échantillon de patients ayant eu un SCI indépendamment de la localisation anatomique des lésions (Article 2 ). Dans cette seconde étude, les facteurs prédictifs identifiés dans la première étude ont été testés dans une grande cohorte prospective de patients ayant eu un SCI affectant différentes topographies.

Dans la deuxième partie de cette thèse, l'objectif principal est d'évaluer la valeur pronostique de l’atteinte cognitive dans la SEP. Les déficits cognitifs ont été associés à l’atteinte cérébrale diffuse, et sont importants dans l'évaluation de la prise en charge des patients ayant une SEP. Nous proposons d’utiliser les troubles cognitifs en tant que marqueur pronostique fiable pour la prédiction du handicap physique à long terme et ses conséquences pour les patients vivant avec cette maladie aux multiples facettes.

Cinq étapes seront envisagées:

1. La première étape était de confirmer la relation entre les paramètres d’IRM reflétant l’atteinte diffuse du cerveau à un stade précoce de la SEP avec les déficits cognitifs et leur détérioration au cours du temps chez les patients atteints de SEP-RR (Article 3 ).

2. La deuxième étape était d’étayer la valeur pronostique de l'atteinte cognitive détectée tôt après le diagnostic de SEP-CD par sa capacité à prédire la progression du handicap physique au cours du temps chez les patients atteints de SEP-RR (Article 4 ).

3. Pour évaluer la valeur pronostique des troubles cognitifs, le fonctionnement cognitif a été comparé entre deux phénotypes cliniques de SEP dont le pronostic est connu comme étant différent (patients ayant une SEP-RR et ceux ayant une SEP-PP), en utilisant la même méthodologie et en prenant en compte les variables démographiques, la durée de la maladie, et le degré de handicap physique (Article 5 ).

16

4. Pour confirmer la valeur pronostique de l'atteinte cognitive, la valeur prédictive des troubles cognitifs sur la qualité de vie (QdV) et le statut professionnel a été évalué dans une étude longitudinale de patients ayant un diagnostic de SEP (Article 6 ).

5. Après avoir démontré la valeur pronostique des troubles cognitifs dans la SEP, le défi reste d'améliorer sa détection. Pour détecter un ralentissement de la VTI des patients ayant une SEP, nous allons dévopper un nouvel outil cognitif informatisé. Pour caractériser les critères psychométriques de ce test et le valider, nous avons effectué une étude de validation en utilisant différents échantillons de patients ayant une SEP et un large groupe de sujets sains contrôles (Article 7 ).

17

Chapitre 2:

Prédiction du diagnostic de la SEP après un syndrome clinique isolé

18

ARTICLE 1:

PREDICTIVE FACTORS FOR MULTIPLE SCLEROSIS IN PATIEN TS WITH CLINICALLY ISOLATED SPINAL CORD SYNDROME

Article accepted in Multiple Sclerosis Journal (Rue t et al., 2011a)

PREDICTIVE FACTORS FOR MULTIPLE SCLEROSIS IN PATIEN TS WITH CLINICALLY ISOLATED SPINAL CORD SYNDROME

Aurélie Ruet, MD Mathilde S.A. Deloire, PhD, Jean-Christophe Ouallet, MD, PhD Sandrine Molinier, MD Bruno Brochet, MD.

Services de Neurologie* et de Neuro-imagerie**, Hôpital Pellegrin, CHU de Bordeaux, France.

Corresponding author : Pr Bruno Brochet, Services de Neurologie* Hôpital Pellegrin, CHU de Bordeaux, 33076, Bordeaux cedex, France. bruno.brochet@chu-bordeaux.fr

Tel : +33556795521

Fax : +33556796025

The statistical analysis was performed by Mathilde Deloire and supervized by Regis Lassale (INSERM U657, University of Bordeaux 2 (Victor Segalen), Bordeaux, France).

Key words: Multiple Sclerosis, Myelitis, Clinically isolated Syndrome, MRI, CSF

Running Title: Predicting MS in cord syndromes

Erratum submitted

19

ABSTRACT

Objectives: To identify predictors of conversion to multiple sclerosis (MS) in patients with a cord clinically isolated syndrome.

Methods: The predictive values for conversion to MS of clinical, MRI and CSF variables in 114 patients with acute partial myelitis confirmed by a spinal cord lesion on MRI were studied. Other causes of cord syndromes were excluded.

Results: MS was diagnosed in 78 patients (68.4%) during 4.0 ± 1.9 years of follow-up. Sixty-seven of these patients (85.9%) had a second clinical episode. The diagnosis of isolated myelitis was maintained for 36 patients, 78% of whom (28 cases) were followed for at least 2 years, comparable to the MS patients. Age, bladder involvement, ≥ 2 cord lesions on MRI, ≥ 9 brain lesions, ≥3 periventricular lesions and intrathecal IgG synthesis predicted conversion to MS. Multivariate logistic analysis identified three predictors of MS diagnosis: age ≤ 40 years, inflammatory CSF and ≥3 periventricular lesions on brain MRI.

Conclusion: Two out of three baseline factors (age, periventricular lesions and inflammatory CSF) predicted conversion to MS with better accuracy than the revised McDonald criteria for dissemination in space.

20

Introduction

Relapsing-remitting multiple sclerosis (RRMS) frequently begins with a clinically isolated syndrome (CIS) characterised by an acute or subacute episode of neurological disturbance due to a single white-matter lesion located in the optic nerve, the brainstem or the spinal cord.1 One of the most frequent CIS lesion location is the spinal cord.1 Among patients with a CIS, a proportion will never have a second episode and cannot be diagnosed with clinically definite MS (CDMS). It may be very important for the patient to know the likelihood of developing a second episode and fulfilling the criteria for CDMS after a single episode. Previously, the International Panel on the Diagnosis of Multiple Sclerosis proposed criteria to establish the diagnosis of MS.2 Specifically, they proposed criteria to identify patients at high-risk for developing CDMS after presenting with a CIS. The proposal was later revised.3 These criteria underlined the value of magnetic resonance imaging (MRI) and incorporated a requirement for dissemination in space (DIS).4,5 Additionally, the recommendations included the use of cerebrospinal fluid (CSF) criteria in patients with ≥2 lesions on MRI. Another set of DIS MRI criteria has been proposed recently.6 Most studies designed to evaluate the utility of MRI or CSF criteria at the first episode included mixed CIS populations, with a majority of patients having optic neuritis. Little is known about the clinical, CSF and MRI predictors for diagnosing CDMS in patients with a spinal cord presentation. The purpose of this study was to determine the clinical, CSF and magnetic resonance imaging (MRI) characteristics observed during an initial episode of acute partial myelitis that may differentiate patients who convert to MS from those who do not.

21

Material and methods

Patients: Medical files of patients admitted to the department of Neurology of the University teaching hospital of Bordeaux (CHU) between 01/01/2000 and 03/01/2007 were screened using the hospital databases. All files of patients with a diagnosis of myelopathy, myelitis and encephalomyelitis were scrutinised for inclusion criteria. We also examined the files of patients seen at our outpatient MS clinic. Complete medical files of referred patients were retrieved from the hospital to which they were admitted during the acute phase of the disease. All patients were admitted as inpatients.

Inclusion criteria:

- Positive diagnosis of acute partial transverse myelopathy (APTM) modified from Ford and colleagues,7 (i) acute or subacute motor or sensory symptoms with or without sphincter dysfunction; (ii) occurrence of symptoms over no more than a 4-week period, sustained for at least 48 hours; (iii) neither clinical nor radiological evidence of spinal compression; (iv) no previous history of a neurological episode.

- Presence of at least one hyperintense T2 lesion on spine MRI fulfilling the description of Mc Donald’s criteria3: little or no swelling of the cord; unequivocally hyperintense lesion if detected with T2-weighted imaging; at least 3 mm in size and occupying only part of the cord cross section. We did not exclude lesions with longitudinal extension of cord lesion ≥ 2 vertebral segments if they occupy only part of the cord cross section.

Exclusion criteria:

- Vascular myelopathy suspected in patients whose symptoms reach maximal severity in <4 hours from onset according to published recommendations.8

- Severe bilateral transverse myelopathy characterised by complete paraplegia or tetraplegia because the focus of this study was acute partial myelitis and not acute transverse myelitis which is a very rare feature of MS.

- MRI abnormality suggestive of another cause of myelopathy (compression, arterio-venous fistulae, infarct in anterior spinal artery).

- History of spine trauma or irradiation.

- Other causes of myelitis, such as systemic disease, neuromyelitis optica according to Wingerchuk criteria9, or infection determined by extensive workup.10

- Previous documented neurological episode lasting more than 24h.

Initial clinical, laboratory and MRI investigations :

Initial data retrieved from medical files included demographics, clinical description of symptoms and signs, past and recent medical history (recent infection), the history of

22

previous transient neurological symptoms and the residual disability assessment using an expanded disability status scale (EDSS). 11 Brain and spine MRI scans were all reviewed. Cervical, dorsal and lumbar levels of the spinal cord were evaluated. In the sagittal plane, T1-weighted images (w.i.) with and without DTPA-gadolinium and T2-w.i were available. In many cases these sequences have been done in the axial plane if a lesion was detected in the sagittal plane. The number of lesions, their extent and localisation in the sagittal plane (number of vertebral levels, cervical, dorsal or lumbar levels) were assessed.

Brain MRI typically included a sagittal T1-w.i., an axial FLAIR and/or T2-w.i. and axial pre-and post-gadolinium T1-w.i.. Technical parameters varied according to sites. The following parameters were evaluated: presence of lesions with a diameter exceeding 3 mm and individual Barkhof/Tintoré DIS criteria4,5 as adapted in the McDonald criteria.2-3

The following CSF measures were analysed when available: CSF cell count, protein level, IgG-index and presence of oligoclonal bands. The techniques used for band detection varied according to sites (sensitised immunofixation or isoelectric focusing). Immunoglobulin G intrathecal synthesis (IgITS) was diagnosed by the presence of ≥ 2 bands not present in the serum and/or raised IgG index.

Follow-up:

Retrospective follow-up data were collected from medical files. General practitioners, hospital and/or community practicing neurologists were contacted to obtain the most recent data. An attempt was made to contact by phone all patients without a recent clinical follow-up. A diagnosis of MS could be made according to recent criteria in three different ways according to Polman et al.3:

(i) if a second clinical neurological episode occurred with clinical evidence of two or more lesions;

(ii) (ii) in case of two attacks but objective clinical evidence of only one lesion if DIS criteria are fulfilled demonstrated by MRI and these patients were identified as McDonald MS;

(iii) (iii) in case of one attack and objective clinical evidence of one lesion (CIS), if DIS and DIT criteria are fulfilled demonstrated by MRI and these patients were identified as McDonald MR MS.

None of the patients received disease-modifying therapy before clinical or MRI conversion to MS. Because the study was based on existing data and no additional burden was placed on the patient, no approval from the institutional review board was needed. Patients’ consent was sought orally at each follow-up appointment. According to French law, all patients admitted in our institution received information about the capture of personal data into the hospital database and their right to access these data. The hospital databases have received approval from the Commision Nationale Informatique et Libertés.

Statistics:

Statistic analysis was performed using STATVIEW ® 5.0 for Windows®.

23

Comparison of clinical, CSF and MR baseline data were made between patients fulfilling MS diagnostic criteria and patients retaining a diagnosis of isolated myelitis at the end of follow-up (if follow-up was ≥2 years) using univariate logistic analysis. The latter group was called CIS2y. Univariate logistic comparison tests were used to compare MS and CIS2y patients for predictability of an MS diagnosis at the end of follow-up.

Several multivariate logistic regression models were used to determine the predictive value of baseline parameters for diagnosis outcome (MS or CIS2y). Independent clinical factors with a conservative significance level of p<0.25 obtained in the univariate analysis were entered simultaneously into a clinical multivariate model. In addition, independent MRI parameters with p<0.25 were entered in a MRI multivariate model. Since only one biological parameter had a p value <0.25, no biological model was set-up. In each model, non-significant factors at the 0.05 level were removed by backward elimination. Odds-ratio (OR) were calculated.

Lastly, a general multivariate logistic regression was performed, including all independent factors with a conservative significance level of p<0.25 obtained in the univariate analysis.

ROC (receiver operating characteristic) curves were used to determine the optimal cutoff point for dichotomisation of the continuous covariates such as age at onset.

Finally, predictive values and accuracy for MS diagnosis of variables identified in the multivariate analysis were compared with the performance of the DIS revised McDonald criteria3 and the Swanton criteria.6 According to these latter criteria DIS was achieved with ≥1 T2 lesion in at least two of three locations defined as characteristic for MS: juxtacortical, periventricular, infratentorial but not spinal cord (CIS location).

Results:

From a total number of 234 selected files, 114 patients fulfilled inclusion criteria for APTM. The mean clinical follow-up duration of these 114 patients was 4.0 ± 1.9 years. All patients had baseline whole cord MRI and all except one had brain MRI. During follow-up, 75.4% had one or several additional cord MRI scans, whereas 90.4% had one or several additional brain MRI scans.

Outcome:

At the end of follow-up patients were classified as MS (78 patients) or isolated myelitis (36 patients). Among the 78 MS patients, 60 fulfilled criteria for CDMS, seven had a relapse affecting the cord and fulfilled McDonald MS criteria as defined above, and 11 patients fulfilled only dissemination in time criteria by MRI (McDonald MRMS). Among MS patients, the diagnosis was established in 78.2% during the first two years of follow-up. The mean diagnosis time for MS from onset was 1.3 ± 1.5 years. The mean follow-up of patients not fulfilling MS diagnostic criteria (36 patients) was 3.6 ± 1.9 years and 78% (28 patients) were followed-up more than two years (CIS2y). The 2-year-rate of multiple sclerosis conversion was 53.5% in the whole sample (n=61 among 114 myelitis).

24

Nine patients lost to follow-up and all had a CDMS diagnosis. No isolated myelitis patients called as CIS2y lost to follow-up.

One hundred six patients, including all MS patients (78) and CIS2y patients (28) were analysed for predictive factors. The mean duration of follow-up in these 106 patients was 4.3 ± 1.7 years without significant difference between CIS2y and MS patients.

Clinical variables:

There were no differences between the two groups with respect to previous history, particularly recent infections (25.0% in CIS2y; 10.4% in MS; p=0.06) and recent immunisation (3.8% in CIS2y, 1.3% in MS, p= 0.4). Other demographics and clinical characteristics are presented in Table 1. According to the univariate analyses, CIS2y patients were associated with older age at onset (OR= 0.92 [0.87-0.96]) and a more frequent sphincter involvement at onset (OR= 0.20 [0.06-0.71]), compared with MS patients. An ROC curve determined an age cut-off of 40 years.

Clinical variables with a p value <0.25 in the univariate models (Table 1) were entered into the multivariate logistic model. After backward elimination only age < 40 years was maintained in the model (p=0.0002; OR = 6.35 [2.42-16.65]).

25

Table 1: Demographics and clinical data in CIS 2y (n= 28) and MS (n= 78) at baseline.

Variables

CIS2y n=28

MS

n = 78

p value

Sex ratio F/M 2.5 5.0 0.18

Mean age ±±±± SD at onset (years) [range]

40 ± 9

[22-52]

32 ± 9

[15-56]

0.0007

Age ≤ 40 years 46.4 84.6 0.0002

Motor deficit at onset 16.7 28.2 0.15

Sensory deficit at onset 92.9 97.4 0.29

Sphincter dysfunction at onset 25 6.4 0.01

Motor deficit at peak 35.7 47.4 0.29

Sensory deficit at peak 100 100

Sphincter d ysfunction at peak 39.3 21.8 0.07

All results (except sex ratio and mean age) are expressed as the percentage of patients.

SD: standard deviation; F: female, M: male.

P values were calculated by univariate logistic analysis for prediction of final diagnosis (CIS2y

or MS).

26

Imaging variables:

Table 2 shows characteristics of cord and brain MRI scans in the study sample. All patients underwent whole cord MRI scans but only 76.4% (81 patients: 55 MS and 26 CIS2y) had post-gadolinium sequences. The only cord variable determined by univariate analysis as a predictor of MS diagnosis was the presence of at least two cord lesions (OR=2.83 [1.03-7.76] p=0.04). All except one patients underwent brain MRI with gadolinium contrast. Among individual Barkhof/Tintoré criteria detected on brain MRI, two were predictive of MS diagnosis: ≥ 9 lesions on T2 sequences (OR= 6.62 [1.46-30.08]) and ≥ 3 periventricular lesions (OR= 5.750 [1.82-18.18]) (Table 2).

Independent MRI variables with a p value <0.25 in univariate models (Table 2) were entered into the multivariate logistic model. Since the two parameters found to be significant in the univariate analyses were not independent (≥ 9 brain lesions and ≥ 3 periventricular lesions), two different models were constructed. In the first model, only the presence of ≥3 periventricular lesions on brain MRI was found to predict MS diagnosis (p=0.0029; OR = 5.75 [1.82-18.18]) and in the second model only the presence of ≥ 9 brain lesions predicted MS diagnosis (p=0.01; OR =6.62 [1.46-30.08] ).

27

Table 2: Cord and brain MRI characteristics in CIS 2y (n=28) and MS (n= 78) during the initial phase.

Variables

CIS2y

n=28

MS

n=78

p value

≥ 2 focal cord lesions 21.4 43.6 0.04

Enhanced cord lesion 53.8 54.5 0.9

Longitudinal extension of cord lesion

≥ 2 vertebral segments

25.9 16.7 0.3

≥ 1 cervical cord lesion (s) 64.3 78.2 0.15

≥ 1 thoracic and/or lumbar cord lesion (s) 42.9 50.0 0.52

Normal brain MRI 25.9 11.5 0.08

≥ 9 brain lesions 7.4 34.6 0.01

≥ 1 enhanced brain lesion 14.8 34.6 0.06

≥ 1 juxtacortical lesion 29.6 32.1 0.8

≥ 3 periventricular lesions 14.8 50.0 0.003

≥ 1 infratentorial lesion 7.4 25.6 0.06

All results are expressed as the percentage of patients.

P values were calculated by univariate logistic analysis for prediction of final diagnosis (CIS2y

or MS).

CSF parameters

CSF results were available for 95.3% (101 patients: 74 MS patients and 27 CIS2y).

No statistical difference was seen in the mean number of cells between the CIS2y group (23 ± 8) and the MS group (9 ± 9) (p=0.32), or for CSF protein levels (0.43 ± 0.16 in CIS2y and 0.41 ± 0.16 in MS patients) (p=0.54).

28

IgITS (presence of oligoclonal bands and/or raised IgG index) was more frequent in MS patients (78.4%) compared with CIS2y patients (51.9%) (OR=3.37 [1.32-8.58] p=0.01).

Final multivariate logistic analysis:

Three variables were found to be independent predictors of MS diagnosis: age (≤ 40 years) (OR=7.82 [2.51-24.39]; p=0.0004), IgITS (OR = 4.09 [1.32-12.67]; p=0.0147) and the presence of ≥3 periventricular lesions (OR= 6.52 [1.74-24.49]; p=0.005).

Logistic regression analysis showed that the predictive value of two out of the three criteria was highly significant (OR= 10.18 [3.71-27.96] p<0.0001).

Table 3 shows performance of these variables for predicting the diagnosis of MS compared with the performance of modified DIS criteria,3 as well as Swanton MR criteria.6

29

Table 3: Predictive values of MS diagnosis of diffe rent parameters (n= 106).

Variables Sensitivity

[IC 95%]

Specificity

[IC 95%]

PPV

[IC 95%]

NPV

[IC 95%]

Accuracy

[IC 95%]

Age

≤ 40 years

84.6

[76.6-92.6]

53.5

[35.5-71.5]

83.5

[75.5-91.5]

55.6

[36.6-74.6]

76.4

[68.4-84.4]

≥ 3 peri -

ventricular

lesions

50.0

[39.0-61.0]

85.1

[71.7-98.5]

90.7

[82.1-99.3]

37.1

[25.1-49.1]

59.0

[50.0-68.0]

IgITS 78.4

[69.0-87.8]

48.1

[29.3-66.9]

80.6

[71.6-89.6]

44.8

[26.8-62.8]

70.3

[61.3-79.3]

2 out of

the 3 variables

81.1

[72.2-90.0]

70.4

[53.2-87.6]

88.2

[80.6-95.8]

57.6

[40.7-74.5]

78.2

[70.2-86.2]

DIS Polman a 65.4

[54.9-75.9]

53.6

[35.1-72.1]

79.7

[69.7-89.7]

35.7

[21.7-49.7]

62.3

[53.3-71.3]

DIS Swanton b 42.3

[31.3-53.3]

71.4

[51.6-91.2]

80.5

[68.5-92.5]

30.8

[19.8-41.8]

50.0

[41.0-59.0]

IgITS: Immunoglobulin intrathecal synthesis (presence of oligoclonal bands and/or raised IgG index); DIS: dissemination in space; PPV = positive predictive value; NPV = negative predictive value.

a: dissemination in space demonstrated by MRI or two or more MRI-detected lesions consistent with MS plus positive CSF.

b: dissemination in space demonstrated with at least one T2 lesion in at least two out of three locations defined as characteristic for MS: juxtacortical, periventricular and infratentorial (but not spinal cord).

30

Discussion

Several recent studies have analysed the predictive values of baseline variables or previously proposed criteria for MS diagnosis. Most of these studies included limited samples of patients with a spinal cord CIS.4,6,12-18 Only two studies included a larger sample (>100) of CIS patients,14,19 but they did not analyse this group separately. Moreover, none of these studies included spinal cord MRI in their analysis; CSF analysis was performed by only a few.13,15,19 The present study was designed to determine which clinical, CSF and imaging variables collected during the acute initial stage could predict conversion to MS. We limited this study to patients with acute partial myelitis and excluded patients with transverse myelitis, which is an uncommon mode of onset of MS.10 We included only patients with the presence of MRI lesion in their spinal cord to ensure the validity of myelitis diagnosis. Therefore the presence or absence of one spinal cord lesion on MRI was not studied as a predictive factor. We only evaluated the predictive value of the following cord MRI parameters: presence of several cord lesions, location and longitudinal extension of cord lesions and gadolinium-enhancement.

The two-year rate of conversion to MS was 53.5% in the whole sample (61 among 114 myelitis), less than the conversion rate observed in the placebo-treated group of the CHAMPS trial (78.6%).20 In that study, MS could be diagnosed clinically or by detecting the development of ≥1 new or enlarging T2 lesions at 6, 12, or 18 months. In two prospective follow-up studies of ATPM, with shorter follow-up and varied diagnostic criteria, the rates of conversion to CDMS were 80% (n=15)7 and 57.7% (n=52).21

Among the clinical parameters analysed in our study, only a few significantly predicted MS outcome by univariate analyses: sphincter involvement at onset and age at onset. We did not confirm the previously reported highest frequency of sensory symptoms during the acute spinal cord syndrome in patients converting to MS. 21 Age of onset appears to be the most relevant clinical predictor, since it remains significant in the multivariate logistic regression analysis of clinical variables as well as in the multivariate analysis that includes biological parameters. To our knowledge this clinical variable has not been identified previously as an independent predictor of conversion from CIS to MS. It is well established that the mean age of onset of RRMS is around 30 years.22 The mean age of patients with a cord CIS who do not convert to MS appears to be >40, which is more similar to patients with acute transverse myelitis.9,23,24 We can assume from this result that studies of MRI diagnostic criteria in MS need to be adjusted according to the age of onset.

As expected, the Barkhof/Tintoré DIS criteria are helpful predictors of conversion to MS. Nine or more brain lesions and ≥3 periventricular lesions emerged as significant predictors but are not independent. The presence of ≥1 gadolinium enhancing lesion and ≥1 infratentorial lesion approached the level of significance. However, 15% of patients who did not convert to MS nevertheless had an enhancing lesion seen on brain MRI. In a recent paper about early MRI in optic neuritis, the multivariate analysis of baseline parameters revealed gender, periventricular and gadolinium-enhancing lesions as independent predictors of CDMS.25

Spinal cord MRI is essential for the diagnostic workup of patients with cord syndrome; however, little is known about the additional value of spinal cord MRI for predicting conversion to CDMS. Revised MS diagnostic criteria3 allow a spinal cord lesion to be

31

considered equivalent to a brain lesion for establishing DIS. Recently proposed new DIS criteria6 include cord lesions among the 4 sites to be considered, but specify that in cases of brain stem and spinal cord syndromes, all lesions within the symptomatic region must be excluded. It is therefore important to examine the predictive value of spinal cord MRI. Neither lesion level in the spinal cord nor lesion longitudinal length or lesion enhancement were predictive. The only variable predicting MS diagnosis was the presence of ≥2 cord lesions. When brain and cord MRI variables are included together in multivariate analysis, the cord variable loses its value as a significant independent predictor. This suggests that, although cord MRI is helpful for the positive diagnosis of APTM, it is less helpful for predicting conversion to MS.

Lastly, the present study showed that CSF analysis improves the prediction of MS outcome in patients with spinal cord CIS. Several studies previously showed the added value of CSF in CIS in general.13,15,19 The final multivariate analysis of the present study showed that CSF status is an independent predictor with very good accuracy for predicting MS diagnosis.

We have presently identified 3 variables, easily detectable in patients who have a spinal cord CIS, aged ≤ 40 years, ≥3 periventricular lesions and IgITS. An accurate prediction of MS conversion was obtained in 78.2% of patients when two out of the three variables were positive. In our cohort, positive predictive value, negative predictive value and accuracy of these factors were better than the previously proposed DIS criteria3,6

This study has some limitations. In particular the retrospective nature of this study could have induced recruitment bias.

These factors must be tested in a new prospective large cohort of CIS patients in order to establish their value as diagnostic criteria in clinical practice.

Acknowledgements:

No external funding was received for the design or the conduct of this study. We thank the clinical neurologists for referring patients.

Role of the funding source : No funding source.

32

References

1 Miller D, Barkhof F, Montalban X, et al. Clinically isolated syndromes suggestive of multiple sclerosis, part I: natural history, pathogenesis, diagnosis, and prognosis. Lancet Neurol 2005; 4:281–88.

2 McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001; 50:121–27.

3 Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald Criteria”. Ann Neurol 2005; 58:840–46.

4 Barkhof F, Filippi M, Miller DH, et al. Comparison of MRI criteria at first presentation to predict conversion to clinically definite multiple sclerosis. Brain 1997; 120:2059–69.

5 Tintoré M, Rovira A, Martinez M, et al. Isolated demyelinating syndromes: comparison of different MR imaging criteria to predict conversion to clinically definite multiple sclerosis. Am J Neuroradiol 2000; 21:702–06.

6. Swanton JK, Fernando KT, Dalton CM, et al. Modification of MRI criteria for multiple sclerosis in patients with clinically isolated syndromes. J Neurol Neurosurg Psychiatry 2006; 77:830–33.

7 Ford B, D. Tampieri D, Francis G. Long-term follow-up of acute partial transverse myelopathy. Neurology 1992; 42:250-2.

8 Transverse Myelitis Consortium Working Group. Proposed diagnostic criteria and nosology of acute transverse myelitis. Neurology 2002; 59:499-505.

9 Wingerchuk DM, Lennon VA, Pittock SJ, Lucchinetti CF, Weinshenker BG. Revised diagnostic criteria for neuromyelitis optica. Neurology 2006; 66:1485–89.

10 de Seze J, Stojkovic T, Breteau G, et al, Acute myelopathies: Clinical, laboratory and outcome profiles in 79 cases. Brain 2001; 124:1509-21.

11 Kurtzke, J. F. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444-52.

12 Dalton CM, Brex PA, Miszkiel KA, et al. Application of the new McDonald criteria to patients with clinically isolated syndromes suggestive of multiple sclerosis. Ann Neurol 2002; 52:47-53.

13 Tintoré M, Rovira A, Río J, et al. New diagnostic criteria for multiple sclerosis: application in first demyelinating episode. Neurology 2003; 60:27-30.

14 Tintoré M, Rovira A, Rio J, et al. Baseline MRI predicts future attacks and disability in clinically isolated syndromes. Neurology 2006; 67:968–72.

15 Villar LM, García-Barragán N, Sádaba MC, et al. Accuracy of CSF and MRI criteria for dissemination in space in the diagnosis of multiple sclerosis. J Neurol Sci 2008; 266:34-7.

33

16 Lo CP, Kao HW, Chen SY, et al. Prediction of conversion from clinically isolated syndrome to clinically definite multiple sclerosis according to baseline MRI findings: comparison of revised McDonald criteria and Swanton's modified criteria. J Neurol Neurosurg Psychiatry 2009; 80:1107-9.

17 Swanton JK, Rovira A, Tintoré M, et al. MRI criteria for multiple sclerosis in patients presenting with clinically isolated syndromes: a multicentre retrospective study. Lancet Neurol 2007; 6:677-86.

18 Korteweg T, Tintoré M, Uitdehaag B, et al. MRI criteria for dissemination in space in patients with clinically isolated syndromes: a multicentre follow-up study. Lancet Neurol 2006; 5:221-7.

19 Tintoré M, Rovira A, Río J, et al. Do oligoclonal bands add information to MRI in first attacks of multiple sclerosis? Neurology 2008; 70:1079-83.

20 Beck RW, Chandler DL, Cole SR, et al. Interferon beta-1a for early multiple sclerosis: CHAMPS trial subgroup analyses. Ann Neurol 2002; 51:481-90.

21 Cordonnier C, de Seze J, Breteau G, et al. Prospective study of patients presenting with acute partial transverse myelopathy. J Neurol 2003; 250:1447-52.

22 Confavreux C, Vukusic S, Adeleine P. Early clinical predictors and progression of irreversible disability in multiple sclerosis: an amnesic process. Brain 2003; 126:770–82.

23 de Seze J, Lanctin C, Lebrun C, et al. Idiopathic acute transverse myelitis: application of the recent diagnostic criteria. Neurology 2005; 65:1950-3.

24 Bruna J, Martínez-Yélamos S, Martínez-Yélamos A, et al. Idiopathic acute transverse myelitis: a clinical study and prognostic markers in 45 cases. Mult Scler 2006; 12:169-73.

25 Swanton JK, Fernando KT, Dalton CM, et al. Early MRI in optic neuritis: the risk for clinically definite multiple sclerosis. Mult Scler 2010; 16 (2): 156-165.

34

Discussion de cet article

1. Résumé des résultats de cet article

Nous rapportons l'une des séries les plus importantes de patients ayant présenté une myélite aiguë partielle (MAP) avec analyse rétrospective, sans a priori, de facteurs à la fois cliniques, biologiques, et d’imagerie par IRM pour prédire le diagnostic de SEP. Il a été identifié 3 facteurs prédictifs indépendants du risque d'avoir après une MAP un diagnostic de SEP: ≤40 ans, la présence d'une inflammation du LCR (présence de BOC et/ou élévation de l’index IgG), et ≥3 lésions périventriculaires visibles sur l'IRM cérébrale initiale. La présence d’au moins 2 de ces 3 facteurs permettait de prédire le diagnostic de SEP avec une meilleure exactitude que l’utilisation des critères actuels de DIS selon les critères révisés de McDonald.

2. Intégration des résultats aux données actuelles de la littérature

Une fois qu'un patient a présenté un SCI typique, il est important d'estimer la probabilité qu’il ait un diagnostic de SEP au cours du temps. A notre connaissance, aucune étude de grande envergure n’a jusqu'à présent déterminé les facteurs pouvant prédire le diagnostic de SEP après un SCI typique à partir d’un échantillon homogène de patients en analysant sans à priori les données cliniques et paracliniques (données du LCR et de l’IRM). Les précédents facteurs prédictifs issus des différentes versions des critères de McDonald, dits critères de DIS, sont une combinaison de marqueurs d’imagerie identifiés à partir de l’étude princeps de Barkhof et collaborateurs (Barkhof et al., 1997) ayant inclus uniquement des données d’IRM; puis de nouveaux critères IRM ont été proposés à partir d’avis d'experts (Swanton et al., 2006; Swanton et al., 2007). Les critères de DIS sont utilisés pour estimer la probabilité d'avoir une SEP après un SCI typique, et peuvent être considérés davantage comme des facteurs de risque de SEP plutôt que des critères diagnostic. Notre étude a révélé que la présence d’au moins 2 des 3 facteurs prédictifs identifiés était plus sensible, spécifique et précise pour prédire le diagnostic de SEP que les critères de DIS de 2005.

Les SCI médullaires sont fréquents parmi les SCI typiques évocateurs de SEP. Les études antérieures axées uniquement sur cette topographie ont souvent inclus des échantillons hétérogènes de patients (Bruna et al., 2006, de Seze et al., 2001; Cordonnier et al., 2003, Harzheim et al., 2004, Sellner et al., 2008; Debette et al., 2009). Les deux diagnostics les plus fréquents après un SCI typique médullaire sont la sclérose en plaques et la myélite idiopathique. Parmi les données cliniques, les facteurs prédictifs de SEP déjà identifiés sont le jeune âge (Bruna et al., 2006.), le sexe féminin (Bruna et al., 2006.), les symptômes sensitifs (de Seze et al., 2001.; Cordonnier et al., 2003), l'absence de choc spinal (Jeffery et al., 1993), un score EDSS initial élevé (Sellner et al., 2008), l'absence d'atteinte du système nerveux périphérique (Harzheim et al., 2004), et des antécédents familiaux de SEP (Sellner et al., 2008). Dans notre étude, après une analyse multivariée, le seul facteur prédictif clinique de SEP était l'âge avec un seuil calculé à ≤40 ans. Les caractéristiques cliniques, tels que le type et la sévérité des symptômes précédemment

35

identifiés, correspondent plus à la description du syndrome médullaire, et sont plus utiles pour distinguer une myélite aiguë partielle transverse (MAPT) d’une myélite aiguë transverse (MAT).

En 1992, Ford et collaborateurs ont introduit le terme de «myélopathie aiguë partielle transverse» (MAPT) pour décrire un groupe de patients présentant une «atteinte incomplète d'au moins un segment médullaire, avec un déficit moteur d’intensité légère à modérée, des symptômes sensitifs divers et dissociés, et pouvant être parfois associés à une atteinte vésico-sphinctérienne» (Ford et al., 1992). La plupart de ces patients avaient une IRM cérébrale anormale avec des lésions évocatrices de SEP, et ont eu un diagnostic de SEP-CD dans les 3 ans. Ces affections de type MAPT ou plus simplement myélite aiguë partielle (MAP) font partie des SCI médullaires les plus évocateurs de SEP.

Les MAT et les MAPT sont deux types distincts d’atteinte de la moelle épinière. Par définition, les deux syndromes se distinguent par l’étendue des dysfonctions au niveau de la moelle épinière. Les MAT font référence à une atteinte complète au niveau spinal, tandis que les MAPT se réfèrent à une atteinte d’intensité plus légère, le souvent unilatérale ou, si elles sont bilatérales, une atteinte asymétrique de la moelle épinière. Cette distinction est cliniquement significative et pertinente en termes de diagnostic étiologique, de pronostic et de prise en charge thérapeutique. En 2002, des critères de MAT ont été proposés par un consortium (TMCWG, 2002). Les MAPT représentent l’est des manifestations courantes de SEP, contrairement aux MAT qui sont une affection rare dans la SEP et doivent faire évoquer des diagnostics différentiels comme les autres maladies non-inflammatoires du SNC (par exemple: les infections, les atteintes vasculaires), d'autres maladies inflammatoires du SNC primaires ou secondaires (par exemple, les maladies systémiques comme la sarcoïdose, les vascularites, la maladie de Behcet, le lupus érythémateux disséminé; l’encéphalomyélite aiguë disséminée, la neuromyélite optique) et les pathologies idiopathiques démyélinisantes inflammatoires du SNC. Les patients atteints de MAT n'ont volontairement pas été inclus dans notre étude, puisque nous avons choisi d’axer l’étude sur les syndromes médullaires les plus évocateurs de SEP.

Les principales anomalies paracliniques associées à un risque accru de SEP-CD sont la présence d’une synthèse intrathécale d'IgG dans le LCR _ soit la présence de bandes oligoclonales (BOC) dans le LCR qui ne sont pas présentes dans le sérum, soit l’élévation de l’index IgG_ et la présence de lésions cérébrales multifocales au niveau de la substance blanche (SB) visibles sous forme d’hyperintensités sur les séquences IRM en T2. L'examen du LCR peut étayer le diagnostic positif de SEP en mettant en évidence la présence de BOC surnuméraires, et aider à rechercher des diagnostics différentiels. Plusieurs études ont rapporté que la présence de BOC dans le LCR augmente le risque de SEP après une myélite (Jeffery et al., 1993; Cordonnier et al., 2003; Sellner et al., 2008; Debette et al., 2009; Bourre et al., 2012).

Le potentiel de l'IRM de la moelle épinière pour étayer le diagnostic de SEP a récemment gagné beaucoup d'intérêt. La détection des lésions asymptomatiques de la moelle épinière est importante pour la prédiction du diagnostic de SEP et a été intégrée dans les critères révisés de DIS (Polman et al., 2005; 2011). De façon générale, les patients ayant présenté un SCI avec une IRM médullaire initiale anormale ont un risque élevé d’avoir un diagnostic de SEP-CD indépendamment des lésions cérébrales et de la présence de BOC dans le LCR (Patrucco et al., 2012). En outre, Okuda et collaborateurs ont souligné la valeur

36

des lésions asymptomatiques de la moelle épinière au stade de syndrome radiologiquement isolé («radiologically isolated syndrome»: «RIS») (2011). Néanmoins, l'IRM de la moelle épinière semble moins utile pour prédire le diagnostic de SEP après un SCI médullaire, même si cet examen reste toujours utile pour établir le diagnostic positif de MAP et l'élimination des diagnostics différentiels. Les nouveaux critères de DIS requièrent l'exclusion des lésions symptomatiques pour les SCI touchant notamment la moelle épinière (Polman et al., 2011).

La présence d’anomalies cérébrales visibles en IRM lors de l’épisode de myélite est fortement prédictive du diagnostic de SEP (Ford et al., 1992; Campi et al., 1995; Cordonnier et al., 2003; Sellner et al., 2008; Debette et al., 2009; Bourre et al., 2012). Lorsque les patients ayant présenté une MAP ont des lésions multiples au niveau cérébral évocatrices de SEP, le taux de transition vers une SEP-CD s’élève à au moins 80-90% dans les premières années suivant le SCI (Ford et al., 1992; Miller et al., 2008). Seules quelques études ont abordé spécifiquement le risque de SEP cliniquement certaine (SEP-CD) chez les patients ayant une IRM cérébrale normale au début de la MAP. Dans ce cas, le taux de transition vers une SEP-CD est estimé entre 10 à 30% dans les 5 années suivant le SCI (Bashir et al., 2000, Scott et al., 2005). En revanche, pour les patients ayant présenté une MAT avec une IRM cérébrale négative, le risque de SEP-CD semble très faible, avec un taux de transition qui serait ≤2% (Berman et al., 1981; Scott et al., 1998.). Dans notre étude incluant uniquement des MAP, nous avons observé que les patients avec une IRM cérébrale initiale normale avaient toutefois un risque non nul de SEP s'ils remplissaient au moins 2 des autres critères (âge ≤40 ans et LCR inflammatoire).

3. Défis et difficultés

Une des limites de cette étude provient de sa nature rétrospective. Une confirmation de nos résultats était donc nécessaire à partir d’une étude prospective et est présentée dans l’article suivant (Article 2 ).

Les critères d'inclusion de cette étude comportaient la démonstration d'au moins une lésion de la moelle épinière détectée par IRM pour assurer le diagnostic de myélite. Ainsi, la présence ou l'absence d'une lésion de la moelle épinière n'a pas été jugé pour prédire le diagnostic de SEP après la MAP. Dans les nouveaux critères proposés pour définir les MAPT, la démonstration du inflammation pouvant être visualisée par une prise de contraste au niveau de la lésion médullaire ou par un LCR inflammatoire n'est plus nécessaire, ce qui différe des critères diagnostiques de MAT idiopathique proposées par «the Consortium Working Group» (Scott et al., 2007). En outre, la région symptomatique doit être exclue en cas de SCI affectant la moelle épinière et le tronc cérébral d’après les critères de Swanton (Swanton et al., 2006). Ainsi, les lésions symptomatiques de la moelle épinière ne sont pas prises en compte dans la démonstration de la DIS dans ces cas.

Pour les rares cas dans lesquels il existait une suspicion clinique de myélite sans visualisation de la lésion à l’IRM médullaire, les patients n'ont pas été inclus dans cette étude. La répétition des examens par IRM peut aider à visualiser la lésion spinale. Comme mentionné ci-dessus, les critères de MAT idiopathiques incluent l'inflammation de la moelle épinière qui peut être démontrée par un rehaussement des lésions après injection de gadolinium, une pléiocytose du LCR ou un index IgG élevé. L’examen par IRM doit être

37

répété dans les 7 jours si le premier examen est considéré comme normal chez les patients ayant une suspicion de myélite. Le moment optimal de réalisation du contrôle IRM après l’épisode de myélite n'a pas été établi. Par ailleurs, la normalité d’une IRM médullaire a été associée à un bon pronostic chez des patients ayant présenté un épisode sévère de MAT (Scott et al., 1998).

Compte tenu de la relative courte période de suivi, l'absence d’établissement du diagnostic de SEP peut être en partie due à une activité moins marquée de la maladie au cours de la période d'observation plutôt que l'absence de SEP pour un patient donné. Un suivi suffisamment long nous permettra de préciser ou non si un diagnostic de SEP peut être posé après l’épisode de MAP. Dans notre étude, les facteurs identifiés étaient prédictifs d’un haut risque d'avoir un diagnostic précoce de SEP après un SCI médullaire (MAP). Même si la question est, en définitive, de savoir si un diagnostic de SEP peut être établi, le défi principal consiste à sélectionner les patients qui bénéficieront le plus d'un traitement spécifique de la maladie. Ainsi, l'identification précoce des patients à haut risque de SEP après un SCI médullaire a des implications importantes dans la gestion des patients, pour établir la fréquence des suivis et aider aux décisions thérapeutiques. L'initiation précoce d’un traitement de fond concerne principalement les patients atteints d'une maladie agressive, définie par un court intervalle entre les poussées, un taux élevé de poussées dans les premières années, et une progression rapide du handicap.

4. Conclusion et perspectives

En conclusion, ces 3 facteurs prédictifs indépendants, âge ≤40 ans, LCR inflammatoire, et ≥3 lésions périventriculaires, sont pertinents pour prédire un diagnostic précoce de SEP après une MAP. Ces facteurs ont cependant besoin d'être validés dans une étude prospective incluant un échantillon différent de patients avec différentes topographies des SCI.

Par conséquent, nous avons testé les performances de ces facteurs pour prédire un diagnostic précoce de SEP après un SCI typique dans une cohorte prospective comportant un large échantillon de patients ayant présenté un SCI de topographie différente. Nous avons également comparé les performances de ces facteurs prédictifs avec celles des critères actuels de DIS (Article 2 ).

38

ARTICLE 2:

EARLY PREDICTORS OF MULTIPLE SCLEROSIS AFTER A CLIN ICALLY ISOLATED SYNDROME

In preparation

EARLY PREDICTORS OF MULTIPLE SCLEROSIS AFTER A CLIN ICALLY ISOLATED SYNDROME

Aurélie Ruet1,2 MD, Georgina Arrambide3 MD, Bruno Brochet1,2 MD, Cristina Auger4 MD, Eva Simon3 BS, Àlex Rovira4 MD, Xavier Montalban3 MD, PhD, MarTintoré3MD, PhD

1 Université de Bordeaux, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, F-33076 Bordeaux, France 2 CHU de Bordeaux, INSERM-CHU CIC-P 0005, & Service de Neurologie, F-33076 Bordeaux, France 3 Neuroimmunology Center Unit, Multiple Sclerosis Center of Catalonia (CEM-Cat), Vall d’Hebron University Hospital and Research Institute, Universitat Autònoma de Barcelona, Barcelona, Spain 4 Magnetic Resonance Unit (IDI), Vall d’Hebron University Hospital Barcelona, Spain Corresponding author: Aurélie Ruet, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, case 78, 146 rue Léo Saignat, 33076 Bordeaux cedex, France. Tel: +33557574817 Fax: +33557574818 aurelie.ruet@hotmail.fr

39

Abstract

Objective— To assess factors, such as age less than 40 years, positive oligoclonal bands (OBs) in cerebrospinal fluid, and 3 or more periventricular (PV) lesions, for predicting clinically definite multiple sclerosis (CDMS) in patients with a clinically isolated syndrome (CIS) affecting different anatomical locations.

Design— A cohort of CIS patients with a median follow-up time of 45 months.

Patients— 652 consecutive CIS patients who presented a typical CIS regardless of the topography.

Interventions— CIS patients underwent regular neurological examinations and brain MRI within 3-5 months of symptom onset and at 1 and 5 years after.

Main Outcome Measures— The predictive value of having at least 2factors for conversion to CDMS was compared with the 2005 dissemination in space (DIS) criteria.

Results— CDMS was diagnosed in 201 patients (31%); 2005 McDonald multiple sclerosis, in 279 (43%). Adjusted hazard ratios for developing CDMS were 4.7 in patients with at least 2 predictive factors and 3.9 in patients with at least 3 Barkhof-Tintoré criteria (DIS 1). Patients who did not meet the DIS 1 criteria but displayed at least 2 predictive factors (N=82) had a 3-fold increased risk of a second attack. Having at least 2 predictive factors was more sensitive (83.2% versus 71.4%) and less specific (48.0% versus 61.1%) than DIS 1, and both criteria showed a similar level of accuracy (56.9% versus 63.7%) for predicting early CDMS.

Conclusions— Age, OBs, and PV lesions determined the risk of CDMS after a typical CIS especially in patients not fulfilling the DIS 1 criteria.

40

Introduction

In patients with a typical clinically isolated syndrome (CIS) suggestive of multiple sclerosis (MS), identification of those with a high risk of developing MS can be accomplished using magnetic resonance imaging (MRI) and examination of the cerebrospinal fluid (CSF) according to the different versions of the McDonald criteria.1–3 High-risk patients can be identified by the presence of at least 3 Barkhof-Tintoré criteria (BTC)4,5 or at least 2 lesions plus oligoclonal bands (OBs). To determine whether clinical, CSF, and imaging parameters could be combined to predict conversion to MS, Ruet et al analyzed a subgroup of patients presenting with spinal cord CIS and showed that an age of onset less than or equal to 40 years, inflammatory CSF, and presence of at least 3 periventricular (PV) lesions were independent predictive factors for MS.6 The presence of 2 of these factors predicted conversion to clinically definite multiple sclerosis (CDMS) with better accuracy than the 2005 dissemination in space (DIS) criteria. Limitations of this study included its retrospective nature and focus on only CIS cases involving the spinal cord. Thus, the main aim of the present study was to assess the performance of the proposed factors for predicting conversion to CDMS in a large prospective cohort of patients with CIS affecting different anatomical locations. The secondary objectives were to determine the added value of these factors in predicting conversion to MS in patients not fulfilling at least 3 BTC and assess the performance of these 3 factors in predicting MS according to the 2005 McDonald MS diagnosis criteria.

Materials and Methods

This study is based on longitudinal clinical, CSF, and MRI data acquired from a cohort of patients with CIS recruited consecutively at the Vall d’Hebron University Hospital in Barcelona between 2001 and 2011.

Selection of patients

Patients presenting with neurological symptoms of the types observed in MS for the first time were recruited at the Vall d’Hebron University Hospital in Barcelona, Spain. The inclusion criteria were as follows: a typical CIS involving the optic nerve, brainstem, spinal cord, or other anatomical area involving hemispheric, multifocal or undetermined locations that was not attributable to other diseases; less than 50 years in age; and onset of symptoms within 3-5 months of clinical, CSF, and MRI examinations. The exclusion criteria were as follows: alternative diagnoses (symptoms attributable to systemic diseases, neuromyelitis optica, and vascular or infectious diseases), reconsideration of CIS diagnosis if patients had previous neurological symptoms (diagnosis of relapsing/remitting MS), and death not related to MS.

41

This study received approval from the local ethicscommittee, and patients signed a written informed consent.

Clinical, CSF, and MRI assessments

Clinical data —Patients were initially asked about any previous history of neurological disturbances. Clinical data included sex, age, and presenting symptoms at baseline. Following the diagnosis of CIS, patients were seen every 3 to 6 months for clinical reassessment and in the event of a suspected relapse. According to the Poser criteria,7 the occurrence of a second clinical event in other parts of the central nervous system after an interval of at least 1 month indicated the development of CDMS only when other diagnoses had been excluded. CDMS was diagnosed when there was a second attack with a new neurological abnormality that was confirmed by examination. If this second event occurred, the time to CDMS was calculated as the interval between the onset of the first symptom and onset of the second event. The time of follow up was calculated as the interval between the date of the last visit and date of the CIS.

Serum and CSF analyses —Biological analyses were performed to exclude alternative diagnoses. IgG OBs were examined using agarose isoelectric focusing combined with immunoblotting.

MRI acquisition and analysis— Baseline brain MRI was performed within 5 months of disease onset and repeated at 1 and 5 years of follow up. MRIs were obtained on a 1.5-Tesla magnet between 2001 and 2010 and on a 3-Telsa since 2010 with a standard head coil. The following sequences of the brain were performed in each patient: 1) Transverse proton density/T2-weighted fast spin-echo; 2) Transverse T2-weighted fast-fluid-attenuated-inversion recovery; and 3) Transverse T1-weighted spin-echo [600/12/2 (TR/TE/ acquisitions)]. For all sequences, 46 interleaved contiguous axial sections were acquired with a 3-mm section thickness covering the whole brain, 192 x 256 matrix, and 188 x 250 mm field of view, giving an in-plane spatial resolution of approximately 1 x 1 mm. The transverse T1-weighted sequence was repeated in those patients who demonstrated focal white matter lesions on T2-weighted sequences after gadolinium injection (0.1 mmol/kg; scan delay, 5 minutes). The MRI scans were assessed by 2 neuroradiologists who were blinded to clinical follow up. Lesions on MRI scans were defined as areas of increased signal intensity larger than 3 mm intensity on both proton-density and T2-weighted images.

Baseline brain MRI scans were considered abnormalif at least one lesion suggestive of demyelination was observed in the proton density/T2-weighted images (except when only one infratentorial lesion was observed in CIS cases presenting as brainstem syndromes). The BTC4,5 were applied on baseline and followup brain MRI scans. The presence of new lesions and/or gadolinium-enhancing lesions at the follow up was also scored to establish dissemination in time (DIT) for MS diagnosis according to the 2005 McDonald criteria.2 According to the MRI component of the 2005 McDonald criteria,2 evidence of DIS was provided in 1 of 2 ways: presence of at least 3 BTC and/or presence of at least 2 T2 lesions plus OBs. DIT was fulfilled when there was a gadolinium-enhancing lesion at least 3 months after CIS onset or if at least one new T2 lesion appeared in the followup scan compared to a reference scan performed at least 1 month after the CIS. The MRI criteria for MS diagnosis

42

were met when patients fulfilled the MRI definitions for DIS and DIT. In addition, patients with a second clinical attack also fulfilled the 2005 DIT criteria. Classification of patients

Each patient was classified according to age (≤ 40 or > 40), CSF (presence or absence of OBs), PV lesions (≥ 3 or < 3 PV lesions), and combinations of 3 predictive factors. Each patient was also classified according to the 2005 DIS criteria in 3 ways. The first way to demonstrate DIS (DIS 1) was through the use of brain MRI scans alone. DIS 1 was positive only if the patient had at least 3 BTC4,5. The second way to demonstrate DIS was considered positive only if the presence of 2 lesions plus OBs was demonstrated (DIS 2). In the third option, a patient can fulfill the 2005 DIS criteria (DIS 3) if the brain MRI scans showed at least 3 BTC or 2 lesions plus OBs in CSF. The primary outcome was the clinical status (CDMS or not CDMS) at the end of the follow up. The secondary outcome was the MS diagnosis according to the 2005 McDonald criteria2

at the last evaluation.

Statistical analyses

Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL, USA) version 19.0. Nonparametric descriptive statistics were performed. The results are shown as the median and interquartile percentiles for continuous variables and as percentages for categorical variables.

The time from CIS to CDMS was analyzed using Kaplan-Meier survival curves in all CIS cases. For patients without a second event, time was recorded using the last known follow-up date. The main outcome in all analyses was conversion to CDMS according to the Poser criteria during the follow up. The secondary outcome was the establishment of a diagnosis of MS according to McDonald 2005 revised criteria. Aside from whole group survival, subgroup analysis was performed according to a number of criteria (age ≤ 40, OBs, and ≥ 3 PV lesions) and categorized as follows: 0/1, 2, and 3 criteria. Log-rank tests were used to compare these categories.

Cox proportional hazard regressions were performed to assess the risk of developing CDMS or MS according to the 2005 McDonald criteria in all CIS patients. Two models were used with or without covariates to adjust for the effects of the predictive variables. Covariates were CIS topography (categorical), sex, and disease-modifying drugs that were started before MS diagnosis (binary). As treatment was the only term with a significant effect, it was included as a covariate in a second Cox model to obtain an adjusted hazard ratio (aHR) for prediction of CDMS or MS by McDonald. Patients who received disease-modifying drugs after CDMS or MS by McDonald diagnosis were included in the group not receiving treatment.

Finally, the performance of the predictive factors was assessed by an analysis of the clinical status (CDMS or not CDMS) during the first 2 years in patients followed for at least 2

43

years. The numbers of true positives (patients fulfilling the criteria and developing CDMS), false positives (patients fulfilling the criteria but not developing CDMS), true negatives (patients not fulfilling the criteria and not developing CDMS), and false negatives (patients not fulfilling the criteria but developing CDMS) were determined. From this, the sensitivity (true positives/[true positives + false negatives]), specificity (true negatives/[true negatives + false positives],) and accuracy ([true positives + true negatives]/total tests), all with exact binomial 95% confidence intervals (CI), were calculated.

Results

Baseline characteristics of subjects and follow up

There were 652 patients included in this study. The median age was 31 (26–38), and the female/male ratio was 2:0 (Table 1). Thirty patients were excluded for various reasons, including the existence of previous neurological symptoms, age of onset older than 50, alternative diagnoses, and death not related to MS.

44

Table 1. Clinical, CSF and MRI Findings in the CIS Cohort.

652 patients

n (%)

401 patients followed for at least 2 years

n (%) Age at onset (years) 31 (26–38) 31 (26–38) Sex rati o

Male Female

2.0 220 (33.7) 432 (66.3)

2.2 125 (31.2) 276 (68.8)

Presenting clinical syndrome Optic neuritis Brainstem syndrome Spinal cord syndrome Other

232 (35.6) 180 (27.6) 165 (25.3)

75 (11.5)

127 (31.7) 115 (28.7) 110 (27.4)

49 (12.2)

Predict ive factors Age ≤ 40 536/651 (82.2) 331 (82.5) ≥ 3 PV 259/503 (51.5) 204 (56.8) +OBs 272/492 (55.3) 191 (58.8) ≥ 2/3 323/651 (49.6) 240 (59.9)

DIS 2005 criteria DIS 1 219/505 (43.4) 181/505 (35.8) DIS 2 49/253 (19.4) 23/174 (13.2) DIS 3 268/472 (56.8) 213/472 (45.1)

Follow up (months) 44.6 (13.7–76.0) 66.7 (51.1-92.7) Diagnosis of MS

CDMS 2005 McDonald MS

201 (30.8) 279 (42.8)

101 (25.2) 191 (47.6)

Data are the median (interquartile percentiles) or number (and percentage) of patients with clinically isolated syndromes (CIS). Abbreviations: CDMS, clinically definite multiple sclerosis; DIS 1, ≥ 3 Barkhof-Tintoré criteria; DIS 2, 2 lesions plus oligoclonal bands in patients not fulfilling ≥ 3 Barkhof-Tintoré criteria; DIS 3, Polman 2005 DIS (≥ 3 Barkhof-Tintoré criteria or 2 lesions with oligoclonal bands).

45

Diagnosis of multiple sclerosis

There were 201 patients (30.8%) that developed a second attack (CDMS) after a median conversion time of 13.1 months (9.8–16.5) (Table 1) and 279 patients (42.8%) had a 2005 McDonald MS diagnosis after a median conversion time of 7.4 months (4.6–10.2).

Baseline MRI scan

There were 343 out of 498 patients (68.9%) that had an abnormal baseline brain MRI scan. Among these, 166 patients (48.4%) had a CDMS diagnosis during the follow up, and 243 (70.8%) fulfilled the 2005 McDonald diagnostic criteria of MS. Of the 155 (31.3%) patients with no brain lesions suggestive of demyelination (except for those with one infratentorial lesion presenting with brainstem syndromes), 11 (7.1%) developed a second episode, and 12 (7.7%) fulfilled the 2005 McDonald diagnosis of MS.

Cerebrospinal fluid examination

There were 492 patients (75.5%) that had a lumbar puncture, and 272 patients (55.3%) were positive for IgG OBs (Table 1).

Disease-modifying drugs

During the study, 236 patients (36.2%) were on disease-modifying drugs, mainly 1 of 3 available beta-interferons. Of these patients, 136 out of 236 (57.6%) started treatment before the second attack, and 77 (32.6%) began treatment at the time of CIS.

Kaplan-Meier survival curves performed for all pati ents

Kaplan-Meier survival curves estimate the time from CIS to CDMS in all patients (Figure 1). The cumulative probability of CDMS was higher in patients having 2 or 3 predictive factors at baseline in comparison with patients having none or one predictive factors (P< .001). Kaplan-Meier curves showed the same result after accounting for the different CIS topographies (data not shown).

46

Figure 1. Time from CIS to CDMS according to baseli ne predictive factors.

Criteria are ≤ 40 years of age, positive oligoclonal bands in CSF and ≥ 3 periventricular lesions at the time of the CIS. Data were obtained from 651 CIS patients. Log Rank: P < .001 Abbreviations: CIS, clinically isolated syndrome; CDMS, clinically definite multiple sclerosis.

47

Cox regression models for all patients

Age, OBs, and PV lesions in predicting MS after CIS —HRs (95% CI) were calculated for each individual baseline predictive factor (age ≤ 40, OBs, and ≥ 3 PV lesions), for different combinations of predictive factors, and for having at least 2 predictive factors regardless of the combination (Table 2). aHRs are presented by considering the effect of disease-modifying drugs in patients starting treatment before the diagnosis of MS (CDMS in Table 2 and 2005 McDonald MS in Supplementary Table 1). Patients under 40 years of age had an aHR of 1.8 (1.2–2.7) for CDMS (P < .01). Having positive OBs or at least 3 PV lesions at baseline was associated with an aHR for developing CDMS of 2.3 (1.6–3.2) or 3.8 (2.6–5.5), respectively. Age at onset under 40 years old plus 3 PV lesions was associated with a higher aHR than OBs plus PV lesions [6.3 (2.6–15.6) versus 4.8 (2.9–8.0)] for developing CDMS. In patients with 3 predictive factors, the estimated aHR for CDMS was 5 times higher, and the aHR for the 2005 McDonald diagnosis was more than 7 times higher than in patients with fewer factors [5.0 (2.2–11.5), P< .001, and 7.4 (5.4–10.2), P < .001]. The estimated aHR was higher in patients with at least 2 predictive factors for MS than in patients having DIS 1 (4.7 (3.2–6.8) versus 3.9 (2.8–5.7); Table 2).

48

Table 2. Hazard Ratios of Baseline Factors and DIS Criteria for Prediction of CDMS.

Patients n (%) HR (95% CI) aHR1 (95% CI)

Predictive factors Age ≤ 40 536/651 (82.2) 1.8 (1.2–2.8)† 1.8 (1.2–2.7)* ≥ 3 PV 259/503 (51.5) 3.0 (2.2–4.2)‡ 3.8 (2.6–5.5)‡ +OBs 272/492 (55.3) 2.2 (1.6–3.1)‡ 2.3 (1.6–3.2)‡ Age ≤ 40 and ≥ 3 PV 212/503 (42.1) 5.3 (2.2–12.9)‡ 6.3 (2.6–15.6)‡ Age ≤ 40 and +OBs 222/492 (45.1) 2.8 (1.3–5.6)* 2.7 (1.3–5.6)* ≥ 3 PV and +OBs 171/407 (42.0) 3.9 (2.4–6.4)‡ 4.8 (2.9–8.0)† 0

48/652 (7.4)

12

12

1 281 /652 (43.1) 1.1 (0.4–2.5) 0.9 (0.4– 2.3) 2 182/652 (27.9) 3.7 (1.6–8.5)† 4.0 (1.8– 9.3)‡ 3 141/652 (21.6) 4.5 (2.0–10.3)‡ 5.0 (2.2–11.5)‡ ≥ 2/3 323/652 (49.5) 3.9 (2.8–5.5)‡ 4.7 (3.2–6.8)‡

DIS criteria DIS 1 219/505 (43.4) 3.1 (2.2–4.2)‡ 3.9 (2.8–5.7)‡ DIS 2 49/253 (19.4) 2.5 (1.4–4.4)‡ 2.6 (1.5–4.5)‡ DIS 3 268/472 (56.8) 3.8 (2.56–5.6)‡ 4.5 (3.0–6.8)‡

Predictive factors without DIS 1

≥ 2/3

82/286 (28.7)

2.8 (1.7–4.7)‡

3.0 (1.8–5.1)‡

0, 1, 2 or 3 are different combinations of predictive factors regardless of the type of the combination. aHR1 is adjusted HR, obtained after adjustment for treatment initiated before clinically definite multiple sclerosis diagnosis. The data for the aHRs were calculated according to the number of patients with available treatment information. 12: Reference Abbreviations: PV, periventricular lesions; +OBs, presence of oligoclonal bands; DIS 1, ≥ 3/4 Barkhof-Tintoré criteria; DIS 2, 2 lesions plus oligoclonal bands in patients not fulfilling ≥ 3/4 Barkhof-Tintoré criteria; DIS 3, DIS according to McDonald 2005 (≥ 3/4 Barkhof-Tintoré criteria or 2 lesions plus oligoclonal bands). P value: * < .01; † < .005; ‡ ≤ .001.

49

eTable 1. Hazard Ratios of Baseline Factors and DIS Criteria for Prediction of 2005 McDonald MS.

Patients n (%) HR (95% CI) aHR1 (95% CI) Predictive factors

Age ≤ 40 536/647 (82.8) 1.6 (1.1–2.3)* 1.6 (1.1–2.2)* ≥ 3 PV 259/503 (51.5) 5.2 (3.9–7.0)‡ 6.7 (4.9–9.1)‡ +OBs 272/492 (55.3) 3.3 (2.5–4.5)‡ 3.6 (2.6–4.9)‡ Age ≤ 40 and ≥ 3 PV 212/503 (42.1) 7.2 (3.5–14.8)‡ 8.7 (4.3–17.9)‡ Age ≤ 40 and +OBs 222/492 (45.1) 3.4 (1.9–6.4)‡ 3.5 (1.9–6.6)‡ ≥ 3 PV and +OBs 171/407 (42.0) 7.7 (5.0–11.9)‡ 9.7 (6.2–15.2)‡ 0

48/647 (7.4)

12

12

1 280/647 (43.3) 1.1 (0.5–2.5) 1.1 (0.5–2.4) 2 182/647 (28.1) 5.7 (2.7–12.3)‡ 6.3 (2.9–13.6)‡ 3 141/647 (21.8) 9.2 (4.3–19.7)‡ 10.0 (4.6–21.5)‡ ≥ 2/3 323/647 (49.9) 6.3 (4.6–8.5)‡ 7.4 (5.4–10.2)‡

DIS criteria DIS 1 219/505 (43.4) 6.3 (4.8–8.3)‡ 8.3 (6.2–11.1)‡ DIS 2 49/253 (19.4) 4.7 (2.9–7.6)‡ 4.8 (2.9–7.8)‡ DIS 3 268/472 (56.8) 8.4 (5.8–12.0)‡ 10.1 (7.0–14.6)‡

Predictive factors without DIS 1

≥ 2/3

82/286 (28.7)

3.4 (2.1–5.3)‡

3.6 (2.2–5.7)‡

0, 1, 2 or 3 are different combinations of predictive factors regardless of the type of the combination. aHR1 is adjusted HR obtained after adjustment for treatment initiated before 2005 McDonald multiple sclerosis diagnosis. The data for the aHRs were calculated according to the number of patients with available treatment information. 12: Reference Abbreviations: PV, periventricular lesions; +OBs, presence of oligoclonal bands; DIS 1, ≥ 3/4 Barkhof-Tintoré criteria; DIS 2, 2 lesions with oligoclonal bands in patients not fulfilling ≥ 3/4 Barkhof-Tintoré criteria; DIS 3, DIS Polman 2005 (≥ 3/4 Barkhof-Tintoré criteria or 2 lesions with oligoclonal bands). P value: * < .01; ‡ ≤ .001.

50

Application of the predictive factors in patients n ot fulfilling at least 3 BTC

Patients with at least 2 predictive factors without DIS 1 at baseline had a higher risk of developing a second attack [aHR 3.0 (1.8–5.1), P < .001]. Of the individuals tested, 191 patients that were followed up for at least 2 years and did not meet DIS 1 criteria; 66 (34.6%) of them presented at least two predictive factors (Figure 2). Of those, 16 (24.2%) developed a second attack in the first 2 years.

Figure 2. Predictive factors in CIS patients not fu lfilling at least 3 BTC.

Predictive factors are ≤ 40 years of age, positive oligoclonal bands and ≥ 3 periventricular lesions.

CIS, clinically isolated syndrome; BTC, Barkhof-Tintoré criteria; CDMS, clinically definite multiple sclerosis.

51

Performance of the predictive factors for MS in the first 2 years after CIS

There were 401 CIS patients (61.5%) that were followed for at least 2 years, and their data were used for the performance analysis of the predictive factors and DIS criteria. The clinical, CSF, and MRI characteristics of the CIS cohort followed for at least 2 years were similar compared to those of the whole cohort (Table 1; chi-square test). A total of 125 patients (62.2%) had a CDMS diagnosis within 2 years of their CIS.

The presence of at least 2 predictive factors (age ≤ 40, OBs, ≥ 3 PV) was more sensitive (83.2% versus 71.4%) and less specific (48.0% versus 61.1%) than DIS 1 for early CDMS, whereas the accuracy was similar (56.9% versus 63.7%) (Table 3). In patients not fulfilling DIS 1, having at least 2 predictive factors was more sensitive (61.5% versus 41.7%) and less specific (69.7% versus 78.7%) than meeting DIS 2 criteria, whereas a similar accuracy was observed (68.6% versus 73.6%). The specificity of at least 2 predictive factors was higher in patients not fulfilling DIS 1 than in patients regardless of DIS1 (69.7% versus 37.3%) without compromising the sensitivity (61.5% and 76.7%, respectively). The accuracy of at least 2 predictive factors was higher in patients without DIS 1 than in patients regardless of DIS1 (68.6% versus 52.7%).

The performance of the predictive factors was similar for the 2005 McDonald MS and CDMS criteria (Supplementary Table 2). The sensitivity of having at least 2 predictive factors without DIS 1 was 68.3%, the specificity was 74.7%, and the accuracy was 73.3%.

52

Table 3. Performance of Predictive Factors for CDMS in CIS Patients Followed for at Least 2 Years

Estimate (95% CI)

Sensitivity Specificity Accuracy Predictive factors

Age ≤ 40 86.1 (77.8–92.2)

18.7 (14.4–23.5)

35.7 (31.0–40.6)

+OBs 72.2 (61.8–81.1)

46.4 (39.9–53.0)

53.5 (48.0–59.1)

≥ 3 PV 78.9 (69.0–86.8)

50.6 (44.4–56.7)

57.7 (52.4–62.8)

≥ 2/3 83.2 (74.4–89.9)

48.0 (42.2–53.8)

56.9 (51.9–61.8)

DIS criter ia

DIS 1 71.4 (61.0–80.4)

61.1 (55.0–67.0)

63.7 (58.5–68.7)

DIS 2 41.7 (22.1–63.4)

78.7 (71.2–84.9)

73.6 (66.4–80.0)

DIS 3 84.3 (75.0–91.1)

46.3 (40.0–52.6)

56.1 (50.7–61.4)

Predictive factors without DIS 1

≥ 2/3 61.5 (40.6–79.8)

69.7 (62.1–76.6)

68.6 (61.5–75.1)

Abbreviations: CDMS, clinically definite multiple sclerosis; CIS, clinically isolated syndrome; 95% CI, 95% confidence interval; PPV, positive predictive value; NPV, negative predictive value; OBs, presence of oligoclonal bands; PV, periventricular lesions; DIS criteria, dissemination in space criteria; DIS 1, ≥ 3/4 Barkhof-Tintoré criteria; DIS 2, 2 brain lesions plus oligoclonal bands in patients not fulfilling ≥ 3/4 Barkhof-Tintoré criteria; DIS 3, Polman 2005 DIS (≥ 3/4 Barkhof-Tintoré criteria or 2 lesions plus oligoclonal bands).

53

eTable 2. Performance of Predictive Factors for 200 5 McDonald MS in CIS Patients Followed for at Least 2 Years.

Abbreviations: CIS, clinically isolated syndrome; 95% CI, 95% confidence interval; PPV, positive predictive value; NPV, negative predictive value; OBs, oligoclonal bands; PV, periventricular lesions; DIS criteria, dissemination in space criteria; DIS 1, ≥ 3/4 Barkhof-Tintoré criteria; DIS 2, 2 brain lesions plus oligoclonal bands in patients not fulfilling ≥ 3/4 Barkhof-Tintoré criteria; DIS 3, DIS Polman 2005 (≥ 3/4 Barkhof-Tintoré criteria or 2 lesions plus oligoclonal bands).

Estimate(95% CI) Sensitivity Specificity Accuracy Predictive factors

Age ≤ 40 86.9 (81.3- 91.3)

21.4 (16.1- 27.6)

52.6 (47.6- 57.6)

+OBs 75.9 (68.6- 82.3)

58.3 (50.3 - 65.9)

67.1 (61.7- 72.2)

≥ 3 PV 84.4 (78.3- 89.4)

70.9 (63.7- 77.5)

77.7 (73.1- 81.9)

≥ 2/3 88.0 (82.5- 92.2)

65.7 (58.9- 72.1)

76.3 (71.8- 80.4)

DIS criteria

DIS 1 77.3 (70.6- 83.2)

83.3 (77.1- 88.5)

80.3 (75.9- 84.3)

DIS 2 59.0 (42.1- 74.4)

85.9 (78.9- 91.3)

79.9 (73.2- 85.6)

DIS 3 91.1 (85.9- 94.8)

70.3 (62.7- 77.2)

81.1 (76.6- 85.1)

Predictive factors without DIS1

≥ 2/3 68.3 (51.9- 81.9)

74.7 (66.9- 81.4)

73.3 (66.4- 79.4)

54

Discussion

Once a typical CIS occurs, it is important to estimate the risk of developing a second attack. In this large prospective cohort of patients with CIS affecting different topographies, we report that age at onset, OBs in the CSF, presence of 3 or more PV lesions and, more importantly, the combination of at least 2 of these factors were associated with an increased risk of developing MS.

Kaplan-Meier curves clearly differentiate 2 groups of CIS patients with a low and high risk of developing CDMS (Figure 1). Patients with zero or one predictive factors have a low risk of developing CDMS, whereas patients with at least 2 predictive factors will likely develop CDMS in a short time. Thus, the cutoff established in the Ruet et al study6 for the predictive factors (the presence of 2 or more predictive factors) appears to be useful in clinical practice. Nevertheless, patients fulfilling zero or one predictive factors occasionally develop CDMS after a longer follow up.

Predictive factors, age, OBs, and 3 PV lesions appear clinically meaningful in identifying CIS patients with a high risk of developing MS. Having at least 2 predictive factors at baseline increased the risk of developing CDMS 5-fold in comparison to patients with zero or one predictive factors.

In this cohort with different types of CIS, we confirmed that being under 40 years old was associated with an increased risk of MS (CDMS and 2005 McDonald MS), which has been previously shown in spinal cord syndromes.6 A large epidemiological study8 reported that MS preferentially affects young people with an average age of onset that is typically less than 40 years old.

We also confirmed the relevance of CSF findings in estimating the risk of a second relapse in inflammatory demyelinating diseases, such as MS. In our study, patients with positive OBs had an aHR of 2.3 for developing CDMS. This is consistent with the results of a study of 415 CIS patients in whom the risk of CDMS increased 1.7 times with the presence of OBs and was independent of MRI abnormalities.9 As recommended by the 2005 McDonald diagnostic criteria, positive OBs were also useful when the baseline MRI was equivocal or did not demonstrate at least 3 BTC.

Although CSF analysis adds useful information, MRI findings remain the main prognostic factor in this disease in which most visible lesions affect the white matter. MRIs appear as the most informative surrogate marker in predicting MS after a typical CIS. The number of brain lesions and number of BTC have been associated with an increased risk of CDMS in comparison with patients with normal brain MRIs.10–13 Interestingly, the PV lesion criterion has been used in all previous proposed MS criteria. Paty’s criteria14 required at least 4 lesions or 3 lesions, with one lesion being PV. Fazeka’s criteria15 required 3 lesions with 2 of the following properties: one PV lesion, an infratentorial lesion, or a lesion of at least 6 mm. In Barkhof’s criteria,4 the cutoff of at least 3 PV lesions best predicted conversion to

55

CDMS. In our study, the estimated aHR of 3 PV lesions was closer to the aHR of at least 3 BTC (3.8 versus 3.9). Interestingly, 3 PV lesions in patients 40 or younger were associated with a higher aHR compared to the presence of at least 3 BTC (6.3 versus 3.9).

We confirmed the utility of the 3 predictive factors for identifying patients at high risk of developing an early second attack because the sensitivity of the combination of at least 2 predictive factors was higher than the sensitivity of at least 3 BTC (83.2% versus 71.4%). We also extended the results of Ruet et al (2011)6 in this prospective CIS cohort with different topographies and confirmed the good predictive value of these factors for developing 2005 McDonald MS. Moreover, we illustrated the additional value of these 3 predictive factors in CIS patients when their MRI scans do not fulfill at least 3 BTC. The predictive factors studied in this work, as well as the established DIS criteria, are prognostic factors for patients with typical CIS. Formal MS diagnosis still needs the additional demonstration of DIT.

Disease-modifying drugs have consistently demonstrated a reduction in the relapse rate after a CIS. The influence of this covariate was accounted for in the Cox regression models, and the clear effect of the predictive factors on the risk of developing MS was slightly higher after adjusting for treatment.

There are several limitations to our study. In the inclusion criteria, the maximum age was less than 50 years. Thus, the specificity and accuracy of the factor “age ≤ 40” could have been underestimated. Moreover, MAGNIMS DIS criteria16 were not analyzed in this cohort. Future studies could assess the performance and benefit of our predictive factors in comparison to these criteria.

In conclusion, we propose that these 3 predictive factors should be applied after a typical CIS to assess the risk of developing CDMS. Our data demonstrate their good performance in clinical practice in predicting the occurrence of a second attack after a typical CIS and their added value in patients not fulfilling at least 3 BTC. Early identification of CIS patients at high risk of having a second attack is important for managing the therapeutic strategy.

56

Acknowledgement

The authors thank Xavier Vidal for statistical analyses (Department of Pharmacology, Hospital Vall d’Hebron, Barcelona, Spain).

References

1. McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for multiple sclerosis:guidelines from the International Panel on the Diagnosis of Multiple Sclerosis. Ann Neurol.2001;50:121-127.

2. Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions tothe‘‘McDonald Criteria.’’ Ann Neurol.200;58:840-846.

3. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: 2010 Revisions to the McDonald Criteria. Ann Neurol.2011;69:292-302.

4. Barkhof F, Filippi M, Miller DH, et al. Comparison of MRI criteria at first presentation to predict conversion to clinically definite multiple sclerosis. Brain.1997;120:2059-2069.

5. Tintoré M, Rovira A, Martinez M, et al. Isolated demyelinating syndromes: comparison of different MR imaging criteria to predict conversion to clinically definite multiple sclerosis. AJNR Am J Neuroradiol. 2000;21:702-706.

6. Ruet A, Deloire MS, Ouallet JC, et al. Predictive factors for multiple sclerosis in patients with clinically isolated spinal cord syndrome.Mult Scler. 2011;17(3):312-318.

7. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research proposals. Ann Neurol.1983;13:227-231.

8. Confavreux C, Vukusic S and Adeleine P. Early clinical predictors and progression of irreversible disability in multiple sclerosis:an amnesic process. Brain. 2003;126:770-782.

9. Tintoré M, Rovira A, Río J, et al. Do oligoclonal bands add information to MRI in first attacks of multiple sclerosis? Neurology.2008;70:1079-1083.

57

10. Tintoré M, Rovira A, Río J, et al. New diagnostic criteria for multiple sclerosis: application in first demyelinating episode. Neurology. 2003;60:27-30.

11. Tintoré M, Rovira A, Río J, et al. Baseline MRI predicts future attacks and disability in clinically isolated syndromes. Neurology.2006;67:968-972.

12. Korteweg T, Tintoré M, Uitdehaag B, et al. MRI criteria for dissemination in space in patients with clinically isolated syndromes: a multicentre follow-up study. Lancet Neurol. 2006;5:221-227.

13. Swanton JK, Rovira A, Tintoré M, et al. MRI criteria for multiple sclerosis in patients presenting with clinically isolated syndromes: a multicentre retrospective study. Lancet Neurol.2007 6:677-686.

14. Paty DW, Oger JJ, Kastrukoff LF, et al. MRI in the diagnosis ofMS: a prospective study with comparison of clinical evaluation,evoked potentials, oligoclonal banding, and CT. Neurology 1988;38:180-185.

15. Fazekas F, Offenbacher H, Fuchs S, et al. Criteria for an increased specificity of MRI interpretation in elderly subjects with suspected multiple sclerosis. Neurology. 1988; 38:1822-1825.

16. Swanton JK, Fernando KT, Dalton CM, et al. Modification of MRI criteria for multiple sclerosis inpatients with clinically isolated syndromes. J Neurol Neurosurg Psychiatry.2006;77:830-833.

58

Discussion de cet article

1. Résumé des résultats de cet article

Dans une grande cohorte prospective de 652 patients avec un SCI affectant différentes topographies, nous avons confirmé la performance des 3 facteurs prédictifs pour le diagnostic de SEP: l'âge de début (≤40 ans), la présence de BOC dans le LCR, et la présence d’au moins 3 lésions périventriculaires. Un patient ayant au moins 2 de ces 3 facteurs lors du SCI présente un risque accru d'avoir un diagnostic précoce de SEP (SEP-CD et diagnostic de SEP selon les critères de McDonald 2005).

2. Intégration des résultats aux données actuelles de la littérature

L’âge au début du SCI peut être utilisé comme un facteur prédictif du diagnostic de SEP après un SCI typique. Dans cette étude, avoir ≤ 40 ans était associé à un risque accru de SEP (SEP-CD et SEP selon les critères de McDonald 2005); cette association avait déjà été démontré pour les SCI médullaires (Ruet et al., 2011a.). Cette constatation correspond à l'âge moyen de début de la SEP, qui est de 30 ans avec un intervalle de 20 à 40 ans, même si les patients peuvent présenter un SCI typique évocateur de SEP à des âges plus extrêmes (Confavreux et al., 2003). Dans une étude d’IRM multicentrique, il n'y avait pas d'association entre l'âge et le risque de SEP (Swanton et al., 2007). Dans une autre étude qui portait uniquement sur les SCI de type névrite optique, le genre, mais pas l’âge, était un facteur prédictif indépendant du diagnostic de SEP-CD après une analyse multivariée (Swanton et al., 2010).

Parmi les facteurs paracliniques prédictifs du diagnostic de SEP, les plus importants sont la présence de lésions en IRM cliniquement silencieuses et de BOC dans le LCR différentes du sérum.

Plusieurs études ont rapporté que les anomalies cérébrales observées sur les examens initiaux d’IRM lors du SCI ont une valeur pronostique avec établissement précoce d’un diagnostic de SEP (Brex et al., 2002, ONTT, 1997; ONTT, 2003; ONTT 2008; Tintoré et al., 2005; Tintoré et al., 2006, Tintoré et al., 2010;. Minneboo et al., 2004; Korteweg et al., 2006; Korteweg et al., 2009;. Swanton et al., 2007; Fisniku et al., 2008).

Le critère d’imagerie «≥3 lésions PV» fait partie des critères de Barkhof. Une lésion périventriculaire (PV) est également incluse parmi plusieurs critères d'IRM pour le diagnostic de SEP (Paty, Fazekas, Barkhof-Tintoré et Swanton). Fazekas et collaborateurs ont mis en évidence cette localisation comme étant spécifique dans la SEP. L'inclusion d'une lésion PV peut augmenter la spécificité, sans diminuer la sensibilité, pour l'établissement du diagnostic chez les patients âgés présentant une suspicion de SEP (Fazekas et al., 1988). La localisation des lésions est importante pour déterminer leur probabilité à être en lien avec une SEP. La différenciation entre les lésions périventriculaires et paraventriculaire est utile pour le diagnostic différentiel de la SEP (ex pour les pathologies vasculaires telle que la microangiopathie). Pour les patients avec apparition tardive de la SEP (après 50 ans), la spécificité des critères de Fazekas était d'environ 69% (de Seze et al., 2005).

59

De plus, les lésions PV restent un facteur prédictif significatif de la SEP après un SCI (Ruet et al., 2011a; Swanton et al., 2010) parce que ces lésions situées à cet emplacement caractéristique correspondent réellement à une dissémination dans l'espace. En effet, les lésions PV ne sont pas symptomatiques dans les SCI typiques de type myélite, névrite optique, et les syndromes du tronc cérébral. Dans un essai clinique, les caractéristiques initiales de l'IRM cérébrale ont été reportées pour 468 patients avec un SCI (Moraal et al., 2009). Parmi les critères de Barkhof-Tintoré, avoir ≥3 lésions PV au moment du SCI était significativement associé à un risque plus élevé de SEP-CD à 3 ans (HR = 1,66, IC 95% 1,14 à 2,41, p<0,01).

La lésion PV est une topographie sélective dans la SEP qui a été observée dans différentes études qui ont porté sur la distribution spatio-temporelle des lésions de la substance blanche chez les patients atteints de SEP-RR (Ceccarelli et al., 2012; Filli et al., 2012). Dans la plus grande étude longitudinale réalisée à ce jour incluant des patients après un SCI (n = 74) (Dalton et al., 2012), la distribution spatiale des lésions cérébrales a été étudiée en lien avec les données cliniques collectées 20 ans après la survenue du SCI évocateur de SEP. Les patients atteints de SEP-CD étaient plus susceptibles que les patients ayant un diagnostic final de SCI à avoir eu des lésions T1 et T2 PV dans le corps calleux, la corona radiata, et les radiations optiques.

Environ 30% des patients ayant présenté un SCI ont une IRM cérébrale considérée normale en dehors de la lésion symptomatique (Barkhof et al., 1997; Tintoré et al, 2006; 2008). Dans la cohorte de l'ONTT, la proportion de patients ayant développé une SEP-CD après une névrite optique sans lésions cérébrales initiales était de 9% après 6 à 10 ans de suivi, et de 2% au cours de la période entre 10 et 15 de suivi (ONTT, 2008). Par ailleurs, 58% des patients après une névrite optique et ayant au moins une lésion cérébrale n'ont pas été diagnostiqués SEP à 5 ans, et seulement 30 à 32% de ces patients ont reçu un diagnostic de SEP-CD au cours de la période entre 6 et 15 ans de suivi (ONTT, 2008).

Près des deux tiers des patients après un SCI typique avaient des BOC dans le LCR non présentes dans le sérum, quels que soient les résultats de l'IRM (Miller et al., 2012). En particulier, la présence de BOC dans le LCR a été signalée chez près d'un tiers des patients ayant une IRM cérébrale normale (sans lésions évocatrices de SEP et sans critères de Barkhof-Tintoré), ce qui suggère la valeur supplémentaire du LCR par rapport à l'IRM dans ces cas (Tintoré et al., 2000;. Tintoré et al., 2008;. Villar et al., 2008; Miller et al., 2008). Dans la dernière version des critères révisés de McDonald (Polman et al., 2011), l'examen du LCR n'est pas nécessaire pour le diagnostic de SEP. Cependant, l'examen du LCR pourrait fournir des informations précieuses; la négativité de l'IRM cérébrale et l'absence de BOC ont une bonne valeur prédictive négative (Masjuan et al., 2006; Zipoli et al., 2009).

60

3. Défis et difficultés

Après un SCI typique, le principal défi est d'identifier les patients à haut risque de survenue d'un deuxième épisode clinique. Dans la première étude, le critère de jugement principal était le diagnostic de la SEP selon les critères de McDonald, 2005 (Polman et al., 2005). Dans cette seconde étude, le critère de jugement principal était le diagnostic de la SEP-CD (Poser et al., 1983), et le critère de jugement secondaire était le diagnostic de la SEP selon les critères révises de McDonald, 2005 (Polman et al., 2005). Certains patients ne bénéficieront pas d'un diagnostic de la SEP-CD au fil du temps même au décours d'un SCI typique et évocateur de SEP. En fait, certains patients peuvent avoir un diagnostic de SEP selon les critères IRM de McDonald (avec une DIT démontrée en IRM), mais ne jamais présenter cliniquement de seconde attaque même après un long suivi (Chard et al., 2011). Ainsi, il reste difficile d'établir un pronostic à titre individuel au stade de SCI en raison ce l'hétérogénéité présente dans la SEP.

4. Conclusions et perspectives

En conclusion, cette grande étude prospective a permis de confirmer la validité des 3 facteurs précédemment identifiés pour prédire le diagnostic de SEP chez les patients après un SCI. Ces facteurs étaient associés à la même exactitude que celle des critères de DIS proposés dans les critères révisés de McDonald. Les patients ayant au moins 2 de ces facteurs prédictifs peuvent être considérés à haut risque de SEP, et une surveillance étroite doit être effectuée jusqu'à la démonstration de la DIT.

Le prochain défi consiste à prédire l'évolution de la maladie chez les patients ayant un diagnostic établi de SEP. Les données récentes sont en faveur de l'utilisation précoce des traitements de fond de la maladie (Jacobs et al., 2000; Kappos et al., 2006; Comi et al., 2009; Comi et al., 2012). Cependant, l'instauration de ces traitements au stade de SCI avant que le diagnostic de SEP soit posé est encore controversée, parce qu'il s'agit de thérapeutiques contraignantes. L'identification de marqueurs pronostiques à ces stades très précoces de la maladie serait utile pour adapter les stratégies thérapeutiques. Des études épidémiologiques récentes suggèrent que le taux de progression de l'invalidité dans les premières années de la maladie est un facteur prédictif d'invalidité à long terme. La prédiction d'une telle évolution pourrait donc être utile pour la gestion des patients ayant une SEP. Le substratum anatomique du handicap accumulé dans la SEP provient surtout de la perte axonale cumulative. Les paramètres d'IRM reflétant l'atteinte cérébrale diffuse ont été étudiés chez les patients après un SCI à travers différentes méthodes mentionnées dans l'introduction de cette thèse. En plus des lésions focales au niveau de la SB, l'atrophie cérébrale peut prédire la conversion après un SCI vers une SEP à court terme (Sbardella et al., 2011). De plus, les lésions corticales ont été observées lors des premiers stades de la SEP (Lucchinetti et al., 2011). Les améliorations dans la détection des lésions corticales par de nouvelles techniques d'IRM, comme les séquences type DIR (Geurts et al., 2005b), ont justifié la proposition de nouveaux critères de SEP incluant la détection des lésions corticales (Filippi et al., 2010b). Fait intéressant, les lésions corticales ont été récemment associées à la synthèse intrathécale d'immunoglobulines avec une forte valeur prédictive pour une conversion vers une SEP après un SCI, ainsi qu'une plus grande activité de la maladie (Calabrese et al., 2012). En outre, la démyélinisation corticale a été associée à des follicules

61

ectopiques de cellules B au niveau sous-pial (Serafini et al., 2004; Magliozzi et al., 2007), et il existe de plus en plus de preuves en faveur du rôle de la réponse immunitaire humorale dans la SEP. Récemment, la présence de lésions corticales lors de RIS a été rapportée (Giorgio et al., 2011), et a été suggéré comme un marqueur IRM pertinent pour l'apparition de symptômes cliniques dus à la SEP. Ces patients avaient en effet également des lésions cervicales sur l'IRM médullaire (Okuda et al., 2011) et une DIT à partir des données des IRM cérébrales (Lebrun et al., 2009; De Stefano et al., 2011), connus pour leur valeur prédictive quant à l'évolution future vers une SEP. Parmi les techniques IRM non conventionnelles, la spectroscopie IRM permet la quantification des lésions axonales par la détection de taux abaissés de N-acétylaspartate (NAA) et l'augmentation de la créatine lors des premiers stades de la maladie (De Stefano et al., 2007). Ces résultats suggèrent la valeur pronostique du niveau d'expression du NAA pour refléter l'intégrité neuronale dans le cerveau.

Les marqueurs prédictifs cliniques de l’invalidité à long terme font toujours défaut au début de la SEP. Les troubles cognitifs ont été observés aux stades précoces de la SEP, et ont été associés à des marqueurs IRM d'atteinte cérébrale diffuse même dès le début de la maladie. Par conséquent, l'atteinte cognitive pourrait être un candidat intéressant comme facteur pronostique détectable aux stades précoces de la SEP. La validation de cette hypothèse sera le but de la deuxième partie de cette thèse.

62

Chapitre 3 :

L’atteinte cognitive aux stades précoces de la Sclé rose en plaques

comme marqueur pronostique de la maladie

63

ARTICLE 3:

MRI PREDICTORS OF COGNITIVE OUTCOME IN EARLY MULTIP LE SCLEROSIS

Article accepté dans Neurology (Deloire et al., 201 1)

64

MRI PREDICTORS OF COGNITIVE OUTCOME IN EARLY MULTIP LE SCLEROSIS

Mathilde S.A. Deloire1,2 PhD, Aurélie Ruet 1,2 MD, Delphine Hamel1, Msc, Melissa Bonnet1 PhD, Vincent Dousset1,3 MD and Bruno Brochet1,2 MD.

1EA 2966, Université de Bordeaux, Bordeaux, France. ; Services de Neurologie2 et Neuroimagerie 3, CHU Bordeaux, France.

Corresponding author: Prof Bruno Brochet, EA 2966, Neurobiology of Myelin Disorders Laboratory, University Victor Segalen, case 78, 146 rue Léo Saignat, 33076 Bordeaux cedex, France. Telephone: +33(0)557571552; Fax: +33(0)5 57574818; e-mail: bruno.brochet@chu-bordeaux.fr

Mathilde Deloire: mathilde.grassin@bb-luni.u-bordeaux2.fr

Aurélie Ruet: rueta@chu-bordeaux.fr

Delphine Hamel : bealdelo@gmail.com

Melissa C. Bonnet: melissa.bonnet@u-bordeaux2.fr

Vincent Dousset: vincent.dousset@chu-bordeaux.fr

Statistical analysis were performed by Mathilde Deloire, PhD and Jeremy Jove, statistician, INSERM U657, University de Bordeaux, Bordeaux, France.

Running Title: MR and cognitive decline in RRMS

40 references; 4 tables.

Keywords: Multiple sclerosis, Brain atrophy, Ventricular fraction, Cognition, Magnetic resonance imaging.

Word count: 2975 words excluding abstract, tables, figure legends and references

65

Supported in part by research grants from Association pour la Recherche contre la Sclérose en Plaques, (ARSEP, France) and Bayer Healthcare France SA.

Sponsors did not participate in design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.

Dr. Melissa C. Bonnet has participated as speaker to a symposium organized by Bayer.

Dr Ruet reports no disclosure in relation with this study. She has been member of advisory boards or participate as speaker to symposia organized by Biogen-Idec.

She is or has been investigator for studies promoted by Novartis, Bayer-Schering, Roche, Lilly, Peptimmune and Merck-Serono and has received subventions for this activity.

Dr Dousset reports no disclosure in relation with this study. He has been member of advisory boards or participate as speeker or chairman to symposia organized by Biogen-Idec. He is or has been investigator for studies promoted by Biogen-Idec, Novartis, Bayer-Schering, Teva, Peptimmune, Lilly and AB sciences and his institution received subventions for this activity

Dr Deloire and Ms Hamel report no disclosure.

Dr Brochet reports no disclosure in relation with this study. He has been member of advisory boards or participates as speaker or chairman to symposia organized by Merck-Serono, Bayer-Schering, Novartis, Biogen-Idec, Teva and Sanofi/Aventis. He is or has been investigator for studies promoted by Biogen-Idec, Novartis, Roche, Sanofi-Aventis, Bayer-Schering, Teva, Peptimmune, Lilly and AB sciences and his institution received subventions for this activity.

.

66

ABSTRACT

Objective : To determine MRI predictors for cognitive outcome in early relapsing remitting multiple sclerosis (MS) patients.

Methods: Inception Cohort. Forty-four patients recently diagnosed with clinically definite MS, were followed-up with clinical and cognitive evaluations at 1, 2, 5 and 7 years and underwent brain magnetic resonance imaging (MRI) including magnetization transfer (MT) imaging at baseline and two years. Cognitive evaluation was also performed in 56 matched healthy subjects at baseline. Cognitive testing included the Brief Repeatable Battery (BRB). Imaging parameters included lesion load, brain parenchyma fraction, ventricular fraction (VF), mean MT ratio (MTR) of lesion and normal appearing brain tissue (NABT) masks.

Results: At baseline, patients presented deficits of memory, attention and information processing speed (IPS). Over 2 years all MR parameters deteriorated significantly. Over 7 years, EDSS deteriorated significantly. Fifty percent of patients deteriorated on memory cognitive domain and 22.7% of patients on IPS domain. Seven-year change of memory scores was significantly associated with baseline diffuse brain damage (NABT MTR). IPS z score change over 7 years was correlated with baseline global atrophy (BPF), baseline diffuse brain damage and central brain atrophy (VF) change over 2 years.

Conclusion: The main predictors of cognitive changes over 7 years are baseline diffuse brain damage and progressive central brain atrophy over the 2 years after MS diagnosis.

67

INTRODUCTION

Cognitive impairment is an important feature of Multiple Sclerosis (MS) and may affect everyday activities. Deficits in memory, information processing speed (IPS), attention, working memory, and executive functions are frequently seen1. Cognitive dysfunction has long been considered to be confined to patients at later stages of the disease but such impairments have been described in clinically isolated syndromes (CIS)2 and in early relapsing-remitting MS (RRMS)3.

The association between cognitive deficiencies and several magnetic resonance imaging (MRI) markers, including lesion load (LL), diffuse brain abnormalities and brain atrophy, has been investigated1. However, little is known about the value of these parameters at the early stages of MS to predict cognitive decline over time and the value of magnetization transfer (MT) imaging parameters. We designed a longitudinal study in newly diagnosed clinically definite MS (CDMS) patients to investigate MRI markers predictive for deteriorating cognitive function. Cross-sectional data at baseline showed that cognitive impairment was associated with MR parameters evaluating diffuse brain damage3. We hypothesized that diffuse brain damage outside lesions in early RRMS detected at baseline and their change over 2 years predict the deterioration of cognitive performance over 7 years. Baseline values of various MR parameters, including LL, MT ratio (MTR) metrics, and global and central atrophy, and their change over 2 years were investigated for predicting the deterioration of cognitive performance over 7 years.

METHODS

Patients and controls

Fifty-eight patients, diagnosed with RRMS within the previous six months without other selection criteria, were included by the coordinating centre between November 2000 and November 2001. Patients were consecutively referred by practicing neurologists from a neurological network in south western France for the purpose of that study. None of these patients received disease-modifying therapies before the second exacerbation.

One patient was excluded from the study due to reconsideration of the MS diagnosis. Forty-four patients (77.2%) completed all evaluations during the 7-year study period and were used for the analysis. All patients underwent clinical and cognitive assessment at baseline (year 0) and after years (y) 1, 2, 5 and 7, and underwent brain MRI at y0 and y2. A sample of 56 healthy subjects matched for sex, age, and educational level was also studied at baseline. Baseline clinical, imaging, and cognitive characteristics of patients and controls have been published previously3.

Standards Protocol Approvals and Patient consents:

The study was approved by the institutional review board (CPP Bordeaux #2000/28) and patients gave written informed consent.

68

Clinical evaluation

Clinical evaluation included Expanded Disability Status Scale (EDSS), Multiple Sclerosis Functional Composite (MSFC), and Montgomery and Asberg Depression Rating Scale (MADRS).

Neuropsychological assessment

The neuropsychological assessment was previously described in detail3. It included the similarities test and the Brief Repeatable Battery (BRB-N) with the Selective Reminding Test (SRT, and 3 subscores: SRT- LTS = Long term storage; SRT-CLTR = Consistent Long term retrieval ; SRT-DR = delay recall; verbal memory), the 10/36 spatial recall test testing short- (SPART) and long-term visuo-spatial memory (SPART DR), the Symbol Digit Modalities Test (SDMT, attention and IPS), the PASAT 3 and 2 second versions (PASAT 3s and PASAT 2s; working memory and IPS) and the Word List Generation test (WLG; verbal fluency). At y7, we applied alternate forms for all these neuropsychological tests to limit practice effect. Data concerning other tests used at baseline (Go/No-Go; Stroop, similarities, Boston and RFF tests) and for all tests at y1, 2 and 5 follow-ups were not used in this report, because no alternate forms were used.

At baseline, patients were classified as patient cognitively impaired (PCI) if they performed less than the 5th percentile of matched controls on at least two tests of the battery (n=22) and patient cognitively unimpaired (PCU) if they did not (n=22).

At each time point, the raw scores of each cognitive score were transformed to z scores using the mean and standard deviation (SD) baseline scores of the 44 patients as the reference data. Z scores were used to study longitudinal changes of cognitive tests and correlations. Additionally, two domain-specific z-scores were created for memory (SRT-LTS, SRT-CLTR, SRT-DR, SPART and SPART-DR) and IPS (SDMT, PASAT 3s and 2s).

MRI evaluation

Image acquisition: Methods for image acquisition and analysis have been detailed previously3. Brain MRI scans were obtained on a Philips Gyroscan ACS-NT 1.5 T scanner, including Fast Fluid-Attenuated Inversion-Recovery (FLAIR) images (TR/TE/TI: 11000/140/2725), MT images using a proton density sequence (TR/TE: 37/2.3 and flip angle = 8) both with and without an MT saturation pulse, and T1-weighted images (TR/TE: 450/12) before and after administration of gadolinium-DTPA (0.1 mmol/Kg).

For all sequences, 26 contiguous interleaved axial slices were acquired with 5 mm slices, 256x256 matrices and 230 x230 fields of view. The slices were positioned to run parallel to a line joining the most inferoanterior and inferoposterior parts of the corpus callosum.

Native and pre-processed data were displayed on workstations running software developed by Bio-Clinica, Inc. (Lyon, France). Following parameters were considered in the analysis: LL, Brain Parenchymal Fraction (BPF), which is defined as the ratio of whole brain parenchyma volume to the intracranial volume4, Ventricular Fraction (VF), defined as the ratio of ventricular volume to the intracranial volume5, and means of lesion and normal-appearing brain tissue (NABT) MTR masks.

69

Statistical analysis

Statview version 5.0 (Windows) was used for statistical analysis. Matching between MS patients (n= 44) and healthy controls (n= 56) was verified using the Chi square test for gender and the t test for age and years of education. For each score, the percentages of patients deteriorating or improving more than 1.0 SD of controls scores or stable over 7 years were calculated in whole group, PCI and PCU groups. For IPS and memory domains, patients were classified in the deterioration group if they have more than one SD decrease based on baseline scores of healthy controls. Additionally, at the follow-up evaluations, the patients were considered cognitively deteriorated according to their deterioration score modified from Kujala6. For each domain (IPS , memory), if the subject scored below –1.0 SD compared with the norms, he/she received one deterioration point; if below –1.5 SD, two; if below -2.0 SD, three.

All comparisons were performed using the t test when distribution of values was normal or by non parametric Wilcoxon test when it was not.

Participants who fulfilled all follow-ups were compared to the original sample with regard to demographic data (age, gender, and years of education), baseline neuropsychological scores, EDSS, MSFC, MADRS, disease duration; LL, MTR means, BPF, and VF.

Differences were considered significant, for all analyses, when p values were less than 5%.

Correlations between change in cognitive scores and MADRS scores were studied by Pearson’s correlation coefficient.

Univariate linear regressions were performed to assess correlation between individual baseline MR variables or changes over the first two years (LL, mean lesion MTR, BPF, VF and mean NABT MTR) and changes over 7 years of the domain specific cognitive z scores (memory and IPS) in the whole group. To determine which MRI parameters were the best predictors of cognitive change over 7 years in the whole group, we performed two stepwise multiple linear regression models for each cognitive change over 7 years (memory and IPS domains) as dependent variable. Independent variables were baseline MRI parameters (LL, mean lesion MTR, BPF, VF and mean NABT MTR) in the two first model and MRI parameters changes over 2 years (LL, mean lesion MTR, BPF, VF and mean NABT MTR) in the other models. In all models, we entered age, education level, gender and MADRS scores as independent variables and baseline EDSS was forced in all models as a covariate. Only independent variables with a conservative significance level of p<0.25 at the univariate analysis were entered simultaneously in the linear regression model. All parameters with p>0.05 were removed from the model by stepwise elimination.

Logistic regression analyses were also performed (see Table e-1 at www.neurology.org).

70

Table e-1: Logistic regression analyses to determine the predictive value of the baseline MRI parameters or the changes of MR parame ters over two years for cognitive outcome:

Predictor B-coefficient OR (95 % CI) P value

Baseline mean NABT MTR

Baseline EDSS

- 0.1744

0.005

0.840 (0.047-0.655)

1.005 ( 0.485-2.08)

<0.01

0.99

BPF change over two years

Baseline EDSS

-0.442

-0.158

0.642 (0.417-0.989)

0.854 (0.407-1.794)

<0.05

0.68

VF change over two years

Baseline EDSS

2.517

-0.041

12.39 (1.49-102.9)

0.960 (0.471-1.960)

<0.05

0.91

Cognitive outcome (deteriorated or not) was determined according to the deterioration score (sum of IPS and memory scores): >4 or not.

Co-linearity was found between BPF change over two years and VF change over two years (r = -0.81). Three multivariate logistic regression models were performed with (i) baseline MRI parameters (LL, mean lesion MTR, BPF, VF and mean NABT MTR) as independent variables, (ii) MRI parameters changes over 2 years (LL, mean lesion MTR, BPF, and mean NABT MTR) and (iii) MRI parameters changes over 2 years with LL, mean lesion MTR, BPF, and mean NABT MTR. P value for entry in model was p< 0.25.

RESULTS

Demographics, clinical and MRI data

Table 1 summarizes demographics and disease characteristics of patients and controls at baseline. MS patients were matched to healthy controls according to gender (p=0.16,χ2 = 1.98), age (p=0.67, t= 0.42), and mean number of years of education (p=0.72, t =-0.36 ). No statistical difference was observed for baseline demographics, clinical, and cognitive scores and MRI parameters between the 44 patients who completed all follow-ups over 7 years and the entire original cohort of 56 patients (all p values >0.05).

The percentage of patients receiving disease-modifying therapies at any time during the study was 95.6%. During the follow-up period, 9 of the 44 patients followed over 7 years converted from RRMS to secondary progressive MS.

None of the patients reached the MADRS threshold for severe depression (MADRS>34)7 at baseline and only one patient at 7 years (MADRS = 46). Mild depression

71

(MADRS score between 7 and 19) was diagnosed in 9 patients at baseline and in 16 patients at y7. MADRS median [range] increased over the 7 years (3.0 [0-21] at baseline and 10.5 [0.0-46] at y7, p <0.001, z = -3.712), and deteriorated in both groups (in CI : 3.0 [0-18] at baseline and 11[0-46] at y7, p=0.017, z = -2.39, in CU : 4.0 [0-21] at baseline and 10.5[0-31] at y7, p=0.004, z = -2.80).

Median EDSS deteriorated between y0 (2.0 [0.0-5.5]) and y7 (2.5 [0.0-8.0]) (p<0.001, z= -3.586) but the MSFC did not change during seven years (p= 0.20, t =1.31 ). The number of patients with EDSS <3 was 77.3 % (34/44) at baseline and 59.1% (26/44) at y7. The number of patients with EDSS≥ 6 was 0 at baseline and 3 at y7 (6.8%). Median EDSS increase is 1.0 [-2.0- +5.0]. During follow-up, disability progression was noted in 21 (48%) patients.

Over 2 years, all MR parameters deteriorated (p<0.05) in the whole group (Table 2).

72

Table 1: Demographic and clinical baseline characte ristics of RRMS patients as well as separately for PCI and PCU groups included in th e follow-up

Patients Controls

n= 44

PCI

n= 22

PCU

n=22

n= 56

Gender (male/female) 10/34 5/17 5/17 20/36

Age (years) 39.0 (8.7) 40.4 (9.9) 37.7 (7.2) 38.2 (10.5)

Education level (years of schooling)

12.2 (2.5) 11.3 (2.3) 13.2 (2.3) 13.05 (2.6)

Disease duration (months) 24.3 (27.4) 22.5(21.7) 26.1(32.6)

EDSS 2.0 [0.0-5.5] 2.0 [0.0-3.5] 1.5 [0.0-5.5]

MSFC -0.053 (0.701) -0.409(0.681) 0.339 (0.490)

MADRS 3.0 [0-21] 2.0 [0-18] 4.0 [0-21]

PCI= patients cognitively impaired; PCU: patients cognitively unimpaired. For all clinical data, scores are expressed as mean (SD) except for EDSS and MADRS which are median (range). EDSS: Expanded Disability Status Scale; MSFC: Multiple Sclerosis Functional Composite; MADRS: Montgomery and Asberg Depression Rating Scale.

73

Table 2: MRI parameters.

Patients

MRI parameters

n=44 PCI n=22 PCU n=22

Baseline Y2 t-score Baseline Y2 t-score Baseline Y2 t-score

LL (cm3) 9.57±11.26 11.26±12.0** -2.967 11.76±13.43 12.78±12.59NS -1.313 7.38±8.34 9.77±11.52** -2.851

BPF 0.881±0.032 0.865±0.045*** 3.895 0.878±0.037 0.861±0.052** 2.879 0.883±0.027 0.869±0.038* 2.565

VF 0.024±0.010 0.027±0.012*** -5.051 0.025±0.010 0.028±0.013** -3.790 0.023±0.009 0.025±0.011** 3.07

Mean lesion MTR 46.66±2.29 46.07±3.11* 1.973 46.24±2.64 45.84±3.35NS 1.153 47.08±1.84 46.29±2.93NS 1.584

Mean NABT MTR 48.44±0.865 47.64±0.954*** 5.841 48.39±0.874 47.64±0.927 *** 6.494 48.49±0.873 47.63±1.00** 3.413

PCI= patients cognitively impaired; PCU: patients cognitively unimpaired. MRI parameters are expressed as mean (SD). LL: lesion load; BPF: Brain Parenchymal Fraction; VF: Ventricular Fraction; NABT: Normal-Appearing Brain Tissue. MTR: Magnetization Transfer Ratio. P values and t-scores concerned comparisons between baseline and y2. NS: p > 0.05; * p≤ 0.05; ** p< 0.01; *** p < 0.001.

74

Cognitive evolution during follow-up:

The baseline results have been published3. Differences were observed between RRMS patients and matched controls for memory, attention, IPS, inhibition, and conceptualization.

Nineteen out of 44 patients (43.18%) had a global deterioration score >2 (sum of memory and IPS domains deterioration scores) and 9/44 (20.45%) a score >4. The proportions of patients deteriorating by cognitive domains more than 1.0 SD of controls scores, improving more than 1.0 SD of controls scores and stable over 7 years are presented in Table 3 for the three groups (all patients, PCI and PCU) without any difference (Fisher test) between the PCI and PCU groups, for memory (p= 0.37, OR = 2.05; 95% CI = 0.57-8.2) and IPS (p= 1.0, OR = 1.0, 95% CI = 0.19-5.2). The proportion of patients deteriorating was 50% for memory (9 patients deteriorated for one score, 8 patients for two score and 5 patients for more than 2 scores), 22.7% for IPS (5 for one score and 5 for at least 2 scores) and 13.6% for the WLG at y7.

Changes over 7 years in cognitive scores were not correlated with baseline MADRS scores or MADRS change over 7 years (whole, PCI, and PCU groups) (Pearson’s correlation coefficient r <0.30, p >0.05).

75

Table 3: Number of patients with deterioration, no change or improvement for the two main cognitive do mains over 7 years.

76

Regression analyses (Table 4)

For memory or IPS z scores change over 7 years as dependent variable, two multiple regression models were performed with (i) baseline MRI parameters (LL, mean lesion MTR, BPF, VF and mean NABT MTR) as independent variables and (ii) MRI changes over 2 years (LL, mean lesion MTR, BPF, VF and mean NABT MTR). In univariate linear regression analyses, memory z score change over 7 years was associated with baseline LL and mean NABT MTR and with VF changes over 2 years. IPS z score change over 7 years was correlated with baseline MRI parameters (LL, BPF, VF and mean NABT MTR) and with VF and BPF changes over 2 years (all p value <0.05, see Table 4). Parameters with p<0.25 in univariate analyses and the baseline EDSS were entered in the two multivariate models (Table 4). After stepwise elimination, the only independent predictor of memory decline was the baseline NABT MTR. After stepwise elimination, baseline NABT MTR and BPF and 2 years change of VF remained in the models for prediction of IPS decline. No MR parameter predicted WLG z score change in multivariate analysis.

77

TABLE 4: Linear regression models predicting change in cognitive z scores over 7 years from MR parameters at baseline or changes over 2 ye ars.

Memory domain includes the Selective Reminding Test (long-term storage; consistent long-term retrieval; delayed recall) and the Spatial Recall Test (recall score and Delayed recall). IPS domain includes the Symbol Digit Modalities Test, the Paced Auditory Serial Addition test (3 and 2 sec versions). LL : Lesion Load; MTR: Magnetization Transfer Ratio; NABT: normal appearing brain

Cognitive z scores change (y7-baseline)

MR parameters Included in models

R Univariate analyses

P values Univariate analyses

Adjusted R2 model

P values Multivariate analyses

Memory domain

Baseline EDSS Baseline LL Baseline BPF Baseline VF Baseline mean lesion MTR Baseline mean NABT MTR

-0.095 -0.333 0.126 -0.175 0.253 0.377

0.54 <0.05 0.42 0.26 0.10 <0.05

0.144 <0.05

Baseline E DSS LL change over 2 years BPF change over 2 years VF change over 2 years Mean lesion MTR change over 2 years Mean NABT MTR change over 2 years

-0.095 -0.085 0.200 -0.290 - 0.090 -0.057

0.54 0.59 0.19 <0.05 0.57 0.72

0.009 0.54

IPS domain

Baseline EDSS Baseline LL Baseline BPF Baseline VF Baseline mean lesion MTR Baseline mean NABT MTR

-0.273 -0.425 0.551 -0.414 0.148 0.511

0.07 <0.01 <0.0001 <0.01 0.34 <0.001

0.416 <0.0001

Baseline EDSS LL change over 2 years BPF change over 2 years VF change over 2 years Mean lesion MTR change over 2 years Mean NABT MTR change over 2 years

-0.273 0.117 0.435 -0.545 0.163 -0.016

0.07 0.45 <0.01 <0.0001 0.29 0.92

0.326 <0.001

78

tissue; BPF; Brain Parenchymal Fraction; VF: Ventricular Fraction. Univariate analyses for age, MADRS scores, gender and educational levels showed not significant results (p>0.05). Variables with a p value <0.25 in univariate analyses were included in multivariate analyses. Variables included in multivariate analyses but not significantly correlated with cognitive Z scores changes in final models were in italic police. Variables significantly correlated to cognitive cha nge in multivariate analyses are in bold police. For statistical analysis and further details, see methods.

79

DISCUSSION

This study emphasizes the frequency of cognitive deterioration in the early stages of MS. This deterioration, which may have many impacts in the daily life of patients, including working and social activities, needs to be detected in order to adapt therapeutic strategies. The present study was designed to determine which MR parameter, measured early in MS course, could help predict subjects who will deteriorate. The results showed that MRI parameters reflecting diffuse brain damage (BPF, VF and mean NABT MTR) predict more strongly cognitive deterioration than lesions.

The mechanisms of cognitive dysfunction in MS are not fully understood, but convergent evidence suggests that disconnection between cortical areas might be the basis of this dysfunction9. Although focal lesions might contribute to this disconnection, several cross-sectional studies using diffusion and MT imaging suggested that axonal damage largely contributes to cognitive alterations by interrupting critical networks3, 9-15. In the present longitudinal study in early MS we observed that baseline NABT MTR, which reflects diffuse brain injury accumulated so far, but no MR parameters studying focal lesions, predicts cognitive changes over 7 years in the two main domains memory and IPS, in multivariate analyses. Although, LL was associated with cognitive outcome in univariate analyses, it did not remain in multivariate models. These results confirmed the importance of diffuse brain damage in cognition alteration in RRMS. Such disconnections could interfere to the allocation of brain resources for a specific task and may explain that the most common cognitive impairment in MS concerned IPS16.

Recent reports have shown correlations between cognitive disabilities and central atrophy and are in agreement with the hypothesis that disconnection between the sub-cortical and cortical structures involved in these networks may contribute to cognitive dysfunction 17-23. Although atrophy measures did not correlate with cognitive functions at baseline3, baseline BPF, a measure of global atrophy and two-year VF change, a measure of central brain atrophy predict decline in IPS performance over 7 years. Although periventricular white-matter atrophy could contribute to VF change, and in particular corpus callosum atrophy, which has been implicated in cognitive impairment in MS24, it is likely that damage of the deep gray-matter and the surrounding white matter is determinant in cognitive deterioration affecting IPS in early RRMS. These results are in agreement with a recent study that analyzed the rates of atrophy in different parts of the brain and correlated them to various clinical indexes including the PASAT as part of the MSFC25. The authors found that a worse performance on the PASAT at follow-up was predominantly related to atrophy development around the ventricles and, to a lesser extent, brainstem atrophy.

Several studies investigated the association between MR parameters and cognition evolution in RRMS but none used MT imaging. One study showed that one-year global atrophy rate was associated with cognitive deficiencies measured several years later26. In that study, prediction of cognitive deterioration could not be evaluated since cognitive data were only measured at the last follow-up. Another study that examined the relation between changes of LL and brain parenchymal volumes over 2 years together with change in cognition during the same time-frame and found that the best predictor was brain parenchymal volumes27. These two studies did not specifically evaluate central atrophy. Another study found that only neocortical volume changes, but not deep gray matter volume changes, were significantly associated with deteriorating cognitive performance over 2.5 years28. It is possible that atrophy of the cerebral cortex accounts for more of the cognitive

80

deterioration when the disease progresses, since the mean disease duration was longer in that later study than in ours (4.0 ± 2.8 vs. 2.0 +/- 2.2 years). It has been suggested that deep central atrophy may peak at earlier stages compared to whole brain atrophy29 and that central atrophy progresses preferentially in the early stages of the disease30. These data can explain that, although baseline BPF which reflects brain whole atrophy developed so far is predictive of cognitive deterioration, it is only the central atrophy change which account for cognitive deterioration in attention/IPS domain. We hypothesized that disruption of deep central brain cognitive networks involved in attentional occurs in our population. It is worth mentioning that cortical atrophy also occurs in early MS31,32 but appears to progress less rapidly than central atrophy29. In our study, we did not specifically study neocortical volume.

Interestingly, it was recently shown that memory performance and mesial temporal volumes, in particular hippocampal atrophy, were correlated in MS33,34 This suggests regional specificity of correlations between atrophy and alterations of a given cognitive domain. In our study, we did not find correlation between memory change and atrophy measures although a trend was observed with VF change (data not shown). However baseline NABT MTR predicted memory changes over 7 years suggesting that diffuse alterations contribute also to the memory deficits by interrupting the relevant networks.

There are some limitations to this study. When performing imaging, it would have been informative to study regional changes in brain volume and MTR, for example, using voxel-based morphometry. Unfortunately those techniques were not available at the beginning of this study. Concerning cognition, the main limitation of longitudinal studies in MS is that relatively few patients deteriorate for cognitive scores and that a relatively large number of them remain stable. One possible reason is practice effect. To limit practice effect we used alternate forms at the 7 year evaluation, never used in the previous evaluations, but we cannot exclude that some practice effect remains. Another explanation for the limited changes in cognitive performances is that cognitive skills deteriorate slowly at the early stages of MS because of the presence of compensatory mechanisms35.

Further studies are needed to determine which therapeutics strategies could limit diffuse brain damage at the early stage of MS and therefore ongoing cognitive deterioration.

ACKNOWLEDGEMENTS

The study received support from ARSEP (Association pour la Recherche contre la Sclérose en Plaques) and Schering France. Work was made possible by the neurologists of the AQUISEP network.

81

REFERENCES

1. Hoffmann S, Tittgemeyer M, von Cramon DY. Cognitive impairment in multiple sclerosis. Curr Opin Neurol 2007; 20:275-280.

2. Feuillet L, Reuter F, Audoin B et al. Early cognitive impairment in patients with clinically isolated syndrome suggestive of multiple sclerosis. Mult Scler 2007;13:124-127.

3. Deloire MS, Salort E, Bonnet M et al. Cognitive impairment as marker of diffuse brain abnormalities in early relapsing-remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2005; 76:519-526.

4. Rudick RA, Fisher E, Lee JC, Simon J, Jacobs L. Use of the brain parenchymal fraction to measure whole brain atrophy in relapsing-remitting MS. Multiple Sclerosis Collaborative Research Group. Neurology 1999; 53:1698-1704.

5. Kalkers NF, Bergers E, Castelijns JA et al. Optimizing the association between disability and biological markers in MS. Neurology 2001;57:1253–1258.

6. Kujala P, Portin R, Ruutiainen J. The progress of cognitive decline in multiple sclerosis A controlled 3-year follow-up. Brain 1997; 120:289–297.

7. Snaith RP, Harrop FM, Newby DA, Teale C. Grade scores of the Montgomery–Asberg depression and the clinical anxiety scales. Br. J. Psychiatry 1986;148:599–601.

8. Calabrese P, Penner IK. Cognitive dysfunctions in multiple sclerosis-a "multiple disconnection syndrome"? J Neurol 2007;254 Suppl 2:18-21.

9. Rovaris M, Iannucci G, Falautano M at al. Cognitive dysfunction in patients with mildly disabling relapsing-remitting multiple sclerosis: an exploratory study with diffusion tensor MR imaging. J Neurol Sci 2002;195:103-109.

10. Cox D, Pelletier D, Genain C et al. The unique impact of changes in normal appearing brain tissue on cognitive dysfunction in secondary progressive multiple sclerosis patients. Mult Scler 2004;10:626-629.

11. Ranjeva JP, Audoin B, Au Duong MV et al. Structural and functional surrogates of cognitive impairment at the very early stage of multiple sclerosis. J Neurol Sci 2006;245:161-167.

12. Benedict RH, Bruce J, Dwyer MG et al. Diffusion-weighted imaging predicts cognitive impairment in multiple sclerosis. Mult Scler 2007;13:722-30.

13. Roca M, Torralva T, Meli F et al. Cognitive deficits in multiple sclerosis correlate with changes in fronto-subcortical tracts. Mult Scler 2008;14:364-369.

14. Lin X, Tench CR, Morgan PS, Constantinescu CS. Use of combined conventional and quantitative MRI to quantify pathology related to cognitive impairment in multiple sclerosis. J Neurol Neurosurg Psychiatry 2008;79:437-441.

82

15. Rovaris M, Riccitelli G, Judica E et al. Cognitive impairment and structural brain damage in benign multiple sclerosis. Neurology 2008;71:1521-6.

16. Forn C, Belenguer A, Parcet-Ibars MA, Avila C. Information-processing speed is the primary deficit underlying the poor performance of multiple sclerosis patients in the Paced Auditory Serial Addition Test (PASAT). J Clin Exp Neuropsychol 2008;13:1-8.

17. Bermel RA, Bakshi R, Tjoa C, Puli SR, Jacobs L. Bicaudate ratio as a magnetic resonance imaging marker of brain atrophy in multiple sclerosis. Arch Neurol 2002;59:275-280.

18. Christodoulou C, Krupp LB, Liang Z et al. Cognitive performance and MR markers of cerebral injury in cognitively impaired MS patients. Neurology 2003;60:1793-1798.

19. Lazeron RH, Boringa JB, Schouten M et al. Brain atrophy and lesion load as explaining parameters for cognitive impairment in multiple sclerosis. Mult Scler 2005;11:524-531.

20. Hildebrandt H, Hahn HK, Kraus JA, Schulte-Herbrüggen A, Schwarze B, Schwendemann G. Memory performance in multiple sclerosis patients correlates with central brain atrophy. Mult Scler 2006;12:428-436.

21. Benedict RH, Bruce JM, Dwyer MG et al. Neocortical atrophy, third ventricular width, and cognitive dysfunction in multiple sclerosis. Arch Neurol 2006;63:1301-1306.

22. Houtchens MK, Benedict RH, Killiany R et al. Thalamic atrophy and cognition in multiple sclerosis. Neurology 2007;69:1213-1223.

23. Tekok-Kilic A, Benedict RH, Weinstock-Guttman B et al. Independent contributions of cortical gray matter atrophy and ventricle enlargement for predicting neuropsychological impairment in multiple sclerosis. Neuroimage 2007;36:1294-1300.

24. Pelletier J, Suchet L, Witjas T et al. A longitudinal study of callosal atrophy and interhemispheric dysfunction in relapsing-remitting multiple sclerosis. Arch Neurol 2001;58:105-11.

25. Jasperse B, Vrenken H, Sanz-Arigita E et al. Regional brain atrophy development is related to specific aspects of clinical dysfunction in multiple sclerosis. Neuroimage 2007;38:529-537.

26. Summers M, Fisniku L, Anderson V, Miller D, Cipolotti L, Ron M. Cognitive impairment in relapsing-remitting multiple sclerosis can be predicted by imaging performed several years earlier. Mult Scler 2008;14:197-204.

27. Zivadinov R, Sepcic J, Nasuelli D et al. A longitudinal study of brain atrophy and cognitive disturbances in the early phase of relapsing-remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2001;70:773-780.

28. Amato MP, Portaccio E, Goretti B et al. Association of neocortical volume changes with cognitive deterioration in relapsing-remitting multiple sclerosis. Arch Neurol 2007;64:1157-1161.

29. Simon JH. Brain atrophy in multiple sclerosis: what we know and would like to know. Mult Scler 2006;12:679-687.

83

30. Audoin B, Davies GR, Finisku L, Chard DT, Thompson AJ, Miller DH. Localization of grey matter atrophy in early RRMS : A longitudinal study. J Neurol.2006;253:1495-1501.

31. Morgen K, Sammer G, Courtney SM et al. Evidence for a direct association between cortical atrophy and cognitive impairment in relapsing-remitting MS. Neuroimage 2006;30:891-898.

32. Calabrese M, Agosta F, Rinaldi F et al. Cortical lesions and atrophy associated with cognitive impairment in relapsing-remitting multiple sclerosis. Arch Neurol 2009;66:1144-50.

33. Sicotte NL, Kern KC, Giesser BS et al. Regional hippocampal atrophy in multiple sclerosis. Brain 2008;131(Pt 4):1134-41

34. Benedict RH, Ramasamy D, Munschauer F, Weinstock-Guttman B, Zivadinov R. Memory Impairment in Multiple Sclerosis: Correlation with Deep Gray Matter and Mesial Temporal Atrophy. J Neurol Neurosurg Psychiatry 2009;80:201-6.

35. Bonnet M, Allard M, Dilharreguy B, Deloire M, Petry KG, Brochet B. Cognitive compensation failure in multiple sclerosis. Neurology 2010; 75:1241–1248.

84

Discussion de cet article

1. Résumé des résultats de cet article

A partir d’une étude longitudinale réalisée sur 7 ans, les paramètres IRM reflétant l'étendue et la sévérité de l'atteinte diffuse au niveau du parenchyme cérébral d’apparence normale (PCAN) et la conséquence des lésions cérébrales diffuses évaluée par des mesures d’atrophie cérébrale (atrophie globale et atrophie centrale) prédisaient mieux les troubles cognitifs des patients ayant une sclérose en plaques rémittente récurrente (SEP-RR) que les lésions T2 visibles au niveau de la substance blanche (SB). L'aggravation des fonctions cognitives au cours des années suivant le deuxième épisode a été corrélée avec les paramètres IRM reflétant principalement l’atteinte initiale cérébrale axonale diffuse et son changement précoce dans les deux premières années après le diagnostic de SEP.

2. Intégration des résultats aux données actuelles de la littérature

Comme nous l'avons mentionné dans l'introduction de cette thèse, le dysfonctionnement cognitif peut être expliqué à la fois par des changements cérébraux au niveau microscopique et macroscopique. Des anomalies au niveau de la SBAN ont été identifiées en utilisant différentes techniques d'IRM quantitatives (ITD, ITM, et spectroscopie par résonance magnétique). Dans notre étude, la procédure IRM par ITM est abordée plus en détail dans l'annexe 1. En plus des lésions focales, les anomalies diffuses au niveau du PCAN, pouvant être soit secondaires soit indépendantes des lésions, étaient significativement associées à l’atteinte cognitive des patients ayant une SEP, même à des stades précoces de la maladie (Deloire et al., 2005).

Il existe seulement que quelques études longitudinales comprenant un nombre suffisant de patients inclus aux stades précoces de SEP-RR ayant bénéficié de plusieurs évaluations cognitives et d’imagerie par IRM avec un suivi à long terme. Les altérations au niveau du PCAN ont été associés à des troubles cognitifs dans la SEP dans une étude longitudinale sur un an (Filippi et al., 2000). La mauvaise performance aux tests d'attention a été associée à une diminution du RTM au niveau de la SBAN détectée cinq ans plus tôt (Summers et al., 2008). La progression sur un an de l’atrophie globale a été associée à des déficits cognitifs globaux et portant sur des tests de mémoire, de VTI, et d'attention effectués cinq après les examens d’IRM (Summers et al., 2008). En outre, les paramètres d’atteinte diffuse cérébrale étaient indépendamment associés aux déficits cognitifs, alors que la charge lésionnelle en T2 et T1 ne différait pas entre les patients ayant ou non un déficit cognitif dans une étude longitudinale sur 2 ans incluant 53 patients atteints de SEP-RR (Zivadinov et al., 2001). Dans une étude multicentrique longitudinale sur 3 ans incluant 49 patients atteints de SEP-RR ayant débuté récemment un traitement par interféron β, les variations du volume lésionnel en T2 et du pourcentage de volume cérébral n’étaient pas significativement associées aux performances cognitives du patient, même si le volume lésionnel en T2 était modérément corrélé avec le résultat de certaines tâches cognitives (Amato et al., 2010). Cependant, l'atrophie focale et

85

centrale n'ont pas été analysées dans ces études (Zivadinov et al., 2001; Summers et al., 2008; Amato et al., 2010). Nous soulignons l’importance de l'atrophie cérébrale centrale précoce chez les patients atteints de SEP-RR et sa corrélation avec le déclin de la VTI dans la SEP. L’atteinte de la VTI est reconnue comme un déficit cognitif primaire et central dans la SEP (DeLuca et al., 2004; Forn et al., 2008) pouvant être présent dès les premiers stades de la maladie. Ainsi, l'identification précoce des troubles de la VTI pourrait être un marqueur clinique pertinent de l'atrophie cérébrale centrale pouvant prédire précocement la progression de l'invalidité évaluée à l’aide de l'échelle EDSS (Lukas et al., 2010).

3. Défis et difficultés

L'une des difficultés inhérentes aux études longitudinales réalisées chez des patients ayant une SEP réside dans la variabilité inter-patient au niveau de leurs performances cognitives. Le dysfonctionnement cognitif dans la SEP a été principalement étudié dans le cadre d’études transversales et son évolution dans le temps est encore débattue. Les périodes de suivi des études longitudinales analysant les troubles cognitifs dans la SEP varient entre 1 à 10 ans (Jennekens-Schinkel et al., 1990;. Kujala et al., 1997; Amato et al., 2001a; Rosti et al., 2007; Denney et al., 2008; Duque et al., 2008; Amato et al., 2010). L’évolution des performances cognitives chez les patients ayant une SEP est en partie contradictoire. Certaines études ont rapporté la préservation des fonctions cognitives, tandis que d'autres ont observé un déclin cognitif léger à modéré au fil du temps (Portaccio et Amato, dans Amato 2011). En fait, les aspects méthodologiques limitent souvent la comparaison directe des résultats, tels que la différence dans la composition des échantillons de sujets, la durée de la période de suivi, et la définition du déclin cognitif au cours du temps. Kujala et al. ont rapporté l'évaluation des performances cognitives pendant 3 ans dans un échantillon de 42 patients atteints de SEP et ont différencié à l'inclusion un groupe de patients préservés cognitivement (PC) et un groupe de patients ayant des troubles cognitifs (AC) avec des niveaux de handicap physique similaires (Kujala et al., 1997). Les patients du premier groupe sont restés stables au niveau cognitif dans la majorité des cas, sauf pour un tiers d’entre eux qui ont présenté une légère détérioration de leurs performances cognitives. En revanche, 77% des patients considérés AC au départ ont présenté un déclin cognitif pour de nombreux tests NP. Ces résultats suggèrent que le déclin cognitif précoce pourrait prédire la survenue d’une détérioration cognitive progressive et généralisée, alors que les patients ayant des performances cognitives intactes pourraient rester stables au niveau cognitif. Une des limites de cette étude est la courte durée de suivi qui pourrait expliquer l'absence de déclin cognitif chez les patients initialement AC. Dans une étude longitudinale sur 10 ans de 45 patients atteints de SEP, le fonctionnement cognitif se détériorait dans le temps même chez les patients initialement sans troubles cognitifs (Amato et al., 2001a). Dans notre étude, nous confortons ces résultats puisque 40,9% des patients AC et 59,1% des patients PC ont présenté une détérioration pour les tâches de mémoire, tandis que la même proportion de patients AC et PC (22,7%) ont présenté une détérioration aux tests de VTI.

Le rôle de l’atteinte de la SG dans la détermination de l’atteinte cognitive dans la SEP a été abordé dans l'introduction de cette thèse. Une limite de notre étude a été de ne pas pouvoir analyser spécifiquement les structures de la SG. La relation spatio-temporelle entre les atteintes de la SB et

86

de la SG et leur contribution respective à l'apparition et au déclin progressif des déficits cognitifs reste à déterminer dans le cadre d’études longitudinales.

4. Conclusion et perspectives

A partir des données de cette étude longitudinale, nous confirmons la corrélation entre les troubles cognitifs et l’atteinte cérébrale diffuse détectée après le diagnostic de SEP et ses changements au cours des deux premières années suivant le diagnostic. Ces résultats soulignent l’importance de l’étendue de la démyélinisation et ses dommages irréversibles au niveau axonal à l’origine du handicap à long terme. Ces données confortent notre hypothèse selon laquelle l’atteinte cognitive pourrait être un marqueur pertinent de la gravité de la maladie et un bon candidat comme un facteur pronostique utilisable dès les premiers stades de la SEP.

87

ARTICLE 4:

EARLY COGNITIVE IMPAIRMENT IN MULTIPLE SCLEROSIS PR EDICTS DISABILITY OUTCOME SEVERAL YEARS LATER

Article accepté dans Multiple Sclerosis Journal (De loire et al., 2010)

EARLY COGNITIVE IMPAIRMENT IN MULTIPLE SCLEROSIS PR EDICTS DISABILITY OUTCOME SEVERAL YEARS LATER

Mathilde Deloire1,2 PhD, Aurélie Ruet, MD² , Delphine Hamel1 Msc, Melissa Bonnet1 PhD, and Bruno Brochet MD1,2

1EA 2966, Université de Bordeaux, Bordeaux, France; Services de Neurologie2, CHU Bordeaux, France

Corresponding author: Prof Bruno Brochet, EA 2966, Neurobiology of Myelin Disorders Laboratory, University Victor Segalen, case 78, 146 rue Léo Saignat, 33076 Bordeaux cedex, France Telephone: +33(0)557571552; Fax: +33(0)5 57574818; e-mail: bruno.brochet@chu-bordeaux.fr

Statistical analyses were performed by Jeremy Jove, statistician, INSERM U657, University de Bordeaux, Bordeaux, France

Running Title: Cognitive impairment predicts disabi lity in RRMS

Keywords: Multiple sclerosis, Disability, SDMT, Cognition, Memory, Information processing speed.

88

ABSTRACT:

Background: Cognition is frequently impaired at the early stages of Multiple Sclerosis (MS). The predictive value of cognitive impairment on disability is unknown.

Objectives : To correlate cognitive impairment and the progression of disability over seven years.

Methods : Forty-five patients, recruited after MS diagnosis, were followed for seven years by yearly EDSS and MSFC evaluations and were classified as cognitively impaired (CI) or unimpaired (CU) according to neuropsychological testing at baseline.

Results: At baseline, 47.8% of patients had CI, with deficits in mainly memory and information processing speed (IPS). The baseline EDSS correlated significantly with one IPS test. The EDSS, but not the MSFC, deteriorated significantly over the seven years in the whole group and the CI group, but not the CU group. A multivariate analysis showed correlations between the EDSS change over five and seven years and two baseline tests evaluating IPS and verbal memory. The deterioration of the EDSS after seven years was significantly correlated to verbal memory testing at baseline after adjusting for age and baseline EDSS.

Conclusion: In this sample of MS patients early in the disease, the baseline IPS and verbal memory impairments predict the EDSS score five and seven years later.

89

INTRODUCTION

Cognitive impairment is a common feature in multiple sclerosis (MS) [1]. It contributes to the clinical burden of the disease, and it has been shown that it significantly impacts health-related quality of life [2]. Although its mechanisms are not fully understood, increasing evidence from magnetic resonance imaging (MRI) studies suggest that a disconnection between key cortical areas by diffuse brain damage could play an important role [3, 4]. The correlation observed between cognitive impairment and brain atrophy [5, 6] or MRI markers of diffuse brain injury [3,7] suggests that cognitive impairment could reflect the disease processes within the brain. Although the literature is quite conflicting about the relationship between cognitive impairment and disability as measured by the Expanded Disability Status Scale (EDSS), a recent study of a large sample of MS patients showed significant correlations between various measures derived from a cognitive battery and physical disability [8]. Different studies showed that patients with clinically isolated syndromes and early relapsing-remitting MS (RRMS) frequently present dysfunction in several cognitive domains, mainly the information processing speed (IPS) and memory [7, 9, 10]. Therefore, we formulate the hypothesis that early cognitive impairment could be a predictor of the disability status several years later. We conducted the present study to address this question by measuring the correlation between early cognitive impairment detected in the months following the diagnosis of clinically definite MS (CDMS) and the progression of disability in the subsequent years.

PATIENTS AND METHODS

Patients and controls

Fifty-eight patients diagnosed in the previous six months with clinically definite RRMS were consecutively recruited after giving written informed consent. The study was approved by the institutional review board. One patient was excluded from the study due to reconsideration of the MS diagnosis. Forty-six patients (82.1%) completed all evaluations during the five-year study period and were used in the analysis, and 45 completed the year seven evaluation. All patients underwent a standardised clinical assessment at least one month after a relapse and/or steroid course at baseline and after one (y1), two (y2), five (y5), and seven years (y7). All patients and a sample of 56 healthy subjects matched for sex, age, and the number of years of education had a neuropsychological assessment at baseline. Clinical, imaging, and cognitive characteristics of these patients at baseline have been published previously [7].

Clinical evaluation

Clinical evaluation, described previously [7], included the EDSS, the Multiple Sclerosis Functional Composite (MSFC), and the Montgomery and Asberg Depression Rating Scale (MADRS). All evaluations were performed at least 30 days after the end of a relapse.

90

Neuropsychological assessment

The neuropsychological assessment was previously described in detail [7] and included the Brief Repeatable Battery for neuropsychological testing (BRB-N) and other tests. The tests in the BRB-N were the Selective Reminding Test (SRT), which tested verbal memory (three subscores: SRT- LTS= Long term storage; SRT-CLTR = Consistent Long term retrieval; SRT-DR = delay recall), the 10/36 spatial recall test testing short- (SPART) and long-term visuo-spatial memory (SPART DR), the Symbol Digit Modalities Test (SDMT), which investigated attention and IPS, the PASAT 3 and 2 second versions (PASAT 3s and PASAT 2s), which tested working memory, and the IPS and the Word List Generation test (WLG), which assessed verbal fluency. Other tests used were the computerised Go/No-go and Stroop tests, which tested inhibition and flexibility and the WAIS-R Similarities test to evaluate conceptualisation.

Patients were classified as cognitively impaired (CI) if they performed less than the 5th percentile of matched controls on at least two tests in the battery (n=22) and cognitively unimpaired (CU) if they did not (n=24).

The raw scores of each of the 12 individual measures generated from the cognitive battery were transformed into z scores using the mean and standard deviation for the 46 patients at baseline as the reference data. Additionally, different domain-specific z-scores were created for short-term memory (SRT-LTS, SRT-CLTR, and SPART), delayed memory (recall) (SRT-DR and SPART-DR), IPS (SDMT, PASAT 3s and 2s), and inhibition (Go/No-go and Stroop).

Statistical analysis

All comparisons of the clinical scores between the CI and CU groups were performed using a t-test when the distribution of the values was normal or by a non-parametric Wilcoxon test when it was not. Comparisons of paired data were made using a Wilcoxon test. The comparisons of changes in clinical scores over seven years between the CU and CI groups were performed using an ANOVA test. The univariate correlations between the cognitive z scores and the EDSS scores at five years and seven years (EDSS change over five or seven years) were assessed using a Pearson correlation coefficient.

Multivariate linear regressions were performed to assess the baseline cognitive z scores to be associated with the disability changes over five or seven years in the whole group. The model was adjusted for age and the baseline EDSS. Only independent variables with a conservative significance level of p<0.25 in the univariate analysis were entered simultaneously in the linear regression model. Factors not significant at the 0.05 level were removed from the model by backward elimination.

Multivariate logistic regression models were used to determine the predictive value of the baseline cognitive z scores for disability outcome (worsened or stable). At the follow-up evaluations, the patients were considered clinically worsened if they had an EDSS score increase ≥ 1.0, when the baseline EDSS was < 6.0, or an EDSS score increase ≥ 0.5, when the baseline EDSS was ≥ 6.0.

91

RESULTS

Demographics, clinical and cognitive data at baseli ne

Table 1 summarizes demographics and disease characteristics of patients at baseline and controls. MS patients were correctly matched to healthy controls according to gender (p=0.12), age (p=0.83) and mean number of years of schooling (p=0.23). No statistical difference was observed for baseline demographics, clinical and cognitive scores and MR parameters between patients who completed all follow-ups and the whole original cohort (data not shown). The percentage of patients receiving disease-modifying therapies at any time during the study was 95.6 %. They were prescribed by their practicing neurologist according to their clinical status and available recommendations.

None of the patients reached MADRS threshold for severe depression (MADRS>34) neither at baseline nor during follow-up. Cognitive scores were not correlated with MADRS scores.

Table 1: Demographic and clinical characteristics o f RRMS patients at baseline included in the follow-up

Patients (n = 46) Controls (n= 56)

Gender (male/female) 10/36 20/36

Age (years) 38.6 (8.7) 38.2 (10.5)

Educational level (years of schooling) 12.4 (2.6) 13.05 (2.6)

Disease duration (months) 23.5 (27.1)

EDSS 2.0 [0.0-5.5]

MSFC -0.031 (0.693)

MADRS 3.0 [0-21]

For all clinical data, scores are expressed as mean (SD) except for EDSS and MADRS which are median (range). EDSS: expanded disability status scale; MSFC: multiple sclerosis functional composite; MADRS: Montgomery and Asberg depression rating scale.

92

The cognitive performances at baseline were previously reported [7]. Cognitive impairment was diagnosed in 47.8% of patients. At baseline, patients presented deficits in verbal and visuo-spatial memory (SRT/CLTR, SRT/DR, SPART/DR), attention, information processing speed (IPS) (SDMT, PASAT 3s, PASAT2s), inhibition (Stroop), and conceptualisation (similarities sub-test of WAIS-R). The baseline median MADRS did not differ significantly between the CI (2.0 [0.0-18.0]) and CU (4.5 [0.0-21.0]) groups (p=0.22).

Disability scores at baseline, five, and seven year s and scores changes overtime according to cognitive status

The baseline median EDSS did not differ significantly between the CI and CU groups (Table 2).

Table 2: Evolution of clinical scores over 7 years for global population, CI and CU

Baseline Y5 (1) Y7 (2)

EDSS 2.0 [0.0-5.5] 2.25 [0.0-6.5]** 2.5 [0.0-8.0]***

CI 2.0 [0.0-3.5] 2.75 [0.0-6.5]* 2.75 [0.0-8.0]**

CU 1.5 [0.0-5.5] 2.0 [0.0-6.0]* 2.0 [0.0-6.5]

MSFC -0.031 (0.693) 0.029 (1.072) 0.087 (0.899)

CI -0.409 (0.681) -0.529 (1.166) -0.160 (0.837)

CU 0.348 (0.469) 0.535 (0.670) 0.333 (0.911)

(1): p values between Y5 and baseline scores: *: p ≤ 0.01; **: p ≤ 0.001; ***: p≤ 0.0001.

(2) p values between Y7 and baseline scores : *: p ≤ 0.01; **: p ≤ 0.001; ***: p≤ 0.0001.

EDSS = Expanded Disability Status Score; MSFC = Multiple Sclerosis Functional Composite; CI = cognitively impaired patients at baseline; CU= cognitively unimpaired patients at baseline.

The baseline MSFC score was significantly worse in the CI patients (p<0.0001). Among the MSFC components, significant differences were observed at baseline for the nine hole peg test (p=0.01) and the PASAT (p<0.001) but not the walking test.

The median EDSS deteriorated significantly between baseline and the five-year follow up (p=0.006), between the five-year and the seven-year follow-ups (p=0.04), and between baseline and the seven-year follow-up (p=0.001) (Table 2). The MSFC did not significantly change during seven years (Table 2).

93

The median EDSS deteriorated significantly in both the CI (p=0.039) and CU groups (p=0.042) over five years, but taking into account the seven years of follow-up, the deterioration of the median EDSS was statistically significant only in the CI patients (p=0.003) but not in the CU patients (p=0.09) (Table 2). Moreover, the deterioration of the EDSS between year five and seven was significant in the CI patients (p=0.02) but not in the CU patients.

The EDSS at year seven was significantly worse in the CI patients than in the CU patients (p=0.05). However, the ANOVA analysis was not able to distinguish the two groups according to the EDSS change overtime (p=0.2).

94

Correlations:

Baseline EDSS correlated significantly with one cognitive z score, the SDMT (r=-0.32; p= 0.03). The fifth year EDSS correlated significantly with the z score of the SRT-LTS (a test of episodic verbal memory) at baseline (r=-0.34, p=0.02) and with the baseline SDMT, a test of IPS (r=-0.45, p=0.002). When considering z scores of different cognitive domains, the five-year EDSS was significantly correlated with the baseline IPS z score (r=-0.33, p=0.026).

The seventh year EDSS correlated significantly with two z scores derived from the SRT (LTS z score, r=-0.39, p=0.01; CLTR z score, r=-0.44, p=0.005) and with the SDMT (r=-0.48, p=0.002). The correlations were significant with the short term memory z score (r=-0.44, p=0.0047) and the IPS z score (r=-0.34, p=0.02). No correlation was observed between EDSS outcomes and the PASAT.

The deterioration of the EDSS during the first five years was significantly correlated with the LTS z score at baseline (r=-0.30, p=0.04).

The deterioration of the EDSS over seven years was significantly correlated with three z scores derived from the SRT (LTS z score, r=-0.30, p=0.042; CLTR z score, r=-0.30, p=0.044; and SRT DR, r=-0.31, p=0.04).

The multivariate linear regressions showed significant correlations between the change in the EDSS over five years and the baseline SDMT (r²= 0.19, p=0.032) and between the change in the EDSS over seven years and the baseline CLTR score (r²=0.17, p=0.05) (table 3 and figure).

95

Table 3: Linear regression models predicting the ch ange in the EDSS over five and seven years from the baseline cognitive scores.

Dependent variable Independent variables P value R2 model

EDSS change over 5 years Baseline SDMT score

Baseline EDSS

Diagnosis Age

0.0283

0.014

0.8367

0.187

EDSS change over 7 years Baseline SRT CLTR score

Baseline EDSS

Diagnosis Age

0.0494

0.0637

0.5151

0.165

For statistical analysis and further details, see methods.

SRT: Selective Reminding Test; SRT CLTR: Consistent Long-Term Retrieval; SDMT: Symbol Digit Modalities Test.

Logistic regression models after adjustment for age and the baseline EDSS showed that the disability status (deteriorated or not) at seven years is significantly correlated to one SRT baseline score (LTS) (OR = 1.072, IC95% = 1.004-1.145, p=0.04).

96

Figure 1. Predictive value of the baseline Symbol Digit Modalities Test (SDMT) score of the Expanded Disability Status Scale (EDSS) outcome after 5 years and of the baseline consistent long-term retrieval (CLTR) score of the EDSS outcome after 7 years. (A) After adjustment for baseline EDSS and age, the baseline SDMT score was able to explain 19% of the variance in the EDSS progression over 5 years. According to the final model equation (∆EDSS y5-y0= 2.846-0.04 SDMTy0-0.455 EDSS y0+0.05 Age), the baseline SDMT score correctly classified 56% of patients as clinically deteriorated at 5 years. (B) After adjustment for the baseline EDSS and age, 17% of the EDSS progression over 7 years is explained by the baseline CLTR. According to the final model equation (∆EDSS y7-y0= 2.113- 0.032

97

CLTR y0-0.398 EDSS y0 +0.020 Age), the baseline CLTR score correctly classified 71% of the patients as clinically deteriorated at 7 years.

98

DISCUSSION:

The prediction of disability outcome in MS is a huge challenge. Many studies have looked for clinical or MR markers that are able to predict the progression of disability over years. Some predictors have been identified such as age [11, 12], interval between first relapses [13], or the MR lesion load at onset [14]. Clinically relevant predictive markers could be very useful to adapt therapeutic strategies at the early stages of MS. In spite of a limited sample size, different analyses of this longitudinal prospective study consistently demonstrated correlations between cognitive impairment at baseline and disability progression over the following years. Two cognitive tests, the SDMT, a test of IPS, and the SRT, a test of verbal memory, measured very early during the course of RRMS, are significantly correlated with EDSS worsening in the following five to seven years. Although the prediction is possible only at the group level, impairment of these clinical tests could be important warning signals at the individual level.

For many years the relationship between cognitive impairment and disability has been questioned. The results of many studies have been contradictory. Several correlation studies indicated virtually no relationship between cognitive impairment and disease duration [15-17] and weak or non-significant relationships between cognitive impairment and physical disability [15, 18, 19]. Other studies found correlation between the EDSS and cognitive tests [20] and in particular between IPS measures and the EDSS [21, 22] as we observed in this study with a significant correlation at baseline between the SDMT score and EDSS. Recently a study including a large sample of MS patients examined the correlation between individual cognitive scores and disability and found a significant correlation between some of them and disability [8]. Three measures of IPS and one memory measure were independent predictors of the disability status. In the present study, the two cognitive tests predictive of disability several years later were focused on the same cognitive functions.

IPS impairment is considered as a central defect in MS [23]. Many studies have found that IPS is frequently impaired in MS [6, 7, 24, 20, 24, 25]. It has been suggested that the SDMT, a test very sensitive to IPS, could be a potential screening test to detect cognitive impairment in MS [26,27]. Several studies suggest that IPS impairment could be, at least in part, responsible of impairment of other cognitive functions like working memory [23] or executive function [28].

Brain compensation can explain, at least in part, that in spite of a severe brain damage, as assessed by MRI, some patients performed normally on cognitive testing, as demonstrated by functional MRI [29]. These brain compensatory mechanisms can help explain the conflicting results concerning the correlation between cognitive impairment and disability and also the relative weak correlation between cognitive impairment and lesion load in MS [3]. These brain compensatory mechanisms could be illustrated by the effect of educational background on cognitive performances found in the early stages of MS [30]. Early RRMS patients with a low-education level performed significantly worse than healthy controls matched for age, gender, and education, in virtually all cognitive tests. On the contrary, RRMS patients with a high-education background, although their disability and MR measures of the burden of the disease did not differ from the previous group, performed as well as the matched healthy controls on nearly all tests. Interestingly, highly educated patients performed less than controls only on three tests, all of them measuring IPS (the SDMT, the PASAT, and a timed version of the Stroop test). This result suggests that brain compensatory

99

mechanisms are less efficient to protect patients against IPS impairment. It could be hypothesised that the requirement of brain compensation explains at least in part information processing slowness. This is in agreement with the observation that the correlation between disability and cognitive measures concerned mainly the IPS. The mechanisms leading to IPS impairment are not well understood. MRI studies showed a correlation between the SDMT scores and deep brain atrophy [5, 6], especially in the thalamus [31, 32]. This may be due directly or indirectly to disconnection of attentional networks involving the thalamus and deep gray structures [33]. Disconnection of sub-cortical/cortical networks, due to diffuse brain damage, involved in complex cognitive function like working memory have been suspected in MS on the basis of fMRI and diffusion-tensor imaging (DTI) [4]. The use of alternative networks due to brain compensation and the loss of automation procedures [34, 35] could contribute to IPS impairment. The requirement of these time-consuming brain compensatory mechanisms is correlated with the extent of diffuse brain damage assessed by MRI, at least at the early stages of the disease [29, 35]. Therefore, the IPS impairment could be a sensitive marker of this brain damage at least during the first years of the disease.

With the exception of the IPS, memory is the most frequently involved domain in MS [6, 7, 18, 20]. Deficits in episodic memory tests are consistently found in patients with MS. Previous research has suggested that memory impairment in MS may be related to dysfunction of the frontal/subcortical axis, which may contribute to deficient encoding, but also mesial temporal lobe dysfunction could lead to deficient consolidation [32]. In the present study, we used the SRT to evaluate the episodic verbal memory. Three sub-scores can be calculated from this test, two measuring new learning and acquisition (LTS and CLTR) and one delayed recall (SRT-DR). We found correlations between each of them and some markers of disability progression. Interestingly, it was recently shown that new learning impairment is more strongly associated with deep gray matter atrophy in MS, in the same way as IPS impairment [32]. Deep gray matter atrophy seems to be an early process in MS and to evolve more rapidly than whole brain atrophy at the earliest stages of the disease [36].

If prediction of disability outcomes by IPS and learning impairment is confirmed, the evaluation of these deficiencies by simple cognitive tests could be useful in the routine evaluation of patients at the early stages of the disease. The SDMT has already been found to be a valuable test for detecting cognitive impairment in MS [26, 27]. It could be easily completed by a test of episodic memory as the SRT, which has been shown to be as sensitive as other classical tests of memory [37].

Acknowledgements :

The study received support from the Association pour la Recherche contre la Sclérose en Plaques, ARSEP, France and Schering France SA. The sponsors did not participate in: the design and conduct of the study; the collection, management, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript. This work was made possible by the contribution of neurologists from the AQUISEP network, a clinical network of neurologists of Aquitaine and Charente devoted to multiple sclerosis care in this area. We specifically thank C Fabrigoule and H

100

Amieva for their help in the choice of cognitive tests and M Boudineau and P Arese for cognitive evaluations.

REFERENCES :

1. Chiaravalloti, ND, DeLuca, J. Cognitive impairment in multiple sclerosis. Lancet Neurol 2008; 7: 1139-51. 2. Benito-León, J, Morales, JM, Rivera-Navarro, J. Health-related quality of life and its relationship to cognitive and emotional functioning in multiple sclerosis patients. Eur J Neurol 2002; 9: 497-502. 3. Rovaris, M, Comi, G, Filippi, M. MRI markers of destructive pathology in multiple sclerosis-related cognitive dysfunction. J Neurol Sci 2006; 245: 111–116. 4. Audoin, B, Guye, M, Reuter, F, Au Duong, MV, Confort-Gouny, S, Malikova, I, et al. Structure of WM bundles constituting the working memory system in early multiple sclerosis: a quantitative DTI tractography study. Neuroimage 2007; 36: 1324-30. 5. Benedict, RH, Weinstock-Guttman, B, Fishman, I, Sharma, J, Tjoa, CW, Bakshi, R. Prediction of neuropsychological impairment in multiple sclerosis: comparison of conventional magnetic resonance imaging measures of atrophy and lesion burden. Arch Neurol. 2004; 61: 226-30. 6. Benedict, RH, Bruce, JM, Dwyer, MG, Abdelrahman, N, Hussein, S, Weinstock-Guttman, B, et al. Neocortical atrophy, third ventricular width, and cognitive dysfunction in multiple sclerosis. Arch Neurol 2006; 63: 1301-6. 7. Deloire, MSA, Salort, E, Bonnet, M, Arimone, Y, Boudineau, M, Amieva, H, et al. Cognitive impairment as marker of diffuse brain abnormalities in early relapsing remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2005; 76: 519–526. 8. Lynch, SG, Parmenter, BA, Denney, DR. The association between cognitive impairment and physical disability in multiple sclerosis. Mult Scler 2005; 11: 469-76. 9. Achiron, A, Barak, Y. Cognitive impairment in probable multiple sclerosis. J Neurol Neurosurg Psychiatry 2003; 74: 443-6. 10. Feuillet, L, Reuter, F, Audoin, B, Malikova, I, Barrau, K, Cherif, AA, et al. Early cognitive impairment in patients with clinically isolated syndrome suggestive of multiple sclerosis. Mult Scler 2007; 13: 124-7. 11. Trojano, M, Liguori, M, Bosco Zimatore, G, Bugarini, R, Avolio, C, Paolicelli, D, et al.Age-related disability in multiple sclerosis. Ann Neurol 2002; 51: 475-80. 12. Confavreux, C, Vukusic, S. Age at disability milestones in multiple sclerosis. Brain 2006; 129: 595-605. 13. Weinshenker, BG, Bass, B, Rice, GP, Noseworthy, J, Carriere, W, Baskerville, J. et al. The natural history of multiple sclerosis: a geographically based study. 2. Predictive value of the early clinical course. Brain 1989; 112: 1419-28. 14. Fisniku, LK, Brex, PA, Altmann, DR, Miszkiel, KA, Benton, CE, Lanyon, R, et al. Disability and T2 MRI lesions: a 20-year follow-up of patients with relapse onset of multiple sclerosis. Brain 2008; 131: 808-17. 15. Beatty, WW, Goodkin, DE, Hertsgaard, D, Monson, N. Clinical and demographic predictors of cognitive performance in multiple sclerosis: do diagnostic type, disease duration and disability matter? Arch Neurol 1990; 47: 305-308.

101

16. Ivnik RJ. Neuropsychological test performance as a function of the duration of MS-related symptomatology. J Clin Psychiatry 1978; 30: 304-307. 17. Rao, SM, Hammeke, TA, McQuillen, MP, Khatri, BO, Lloyd, D. Memory disturbance in chronic progressive multiple sclerosis. Arch Neurol 1984; 41: 625631. 18. Rao, SM, Leo, GJ, Bernardin, L, Unverzagt, F. Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction. Neurology 1991; 41: 685-691. 19. van den Burg, W, van Zomeren, AH, Minderhoud, JM, Prange, AJA, Meijer, NSA. Cognitive impairment in patients with multiple sclerosis and mild physical disability. Arch Neurol 1987; 44: 494-501. 20. Nocentini, U, Pasqualetti, P, Bonavita, S, Buccafusca, M, De Caro, MF, Farina, D, et al. Cognitive dysfunction in patients with relapsing-remitting multiple sclerosis. Mult Scler 2006; 12: 77-87. 21. Hohol, MJ, Guttmann, CR, Orav, J, Mackin, GA, Kikinis, R, Khoury, SJ, et al. Serial neuropsychological assessment and magnetic resonance imaging analysis in multiple sclerosis. Arch Neurol 1997; 54: 1018-25. 22. De Sonneville, LM, Boringa, JB, Reuling, IE, Lazeron, RH, Adèr, HJ, Polman, CH. Information processing characteristics in subtypes of multiple sclerosis. Neuropsychologia 2002; 40: 1751-65. 23. DeLuca, J, Chelune, GJ, Tulsky, DS, Lengenfelder, J, Chiaravalloti, ND. Is speed of processing or working memory the primary information processing deficit in multiple sclerosis? J Clin Exp Neuropsychol 2004; 26: 550-62. 24. Franklin, G.M., Nelson, L. M., Filley, C. M., & Heaton, R. K. Cognitive loss in multiple sclerosis: Case reports and review of the literature. Archives of Neurology, 1989; 46: 162–167 25. Benedict, RH, Duquin, JA, Jurgensen, S, Rudick, RA, Feitcher, J, Munschauer, FE, et al. Repeated assessment of neuropsychological deficits in multiple sclerosis using the Symbol Digit Modalities Test and the MS Neuropsychological Screening Questionnaire. Mult Scler 2008; 14: 940-6. 26. Deloire, MS, Bonnet, MC, Salort, E, Arimone, Y, Boudineau, M, Petry, KG et al. How to detect cognitive dysfunction at early stages of multiple sclerosis? Mult Scler 2006; 12: 445-52. 27. Parmenter, BA, Weinstock-Guttman, B, Garg, N, Munschauer, F, Benedict, RH. Screening for cognitive impairment in multiple sclerosis using the Symbol digit Modalities Test. Mult Scler 2007; 13: 52-7. 28. Denney, DR, Lynch, SG, Parmenter, BA, Horne, N. Cognitive impairment in relapsing and primary progressive multiple sclerosis: mostly a matter of speed. J Int Neuropsychol Soc 2004; 10: 948-56. 29. Ranjeva, JP, Audoin, B, Au Duong, MV, Confort-Gouny, S, Malikova, I, Viout, P, et al. Structural and functional surrogates of cognitive impairment at the very early stage of multiple sclerosis. J Neurol Sci 2006; 245: 161-7. 30. Bonnet, MC, Deloire, MS, Salort, E, Dousset, V, Petry, KG, Brochet B. Evidence of cognitive compensation associated with educational level in early relapsing-remitting multiple sclerosis. J Neurol Sci 2006; 251: 23-8. 31. Houtchens, MK, Benedict, RH, Killiany, R, Sharma, J, Jaisani, Z, Singh, B, et al. Thalamic atrophy and cognition in multiple sclerosis. Neurology 2007; 69: 1213-23.

102

32. Benedict, RH, Ramasamy, D, Munschauer, F, Weinstock-Guttman, B, Zivadinov, R. Memory impairment in multiple sclerosis: correlation with deep gray matter and mesial temporal atrophy. J Neurol Neurosurg Psychiatry 2009; 80: 201-6. 33. Posner, MI, Sheese, BE, Odludaş, Y, Tang Y. Analyzing and shaping human attentional networks. Neural Netw 2006; 19: 1422-9. 34. Bonnet, MC, Dilharreguy, B, Allard, M, Deloire, MS, Petry, KG, Brochet, B. Differential cerebellar and cortical involvement according to various attentional load: role of educational level. Hum Brain Mapp 2009; 30: 1133-43. 35. Bonnet, MC, Allard, M, Dilharreguy, B, Deloire, M, Dousset, V, Petry K et al. Cognitive compensation failure in RRMS patients. Neurology 2009; 72; suppl 3: A436. 36. Simon, JH. Brain atrophy in multiple sclerosis: what we know and would like to know. Mult Scler 2006; 12: 679-87. 37. Strober, L, Englert, J, Munschauer, F, Weinstock-Guttman, B, Rao, S, Benedict, R. Sensitivity of conventional memory tests in multiple sclerosis: comparing the Rao Brief Repeatable Neuropsychological Battery and the Minimal Assessment of Cognitive Function in MS. Mult Scler 2009 [Epub ahead of print].

103

Discussion de cet article

1. Résumé des résultats de cet article

Dans un échantillon de patients atteints de SEP-RR évalués dans les 6 mois après leur diagnostic et suivis longitudinalement pendant 7 ans, le dépistage précoce des troubles cognitifs pouvait prédire partiellement l'incapacité physique évaluée par le score EDSS 5 et 7 ans plus tard. La détérioration du score de l’EDSS sur 7 ans n'était significative que chez les patients présentant des déficits cognitifs précoces. Ceci suggère que la détection précoce des troubles cognitifs chez les patients ayant une SEP pourrait aider à identifier les patients à haut risque de déclin physique ultérieur et d’évolution vers une maladie plus sévère. Plusieurs corrélations d'intensité légère à modérée ont été observées entre les tâches de VTI et de mémoire épisodique verbale et les scores EDSS à différentes périodes de suivi. Ces résultats confirment notre hypothèse relative à la valeur pronostique des déficits cognitifs chez les patients atteints récemment de SEP-RR, et soulignent la nécessité d'évaluer les fonctions cognitives à un stade précoce de la SEP.

2. Intégration des résultats aux données actuelles de la littérature

La relation entre les troubles cognitifs dans la SEP et le handicap physique évalué en utilisant l'échelle EDSS est contradictoire. Certaines corrélations d’intensité modeste ou d’autres non significatives ont été rapportées entre l’atteinte cognitive et l'incapacité physique évaluée à partir des scores EDSS dans des études transversales (van den Burg et al., 1987; Beatty et al., 1990; Rao et al., 1990; Patti et al., 2009). Dans une étude longitudinale sur 10 ans, l’évolution des paramètres cliniques associés à la sévérité de la maladie et celle des performances neuropsychologiques avaient tendance à converger lors du suivi à long terme (Amato et al., 2001a). D'autres études ont montré une corrélation significative entre l'EDSS et la performance aux tests cognitifs (Nocentini et al., 2006), en particulier entre le score EDSS et les mesures de VTI (Hohol et al., 1997; De Sonneville et al., 2002; Lynch et al., 2005). Dans un échantillon hétérogène de 253 patients ayant une SEP-CD, Lynch et al. Ont rapporté une association significative entre l’atteinte cognitive et l'incapacité physique pour des patients ayant une durée de maladie supérieure à 10 ans et également ceux à un stade plus précoce de la SEP (Lynch et al., 2005). De plus, il est possible que les patients présentent des troubles cognitifs en l’absence de symptômes déficitaires physiques significatifs. Parmi les tests de la batterie cognitive, ceux visant à étudier la VTI (PASAT, word association test (WAT), et la tour de Londres) ont été retrouvés comme facteurs prédictifs indépendants de l’atteinte physique après une analyse de régression multiple (Lynch et al., 2005 ). Deux de ces mesures (WAT et PASAT) étaient les seuls scores NP permettant de distinguer les patients ayant une SEP-RR de ceux ayant une SEP-SP. Hohol et al. ont également mis en évidence une corrélation entre le score EDSS et les performances à un test de VTI (le SDMT) (Hohol et al., 1997). De même, De Sonneville et al. ont rapporté des corrélations entre le score EDSS et diverses mesures de VTI (De Sonneville et al., 2002). Ainsi, même si les relations généralement observées

104

sont modestes entre l’atteinte cognitive et l'incapacité physique dans la SEP, la majorité de ces conclusions portent principalement sur la mesure de la VTI, ce qui est également le cas dans notre étude. Ces résultats importants mettent en évidence la valeur pronostique de l'atteinte de la VTI qui est considérée comme un déficit cognitif central dans la SEP (DeLuca et al., 2004; De Sonneville et al., 2002; Forn et al., 2008), et pourrait expliquer en partie les déficits cognitifs observés dans la SEP.

La valeur pronostique des déficits cognitifs a également été suggérée à partir d’études récentes concernant des patients atteints de SEP dite «bénigne» (SEPB) (EDSS ≤3,0 après ≥15 ans d’évolution de maladie) (Amato et al., 2006; Amato et al., 2008). Dans un échantillon de 63 patients ayant une SEPB, le nombre de tests NP a été associé à une évolution vers une SEP dite non bénigne, ce qui correspond à une augmentation du score EDSS (score EDSS ≥4,0) ou l’évolution vers une forme de SEP secondairement progressive (SEP-SP) (Portaccio et al., 2009).

La capacité de compensation du cerveau est susceptible de jouer un rôle majeur dans l'explication de l'hétérogénéité interindividuelle des manifestations cliniques chez les patients ayant une SEP. Certains patients atteints de SEP-RR ont des performances cognitives normales par rapport aux contrôles en dépit d’altérations cérébrales visibles en IRM. Les mécanismes de compensation cérébrale, qui ont été rapportés à partir d’études en IRM fonctionnelle chez des patients analysés dès les premiers stades de la maladie, pourraient expliquer les résultats contradictoires (Audoin et al., 2005; Bonnet et al., 2010). Ces mécanismes de compensation pourraient en partie expliquer les corrélations d’intensité relativement modeste entre l’atteinte cognitive et l'incapacité physique d’une part, les scores cliniques et les marqueurs d'IRM conventionnelle d’autre part, chez les patients atteints de SEP-RR. En outre, des corrélations d’intensité modérée à importante ont été observées entre les valeurs du RTM du PCAN reflétant la sévérité de l’atteinte extra-lésionnelle et les activations de plusieurs aires corticales situées dans des réseaux sensori-moteurs ou d’intégration multimodale estimées à l'aide d'IRM fonctionnelle chez des patients atteints de SEP-RR (Rocca et al., 2002). Ces résultats suggèrent que non seulement les lésions de SEP visibles macroscopiquement mais aussi les altérations du PCAN peuvent être à l’origine d’une réorganisation corticale pouvant limiter les conséquences fonctionnelles en lien avec les dommages structuraux retrouvés dans la SEP.

3. Défis et difficultés

L'EDSS est l’échelle de référence utilisée pour estimer la progression du handicap des patients ayant un diagnostic de SEP. Mais, cette échelle présente des limites car elle n'est pas linéaire. Il existe une controverse au sujet de la signification clinique des définitions de la progression de l'invalidité couramment utilisées dans les essais cliniques des patients ayant une SEP. Il existe une certaine imprécision au niveau de la cotation de l’EDSS de la plupart des patients atteints de SEP due à une variabilité inter-observateur, en particulier concernant les scores EDSS allant de 1,0 à 3,5. Dans notre étude, chaque score EDSS était centralisé et un neurologue expérimenté a confirmé chaque score en utilisant à la fois les données de l'examen neurologique issues des dossiers de patients et les périmètres de marche effectués lors de chaque visite de suivi. Les patients ont été considérés comme cliniquement détériorés s’il existait une augmentation du

105

score EDSS d’au moins 1,0 point si leur score EDSS initial était <6,0 ou s’il existait une augmentation du score EDSS d’au moins 0,5 point si le score EDSS initial était ≥ 6,0. L'aggravation d'un score EDSS d’au moins 1,0 point persistante pendant au moins 3 (à 6) mois est souvent utilisée pour définir une progression de l'invalidité dans les essais cliniques de patients atteints de SEP-RR. Cette convention a été en partie basée sur les recommandations de Ellison et ses collègues, qui ont rapporté que les taux de récupération clinique sont faibles pour les patients s’étant aggravés d'au moins 1,0 point sur l'échelle EDSS (Ellison et al., 1994). Les variations de l'échelle EDSS peut être évaluée à 3 mois ou à 6 mois d'intervalle avant la confirmation de l'invalidité prolongée. Récemment, il a été suggéré la progression d’au moins 1 à 2 points du score EDSS sur une durée d’au moins 1 an pour une définir une aggravation persistante de l'invalidité dans la SEP (Ebers et al., 2008).

Comme nous l'avons souligné, il existe une grande hétérogénéité concernant l'activité et l’évolution de la maladie chez les patients ayant une SEP. Au cours de notre étude, près d'un tiers des patients atteints de SEP ont montré une détérioration soutenue de l'invalidité estimée à partir du score EDSS lors des évaluations cliniques menées à 5 et à 7 ans par rapport à l’évaluation initiale. En outre, près de la moitié de ces patients ont présenté une aggravation du score EDSS entre 5 et 7 ans. En revanche, 15% des patients atteints de SEP ont montré une amélioration du score EDSS à 5 et 7 ans par rapport à l'évaluation initiale. Plus de la moitié des patients atteints de SEP-RR n'a pas présenté de détérioration clinique d’après la définition utilisée et leur score EDSS a été stable dans le temps. L’évolution du score EDSS au fil du temps est illustrée dans la figure 1. Cette grande variabilité entrave en partie la compréhension de l'évolution de la maladie au niveau individuel.

Cette étude comporte certaines limites en raison de sa nature observationnelle. La plupart des patients ont reçu un traitement de fond pouvant influer sur l'évolution du score EDSS, même si le débat persiste sur la capacité de neuroprotection des traitements disponibles actuellement. Bien que les résultats à 3 ans de l'extension d'un essai clinique pivot de patients ayant reçu précocement l’interféron β-1b par rapport à ceux traités de façon différée sont encourageants, avec une réduction du score EDSS, ces données n'ont pas été confirmées lors de l'évaluation clinique à 5 ans (Kappos et al., 2007; Kappos et al., 2009). Le principal objectif des thérapeutiques proposées dans la SEP est d'empêcher la progression du handicap. Nous avons eu l’opportunité d'utiliser les données d’une cohorte de patients issus de la population d’Aquitaine suivis longitudinalement pendant 7 ans après le diagnostic de SEP. Même s'il est évident que les traitements de fond n'ont pas été contrôlés dans cette étude, il est peu probable que leur utilisation seule puisse expliquer la relation observée entre les performances cognitives et les scores EDSS des patients ayant une SEP. L'impact sur le fonctionnement cognitif de la quasi-totalité des traitements de fond n'a pas encore été établi, en partie dû au fait que les performances cognitives n’ont pas été considérées parmi les critères de jugement principaux dans les essais cliniques. Il est encourageant de noter qu'une étude récente ayant inclus des patients après un SCI typique a révélé que l’instauration précoce d’un traitement par IFNβ-1b avait un effet positif sur les performances obtenues au test de la PASAT-3 secondes sur 5 ans (Penner et al., 2012).

106

4. Conclusion et perspectives

Comme l’atteinte cognitive a été identifiée comme un facteur clinique pronostique précoce chez les patients atteints de SEP-RR, l'évaluation cognitive devrait faire partie de l’examen neurologique de routine et dans le cadre de la recherche chez les patients atteints de SEP. Par conséquent, la réalisation d’une large batterie neuropsychologique (NP) devrait être pratiquée pour détecter les troubles cognitifs chez les patients ayant une SEP. Une des limitations est que le fait que l'évaluation cognitive est chronophage et difficile à appliquer au décours des examens neurologiques de suivi. Il existe un intérêt croissant afin d’établir une batterie NP minimale et des outils NP de dépistage dans la SEP (ce qui sera abordé plus loin avec l'article 7).

107

ARTICLE 5:

COGNITIVE IMPAIRMENT DIFFERS BETWEEN PRIMARY PROGRE SSIVE AND RELAPSING-REMITTING MULTIPLE SCLEROSIS

Article submitted in revision

COGNITIVE IMPAIRMENT DIFFERS BETWEEN PRIMARY PROGRE SSIVE AND RELAPSING-REMITTING MULTIPLE SCLEROSIS

A. Ruet 1,2 MD, M. Deloire2 PhD, J. Charré-Morin2 Msc, D. Hamel1 Msc, and B. Brochet1,2 MD.

1 Université de Bordeaux, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, F-33076 Bordeaux, France

2 CHU de Bordeaux, INSERM-CHU CIC-P 0005, & Service de Neurologie, F-33076 Bordeaux, France

Corresponding author: Bruno Brochet, Université de Bordeaux, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, case 78, 146 rue Léo Saignat, F-33076 Bordeaux cedex, France. Tel: +33 557574817; Fax: +33 557574818 bruno.brochet@chu-bordeaux.fr Authors : aurelie.ruet@hotmail.fr; julie.charre@chu-bordeaux.fr; delphinehamel.dh@gmail.com; mathilde.grassin@bb-luni.u-bordeaux2.fr; bruno.brochet@chu-bordeaux.fr Running title: Cognition in PPMS and RRMS Number of characters with spaces in the title: 83 Number of words in the abstract: 249 Number of words in the body of manuscript: 2957 Number of figures, color figures and tables: 2 figures, 3 tables and 2 e-tables (supplemental data) Number of references: 33 Study Funding The study was supported by Bayer Healthcare, France

108

Author contributions

Dr. A Ruet- Study design, analysis and interpretation of data, write the original manuscript. Dr. M Deloire- Study design, acquisition of data, analysis and interpretation, critical revision of the manuscript for important intellectual content. J Charré-Morin-Acquisition of data, critical revision of the manuscript for important intellectual content. D Hamel- Study design, critical revision of the manuscript for important intellectual content. Prof Brochet- Study concept and design, study supervision, critical revision of the manuscript for important intellectual content. Statistical analyses were performed by Dr A.Ruet and Dr MSA.Deloire. Study Funding The study was supported by Bayer Healthcare, France. The sponsor did not participate in any aspect of the design or performance of the study, including the data collection, management, analysis, and interpretation or the preparation, review, and approval of the manuscript. Disclosures Dr Ruet participated as a speaker to symposia organized by Biogen-Idec and Teva. She is or was investigator for studies promoted by Novartis, Bayer Healthcare, Roche, Lilly, Peptimmune, and Merck-Serono and has received subventions for this activity. She received research support from Novartis. She is also a recipient of a fellowship-grant from Fondation pour la Recherche Médicale. Dr Deloire and D.Hamel report no disclosure. J Charré-Morin received honorarias for speaking at scientific meetings supported by Merck Serono and support for participation to congresses from Biogen-Idec.

Prof Brochet or his institution received honorarias for speaking at scientific meetings and serving as member of scientific advisory boards for Bayer Healthcare, Biogen-Idec, Merck Serono, Genzyme, Novartis, and Teva and his institution received research grants from Bayer Healthcare, Teva, Merck Serono, Novartis, Biogen-Idec, Sanofi-Aventis, ARSEP, and Roche.

109

Abstract Objectives: To characterize the cognitive abilities of primary progressive and relapsing-remitting multiple sclerosis patients (PPMS and RRMS) in comparison with healthy controls (HCs) matched for age, sex, and education level while taking into account the different clinical characteristics of PPMS and RRMS and to compare the cognitive patterns of these two types of MS.

Methods: Forty-one PPMS patients, 60 RRMS patients, and 415 HCs were recruited. Controls were divided into 20 groups according to age, sex, and education level. Participants were assessed with a large battery of neuropsychological (NP) tests that included a modified version of the Brief Repeatable Battery, the Stroop test, computerized tests from the Test for Attentional Performance battery, the numerical span test, and the Rey Complex Figure.

Results: PPMS patients performed worse than their matched HCs on nearly all NP tests. RRMS patients performed worse than matched HCs on a computerized digit-symbol substitution task and the alertness test, reaction time for visual scanning, and Paced Auditory Serial Addition Test 3 seconds. PPMS patients had worse NP scores and were more impaired in cognitive domains than RRMS patients. After controlling for Expanded Disability Status Scale score, the results remained unchanged.

Interpretation: The PPMS patients presented with a wide range of cognitive deficits in information processing speed (IPS), attention, working memory, executive function, and verbal episodic memory, whereas the impairment in RRMS patients were limited to IPS and working memory when compared with their respective matched HCs. Cognitive deficits were more severe in PPMS than RRMS patients.

110

INTRODUCTION

Little information is available on the cognitive dysfunction that occurs in primary progressive multiple sclerosis (PPMS) as compared with relapsing-remitting MS (RRMS). This is largely due to the methodological flaws of some studies that used inappropriate control groups that did not take into account differences in age, sex, and education level, which are frequent between RRMS and PPMS patient groups. The first study comparing cognitive performance in a homogeneous sample of PPMS patients with cognitive performance in secondary-progressive MS (SPMS) patients found cognitive impairment in 7% of PPMS patients versus 53% of SPMS patients with similar physical disabilities.1 In the MAGNIMS study2 , 63 PPMS or transitional progressive MS patients were paired with controls and 28.6% of patients were diagnosed as cognitively impaired, suggesting less frequent impairment in PPMS patients than in RRMS patients. In contrast, some studies that have compared, with similar methodologies, selected samples of patients with PPMS and RRMS, found more frequent impairment in PPMS patients than in RRMS patients.3-4 However, these studies did not use separate control groups.

The aim of that study was to compare the cognitive performance of PPMS and RRMS patients using a unique methodology while taking into account the different clinical characteristics of the patients with these two forms of MS by recruiting a wide sample of healthy control subjects that are strictly matched for age, sex, and education level to each MS patient group. We hypothesized that the PPMS patients would have more extensive cognitive dysfunction than the RRMS patients.

SUBJECTS AND METHODS

Participants

Patients

Persons with MS (PwMS) were recruited from the MS Centre of Bordeaux between April 2009 and April 2011. The eligibility criteria were as follows: RRMS diagnosis according to the Poser criteria5 or PPMS diagnosis according to the 2005 McDonald criteria6; an elapsed time since the first MS symptoms of less than 10 years for RRMS patients and less than or equal to 14 years for PPMS patients; greater than 18 years of age; and French speaker. The exclusion criteria were as follows: other neurological diseases that could explain the symptoms; SPMS; a history of psychiatric illness with the exception of stable depressive symptoms; starting or stopping antidepressants in the previous two months; alcohol, drug, or substance abuse in the previous two years; steroid treatment within the last 30 days; and recent cognitive assessment (within less than one year). Each PwMS underwent a full neurological examination. Disability was measured using a French-adapted version7 of the Expanded Disability Status Scale (EDSS).8

Healthy Controls

Healthy controls (HCs) were recruited and divided into 20 groups according to age (18-34, 35-44, 45-54, 55-64, and more than 65 years), sex, and education level (secondary education (usually 12 years of schooling) and graduated (at least “baccalaureate”)). Individuals at least 18 years old who were French speakers were eligible to participate in this study. The exclusion criteria were as follows: a personal history of neurological disease or psychiatric illness with the exception of stable

111

depressive symptoms; serious head injury; a familial history of MS; starting or stopping antidepressants in the previous two months; alcohol, drug, or substance abuse in the previous two years; and recent cognitive assessment (within less than one year). HCs received compensation for participating in the study.

This study was approved by the ethics committee, including the institutional review board for human subject research of Bordeaux. All subjects gave written informed consent to participate in the study before their inclusion.

Neuropsychological Assessment

PwMS and HCs were assessed with a comprehensive neuropsychological (NP) battery that included some tests from the Brief Repeatable Battery (BRB-N),9-11 the Paced-Auditory Serial Addition Test–3 seconds (PASAT 3s; testing working memory), the Selective Reminding Test (the SRT and its three subscores, SRT-LTS = long-term storage; SRT-CLTR = consistent long-term retrieval and SRT-DR = delay recall, which test episodic verbal memory), the 10/36 spatial recall test for short visuo-spatial memory (SPART), the delayed recall of visuo-spatial memory (SPART-DR) test, and the Word List Generation test (WLG; assessed verbal fluency). A computerized digit-symbol substitution task, called the Computerized Screening Cognitive Test (CSCT),12 was used instead of the Symbol Digit Modalities Test (SDMT) to assess information processing speed (IPS). In this test, the answer is given orally, and the task lasts 90 seconds. In contrast to the keys for the classical digit/symbol substitution tests like the SDMT and the digit/symbol subtest of the Weschler Adult Intelligence Scale,13 the key is generated automatically by the software for each test session, and it differs with each presentation to prevent learning of the key. The score is the number of accurate answers given in 90 seconds. Computerized tests from the Test of Attentional Performance (TAP) of Zimmermann et al.14 consisted of the subtests for alertness, visual scanning, flexibility, and visual and auditory divided attention. For alertness, flexibility, and visual scanning, reaction times (RTs) in milliseconds were evaluated IPS. For divided attention, the RTs and AA ratios of the simple task (auditory or visual divided attention) to the double task (auditory and visual divided attention) were used to assess attention. The Stroop 45 seconds test,15 the numerical span test (forward and backward),16 and the Rey Complex Figure (RCF) task17 were also performed. Seven cognitive domains were categorized: IPS, attention, working memory, verbal and visual episodic memory, visuoconstruction, and executive function (Table 1).

112

Table 1. Cognitive domains and neuropsychological t ests included in the battery

IPS Attention Working memory Executive function

Episodic Verbal memory

Episodic Visual memory

Visuo construction

CSCT Alertness (RT) Visual scanning (RT) : -with a target - without target Flexibility (RT)

Ratio divided attention (RT, AA): -auditory - visual Visual scanning (AA) - with a target - without target

PASAT 3s Numerical span test: - forward - backward

Flexibility (AA) Stroop 45 WLG 90

SRT: -SRT-LTS -SRT-CLTR -SRT-DR

10/36 SPART : Immediate recall Delay recall

RCF copy

IPS: Information Processing Speed, CSCT: Computerized Screening Cognitive Test, TAP: Test of Attentional Performance, RT: reaction time, in milliseconds, AA: Accurate Answers, PASAT 3s: Paced Auditory Serial Addition Test 3.0 s, WLG 90: Word List Generation test, SRT: Selective Reminding Test, LTS: Long-Term Storage, CLTR: Consistent Long-Term Retrieval, SRT-DR: Delay Recall, 10/36 SPART-IR: Spatial Recall Test for the immediate recall of short visuo-spatial memory, 10/36 SPART-DR: Spatial Recall Test for long-term visuo-spatial memory, RCF: Rey Complex Figure.

113

Depression, Anxiety and Fatigue

Each subject filled out questionnaires concerning depressive symptoms (Beck Depression Inventory II, BDI II18), anxiety (State-Trait Anxiety Inventory, STAI19), and subjective fatigue (UK Neurological Disability Scale fatigue score, UKNDS20). Subjects were considered to be free of depressive symptoms if their BDI II scores were below 13, to have mild depressive symptoms if their BDI II scores were between 14 and 19, to have moderate depressive symptoms if their scores were between 20 and 28, and to have severe depressive symptoms if their scores were greater than 29. Based on the UKNDS, each subject was considered to have permanent subjective fatigue impacting daily activities if their score was at least three.

Statistical Analyses

Statistical analyses were performed with Statview version 5.0 software for Windows. For age, disease duration, and NP scores, the results are shown as the means ± the standard deviations (SD). For the EDSS, BDI II, anxiety, and fatigue scores, the results are shown as the medians (ranges).

Unpaired t-tests were used to compare clinical characteristics, such as sex, age, education level, and disease duration, between PPMS and RRMS patients. Mann-Whitney U tests were used to compare the median EDSS, BDI II, anxiety, and fatigue scores of the two groups of MS patients.

Each MS patient score was compared with the mean value of that patient’s group of HCs matched for age, sex, and education level. z-scores were calculated for each NP score with the following formula: (patient’s score - mean value of HC group matched for age, sex, and education level)/standard deviation of the matched HC.

For a given NP score, patients were considered impaired if their z-scores were below the fifth percentile for their matched HC group. z-scores were also calculated for each cognitive domain using the following formula: (sum of the patient’s NP z-scores for each domain/ the number of z-scores in each domain). Patients were considered impaired in a given domain if their z-score were below the fifth percentile for their matched HC group. A chi-squared test was used to compare the proportions of PPMS and RRMS patients that were considered impaired in NP and the cognitive domain z-scores of these groups of patients. A t-test was used to compare the mean of the z-scores for the NP tests and for the cognitive domains between PPMS and RRMS patients. As a measure of effect sizes, the Cohen d was calculated. This value indicates the magnitude of the mean difference in SD units. According to Cohen,21 effect sizes can be interpreted as being small (d = 0.2), medium (d = 0.5), or large (d ≥ 0.8).

Because PwMS had higher scores for depressive symptoms, anxiety, and subjective fatigue than their matched HCs, Pearson correlation analyses were performed between each NP score and the BDI II, STAI-S, and subjective fatigue scores in a post-hoc procedure. An analysis of covariance (ANCOVA) was performed for each NP and cognitive domain z-score that differed between PPMS and RRMS patients. EDSS scores were entered into this post-hoc procedure to investigate possible

114

differences in task performance due to this variable. For all analyses, differences were considered significant when the p values were less than 5%.

RESULTS

Clinical Characteristics of PPMS and RRMS Patients and HCs

Forty-one PPMS patients, 60 RRMS patients, and 415 HCs were included in this study. Table 2 shows the clinical characteristics of the PwMS. Based on their demographic and clinical characteristics, 41 PPMS patients and 60 RRMS patients were compared with 263 and 310 HCs, respectively. Thirty-six PPMS patients (87.8%) and 53 RRMS patients (88.3%) were taking disease-modifying drugs at the time of the examination.

PPMS and RRMS patients had higher BDI II, STAI-S and subjective fatigue scale scores than their respective matched HCs (p<0.001). Twenty-four PPMS patients (58.5%) and 267 HCs matched to PPMS patients (93.3%) had no depressive symptoms. Seventeen PPMS patients (41.5%) and 16 of their matched HCs counterparts (0.06%) had mild to moderate depressive symptoms. No PPMS patients and three of their matched HCs (0.01%) counterparts had severe depressive symptoms. Forty RRMS patients (66.7%) and 288 of their matched HCs (92.9%) counterparts had no depressive symptoms. Nineteen RRMS patients (31.7%) and 21 of their matched HC (0.07%) counterparts had mild to moderate depressive symptoms. One RRMS patient (0.02%) and one matched HC (0.003%) counterpart had severe depressive symptoms. There was no difference between the percentages of PPMS and RRMS patients with subjective fatigue, but PPMS patients had worse subjective fatigue z-scores than the RRMS patients did (p<0.05).

115

Table 2. Clinical characteristics of 41 PPMS and 60 RRMS patients

SD: standard deviation, F: female; M: male, EDSS: Expanded Disability Status Scale, BDI II: Beck Depression Inventory II; STAI-S: State-Trait Anxiety Inventory-State. p values from a chi-squared test comparing sex; p values from non-matched t-tests comparing means; p values from Mann-Whitney U tests comparing medians.

PwMS Performed Worse than their Matched Controls, a nd PPMS Patients had More Extensive Cognitive Impairment Compared with HCs than the RRM S Patients did

NP scores were compared between each group of PwMS and the HCs matched for age, sex, and education level. PwMS showed a range of cognitive deficits compared with their matched HCs (Fig. 1). Three PPMS patients and one RRMS patient did not perform the visual scanning and flexibility TAP tests due to hand disabilities.

PPMS patients performed more poorly than HCs on 16 out of 23 (69.6%) NP scores, whereas RRMS patients exhibited lower NP performance on 5 out of 23 scores (21.7%) when compared with their matched HCs (Table 3). PPMS patients performed more poorly than their matched HCs, except for the following scores: RT and AA ratios for auditory divided attention, AA ratio for visual divided attention, AA for visual scanning with a target, 10/36 SPART-IR, 10/36 SPART-DR, and the copy for the RCF (Table 3). RRMS patients had lower performances than their matched HCs on the CSCT, the alertness test, the RT for visual scanning with and without a target, and the PASAT 3s (Table 3). More PPMS than RRMS patients were impaired according to the following scores: CSCT, WLG 90, AA for flexibility, SRT-CLTR, SRT-DR, and 10/36 SPART-DR (Fig. 1).

PPMS patients RRMS patients p-value

Gender F/M, n 24/17 49/11 <0.05

Mean age±SD (years) 52.1±8.7 37.3±9.9 <0.001

High education level: n (%) 19 (46.3) 37 (61.7) NS

Mean disease duration ±SD (years) 4.8±3.9 4.1±3.0 NS

Median EDSS scores (range) 3.5 (1.5-7.0) 1.5 (0-4.5) <0.001

Median BDI II scores (range) 12 (5-27) 9 (0-34) NS

Median STAI-S scores (range) 32 (20-61) 34 (20-60) NS

Median fatigue scores (range) 1 (0-3) 1 (0-5) NS

116

PPMS Patients were Impaired in More Cognitive Domai ns than RRMS patients

PPMS patients performed worse than their matched HCs in almost all cognitive domains, except visual episodic memory and visuoconstruction. RRMS patients performed worse than their matched HCs in the IPS and working memory domains. The performance of PPMS patients was worse than that of RRMS patients for working memory (p<0.05) and verbal episodic memory z-scores (p<0.05). The proportions of patients impaired in at least two domains were 47.4% in PPMS and 18.6% in RRMS (p<0.01).

For the two cognitive domains of executive function and verbal episodic memory, the percentage of PwMS impaired relative to HCs was significantly greater for PPMS than RRMS patients (Fig. 2). Unlike PPMS patients, RRMS patients were impaired in RT for visual divided attention when compared with their matched HCs, but the level of impairment reflected by this score was not different between the PPMS and RRMS patients (Fig. 1).

PPMS Patients Performed Worse than RRMS Patients in Cognitive Testing

Regarding cognitive domains, more than 20% of PPMS patients performed worse than their matched HCs with respect to IPS and verbal episodic memory, whereas more than 20% of RRMS patients exhibited only IPS impairment when compared with their matched HCs (Fig. 2). PPMS patients had lower cognitive z-scores than RRMS patients on the CSCT (p<0.01), the SRT-CLTR (p<0.05), and the numeral forward span test (p<0.05). PPMS patients also had longer RTs in the flexibility TAP test (p<0.05) and lower AAs in the TAP flexibility test (p<0.01).

117

Figure 1 : Percentages of impaired MS patients in comparison with their matched controls for each neuropsychological

score.

CSCT: Computerized Screening Cognitive Test, RT: reaction times, AA: Accurate Answers, PASAT 3s: Paced Auditory Serial Addition Test 3.0 s, WLG 90: Word List Generation test, SRT: Selective Reminding Test, LTS: Long-Term Storage, CLTR: Consistent Long-Term Retrieval, SRT-DR: Delay Recall, 10/36 SPART-IR: Spatial Recall Test for immediate recall for short visuo-spatial memory, 10/36 SPART-DR: Spatial Recall Test for long-term visuo-spatial memory, RCF: Rey Complex Figure. p-values: *p<0.05; † p<0.01 for a chi-squared test comparing proportions of impaired PPMS and RRMS patients.

Ruet Tables

118

Figure 2 : Percentages of impaired MS patients in comparison with their matched controls for each cognitive domain.

IPS: information processing speed, PPMS: primary progressive multiple sclerosis patients, RRMS: relapsing-remitting multiple sclerosis patients.

*p<0.05 for a chi-squared test comparing proportions of impaired PPMS and RRMS patients.

Ruet Tables

119

Table 3 . Mean neuropsychological scores and effect sizes for PPMS and RRMS patients in comparison with their matched controls

PPMS(n=41)ξξξξ HCs(n=263) d RRMS(n=60)ξξξξ HCs(n=310) d CSCT 37.4±10.7‡ 45.6±4.5 1.8 48.3±9.8† 51.7±4.5 0.8

Alertness 344.0±73.7‡ 291.9±24.2 2.2 312.1±71.5‡ 274.7±23.5 1.6 RT visual scanning with a target 3872.3±1123.4‡ 3045.6±161.3 5.1 3420.1±990.0‡ 2905.9±203.7 2.5 RT visual scanning without a target

6880.1±2261.9‡ 5571.3±313.0 4.2 5930.9±1970.5† 5254.9±397.3 1.7

RT flexibility 1086.9±398.2‡ 853.7±85.4 2.7 822.7±264.7 768.9±78.7 0.7 AA visual scanning with a target 36.2±10.9 39.9±7.1 0.5 39.5± 8.1 40.4±7.3 0.1 AA visual scanning without a target

49.7±0.6* 49.9±0.4 0.5 49.0±6.5 49.9±0.4 2.0

Ratio RT Auditory divided attention

1.1±0.1 1.1±0.1 0.2 1.1±0.2 1.1±0.03 0.3

Ratio RT Visual divided attention 1.0±0.1* 0.8±0.3 0.4 0.9±0.1 1.0±0.0 1.4 Ratio AA Auditory divided attention

1.0±0.3 1.0±0.0 0 1.0±0.1 1.0±0.0 0

Ratio AA Visual divided attention 1.0±0.1 1.0±0.0 0 1.0±0.1 1.0±0.1 0 PASAT 3s 37.7±12.4‡ 45.0±3.6 2.0 41.9±13.9* 45.7±4.0 1.0 Numeral forward span test 5.2±0.8† 5.5±0.5 0.6 5.7±1.1 5.7±0.5 0 Numeral backward span test 3.6±0.7† 4.0±0.2 2.0 4.1±0.9 4.2±0.6 0.2 Stroop 45 -31.5±8.6† -36.1±3.1 1.5 -30.5±9.0 -32.7±2.7 0.8 WLG 90 26.2±7.8* 29.2±1.9 1.6 27.7±7.4 29.2±2.9 0.5 AA Flexibility 92.3±9.3* 95.7±1.7 2.0 95.0±13.3 95.8±1.0 0.8 SRT-LTS 44.6±15.1‡ 52.8±3.8 2.2 54.9 ±9.6 57.2±2.5 0.9 SRT-CLTR 36.4±16.5‡ 45.6±4.1 2.2 47.6±12.9 50.3±3.2 0.8 SRT-DR 9.2±2.6† 10.4±0.8 1.5 10.7±1.5 10.7±0.4 0 10/36 SPART- IR 20.9±4.5 21.4±1.2 0.4 23.7±4.8 23.1±1.8 0.3 10/36 SPART- DR 7.0±2.9 7.3±0.5 0.6 8.5±1.8 8.0±0.8 0.7 RCF copy 32.4±3.0 33.1±1.0 0.7 33.4±4.0 33.8±1.0 0.4

Ruet Tables

120

IPS: Information Processing Speed, CSCT: Computerized Screening Cognitive Test, RT: reaction times, in milliseconds, AA: Accurate Answers, PASAT 3s: Paced Auditory Serial Addition Test 3.0 s, WLG: Word List Generation test, SRT: Selective Reminding Test, LTS: Long-Term Storage, CLTR: Consistent Long-Term Retrieval, SRT-DR: Delay Recall, 10/36 SPART IR: Spatial Recall Test for immediate recall of short visuo-spatial memory, 10/36 SPART-DR: Spatial Recall Test for long-term visuo-spatial memory, RCF: Rey Complex Figure. ξ: Visual scanning and flexibility (RT and AA) scores were obtained for only 38 PPMS patients, and flexibility (RT and AA) scores were obtained for 59 RRMS patients. Effect sizes were estimated from the Cohen’s d. p value: * p<0.05; † p<0.01; ‡p<0.001

121

Confounding Factors

EDSS

ANCOVA analyses performed for each NP and cognitive domain z-score that differed between PPMS and RRMS patients showed that after controlling for EDSS, the results remained largely unchanged (see e-Tables 1 and 2).

e-table 1. ANCOVA for NP z-scores differents between PPMS and RRMS patients.

F p

z-score CSCT

EDSS score

EDSS score *form

0.17

0.33

0.69

0.57

z-score SRT-CLTR

EDSS score

EDSS score *form

0.15

0.35

0.70

0.56

z- score numeral forward span

EDSS score

EDSS score *form

0.25

0.01

0.62

0.93

z-score RT flexibility

EDSS score

EDSS score *form

1.05

0.04

0.31

0.85

z-score AA flexibility

EDSS score

EDSS score*form

0.03

1.11

0.87

0.29

CSCT: Computerized Screening Cognitive Test; SRT: Selective Reminding Test; CLTR: Consistent Long-Term Retrieval; RT: reaction time; AA: Accurate Answers; EDSS: Expanded Disability Status Scale; F: procedure of the least significant difference method (PLSD) of Fisher.

122

e-table 2. ANCOVA for cognitive domains z-scores differents between PP and RRMS

F p

z-score working memory

EDSS score

EDSS score *form

0.11

0.05

0.75

0.83

z-score verbal episodic memory

EDSS score

EDSS score *form

0.09

1.66

0.77

0.66

EDSS: Expanded Disability Status Scale; F: procedure of the least significant difference method (PLSD) of Fisher.

Depressive Symptoms, Anxiety, and Fatigue

As PwMS had higher scores for depressive symptoms, anxiety, and subjective fatigue than their matched HCs, correlation analyses between NP scores and the BDI II, STAI-S, and subjective fatigue scores were performed. No correlations were found between these variables.

DISCUSSION

This study allowed the comparison of the cognitive performances of patients with PPMS and patients with RRMS using a unique methodology and adequate HCs. PPMS patients had a wide range of cognitive deficits affecting IPS, attention, working memory, executive function, and verbal episodic memory, whereas the impairment in RRMS patients was limited to IPS, attention, and working memory when compared with their respective matched HCs.

All the demographic and clinical characteristics of the two PwMS groups were similar except for age, sex, and EDSS. We used z-scores based on the data from a large sample of HCs matched for age and sex to account for these differences. To our knowledge, this is the first published study with an HC sample of this size that accounts for age, sex, and education level to adequately characterize the cognitive patterns of PPMS and RRMS patients. Although, as expected, EDSS scores were higher in PPMS than RRMS patients, the differences in NP task performance remained unchanged after controlling for EDSS. This finding is in contrast to the results obtained for 55 PPMS and 108 RRMS patients in a study from Amsterdam.4 After controlling for EDSS, RRMS and PPMS patients were no longer different with respect to the PASAT 3s and SDMT scores in that study. Additionally, the observed differences in NP task performance between the two MS phenotypes could not be due to disease duration, as there was no significant difference in disease duration between the PPMS and RRMS patients in our study.

123

PPMS patients performed significantly worse than their matched HCs on almost all NP tests. In accordance with previous studies, the cognitive impairment of PPMS patients included a wide range of domains such as IPS, attention, working memory, verbal episodic memory, and executive function.2,4,22-26 Although cognitive impairment in PPMS has previously been documented, it is noteworthy that the reported rates vary widely in previous studies (from 7% to 58%).2,23,26 In our study, 47.4% of PPMS patients were impaired in at least two cognitive domains. These data extend the results of Wachowius et al. (2005), who found impairments in at least two cognitive domains in 50% of a sample of PPMS patients when compared with an HC group.23

RRMS patients performed worse than their matched HCs on NP tests concerning IPS and working memory. It is well known that RRMS patients frequently show impairment on the SDMT and the PASAT.27 IPS impairment appears to be a central cognitive defect that is mainly reported in RRMS patients.28-30

The main finding of our study is the difference in frequency and severity of cognitive impairment in PPMS patients in comparison with RRMS patients. PPMS patients were more frequently impaired than RRMS patients in executive function and verbal episodic memory and performed more poorly than RRMS patients on working memory and verbal episodic memory tests. IPS was frequently impaired in PwMS, but the difference in frequency and severity of IPS impairment between PPMS and RRMS was not significant. The poor IPS performance of PPMS patients has been reported in previous studies.4, 30-31 Denney et al. (2004) reported that, after controlling for age, sex, education level, fatigue, and depression, the only cognitive measures in which PPMS and RRMS patients differed from HCs were those related to IPS.31 PPMS patients differed from RRMS patients in verbal episodic memory, unlike the results of Huijbregts et al.4 Gaudino et al. (2001) reported that PPMS patients had greater difficulty acquiring new verbal information than RRMS patients.3 Similar to the findings of Huijbregts et al. (2004), we found no differences in visuospatial memory impairment between RRMS and PPMS patients as measured by overall 10/36 SPART performance.4 Nevertheless, PPMS patients were more frequently impaired than RRMS patients on the 10/36 SPART delayed recall tasks in our study, in contrast to the results of Huijbregts et al.4

Interestingly, the severity of cognitive impairment was more pronounced in PPMS patients than in RRMS patients, as the effect sizes for almost all NP scores were two-fold higher in PPMS than in RRMS patients; this finding is in accordance with the results of Huijbregts et al. (e.g., their SDMT and PASAT 3s results).4

One limitation of this study is the absence of magnetic resonance brain imaging of the PwMS to better understand the mechanisms underlying the cognitive impairments of patients with these different types of MS. We have previously shown that cognitive impairment may be a marker of diffuse brain abnormalities in early RRMS patients.32 The observed group differences in the present study could reflect the fact that PPMS patients have more widespread brain damage; specifically, pathological studies suggest that PPMS patients have diffuse pathology in normal-appearing white matter and gray matter injury (both cortical and deep gray matter damage).33

ACKNOWLEDGEMENTS

We thank Loock T for his contribution to data collection and Ouallet JC for his help recreating the MS patients.

124

References 1. Comi G, Filippi M, Martinelli V, et al. Brain magnetic resonance imaging correlates of cognitive impairment in primary and secondary progressive multiple sclerosis. J Neurol Sci 1995;132:222-227. 2. Camp SJ, Stevenson VL, Thompson AJ. et al. Cognitive function in primary progressive andtransitional progressive multiple sclerosis: a controlled study with MRI correlates. Brain 1999;122(Pt 7):1341-1348. 3. Gaudino EA, Chiaravalloti ND, DeLuca J, Diamond BJ. A comparison of memory performance in relapsing-remitting, primary progressive and secondary progressive, multiple sclerosis. Neuropsychiatry Neuropsychol Behav Neurol 2001;14(1):32-44. 4. Huijbregts SC, Kalkers NF, de Sonneville LM, de Groot V, Reuling IE, Polman CH. Differences in cognitive impairment of relapsing remitting, secondary, and primary progressive MS. Neurology 2004 27;63(2):335-339. 5. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research proposals. Ann Neurol 1983;13:227–231. 6. Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the ‘McDonald Criteria’’. Ann Neurol 2005;58:840–846. 7. Brochet B. Assessing incapacity at early stages of Multiple sclerosis using the EDSS. Rev Neurol (Paris) 2009;165(Suppl 4):S173-S179. 8. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33(11):1444–1452. 9. Rao SM, the Cognitive Function Study Group of the National Multiple Sclerosis Society. A manual for the Brief Repeatable Battery of Neuropsychological Tests in multiple sclerosis. Milwaukee, WI: Medical College of Wisconsin; 1990. 10. Bever CT, Grattan L, Panitch HS, Johnson KP. The brief repeatable battery of neuropsychological tests for multiple sclerosis: a preliminary serial study. Mult Scler 1995;1:165-169. 11. Boringa JB, Lazeron RH, Reuling IE, et al. The brief repeatable battery of neuropsychological tests: normative values allow application in multiple sclerosis clinical practice. Mult Scler 2001;7:263-267. 12. Ruet A, Charré-Morin J, Brochet B, Hamel D, Deloire M. How to detect cognitive impairment in relapsing-remitting and primary-progressive multiple sclerosis. Mult Scler 2011;17(Suppl 10):S277-S505. 13. Harris JG, Wagner B, Cullum CM. Symbol vs. Digit Substitution Task Performance in Diverse Cultural and Linguistic Groups. Clin Neuropsychol 2007;21(5):800-810. 14. Zimmermann P, Fimm B. TAP 2.1 Tests d’évaluation de l’attention Würzelen: Psytest 2009. 15. Golden CJ. A group version of the Stroop Color and Word Test. J Pers Assess 1975; 39, 386-388. 16. Wechsler D. Wechsler Adult Intelligence Scale-III administration and scoring manual. San Antonio, TX: The Psychological Incorporation 1997. 17. Rey A. L’examen psychologique dans les cas d’encephalopathie traumatique. Arch Psychol (Geneve) 1941;28:286–340. 18. Beck AT, Brown G, Steer RA. Beck Depression Inventory II manual. San Antonio, TX: The Psychological Corporation 1996. 19. Spielberger CD. Manual for the state-trait anxiety inventory. Palo Alto: Consulting Psychologists Press; 1983.

125

20. Sharrack B, Hughes RA. The Guy's Neurological Disability Scale (GNDS): a new disability measure for multiple sclerosis. Mult Scler 1999;5(4):223-233. 21. Cohen J. Statistical power analysis for the behavioral sciences, 2nd ed. Hillsdale: Lawrence Earlbaum Associates; 1988. 22. Foong J, Rozewicz L, Chong WK, Thompson AJ, Miller DH, Ron MA. A comparison of neuropsychological deficits in primary and secondary progressive multiple sclerosis. J Neurol 2000;247:97–101. 23. Wachowius U, Talley M, Silver N, Heinze HJ, Sailer M. Cognitive impairment in primary and secondary progressive multiple sclerosis. J Clin Exp Neuropsychol 2005;27:65–77. 24. Ukkonen M, Vahvelainen T, Hämäläinen P, Dastidar P, Elovaara I. Cognitive dysfunction in primary progressive multiple sclerosis: a neuropsychological and MRI study. Mult Scler 2009;15:1055-1061. 25. Penny S, Khaleeli Z, Cipolotti L, Thompson A, Ron M. Early imaging predicts later cognitive impairment in primary progressive multiple sclerosis. Neurology 2010;74:545-552. 26. Potagas C, Giogkaraki E, Koutsis G, et al. Cognitive impairment in different MS subtypes and clinically isolated syndromes. J Neurol Sci 2008;267:100-106. 27. Chiaravalloti ND, DeLuca J. Cognitive impairment in multiple sclerosis. Lancet Neurol 2008; 7:1139–1151. 28. Forn C, Belenguer A, Parcet-Ibars MA, Avila C. Information-processing speed is the primary deficit underlying the poor performance of multiple sclerosis patients in the Paced Auditory Serial Addition Test (PASAT). J Clin Exp Neuropsychol 2008;13:1–8. 29. DeLuca J, Chelune GJ, Tulsky DS, Lengenfelder J, Chiaravalloti ND. Is speed of processing or working memory the primary information processing deficit in multiple sclerosis? J Clin Exp Neuropsychol 2004;26:550–562. 30. De Sonneville LMJ, Boringa JB, Reuling IEW, Lazeron RHC, Adèr HJ, Polman CH. Information processing characteristics in subtypes of multiple sclerosis. Neuropsychologia 2002;40:1751–1765. 31. Denney DR, Lynch SG, Parmenter BA, Horne N. Cognitive impairment in relapsing and primary progressive multiple sclerosis: mostly a matter of speed. J Int Neuropsychol Soc 2004;10:948–956. 32. Deloire M, Salort E, Bonnet M, et al. Cognitive impairment as marker of diffuse brain abnormalities in early relapsing remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2005;76(4):519-526. 33. Kutzelnigg A, Lassmann H Cortical lesions and brain atrophy in MS. J Neurol Sci 2005;233:55–59.

126

Discussion de cet article

1. Résumé des résultats de cet article

Les troubles cognitifs diffèrent entre les patients ayant une SEP-PP et ceux atteints d’une SEP-RR. La figure 2 illustre les pourcentages des domaines cognitifs altérés dans ces deux types de population de patients atteints de SEP par rapport à leurs sujets contrôles appariés pour l'âge, le sexe, et le niveau d'éducation. Il est évident que la VTI est le domaine cognitif le plus fréquemment atteint dans les deux groupes de patients ayant une SEP-PP. Les deux domaines cognitifs, qui diffèrent entre les patients ayant une SEP-PP et les ceux ayant une SEP-RR, concernaient la mémoire épisodique verbale et les fonctions exécutives en terme de différence de fréquence et la mémoire épisodique verbale et la mémoire de travail en terme de différence de sévérité. Près de la moitié des patients avec une SEP-PP et moins d'un cinquième des patients atteints de SEP-RR présentaient une atteinte à au moins deux domaines cognitifs.

Les patients ayant une SEP-PP présentaient des déficits cognitifs plus marqués que les patients atteints de SEP-RR, même après prise en compte des déficiences physiques, évaluées par le score EDSS, et avec des durées moyennes de maladie similaires.

2. Intégration des résultats aux données actuelles de la littérature

De nombreuses études ont rapporté la fréquence de l’atteinte cognitive chez les patients ayant une SEP-RR, alors que seulement quelques études ont porté sur la cognition chez les patients atteints de SEP avec une forme progressive (Comi et al., 1995; Foong et al., 2000; Gaudino et al., 2001; Huijbregts et al., 2004; Wachowius et al., 2005). Il manque d’études analysant le fonctionnement cognitif en prenant en compte les différentes caractéristiques des patients présentant différents phénotypes de SEP. Quelques études ont utilisé une méthodologie semblable pour comparer des échantillons de patients atteints de SEP-RR et de SEP-PP. Ces études ont rapporté une atteinte cognitive plus fréquente chez les patients ayant une SEP-PP par rapport aux patients atteints de SEP-RR, mais souvent sans utiliser de groupes témoins. Ainsi, dans une étude portant sur la mémoire de 18 patients ayant une SEP-PP, 21 patients ayant une SEP-RR, et 25 patients ayant SEP-SP, il a été retrouvé une atteinte cognitive plus fréquente et plus sévère parmi les patients ayant une forme progressive de SEP (Gaudino et al., 2001). Une autre étude incluant une plus large batterie NP et un plus grand échantillon de patients (108 patients avec une SEP-RR, 71 avec une SEP-SP, et 55 avec une SEP-PP) comparativement à 67 sujets contrôles a révélé des résultats similaires (Huijbregts et al., 2004). Les patients atteints de SEP-RR obtenaient de meilleurs résultats que les patients ayant une SEP-PP pour les tâches de VTI (SDMT et PASAT-3 secondes) après ajustement de l'âge et du sexe. Cependant, après ajustement du score EDSS, les performances à ces tests NP (SDMT et PASAT-3 secondes) ne différaient plus entre les patients ayant une SEP-RR et ceux avec une SEP-PP. Dans une autre étude, 50% des patients ayant une SEP-PP présentaient des performance inférieures à au moins deux domaines cognitifs par rapport à des sujets contrôles (Wachowius et al., 2005). Même si ces résultats sont en accord avec les nôtres, il faut souligner le fait que l’interprétation de ces études reste limitée parce qu'il n’existe pas de groupe de sujets contrôles appariés strictement à chaque groupe de patients en fonction de leur forme de SEP et de leurs caractéristiques cliniques contrairement à ce que nous présentons dans notre étude.

127

Le ralentissement de la VTI est un déficit cognitif central chez les patients ayant une SEP (DeLuca et al., 2004; Forn et al., 2008). Dans notre étude, les patients atteints de SEP-RR et ceux ayant une SEP-PP étaient principalement déficitaires dans les tests de VTI sans différence significative entre eux concernant la fréquence, sauf pour le CSCT, et la gravité, sauf également pour le CSCT et le TR pour le test de flexibilité de la TAP. Ces résultats sont cohérents avec les résultats de Denney et al. qui ont rapporté une différence entre les patients ayant une SEP-PP et ceux ayant une SEP-RR comparativement aux sujets contrôles uniquement pour les tâches cognitives relatives à la VTI, après ajustement de l'âge, du sexe, du niveau d'éducation, de la fatigue, et de la dépression (Denney et al., 2004). En outre, aucune différence significative n'a été observée concernant l’atteinte de la VTI et de l’attention entre les deux formes progressives de la maladie (plus de la moitié des patients atteints de SEP progressive étaient déficitaires pour ces fonctions) (Wachowius et al., 2005).

La forme évolutive de la SEP pourrait jouer un rôle majeur dans le type de dysfonctionnement cognitif détecté chez les patients. Ceci pourrait être en partie expliqué par les différences d’atteinte tissulaire cérébrale mise en évidence chez les diverses formes de SEP. Nos résultats pourraient refléter l'étendue des lésions cérébrales présentes chez les patients ayant une SEP-PP et soutenir le rôle pronostique des troubles cognitifs dans la SEP. Récemment, il a été rapporté que l’atteinte à la fois focale et diffuse de la SB et de la SG pouvait contribuer à la prédiction des performances cognitives au niveau de la VTI, de l'attention et des fonctions exécutives dans une étude de 31 patients ayant une SEP-PP suivis pendant 5 ans (Penny et al., 2010). L’importance de la perte axonale pourrait expliquer ces résultats concernant les formes progressives de SEP. L’estimation de l’atrophie cérébrale a été proposé comme meilleur marqueur prédictif des performances cognitives dans la SEP que la charge lésionnelle évaluée chez ces patients (Benedict et al., 2004b). Il a été émis l'hypothèse que l'atrophie chez les patients ayant une SEP-PP pourrait représenter plus une axonopathie diffuse, reflétant des lésions axonales à la fois au niveau de la SBAN (Schmierer et al., 2004) et au niveau de la SG (Cercignani et al., 2001). Récemment, une étude immunohistochimique post-mortem incluant 26 patients ayant une SEP-PP a mis en évidence une inflammation diffuse généralisée méningée pouvant jouer un rôle important dans la pathogenèse des lésions de la SG au niveau cortical et contribuer à l’aggravation clinique dans ces formes progressives de SEP (Choi et al., 2012).

3. Défis et difficultés

Les estimations de la prévalence des déficits cognitifs dans la SEP varient selon la nature et la taille de l'échantillon de sujets étudiés (comparaison de patients à des sujets témoins ou de patients entre eux), la définition de l’atteinte cognitive, et les tests inclus dans la batterie NP. Même si l'échantillon de patients ayant une SEP-PP était plus faible que celui des patients avec une SEP-RR, la force de notre étude a été d'utiliser un échantillon de sujets contrôles strictement appariés et de taille importante pour réaliser la comparaison du fonctionnement cognitif entre les patients atteints de SEP-PP et ceux ayant une SEP-RR avec la même méthodologie.

128

4. Conclusion et perspectives

Nos données suggèrent que le dysfonctionnement cognitif est fréquent non seulement chez les patients atteints de SEP-RR, mais aussi chez les patients ayant une SEP-PP. Nos résultats illustrent la valeur pronostique des troubles cognitifs chez les patients ayant une SEP, et nous insistons sur la nécessité de les détecter à la fois dans la pratique clinique et dans les études de recherche dans la SEP. Nous soulignons le fait que le ralentissement de la VTI est un élément clé de l'atteinte cognitive dans la SEP, et sa détection apparaît comme une priorité dans la gestion des patients ayant une SEP. Dans la plupart des batteries NP, les tests chronométrés apparaissent comme les plus pertinents pour détecter les déficits cognitifs dans la SEP, car ces tests sont particulièrement sensibles pour estimer la VTI. Nous proposons le CSCT comme un bon candidat pour évaluer la VTI. Une étude de validation sera détaillée dans l'article 7 de cette thèse.

129

ARTICLE 6:

COGNITIVE IMPAIRMENT, HEALTH-RELATED QUALITY OF LIF E AND VOCATIONAL STATUS IN THE EARLY STAGES OF MULTIPLE SCLEROSIS: A 7-YEAR LONGITUDINAL STUDY

Article accepté dans Journal of Neurology.

COGNITIVE IMPAIRMENT, HEALTH RELATED QUALITY OF LIF E AND VOCATIONAL STATUS AT EARLY STAGES OF MULTIPLE SCLEROSIS: A 7-YEAR LON GITUDINAL STUDY

Aurélie Ruet 1,2, MD, Mathilde Deloire1,2 PhD, Delphine Hamel1, Jean-Christophe Ouallet1.2, MD, PhD, Klaus Petry1, PhD and Bruno Brochet1,2 MD

1 Université de Bordeaux, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, F-33076 Bordeaux, France

2 CHU de Bordeaux, INSERM-CHU CIC-P 0005, & Service de Neurologie, F-33076 Bordeaux, France

Corresponding author: Prof. Bruno Brochet, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, case 78, Université de Bordeaux, 146 rue Léo Saignat, 33076 Bordeaux cedex, France.

Telephone: +33(0)557571552; Fax: +33(0)5 57574818;

e-mail: bruno.brochet@chu-bordeaux.fr

The authors do not have to disclose any actual or potential conflict of interest including any financial, personal or other relationships with other people or organisations that could inappropriately influence, or be perceived to influence, their work.

130

ABSTRACT

The association between cognitive impairment, health-related quality of life (HRQoL) and vocational status is not well known in multiple sclerosis (MS). This study assesses this association in a sample of 65 newly diagnosed MS patients followed longitudinally. Each patient underwent a standardised clinical assessment, cognitive tests and the HRQoL SEP-59 questionnaire six months after the MS diagnosis (baseline) and seven years later (y7).Vocational status was also established at baseline and at y7 in MS patients. The HRQoL at baseline was severely reduced in MS patients compared with healthy subjects. The independent predictors for HRQoL composite scores at y7 were the baseline depression score and the memory Z-score. Moreover, 81.5% of MS patients worked at baseline and only 54.4% worked at y7. Among the MS patients who did not work at y7, 72.7% of them were cognitively impaired, while 27.3% were unimpaired at baseline. The vocational status at y7 was significantly associated with the baseline IPS Z-score, EDSS and age. Vocational status at y7 and its change over seven years was significantly associated with cognitive deterioration. IPS or memory dysfunction in the early stages of MS is correlated with a decreased level in health perception, independent of fatigue, depression and physical disability. Cognitive impairment at the diagnosis of MS increases the risk of changing vocational status in MS patients seven years later.

KEYWORDS: Multiple Sclerosis; Information Processing Speed; Cognition; Quality of life; Work; Vocational Status.

131

1. Introduction

The goal of health-related quality of life (HRQoL) scales is to provide tools adapted to the evaluation of various dimensions of a given situation, including concerns that are not usually taken into account by standard medical assessment [1]. Many studies have used these scales to assess the perceived health in persons with multiple sclerosis (PwMS) [2]. These studies consistently showed that PwMS have a significant decrease in HRQoL. The key factors for such low HRQoL scores have been studied, and low ratings of HRQoL have been associated with physical disability, fatigue and depression [2]. Physical HRQoL was associated with EDSS, fatigue, and depression, while mental HRQoL was associated with only depression and fatigue [3].

Cognitive impairment is a prominent feature of multiple sclerosis (MS), and deficits occur in information processing speed (IPS), memory, attention, working memory, and executive functions [4]. The contribution of cognitive disturbances to low HRQoL has been hypothesised, but few studies have addressed this question [5-8]. One study including patients with CIS and RRMS with disease duration less than 3 years showed that after controlling for depression, only a few cognitive scores, measuring IPS, were mildly associated with several measures of HRQoL, including the physical composite score of the Short-Form 36 (SF-36) and the modified social support survey [8].

In their pioneering study on cognition in MS, Rao et al. [9] showed that PwMS with cognitive impairment are less likely to be working. The link between vocational status and cognitive impairment has been consistently confirmed by recent studies [3, 10-11]. In a cross-sectional study, HRQoL and vocational status were measured in a sample of 120 MS patients [3]. Vocational status was predicted by objective measures of cognitive function (information processing speed and executive functions) and disease duration. A few studies focused on PwMS at early stages of the disease [12-14] or included a significant sample of PwMS at these early stages [15]. These studies suggested that HRQoL is altered very quickly following the onset of the disease. To date, there is no longitudinal and prospective study focusing on the relationship between HRQoL, vocational status and cognitive impairment in a homogeneous group of MS patients at early stages. We took advantage of a longitudinal cohort of newly diagnosed MS patients [16-17] to study the link between HRQoL, vocational status and early cognitive status. The main objective of this study was to investigate the relationship between cognitive impairment at baseline in MS patients and HRQoL and vocational status at y7. We hypothesised that the cognitive impairment at baseline could predict the HRQoL and the vocational status seven years after the MS diagnosis.

2. Materials and Methods

2.1 Subjects

2.1.1 Patients

A longitudinal cohort study of newly diagnosed MS patients was initiated in the Aquitaine region (south-western France) in November 2000. Participation in the study was suggested to every patient who was newly diagnosed with clinically definite MS, according to Poser et al. criteria, by neurologists in the Aquitaine MS network (AQUISEP, a clinical network which includes private and hospital neurologists involved in MS care in the Aquitaine region). All eligible patients agreed to be included. The exclusion criteria were as follows: patients with other neurological diseases, history of psychiatric illness, head trauma and alcohol or drug abuse. The patients did not receive any compensation. Sixty-nine patients recently diagnosed with MS (less than 6 months) were enrolled, including fifty-eight patients with relapsing-remitting MS (RRMS) [19] and 11 patients with

132

progressive MS (PMS) (either primary progressive MS (PPMS) [19] or transitional progressive MS (TPMS)) [20]. The patients were consecutively referred to the coordinating centre and gave written informed consent before inclusion in the study, according to the Declaration of Helsinki. This study was approved by the local ethics committee.

One patient was excluded from the study due to reconsideration of the MS diagnosis, and another patient was excluded because he was evaluated during an exacerbation. All assessments were performed more than 30 days after the patient’s last exacerbation. The percentage of MS patients receiving disease-modifying therapy according to their clinical status and to the available recommendations at any time during the study was 95.8%.

2.1.2 Healthy subjects

A sample of 65 healthy subjects matched for sex, age and the number of years of education received a neuropsychological assessment at baseline. These subjects served as a reference for the cognitive evaluation of the MS patients. Their data, compared with that of the MS patients, allowed us to classify patients according to their cognitive status (cognitively impaired or unimpaired).

For the HRQoL, data from the French population were obtained from those published by Leplège et al. [21]. The mean value in each category for age- and gender-matched control subjects was compared to that of each patient.

2.2 Measures

2.2.1. Neuropsychological assessment

The neuropsychological assessment included the similarities subtest of the WAIS-R (testing conceptualisation), the Brief Repeatable Battery (BRB-N) with the Selective Reminding Test (SRT and 3 subscores: SRT-LTS = long-term storage; SRT-CLTR = consistent long-term retrieval and SRT-DR = delay recall, which tests verbal memory), the 10/36 spatial recall test for short- (SPART) and long-term visuo-spatial memory (SPART-DR), the Symbol Digit Modalities Test (SDMT, attention and IPS), the PASAT 3 and 2 seconds (PASAT 3s and PASAT 2s; working memory and IPS) and the Word List Generation test (WLG; verbal fluency). The RRMS patients were compared to 54 healthy subjects and the PMS patients were compared to 11 healthy subjects, all matched for sex, age and educational level.

At baseline, the MS patients were classified as patients cognitively impaired (PCI) if they performed less than 1.5 standard deviations (SD) below that of the matched controls on at least two tests of the battery, and as patients cognitively unimpaired (PCU) if they did not. For all neuropsychological tests, Z-scores were created based on the data from the healthy controls matched for sex, age and the number of years of education. The formula we used is as follows: Z-score = (patient’s score at baseline-mean value of matched healthy subjects at baseline)/standard deviation of matched healthy subjects at baseline. Two domain-specific Z-scores were created for memory (SRT-LTS, SRT-CLTR, SRT-DR, SPART and SPART-DR) and IPS (SDMT, PASAT 3s and 2s).

Patients were classified as cognitively deteriorated if their cognitive performances declined between y7 and baseline by at least 1.0 SD of controls in at least one domain (memory or IPS).

133

2.2.2 Quality of life

We used SEP-59 subscores to assess different dimensions of HRQoL. The SEP-59 scale is a specific French HRQoL MS scale, adapted from the MSQoL-54 [1]. It contains 59 items divided into 15 scales and scores, graded from 0% to 100%; higher scores indicate a better HRQoL. The MSQoL-54 was created by adding disease-specific items to those of the Short Form-36 health survey questionnaire (SF-36) [22]. It includes subscales from the SF-36 (physical functioning, role limitations relating to physical health (role physical), bodily pain, general health perceptions, vitality, social functioning, role limitations relating to mental health (role emotional), and mental health) and specific subscales from the SEP-59 (distress, cognitive function, sexual function, satisfaction with sexual function, overall HRQoL, sleep and social support). For the axes bodily pain, social functioning and general health perceptions, some questions are taken into account only for calculating the SEP-59 scores, but not the SF-36 scores. We used two composite scores derived from the SF-36 questionnaire: the Physical Composite Score (PCS/SF-36), which includes physical functioning, role physical, bodily pain, general health perceptions, and the Mental Composite Score (MCS/SF-36), which includes role emotional, mental health, vitality and social functioning. Data from the French population were obtained from Leplège [22]. The values from each patient were compared to the mean value for each category from the age- and gender-matched control subjects.

Then, we calculated two composite scores adapted from MSQoL-54 mental and physical composite scores. The Physical Composite Score/SEP-59 (PCS/SEP-59) included physical functioning, role physical, bodily pain, vitality, general health perceptions, social functioning and distress. The Mental Composite Score/SEP-59 (MCS/SEP-59) included role emotional, mental health, cognitive function, distress and overall QoL. Each subscale is scored from 0% to 100%, with higher scores indicating a better HRQoL.

2.2.3 Vocational status

The employed group consisted of patients who were paid workers, working full-time or with a reduction in their duties, and students. The unemployed group consisted of housewives, volunteers, recipients of disability benefits, persons on prolonged medical leave, and retired workers. Vocational status was established at baseline and at y7. The patients were classified according to their vocational status change between baseline and y7 (loss of employment, reduction in duties, or becoming workers).

2.2.4. Other clinical measures

The patients received a standardised neurological examination performed by a neurologist at baseline and at y7 to establish the Expanded Disability Status Score (EDSS). At y7, the patients were considered to have clinically declined if they had an EDSS score increase ≥1.0 with a baseline EDSS < 6.0 or if they had an EDSS score increase ≥0.5 with a baseline EDSS ≥ 6.0.

At baseline and at y7, other measures included the Montgomery and Asberg Depression Rating Scale (MADRS) to assess depressive symptoms and the 5-graded fatigue subscale of the United-Kingdom Neurological Disability Scale (UKNDS) to assess fatigue.

134

2.3 Statistical analysis

All statistical analyses were performed using the Statview version 5.0 software for Windows. MS patients were compared using the chi-square test for proportion comparisons (gender and cognitive impairment according to vocational status) and the t test for comparisons between MS patients at y7 and all MS patients at baseline (age, disease duration, EDSS, MSFC, MADRS, fatigue and cognitive scores at baseline). Student’s t test was used for SF-36 composite score and cognitive score comparisons between MS patients and healthy subjects. We compared MS patients according to vocational status at y7 based on age, EDSS, MADRS, fatigue scores and cognitive Z-scores using a t test. Deterioration of HRQoL composite scores during 7 years was assessed using the paired t test in MS patients.

Several univariate analyses were performed for prediction of the HRQoL and the vocational status at seven years and over seven years. The variables with a conservative significance level of p<0.25 in the univariate analysis were entered simultaneously in the regression model as independent variables for prediction of the HRQoL composite scores and vocational status. All parameters at the 0.05 level were removed from the model by stepwise elimination. Differences were considered significant for all analyses when the p values were less than 5%.

- Prediction of HRQoL:

Linear regression analyses were performed to assess the correlation between baseline clinical or cognitive variables (age, EDSS scores, depression, fatigue scores and cognitive Z- scores) and each HRQoL composite score (PCS/SEP-59 and MCS/SEP-59). The patients who did not complete an HRQoL questionnaire at baseline (n=2) were excluded from the statistical analyses.

- Prediction of vocational status:

Logistic regression analyses were used to assess the measures that predict vocational status at y7 or its change over 7 years. The dependent variables were vocational status at y7 in the 2 models. In one model, we entered age, the EDSS, MADRS and fatigue scores, and the Z-scores for IPS, memory, and WLG as independent variables. In the other model, we entered cognitive deterioration and EDSS deterioration as independent variables. The dependent variable was the change in vocational status over 7 years in the 2 last models, with an adjustment for the vocational status at baseline; one model used baseline parameters as independent variables, and the other model used cognitive and EDSS deterioration as independent variables. Working status at baseline was taken into account for vocational status change over 7 years because the loss of employment or a reduction in duties was relevant for our analysis of patients who worked at baseline.

3. Results

3.1. Demographic characteristics

Sixty-five MS patients completed the HRQoL questionnaire at baseline, and 48 patients completed it at y7. The demographic and clinical findings are presented in Table 1. There was no significant difference between the followed-up MS group (n=48) and the whole MS group (n=65) for demographics, neuropsychological and clinical parameters at baseline.

135

Table 1: Clinical and demographic findings of MS pa tients

MS patients at baseline

RRMS (n=54)

PMS (n=11)

MS patients followed for 7 years

RRMS (n=41)

PMS (n=7)

Gender (Male/Female) 20/45 12/36

Mean age (SD) 39.0±10.4 39.73±9.95

Mean disease duration (SD) (months) 31.2±38.2 30.69±38.02

EDSS median (range) 2.0 (0-6.0) 2 (0-6)

MSFC mean (SD) -0.003±0.599 -0.043 ±0.60

MADRS median (range) 4 (0-28) 3 (0-28)

Fatigue score median (range) 1 (0-5) 1 (0-5)

RRMS: Relapsing-Remitting Multiple Sclerosis patients ; PMS: Progressive Multiple Sclerosis patients ; SD: standard deviation

3.2. Cognitive evaluation

The patients had lower scores in comparison with matched healthy subjects for episodic memory (SRT-LTS, SRT-CLTR, SRT-DR and SPART-DR, all p<0.01), IPS (SDMT and PASAT 2 s, p<0.001, PASAT 3 s, p<0.01) and executive functions (similarities subtest of the WAIS-R), p<0.001).

Compared with matched healthy subjects, 34 of 65 MS patients (52.3%) were classified as cognitively impaired (PCI) at baseline, and 31 of 65 patients (47.7%) were classified as cognitively unimpaired (PCU).

The proportion of MS patients considered to have cognitively deteriorated over 7 years was 54.2% for the memory domain, 19.1 % for the IPS domain and 72.3% for the memory or IPS domain.

3.3. Neurological examination

The EDSS scores at baseline and at 7 years are detailed in Table 1. Twenty-eight of 48 patients (58.3%) were listed as having deteriorated according to the EDSS at 7 years.

136

3.4. Health related quality of life

3.4.1. HRQoL at baseline

Two patients did not complete the HRQoL questionnaire at baseline. Figure 1 presents the results of the SF-36 questionnaire (scores and composites) for the RRMS group (n=54) and the PMS group (n=11) and their matched controls from the French population. For all axes, the HRQoL was lower in the MS patients than in their matched controls. The results of the SEP-59 scale for each specific axis and the composite scores are also detailed for MS patients in Figure 1.

3.4.2. Evolution of HRQoL over seven years

Physical functioning (p=0.001), cognitive function (p<0.001), sexual function (p<0.05) and overall HRQoL (p<0.05) scales deteriorated during 7 years. The composite scores did not change significantly during 7 years in the MS patients.

3.4. 3.Predictors of HRQo L at seven years

The PCS/SEP-59 and the MCS/SEP-59 at y7 did not differ among MS patients according to their cognitive status at baseline (PCI/PCU).

For the PCS/SEP-59, the baseline variables entered in the multivariate model were age, the EDSS, MADRS and fatigue scores, and the Z-scores for memory and IPS. The PCS/SEP-59 was predicted independently by the MADRS score (p<0.001, R2= 0.230) and the memory Z-score (p<0.05, R2= 0.093) (p model <0.0001; R2 model =0.350).

For the MCS/SEP-59, the baseline variables entered in the multivariate model were age, the EDSS, MADRS and fatigue scores, and the Z-scores for memory and IPS. The MCS/SEP-59 was predicted independently by the MADRS score (p=0.0005, R2= 0.245) and the memory Z-score (p<0.05, R2= 0.124) at baseline (p model<0.0001; R2 model=0.376).

3.4.4. Predictors of HRQoL change over seven years

The PCS/SEP-59 and the MCS/SEP-59 over 7 years did not differ among MS patients according to their cognitive status at baseline (PCI/PCU). Over 7 years, the PCS/SEP-59 and the MCS/SEP-59 were not predicted by any of the baseline variables.

137

Figure 1: Scores of the SF-36 questionnaire and SF- 36 composites scores for patients with RRMS (n=54) and PMS (n=11) and their matched controls from the general population [21]. NS= not significant. PCS/SF-36 = Physical Composite Score including physical functioning, role physical, bodily pain, general health perceptions. MCS/SF-36 = Mental Composite Score including role emotional, mental health, vitality and social functioning. * indicates p<0.05; ** indicates p<0.01; *** indicates p<0.001.

138

3.5. Vocational status

3.5.1. Vocational status at baseline

At baseline, 53 of 65 (81.5%) MS patients worked. Among the workers, four patients were students, whereas among the unemployed patients, one patient was a housewife and one patient was a volunteer. Eight patients received disability benefits at baseline.

3.5.2. Vocational status at seven years

At seven years, 26 of 48 MS patients (54.1%) worked. Two patients had restricted their activities and were included in the group with vocational status change, and twenty-one patients had received disability benefits at seven years.

3.5.3. Vocational status change over seven years

None of the MS patients without work at baseline were working at year 7. Thirteen of 48 patients (27.1%) changed their employment status over the 7-year study period.

3.5.4. Predictors of vocational status

Cognitive status (PCI/PCU) was associated with vocational status at y7. Among the MS patients who did not work at y7, 72.7% of them were PCI at baseline, whereas 27.3% were PCU at baseline (p<0.05, OR=3.64, 95%: CI: 1.08-12.31). The cognitive status (PCI/PCU) in MS patients was not associated with their vocational status change over seven years.

The predictors of vocational status at 7 years and its change over seven years are detailed in Table 2 based on baseline variables (multivariate models with cognitive Z-scores) and in Table 3 based on cognitive and disability deterioration.

139

Table 2: Predictors of vocational status in MS pati ents based on baseline variables

Dependent variables Independent variables B OR [95 % CI]

Vocational status at y7 Ageb*

EDSSb*

IPS Z-scoreb*

WLG Z-scorea

Memory Z-scorea

MADRSa

Fatiguea

0.24

3.28

-3.08

1.28 [1.04- 1.56]

26.54 [1.67-420.93]

20 [1.59-333.33]

Change of vocational status over 7 y Age

EDSSa

IPS Z-scoreb*

Fatiguea

MADRS

Vocational status at yo

-0.837

2.31 [1.04-5.13 ]

*indicates p <0.05, ** indicates p<0.01

PCI: patients cognitively impaired, PCU: patients cognitively unimpaired

a Variables included in multivariate analyses but not significantly correlated with vocational status in final models.

b Variables significantly correlated with vocational status in multivariate analyses.

140

Table 3: Predictors of vocational status in MS pati ents based on cognitive and disability deterioration over seven years

Dependent variables Independent variables B OR [95 % CI]

Vocational status at y7 Cognitive deterioration

over 7 years*

EDSS deterioration over 7 years

1.94 6.97 [1.34- 36.34]

Change of vocational status over 7 years

Cognitive deterioration

over 7 years*

EDSS deterioration over 7 years

Vocational status

at y0

1.70

5.46 [1.01- 29.39]

*indicates p <0.05

141

4. Discussion

According to this longitudinal study, HRQoL is severely altered in newly diagnosed RRMS and PMS patients, and cognitive impairment contributes significantly but weakly to this HRQoL alteration. Early cognitive status, together with physical disability, contributes to vocational status seven years after the MS diagnosis.

This study shows that HRQoL is significantly decreased in this sample of MS patients compared with the general population. The worst SF-36 scores observed in MS patients were for role limitation due to physical problems, vitality and general health perception subscales. In the present study, we used the SEP-59 scale, which is the French version of MSQoL-54, a validated scale for French-speaking PwMS [1], and includes all the axes of the generic scale SF-36 as well as additional axes specifically designed for MS. Here, we used the SEP-59 scale both as a generic scale to compare the results of MS patients to the French general population data (which are available for the SF-36 scale) and as a disease-targeted HRQoL measure. The scores of the two composite scores, generated from the SF-36 survey questionnaire and all SF-36 scales, were significantly lower for MS patients than controls in the French population. This emphasizes that both the whole health perception (general health perceptions) is altered in newly diagnosed MS patients as well as specific aspects covered by the subscales (social functioning and mental health, pain, physical functioning, role physical, role emotional, and vitality). The composite scores were both predicted by the MADRS score at seven years in the MS patients. This is in accordance with the results of the study by Benedict et al. published in 2005 [3]. It is important to consider that the distress axis is part of these two composite scores. Moreover, our results confirm previous studies conducted in the Netherlands and in Germany comparing the scores of recently diagnosed RRMS patients who completed a HRQoL questionnaire with those of the general population [12, 14]. The Dutch study used the SF-36 survey and compared data obtained in newly diagnosed patients in several MS clinics with data from the general Dutch population [12]. They observed that 9% to 38.5% of the patients had “aberrant” scores compared to those of the general population according to the subscales. The most deteriorated scales were role limitation due to physical health problems (38.5%), physical functioning (26.3%) and general health perceptions (25%). The German study included a large sample of RRMS patients with a median disease duration of 13 months and used the EQ-5D questionnaire, which consists of two parts: a visual analogue scale (VAS) and a descriptive part addressing five fields rated with a 3-point graded scale (mobility, self-sufficiency, normal activity, pain/discomfort, and fear/depression) [14]. The VAS value of the RRMS patients was significantly inferior to the mean VAS of the normal German population. However, this study did not analyse specific subscales and did not include patients with PMS. Other studies used HRQoL scales in PwMS in early stages of the disease [12, 15], but did not compare the results to normative data. The analysis of the disease-targeted axes showed poor scores for several domains (sleep, social support, and sex) of the SEP-59 scale and for two SEP-59 composite scores. The cognitive function subscale scores cannot be compared with the general population data because this subscale is not included in the SF-36 survey. The score of this subscale is greater than 70%, which is quite high. However, we have previously demonstrated that this score is not correlated with neuropsychological testing; rather, it is correlated with the depressive score in the RRMS group at baseline [23], suggesting that it reflects psychological distress more than cognitive status.

142

To the best of our knowledge, no other study has compared HRQoL scores of newly diagnosed patients with PMS with normal controls. The scores from these patients, in spite of the very low sample size of this group, were lower than those of the general population for the SF-36 subscales, which showed the severity of this alteration. A comparison between newly diagnosed PMS patients and RRMS patients for these SF-36 scores and for some SEP-59 specific scores (social support, mental health, sexual function and satisfaction) suggest that the newly diagnosed PMS patients have a greater decrease of HRQoL than RRMS patients with a similar disease duration. Over seven years, the composite scores from the HRQoL scale did not change significantly in the MS group, although some subscales deteriorated (physical functioning, cognitive function, overall QoL, and sexual function). In the Dutch longitudinal study, the SF-36 physical function subscale deteriorated significantly over 3 years [12, 24].

Cognitive impairment is frequent in MS, even at an early stage. We previously showed in this sample that approximately half of the newly diagnosed RRMS patients had a cognitive impairment predominantly affecting IPS [16]. We found a correlation between cognitive scores and magnetic resonance imaging (MRI) parameters reflecting diffuse tissue injury [16], suggesting that these cognitive deficiencies could be due to disconnection caused by diffuse axonal pathology. This interpretation was strengthened by the presence of diffuse pathology in newly diagnosed RRMS patients, as evidenced by MRI, which is also predictive of cognitive deterioration during the following seven years [17]. We also observed that cognitive impairment at baseline was an independent predictor of EDSS deterioration over five to seven years [26]. Altogether, these results show that the cognitive deterioration detected in newly diagnosed RRMS patients is a good marker of diffuse brain pathology already present at this stage. However, the level of impairment for most neuropsychologist tests is, at worst, moderate. It is debatable whether this impairment is clinically relevant and whether it has consequences in the daily lives of patients. IPS and memory impairment contribute only slightly to HRQoL degradation in these patients. For the physical composite score of SEP-59 at seven years, the MADRS score assessing depressive symptoms and the memory Z-score explained 23% and 9.3% of the variance respectively. For the mental composite of SEP-59 at seven years, the MADRS score and the memory Z-score explained 24.5% and 12.4% of the variance respectively.

In our study, vocational status seven years after the MS diagnosis was predicted by cognitive deterioration during these years and by early IPS function evaluated 6 months after the MS diagnosis. In PwMS with various disease durations, several studies showed that cognitive impairment could lead to working limitations [5, 10-12]. In the present study, we showed the impact of cognitive deficiencies on vocational status at early stages of the disease. At baseline, a majority of the patients worked (81.5%), but after seven years, 27.1% of patients changed their vocational status. Among MS patients who did not work at y7, 72.7% of them were PCI at baseline, whereas 27.3% were PCU at baseline. It appears that the cognitive status at baseline plays a major role in vocational status after seven years. The vocational status at y7 was associated with the baseline IPS Z-score, age and baseline EDSS. However, cognitive deterioration over seven years significantly predicted vocational status change over seven years, independently of EDSS deterioration, after adjustment of vocational status at baseline.

Strober et al. showed that employment status was related to test scores that assess IPS (SDMT) and learning and memory (SRT) in a sample of 101 MS patients [26].

143

Employment status was also associated with disease duration, MS subtype, the level of neurological impairment and fatigue. In a longitudinal study focusing on vocational status in MS patients, it was reported that many employees with MS left their job mainly because of fatigue, mobility-related symptoms, arm and hand difficulties and cognitive deficits [27]. However, in this study, neuropsychological tests were not performed.

The main limitation of this study is the sample size. However, in spite of this limitation, the population basis of this sample strengthens these results. Every person diagnosed with MS by the neurologists from the network during the inclusion period was referred to be included in the study. One can assume that this community-based sample is less subject to selection bias than a sample recruited from a MS clinic population and more closely reflects the MS population at the time of diagnosis. IPS impairment, the most common cognitive alteration observed in newly diagnosed MS patients, was associated with the risk of quitting work or the reduction of duties for MS patients in the years following diagnosis. Another limitation of our study is that the cognitive test battery did not comprehensively evaluate executive functions, which was been associated with vocational status in prior studies [3]. However, two tests of executive functions were performed, including the WLG, a test of verbal semantic fluency, and the similarities subtest of the WAIS-R. Another limitation is that control subjects did not fill the HRQoL scale. The SEP-59 is a disease specific HRQoL scale and is not appropriate for healthy subjects. However, for the SF-36 part of the scale, we used the normative data published by Leplege et al. [21] obtained form a sample of healthy subjects representative of the French population.

Cognitive impairment in newly diagnosed MS patients increased the risk of quitting their employment or reducing their duties in the years following MS diagnosis.

FUNDING

The study was conducted as part of a study supported by the Association pour la Recherche contre la Sclérose en Plaques, (ARSEP), France, and Schering SA, France. The sponsors did not participate in any aspects of the design or conduct of the study, including data collection, management, analysis, and interpretation, as well as preparation, review, and approval of the manuscript. Dr. Ruet is a recipient of a fellowship-grant from Fondation pour la Recherche Médicale.

144

References

[1] Vernay D, Gerbaud L, Biolay S, et al. (2000) Quality of life and multiple sclerosis: validation of the French version of the self-questionnaire (SEP-59) Rev Neurol (Paris), 156, 247-263.

[2] Mitchell AJ, Benito-León J, Morales González JM, Rivera-Navarro J (2005) Quality of life and its assessment in multiple sclerosis: integrating physical and psychological components of wellbeing. Lancet Neurol, 4, 556-66.

[3] Benedict RH, Wahlig E, Bakshi R, Fishman I, Munschauer F, Zivadinov R et al. (2005) Predicting quality of life in multiple sclerosis: accounting for physical disability, fatigue, cognition, mood disorder, personality, and behavior change. J Neurol Sci; 231, 29-34.

[4] Brochet B (2011) Prevalence, profile and functional impact of cognitive impairment in multiple sclerosis. In Amato MP (Elsevier) Cognitive Impairment in Multiple Sclerosis, Milano, pp. 1-8.

[5] Benito-León J, Morales JM, Rivera-Navarro J (2002) Health-related quality of life and its relationship to cognitive and emotional functioning in multiple sclerosis patients. Eur J Neurol, 9, 497-502.

[6] Barker-Collo SL (2006) Quality of life in multiple sclerosis: does information-processing speed have an independent effect? Arch Clin Neuropsychol; 21, 167-7.

[7] Miller DM, Cohen JA, Kooijmans M, Tsao E, Cutter G, Baier M (2006) Change in clinician-assessed measures of multiple sclerosis and subject-reported quality of life: results from the IMPACT study. Mult Scler, 12, 180-6.

[8] Glanz BI, Healy BC, Rintell DJ, Jaffin SK, Bakshi R, Weiner HL (2010) The association between cognitive impairment and quality of life in patients with early multiple sclerosis. J Neurol Sci, 290, 75-9.

[9] Rao SM, Leo GJ, Ellington L, Nauertz T, Bernardin L, Unverzagt F (1991) Cognitive dysfunction in multiple sclerosis. II. Impact on employment and social functioning. Neurology, 41, 692-6.

[10] Amato MP, Ponziani G, Siracusa G, Sorbi S (2001) Cognitive dysfunction in early-onset multiple sclerosis: a reappraisal after 10 years. Arch Neurol; 58, 1602-6.

[11] Morrow SA, Drake A, Zivadinov R, Munschauer F, Weinstock-Guttman B, Benedict RH. (2010) Predicting loss of employment over three years in multiple sclerosis: clinically meaningful cognitive decline. Clin Neuropsychol, 24, 1131-45.

[12] de Groot V, Beckerman H, Lankhorst GJ, Polman CH, Bouter LM (2005) The initial course of daily functioning in multiple sclerosis: a three-year follow-up study. Mult Scler,11, 713-8.

[13] Kern S, Schrempf W, Schneider H, Schultheiss T, Reichmann H, Ziemssen T (2009) Neurological disability, psychological distress, and health-related quality of life in MS patients within the first three years after diagnosis. Mult Scler,15, 752-8.

145

[14] Putzki N, Fischer J, Gottwald K, Reifschneider G, Ries S, Siever A, et al. (2009) Quality of life in 1000 patients with early relapsing-remitting multiple sclerosis. Eur J Neurol, 16, 713-20.

[15] Patti F, Pozzilli C, Montanari E, Pappalardo A, Piazza L, Levi A, et al. (2007) Effects of education level and employment status on HRQoL in early relapsing-remitting multiple sclerosis. Mult Scler, 13, 783-91.

[16] Deloire MS, Salort E, Bonnet M, Arimone Y, Boudineau M, Amieva H, et al. (2005) Cognitive impairment as marker of diffuse brain abnormalities in early relapsing remitting multiple sclerosis. JNNP, 76, 519-26.

[17] Deloire MS, Ruet A, Hamel D, Bonnet M, Dousset V, Brochet B (2011) MRI predictors of cognitive outcome in early multiple sclerosis. Neurology, 76, 1161-7.

[18] Lublin FD, Reingold SC (1996) Defining the clinical course of multiple sclerosis: results of an international survey. National Multiple Sclerosis Society (USA) Advisory Committee on Clinical Trials of New Agents in Multiple Sclerosis. Neurology, 46, 907-11.

[19] Gayou A, Brochet B, Dousset V (1997) Transitional progressive multiple sclerosis: a clinical and imaging study. JNNP, 63, 396-8.

[20] Keller SD, Ware JE Jr, Bentler PM, Aaronson NK, Alonso J, Apolone G, et al. (1998) Use of structural equation modeling to test the construct validity of the SF-36 Health Survey in ten countries: results from the IQOLA Project. International Quality of Life Assessment. J Clin Epidemiol; 51, 1179-88.

[21] Leplège A, Ecosse E Coste J, Pouchot J, Perneger T (2001) Le questionnaire MOS SF-36: Manuel de l'utilisateur et guide d'interprétation des scores (Editions Estem), Paris.

[22] Vickrey BG. (1995) Multiple Sclerosis Quality of life (MSQOL)-54 Instrument. University of California, Los Angeles.

[23] Deloire MS, Bonnet MC, Salort E, Arimone Y, Boudineau M, Petry KG, et al. (2006) How to detect cognitive dysfunction at early stages of multiple sclerosis? Mult Scler, 12, 445-52.

[24] de Groot V, Beckerman H, Uitdehaag BM, de Vet HC, Lankhorst GJ, Polman CH et al. (2006) The usefulness of evaluative outcome measures in patients with multiple sclerosis. Brain, 129, 2648-59.

[25] Deloire M, Ruet A, Hamel D, Bonnet M, Brochet B (2010) Early cognitive impairment in multiple sclerosis predicts disability outcome several years later. Mult Scler,16, 581-7.

[26] Strober LB, Christodoulou C, Benedict RH, Westervelt HJ, Melville P, Scherl WF, Weinstock-Guttman B, Rizvi S, Goodman AD, Krupp LB (2011). Unemployment in multiple sclerosis: the contribution of personality and disease. Mult Scler, 19.

[27] Simmons RD, Tribe KL, McDonald EA (2010). Living with multiple sclerosis: longitudinal changes in employment and the importance of symptom management. J Neurol, 257 (6): 926-36.

146

Discussion de cet article

1. Résumé des résultats de cet article

Dans une cohorte de patients ayant un diagnostic récent de SEP, la santée liée à la Qdv était altérée par rapport à des sujets sains contrôles appariés en fonction des caractéristiques démographiques. Les performances aux tâches de mémoire et les symptômes dépressifs étaient associés aux scores composites physiques et mentaux de Qdv sur une période de suivi de 7 ans après le diagnostic de la SEP en prenant en compte comme variables l'âge initial, les scores EDSS, MADRS, et ceux estimant la fatigue, les z scores pour les tests de VTI et de la mémoire.

Alors que les résultats aux tests de VTI, l'âge et le score EDSS prédisaient indépendamment le statut professionnel 7 ans après le diagnostic de la SEP, seule la performance initiale aux tests de VTI pouvait prédire le changement du statut professionnel sur 7 ans. De plus, la détérioration cognitive survenant dans les 7 ans après le diagnostic de SEP était associée à la fois au statut professionnel à 7 ans et à son évolution sur 7 ans.

2. Intégration des résultats aux données actuelles de la littérature

Nous confirmons les résultats d'études antérieures montrant qu’il existe une altération de la santé liée à la Qdv chez des patients ayant une SEP (Mitchell et al., 2005) et un impact négatif de l’atteinte cognitive. L’évaluation de la santé liée à la Qdv requiert le remplissage de questionnaires bien validés. Vickrey et ses collègues ont développé un instrument de mesure nommé «MSQoL-54» (Vickrey et al., 1995) pour évaluer la Qdv à la fois de façon générale et spécifique aux patients ayant une SEP (Vickrey et al., 1997). Il a été reporté une excellente fiabilité et consistance interne de cet auto-questionnaire, confortant la validité de son contenu et de sa forme (Vickrey et al., 1995). Peu de questionnaires spécifiques sont disponibles en français pour les patients atteints de SEP. L’auto-questionnaire MSQoL-54 a été traduit, adapté et validé dans la population française sous l’appellation SEP-59 et a été utilisé dans cette étude (Vernay et al., 2000). Le questionnaire «Multiple Sclerosis International Quality of Life» (MusiQoL) (Simeoni et al., 2008) a également été validé dans la population française et il s’agit d’un auto-questionnaire multidimensionnel (Baumstarck-Barrau et al., 2011a).

La relation entre les troubles cognitifs et la Qdv est contradictoire dans la SEP. Le dysfonctionnement cognitif influait sur la Qdv des sujets dans une étude incluant 100 patients atteints de SEP (Rao et al., 1991b). Ces résultats ont été corroborés par des études récentes qui indiquent l'impact des troubles cognitifs sur le bien-être et l'état d'invalidité des patients SEP (Benito-Leon et al., 2002; Chiaravalloti et al., 2008; Clavelou et al, 2009;. Fernandez et al., 2011). Dans notre échantillon de patients ayant une SEP, les performances initiales aux tâches de mémoire étaient indépendamment associées avec les scores composites de Qdv évalués quelques années plus tard. En outre, une faible association a été retrouvée entre la Qdv évaluée à l’aide des questionnaires SF-36 et MusiQoL et l’atteinte

147

cognitive dans une étude transversale de 124 patients ayant une SEP (Baumstarck-Barrau et al., 2011b). En revanche, dans une grande étude nommée COGIMUS incluant 331 patients atteints de SEP-RR suivis pendant 3 ans, aucune différence significative n'a été observée concernant les scores composites de Qdv (MSQoL-54) selon l'état cognitif initial des patients (Patti et al., 2011).

Les symptômes dépressifs et ceux liés à l'anxiété ont été corrélés avec la Qdv subjective (Wang et al., 2000; Clavelou et al., 2009; Chahraoui et al., 2010). La santé physique liée à la Qdv a été associée non seulement au handicap physique, mais aussi à la fatigue et à la dépression, tandis que la santé mentale liée à la Qdv a été associée à la dépression et à la fatigue chez les patients ayant une SEP (Amato et al, 2001b;.Benedict et al., 2005). Dans une étude transversale ayant inclus 103 patients atteints de SEP, il a été retrouvé une corrélation négative d’intensité modérée entre le niveau d'incapacité et le score composite physique du MSQoL-54, et une forte corrélation négative entre la dépression, la fatigue et le score composite mental du MSQoL-54 (Amato et al., 2001b). Dans notre étude longitudinale, nous avons rapporté que le score EDSS initial ne prédisait pas le score composite physique (PCS/SEP-59) à 7 ans, tandis que les symptômes dépressifs initiaux ont été associés de façon significative avec les 2 scores composites (PCS/SEP-59 et MCS/SEP-59) à 7 ans. Il a été rapporté que la dépression était le principal facteur influençant la fatigue chez les patients atteints de SEP, et que la fatigue était corrélée avec des actions de contrôle dans un échantillon de 41 patients atteints de SEP (Penner et al., 2007).

Le statut cognitif initial associé au handicap physique a contribué à la situation professionnelle 7 ans après le diagnostic de la SEP dans notre échantillon de patients. Ce résultat est cohérent avec l'étude précédente de Rao et collaborateurs montrant que les patients atteints de troubles cognitifs étaient plus souvent sans emploi et pratiquaient moins d’activités à la fois sociales et professionnelles (Rao et al., 1991b). La fréquence des patients ayant une SEP sans emploi n'est pas bien définie et peut varier de 24 à 80% selon les études (LaRocca et al., 1985; Kornblith et al., 1986; Grønning et al., 1990;. Julian et al., 2008). Alors que plus de 90% des patients travaillaient avant la survenue des premiers symptômes liés à la SEP (LaRocca et al., 1982; Pompéi et al., 2005), près de 70-80% se retrouvent sans emploi dans les 5 ans suivant le diagnostic de SEP (Kornblith et al., 1986.). Les facteurs de risque de la perte d’emploi pour ces patients ont été identifiés comme étant: le sexe féminin (LaRocca et al., 1985.), le faible niveau d'éducation (LaRocca et al., 1985.), le niveau élevé de handicap physique (LaRocca et al., 1985;. Benedict et al., 2005; Strober et al., 2012), la durée de la maladie (Strober et al., 2012), et une forme progressive de la maladie (Kornblith et al., 1986; Strober et al., 2012). Il est important de considérer que le handicap physique et les caractéristiques démographiques expliquent seulement 14% de la variance dans une étude incluant un large échantillon de 312 patients atteints de SEP (LaRocca et al., 1985). Certains symptômes spécifiques liés à la SEP contribuent également au fait de ne pas travailler, comme les troubles de la marche (O'Connor et al., 2002.), les troubles vésico-sphinctériens, la sensibilité à la chaleur (Simmons et al., 2010), la fatigue (Julien et al., 2008; Smith et al., 2005; Strober et al., 2012), et le type de personnalité (ex : la persistance) (Strober et al., 2012). Récemment, il existe un gain d'intérêt pour les troubles cognitifs dans la SEP, et l’atteinte de la VTI et de la mémoire ont été significativement associées au statut professionnel des patients ayant une SEP (Benedict et al., 2005; Morrow et al., 2010, Strober et al., 2012). Nos résultats confirment le rôle des troubles cognitifs aux

148

stades précoces de la SEP dans la prédiction du statut professionnel dans les années suivant le diagnostic de SEP. La détérioration cognitive était le seul facteur prédictif du statut professionnel à 7 ans et de son évolution au cours de cette période indépendamment de la progression de l'invalidité physique dans notre échantillon de patients ayant une SEP.

3. Défis et difficultés

La validité des auto-questionnaires de Qdv chez les patients souffrant de troubles cognitifs a été débattue, et il a été suggéré que les déficits cognitifs pourraient altérer la perception de la Qdv des patients ayant une SEP. Dans une étude de 117 patients ayant un diagnostic de SEP, l’atteinte cognitive a été associée à une surestimation de l'auto-perception de la Qdv (Gerbaud et al., 2006). Nos données obtenues aux stades précoces de la SEP n’ont pas confirmé ces résultats. Nous n'avons pas observé de différences pour la plupart des axes de Qdv en fonction du statut cognitif des patients ayant une SEP, à l'exception des axes de détresse et de douleur, mais qui étaient associés à des scores plus faibles chez les patients atteints de troubles cognitifs par rapport aux patients cognitivement intacts. En outre, la Qdv a été étudiée en utilisant 2 auto-questionnaires (MusiQoL et SF36) dans une étude transversale de 124 patients atteints de SEP analysés en fonction de leur performance au Stroop qui est un test NP évaluant les fonctions exécutives (Baumstarck et al., 2012). Les scores de Qdv dans cet échantillon de patients ne différaient pas en fonction de leur performance à ce test. Ces résultats suggèrent que le dysfonctionnement exécutif ne semble pas compromettre la fiabilité et la validité des auto-questionnaires de Qdv. D'autres études doivent confirmer si les patients souffrant de troubles cognitifs peuvent apprécier de façon adéquate leur Qdv. Enfin, il semble opportun de prendre en considération les résultats de l'évaluation cognitive et psychologique des patients ayant une SEP dans le cadre de l’interprétation des auto-questionnaires de Qdv, parce que le type et la sévérité de l’atteinte cognitive pourraient influencer ces résultats.

Même si l'évaluation du bien-être des patients, mesuré à l'aide des auto-questionnaires de Qdv, fournit des informations complémentaires intéressantes pour les médecins et les chercheurs, il manque de consensus pour interpréter ces résultats et leur pertinence clinique ainsi que la valeur du changement minimal cliniquement significatif (Brozek et al., 2006).

4. Conclusion et perspectives

Nous avons confirmé la valeur pronostique des troubles cognitifs aux stades précoces de SEP en illustrant les conséquences négatives sur la Qdv et le statut professionnel de ces patients. Ces résultats encouragent le dépistage précoce des déficits cognitifs chez les patients ayant un diagnostic de SEP afin de pouvoir adapter les stratégies thérapeutiques dès les premiers stades de la maladie. Récemment, il a été étudié les relations entre les stratégies d'adaptation, les troubles cognitifs et psychologiques et la Qdv dans la SEP (Goretti et al., 2010). Actuellement, il manque de données en faveur de l'efficacité des traitements symptomatiques et des traitements de fond sur les déficits cognitifs dans la SEP, mais certains essais cliniques sont en cours. Les essais cliniques dans la SEP devraient avoir comme critère de jugement principal à la fois les déficits

149

physiques et cognitifs. Récemment, une étude observationnelle en ouvert a rapporté l'impact bénéfique d'un anticorps monoclonal sur les performances cognitives de patients atteints de SEP-RR (Iaffaldano et al., 2012). De plus, il manque d'études bien conçues pour explorer l'efficacité de la réadaptation cognitive dans la SEP (O'Brien et al., 2008; Rosti-Otajärvi et Hämäläinen, 2011). Comme le ralentissement de la VTI peut prédire le statut professionnel des patients ayant une SEP, sa détection précoce et sa prise en charge semblent essentielles dans la gestion de ces patients. Notamment, l'apparition de la SEP se produit le plus souvent chez des jeunes adultes, qui sont des individus issus de la population active. Par conséquent, il serait cliniquement pertinent d’étudier l'impact de la réadaptation et de la remédiation cognitive sur le statut professionnel et son évolution dans le temps pour les patients ayant une SEP.

150

ARTICLE 7:

VALIDITY AND PREDICTIVE VALUE OF A NEW COMPUTERISED COGNITIVE TEST FOR THE DETECTION OF INFORMATION PROCESSING SPEED IMPAIRMENT IN MULTIPLE SCLEROSIS

Article submitted, in revision

VALIDITY AND PREDICTIVE VALUE OF A NEW COMPUTERISED COGNITIVE TEST FOR THE DETECTION OF INFORMATION PROCESSING SPEED IMPAIRMENT IN MULTIPLE SCLEROSIS

Aurélie Ruet 1,2 MD, Mathilde S.A. Deloire2 PhD, Julie Charré-Morin2 Msc, Delphine Hamel1 Msc, and Bruno Brochet1,2 MD.

1 Université de Bordeaux, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, F-33076 Bordeaux, France

2 CHU de Bordeaux, INSERM-CHU CIC-P 0005, & Service de Neurologie, F-33076 Bordeaux, France

Corresponding author:

Bruno Brochet, Université de Bordeaux, INSERM U.1049 Neuroinflammation, Imagerie et Thérapie de la Sclérose en plaques, case 78, 146 rue Léo Saignat, F-33076 Bordeaux cedex, France. bruno.brochet@chu-bordeaux.fr

Keywords: Cognition, Multiple Sclerosis, Information Processing Speed, Reaction Times, Computerised Neuropsychological Test.

151

ABSTRACT

Background: Cognitive impairment in multiple sclerosis (MS) primarily applies to information processing speed (IPS).

Objective: To validate a new digit/symbol substitution test in healthy subjects and persons with MS (PwMS), and assess its performance to detect IPS impairment.

Methods: A sample of 101 PwMS (60 with relapsing-remitting MS (RRMS) and 41 with primary progressive MS (PPMS)) and 415 healthy controls (HCs) underwent an IPS battery, including reaction times of subtests of the Test of Attentional Performance battery and a newly developed in-house digit/symbol substitution task, the Computerised Speed Cognitive Test (CSCT). Healthy controls were divided into 20 groups according to age, sex, and education level. Additionally, the CSCT was evaluated in a second cohort of 43 PwMS (31 RRMS and 12 progressive MS) for comparison with the Symbol Digit Modalities Test (SDMT).

Results: There was an association between the CSCT scores with the age and with the education level, but not with the gender of the subjects. The CSCT had excellent reliability in both HCs and PwMS, and it showed a weak practice effect at 6-month. This test was clinically relevant and had good ecological validity in PwMS. There was a strong correlation between the CSCT with the SDMT and with other IPS tests in PwMS. The CSCT had the best sensitivity for the prediction of IPS impairment in PwMS, and was one of the best accurate tests among IPS battery.

Conclusion: As the CSCT presented many relevant characteristics, this test could be a good candidate for detecting IPS impairment in PwMS. Further studies should determine the minimal clinically meaningful change for monitoring IPS impairment in MS over time.

152

INTRODUCTION

There is increasing evidence that cognitive impairment is a prominent feature of multiple sclerosis (MS), occurs very early in the disease, and negatively impacts working abilities independently of any other potential confounding factors, such as fatigue, depression, and physical disability (review by Brochet, in Amato, 2011). Information processing speed (IPS) impairment is the most common cognitive alteration observed in persons with MS (PwMS) (Zakzanis et al., 2000). It has been shown to predict subsequent disability in early RRMS patients (Deloire et al., 2010) and their vocational status (Morrow et al., 2010a). Its detection has been recognised as the main objective in routine neuropsychological (NP) evaluation in PwMS (Langdon et al., 2012). Several NP tests are frequently used in clinical practice to detect IPS impairment (review by Benedict, in Amato, 2011). The Paced Auditory Serial Addition test (PASAT) (Rao, 1990) which is part of the Multiple Functional Composite Score (MFSC) (Fisher et al., 1999), is widely used, but it could be considered as more of a measure of working memory than IPS. Several studies have used reaction times (RTs) after visual or auditory computerised stimuli (Jennekens-Schinkel et al., 1988; Schulz et al., 2006; Bergendal et al., 2007). In earlier studies, RTs to the Sternberg Memory Scanning task have been used (Rao et al., 1991). The computerised battery developed by Zimmerman (Zimmermann and Fimm, 1994) to assess attention allowed for calculating RTs for several cognitive tasks, and it has been used to evaluate IPS in PwMS (Tinnefeld et al., 2005). Additionally, IPS has been tested in PwMS using non-verbal substitution tasks, such as the WAIS-R Digit Symbol (Weschler, 1997; Bensa et al., 2006; Dujardin et al., 2004) or the Symbol Digit Modalities test (SDMT) (Smith, 1982). The SDMT was demonstrated to be a better screening test than other cognitive tests for the detection of cognitive impairment in newly diagnosed RRMS patients (Deloire et al., 2006) and in a mixed sample of RRMS and secondary progressive multiple sclerosis (SPMS) patients (Parmenter et al., 2007). Recently, it has been suggested as the test for IPS evaluation in the brief international cognitive assessment for MS (BICAMS) battery, which is proposed as a minimal assessment tool in MS (Langdon et al., 2012).

In different digit/symbol substitution tasks, a unique key showing the association of symbols with digits is provided, and it is similar for each test session. Therefore, the key could be memorised when patients frequently perform the test, and this could affect the performances of the patient, increasing the practice effect. In the same way, the list of symbols during the test is always presented in the same order. The practice effect has been investigated (Benedict et al., 2008; Morrow et al., 2010b), and alternate forms have been proposed, but their equivalence was not demonstrated (Boringa et al., 2001; Benedict et al., 2012a). Therefore, we developed a new NP test called the Computerised Speed Cognitive test (CSCT) (Ruet et al., 2011). This test is based on the same principle of digit/symbol substitution as the previously described tests, but in this test, the key is automatically generated by software for each session. Thus, the key is different for each presentation. Additionally a new list of symbols is displayed for each test session. The main aim of this study is to validate the CSCT in a population of healthy subjects and several groups of PwMS. The secondary objective is to assess its performance to predict IPS impairment in PwMS.

153

MATERIALS AND METHODS

SUBJECTS

- Healthy controls

Twenty groups of healthy controls (HCs) were established according to five age categories (18-34, 34-45, 45-54, 55-64, and ≥65), gender, and levels of education (low education level (LEL) was below secondary education (usually 12 years of schooling), and high education level (HEL) was above secondary education and graduated at least “baccalaureate”). Individuals who were at least 18 years old and French speakers were eligible to participate in this study. Non-inclusion criteria consisted of a history of neurological disease; psychiatric illness, with the exception of stable depressive symptoms; starting or stopping antidepressants in the previous two months; other significant chronic disease; abuse of alcohol or drugs in the current or previous two years; a family history of MS; a cognitive complaint; or having previously participated in a cognitive study in the last year. HCs were paid for participating in the study.

- Multiple sclerosis patients

Two samples of MS patients were used.

One group (MS1) included 60 RRMS patients and 41 primary progressive MS (PPMS) patients. All patients aged at least 18 years old had to fulfil revised McDonald 2005 criteria (Polman et al., 2005) and be French speakers. Exclusion criteria were as follows: patients with SPMS; presence of any disease other than MS that could contribute to neurological symptoms and signs; psychiatric illness, with the exception of stable depressive symptoms; starting or stopping antidepressants in the previous two months; abuse of alcohol or drugs in the current or previous two years; and steroid treatment within the last 30 days.

Another group of 43 MS patients (MS2) consisted of RRMS and progressive MS (SPMS and PPMS) patients followed since 2000 as part of a cohort study in our department. This MS2 group of patients was analysed for comparison between the SDMT and the CSCT, and these patients had not had cognitive evaluation in the previous three years before entering the present study.

Patients could be on disease-modifying therapy but with a steady dosage for at least three months.

This study was approved by the ethics committee, including the institutional review board for human subject research of Bordeaux. All subjects gave written informed consent to participate in the study before their inclusion.

154

METHODS

Neuropsychological assessment

First, patients in the MS1 group and HCs were evaluated by a qualified senior neuropsychologist using a large NP battery, including the following IPS tests:

- The CSCT, which is a new symbol/digit substitution task made up of a key displayed on the upper part of the screen with a list of nine symbols. Under this row, a list of nine digits is displayed. The sequences of symbols and digits of the key are automatically generated for each session of training and testing and are different from one session to another. In the lower part of the screen, symbols appear on consecutive lines, and a blinking vertical line indicates which symbol has to be taken into consideration. The session always begins with a training trial with 15 symbols. When the subject orally provides the answer, the examiner enters the digit on the keypad of the computer, and the blinking signal progress to the next symbol. The visual scanning is limited to the route between the symbol, the key, and the symbol. There is no need for the subject to scan the entire row to find the new symbol or memorise the location. The subject has to give as many possible correct answers as they can during 90 seconds. The test stimuli and instructions that were displayed on the computer screen before training and read by the neuropsychologist were standardised. Consistent stimulus presentation was proposed for each subject. The CSCT was performed at the beginning of the NP evaluation (CSCT-1) and at the end (CSCT-2 after approximately two hours of NP assessment). After six months, MS1 patients and half of the HCs underwent the CSCT again (CSCT-3).

- Four subtests of the TAP (Zimmermann and Fimm, 2009) were also used to assess IPS. The alertness subtest is designed to assess phasic alertness. RT measurements were taken into account without or with a warning. The visual scanning has two subtests, without a target or with a target, that enable testing the ability to shift visual attentional focus. The flexibility subtest contains a verbal condition with a letter or a number and a non-verbal condition with angular and round figures.

Second, patients in the MS2 group underwent the CSCT and the SDMT (Smith, 1982) administered by a senior neuropsychologist. The SDMT consists of single digits paired with abstract symbols. Rows of the nine symbols are arranged pseudo-randomly. The patient must say the number that corresponds to each symbol.

Neurological evaluation

Each MS patient underwent a full neurological examination. Disability was measured using a French-adapted version (Brochet, 2009) of the Expanded Disability Status Scale (EDSS) (Kurtzke, 1983).

155

Other evaluations

Each patient and HC completed questionnaires concerning depressive symptoms (Beck Depression Inventory II, BDI II (Beck et al., 1996)), anxiety (State-State Trait Anxiety Inventory, S-STAI (Spielberg, 1983)), and subjective fatigue (UK Neurological Disability Scale fatigue score, UKNDS (Sharrack and Hughes, 1999)).

Statistical analyses:

Statistical analyses were performed with Statview version 5.0 software for Windows.

For age, disease duration, CSCT and SDMT scores, the results are shown as the means ± the standard deviations (SDs). For the EDSS, BDI II, anxiety, and fatigue scores, the results are shown as the medians (ranges).

Analysis of the MS1 patient group and HCs

Pearson’s correlations were performed between the CSCT scores and age and between different scores of the CSCT over time in HCs and PwMS. Pearson’s correlations were also established between CSCT scores and the scores of other IPS TAP tests in both groups. Differences were investigated according to sex, age, and education level by analysis of variance in each group of subjects. Cohen’s d was calculated to estimate the test-retest effect between the CSCT-2 and CSCT-3 scores in HCs and PwMS.

Each patient score was compared with the mean value of that patient’s group of HCs matched for age, sex, and education level. Additionally, z scores were calculated for each NP score with the following formula: (MS1 patient’s score - mean value of their own HC group matched for age, sex, and education level)/standard deviation of the matched HC. Patients were classified impaired for a given NP test if their z scores were below the fifth percentile compared with the mean value of their own HC group adequately matched by age, sex, and education level. HCs were also classified impaired for a given NP test if their z scores were below the fifth percentile compared with the mean value of their own group. Moreover, patients were classified impaired for the IPS domain if they were impaired for at least two scores, excluding the test studied as a potential predictive test. Regression logistic analysis was performed to predict vocational status according to CSCT impairment. Additionally, bayesian statistics were used to predict IPS impairment using the following equations: sensitivity=/number of impaired patients for the studied test/number of IPS impaired patients; specificity=/number of unimpaired patients for the studied test/number of IPS unimpaired patients; positive predictive value=/number of impaired patients for the studied test/number of positive tests; negative predictive value=/number of unimpaired patients for the studied test/number of negative tests; and accuracy= number of impaired and unimpaired patients for the studied test/all patients studied. All of these measurements were presented with exact binomial 95% confidence intervals.

156

Analysis of the MS2 patient group

Patients were classified impaired for the SDMT based on the norms published by Smith (Smith, 1982). From available normative data in the Smith manual, the cut-off of 1.5 standard deviations (SDs) was chosen to classify MS patients. MS2 patients were considered impaired for the SDMT if their score was below 1.5 SDs of HCs matched by age and education level. In addition, they were considered impaired for the CSCT if their score was below 1.5 SDs compared with the mean value of the CSCT of the HC group matched by age, sex, and education level. The chi-square test was performed between impaired patients on CSCT and impaired patients on SDMT. Pearson’s correlation analysis was performed between CSCT and SDMT scores.

For all analyses, differences were considered significant when the p values were less than 5%.

RESULTS

Performance of the CSCT in HCs

The mean CSCT score was 47.6±9.9 in HCs. Twenty-four out of 415 HCs (5.8%) were classified impaired for the CSCT.

- Reliability

The Pearson’s correlation coefficient between CSCT-1 and CSCT-2 was 0.83 (p<0.001), which supported a good reliability of the CSCT in 415 HCs.

- Effect of age, sex, and education level

The Pearson’s correlation coefficient between CSCT values and age was -0.55 (p<0.001) in HCs, showing a strong age effect. The mean CSCT values did not differ significantly between men and women in this sample of HCs. The CSCT scores were lower in HCs with LEL compared with HCs with HEL (p<0.001). The mean CSCT scores are presented according to age category and education level in HCs (Figure 1).

- Test-retest

The effect size between CSCT-2 and CSCT-3 was small, as Cohen’s d value was approximately 0.1 in 220 HCs evaluated over time.

157

Figure 1. Performances of 101 MS1 patients and 415 HCs on the CSCT-1 according to age and education level.

LEL: low education level; HEL: high education level; HCs: healthy controls; MS: multiple sclerosis patients from the first group (MS1).

158

Performance of the CSCT in MS patients

The clinical characteristics and some NP features of the 101 MS1 and 43 MS2 patients are presented in Table 1 and 2, respectively. Three PPMS patients and one RRMS patient from the MS1 group did not perform the visual scanning and flexibility TAP tests due to hand disabilities.

- Reliability

The reliability of the CSCT in 101 MS1 patients was high, as the correlation coefficient between the CSCT-1 and the CSCT-2 was 0.90 (p<0.001).

- Effect of age, sex, and education level

CSCT scores were correlated with age in the MS1 patient group (r=-0.55; p<0.001). They were not different according to sex (p=0.86), but they differed according to education level (p<0.05). Figure 1 illustrates the scores of the CSCT according to age categories and education level.

- Effect of phenotypes of MS

CSCT scores differed according to the phenotype of MS (p<0.001). There was a significant interaction between age and the type of MS.

- Test-retest

The effect size was small, as the Cohen d value was approximately 0.1 between the CSCT-2 and CSCT-3 scores in 88 MS1 patients evaluated over time.

- Effect of depressive symptoms, anxiety, and subject ive fatigue

There were no significant correlations between the CSCT scores and the BDI-II, STAI-S, or subjective fatigues scores in the MS1 patient group.

159

Table 1. Clinical characteristics of PPMS and RRMS patients (MS1 group) and their performance on the CSCT-1.

MS1 patients

n=101

PPMS patients

n=41

RRMS patients

n=60

Sex F/M 73/28 24/17 49/11

Age mean±SD (years) 43.3±11.9 52.1±8.7 37.3±9.9

Disease duration mean±SD 4.2±3.9 4.8±3.9 4.1±3.0

High education level, n (%)° 56 (55.4) 19 (46.3) 37 (61.7)

CSCT mean score±SD 43.9±11.4 37.4±10.7 48.3±9.8

Impaired CSCT, n (%) 29 (28.7) 17 (41.5) 12 (20.0)

Median EDSS score [range] 2.5 [0-7.0] 3.5 [1.5-7.0] 1.5 [0-4.5]

Median BDI II score [range] 10 [0-34] 12 [5-27] 9 [0-34]

Median STAI-S score [range] 33 [20-61] 32 [20-61] 34 [20-60]

Median fatigue score [range] 1 [0-5] 1 [0-3] 1 [0-5]

MS1: multiple sclerosis patients from the first group; PP: primary progressive; RR: relapsing-remitting; F: female; M: male; SD: standard deviation; CSCT: Computerised Speed Cognitive Test; EDSS: Expanded Disability Status Scale

160

Table 2. Clinical and neuropsychological findings i n PMS and RRMS patients (MS2 group).

MS2 patients

n=43

PMS patients

n=12

RRMS patients

n=31

Sex F/M 30/13 8/4 22/9

Age mean±SD (years) 48.4±9.6 56.0±8.8 45.4±8.3

Disease duration mean±SD 12.8±3.1 14.0±4.5 12.4±2.2

High education level, n (%) 19 (44.2) 4 (33.3) 15 (48.4)

CSCT mean score±SD 38.8±12.4 31.1±11.6 41.8±11.5

Impaired CSCT, n (%) 19 (44.2) 6 (50.0) 13 (41.9)

SDMT mean score±SD 46.7±14.2 36.8±13.1 50.6±12.8

Impaired SDMT, n (%) 14 (32.6) 6 (50.0) 8 (25.8)

Median EDSS score [range] 3.5 [0-8.5] 5.5 [3.5-8.5] 2.5 [0-6.0]

Median BDI II score [range] 12 [0-42] 16 [1-42] 11 [0-33]

Median STAI-S score [range] 40.5 [20-74] 42 [22-69] 40 [20-74]

Median fatigue score [range] 1 [0-5] 2 [0-4] 0 [0-5]

MS2: multiple sclerosis patients from the second group; PMS: progressive multiple sclerosis, including primary progressive and secondary progressive MS; RR: relapsing-remitting; F: female; M: male; SD: standard deviation; CSCT: Computerised Speed Cognitive Test; EDSS: Expanded Disability Status Scale.

161

- Validity of the CSCT

MS1 patients were more impaired on the CSCT than HCs (28.7% versus 5.8%). A difference in performance on the CSCT between MS1 patients and HCs was found after accounting for age and education level (Figure 1).

More unimpaired MS1 patients on the CSCT were employed than impaired MS1 patients (79.2% versus 20.8%, p<0.01). Impairment on the CSCT was associated with an increased risk of unemployment (OR=3.5 [1.4-8.9]; p<0.01).

- Comparison between SDMT and CSCT scores

The Pearson’s correlation coefficient between CSCT and SDMT scores was high in the 43 MS2 patients (r=0.94; p<0.001). Among this group, 44.2% of patients were impaired on the CSCT, and 32.6% were impaired on the SDMT (p<0.001) (Table 2).

- Detection of information processing slowness

Significant correlations were observed between the CSCT scores and the RTs for the flexibility test, visual scanning with a target, and alertness in 97 MS1 patients (Table 3). The CSCT had the best sensitivity (60.0%) for predicting IPS impairment and was one of the most accurate tests (accuracy of 77.3%) among the IPS tests in the MS1 patient group (Table 4).

Table 3. Correlations between the CSCT and other IP S tests in 97 MS1 patients.

Pearson’s coefficient r

Healthy controls MS1 patients

RT flexibility -0.60‡ -0.65‡

Alertness -0.38‡ -0.41‡

RT visual scanning with a target -0.41‡ -0.52‡

RT visual scanning without a target -0.31‡ -0.38†

MS: multiple sclerosis; RTs: Reaction Times.

†: p<0.01; ‡: p<0.001

162

Table 4. Predictive value of the CSCT for IPS impai rment in 97 MS1 patients.

Sensitivity

[95% CI]

Specificity

[95% CI]

Accuracy

[95% CI]

PPV

[95% CI]

NPV

[95% CI]

CSCT 60.0

[50.3-69.7]

85.1

[78.0-92.2]

77.3

[69.0-85.6]

64.3

[54.8-73.8]

82.6

[75.1-90.1]

Alertness 53.3

[43.4-63.2]

74.6

[65.9-83.3]

68.0

[58.7-77.3]

48.5

[38.6-58.4]

78.1

[69.9-86.3]

RT flexibility 43.3

[33.4-53.2]

91.0

[85.3-96.7]

76.3

[67.8-84.8]

68.4

[59.1-77.7]

78.2

[70.0-86.4]

RT VS with a target 52.0

[42.1-61.9]

86.1

[79.2-93.0]

77.3

[69.0-85.6]

56.5

[46.6-66.4]

83.8

[76.5-91.1]

RT VS without a target 29.6

[20.5-38.7]

88.6

[82.3-94.9]

72.2

[63.3-81.1]

50.0

[40.1-59.9]

76.5

[68.1-84.9]

CI: Confidence Interval; PPV: Positive Predictive Value; NPV: Negative Predictive Value; CSCT: Computerised Speed Cognitive Test; RTs: Reaction Times; VS: visual scanning.

163

DISCUSSION

IPS is the central cognitive domain impaired in MS. Its detection has been recognised as the main objective in routine NP evaluation in PwMS (Langdon et al., 2012). We proposed a new IPS test called the CSCT, which could easily be administered in daily clinical practice for testing IPS impairment. This study describes interesting characteristics of the CSCT concerning its standardisation and evaluates its validity in a large sample of HCs and different phenotypes of PwMS. The CSCT is a brief test that requires only a few minutes to administer and score. It should be easily administered and applicable in a wide variety of PwMS, and it has the potential to be used in cross-cultural international studies.

The validation process encompassed the several steps required for establishing psychometric standards that have recently been summarised by Benedict et al (2012b).

The first step is the standardisation of the test to ensure that test stimuli and administration procedures have face validity and consistent stimulus presentation. The CSCT is a digit symbol/substitution task presented on a computer in a consistent manner for each session. It consists of a mandatory training trial, with a given key and a line of symbols to pair, and the testing part with a new key, which lasts 90 seconds. The instructions are standardised and appear on the computer screen. The construction of the test was inspired by other well-known IPS measures, such as the digit-symbol subtest of the WAIS-R (Wechsler, 1997), the SDMT (Smith, 1982). This ensures its face validity.

The second step is the normalisation of raw scores. The test has been evaluated in a large sample of healthy subjects in various age, gender and educational level categories. Interestingly, the normative data has been implemented in the software, allowing the immediate classification of a given patient according to the SD from the mean scores of the group of HCs matched for age, gender and education.

Reliability is the next step of the validation and one of the most critical. Test-retest reliability is considered good if r>0.8 (Benedict et al., 2012b). The reliability of the CSCT was found to be excellent in both HCs (0.83) and MS patients (0.90). It is very close to the values obtained with the SDMT, which have consistently been found to be higher than 0.8 (Benedict et al., 2008; Morrow et al., 2010b; Benedict et al., 2012a). We confirmed this result by showing a small effect-size between two sessions in both HCs and MS patients.

The next step is the validity of the test itself, which refers to the meaning of a test score. We have approached this validity in a criterion-related way. First, we observed that the CSCT is able to discriminate MS patients from HC controls. In the MS group, 28 to 44% of MS patients were impaired in the CSCT. We observed an effect of the MS phenotype, with patients in a progressive stage of MS having worse scores in the CSCT. Age and education influenced the scores in MS patients, but this effect was taken into account by using controls matched for these variables. We tested the effect of fatigue, anxiety and depression as possible confounders. No effect of these parameters was observed. Second, we determined whether the CSCT is actually an IPS test. For that purpose, we used correlations between the CSCT and other established measures of IPS. In the first group of MS patients, we determined correlations with different RTs of subtests of the TAP battery. Indeed, the measurement of RTs from TAP subtests provided the opportunity to assess IPS

164

in several specific tasks, such as alertness and visual scanning, and the CSCT showed a strong correlation with RTs of these subtests. The correlations were good, but interestingly, the lowest r values were obtained for the correlation with the visual scanning task, suggesting that visual scanning does not play a major role in the CSCT score. This result is reassuring in MS patients, in which oculo-motor deficiencies are not rare.

In the second group of MS patients, we tested the correlation with the SDMT, which is currently considered to be the best IPS test in MS (Deloire et al., 2006; Parmenter et al., 2007). This test has recently been proposed to be part of a minimal battery, the BICAMS, for assessing cognitive impairment in PwMS (Langdon et al., 2012). Previously, the SDMT was also proposed as a viable alternative to the PASAT in studies using a brief MS functional neurological composite score (MSFC) (Brochet et al., 2008; Drake et al., 2010). In addition to IPS, visual scanning is involved in this task, in which patients have to rapidly locate the correct symbol/digit pairing on the key and then the symbol in each case. The CSCT scores were highly correlated with the SDMT scores in our second group of MS patients (r=0.94). Interestingly the proportion of patients with significantly abnormal scores compared to HCs was higher with the CSCT than with the SDMT, suggesting that the former is possibly more sensitive.

Finally, the validation of the test was confirmed by its ability to predict IPS impairment using other tests with an accuracy of 73%, suggesting that it captures a large part of the IPS slowness.

To approach the ecological validity of the test, we determined the correlation with vocational status. Several studies showed that cognitive impairment could lead to working limitations (Rao et al., 1991; Amato et al., 2001; Benedict et al., 2005). A decrease in the SDMT was found to be predictive of deterioration in vocational status in a retrospective study of 97 PwMS (Morrow et al., 2010a). It is noteworthy that the CSCT was associated with vocational status in PwMS in our first sample of MS patients.

The last step of the validation process is to consider the practice effect. One limitation of the SDMT is the presence of a practice effect (Bendict et al., 2008, Morrow et al, 2010b). Alternate forms have been proposed by Rao et al (1990) for the SDMT, but it has been reported that these alternate forms were not equivalent (Boringa et al., 2001; Benedict et al. 2012a). The question of alternate forms has been addressed using software to generate new keys and forms at each session. Even if the practice effect could include both item-specific and task-specific learning, changing the stimuli could reduce it (Benedict et al., 2005). At 6-month retesting, we observed a very small size-effect, suggesting a limited practice effect of the CSCT. However, this result needs to be confirmed in studies with longer follow-ups and a greater number of tests. If this low practice effect is confirmed, the CSCT will be a good candidate for monitoring IPS impairment over time in PwMS. The challenge will be to detect the minimal clinically meaningful change and to apply this suitable information to improving the management in PwMS.

165

ACKNOWLEDGMENTS

We thank Loock T for his contribution to data collection and Ouallet JC for his help in recruiting the MS patients.

The study was supported by Schering SA, France. The sponsors did not participate in any aspects of the design or conduct of the study, including data collection, management, analysis, or interpretation or the preparation, review, or approval of the manuscript.

Dr. Ruet is a recipient of a fellowship-grant from Fondation pour la Recherche Médicale.

References

Amato MP, Ponziani G, Siracusa G, Sorbi S. Cognitive dysfunction in early-onset multiple sclerosis: a reappraisal after 10 years. Arch Neurol 2001; 58, 1602-6.

Beck AT, Brown G, Steer RA. Beck Depression Inventory II manual. San Antonio, TX: The Psychological Corporation 1996.

Benedict RH, Wahlig E, Bakshi R, et al. Predicting quality of life in multiple sclerosis: accounting for physical disability, fatigue, cognition, mood disorder, personality, and behavior change. J Neurol Sci 2005; 231: 29-34.

Benedict RH, Duquin JA, Jurgensen S, et al. Repeated assessment of neuropsychological deficits in multiple sclerosis using the Symbol Digit Modalities Test and the MS Neuropsychological Screening Questionnaire. Mult Scler 2008; 14: 940-6.

Benedict RH. Assessment tools. In Amato MP (Elsevier) Cognitive Impairment in Multiple Sclerosis, Milano, 2011; pp. 9-16.

Benedict RH, Smerbeck A, Parikh R, Rodgers J, Cadavid D, Erlanger D. Reliability and equivalence of alternate forms for the Symbol Digit Modalities Test: implications for multiple sclerosis clinical trials. Mult Scler 2012a; 18: 1320-5.

Benedict R, Amato MP, Boringa J, et al. Brief International Cognitive Assessment for MS (BICAMS): international standards for validation. BMC Neurol 2012b, 12: 55. [Epub ahead of print].

Bensa C, Bertogliati C, Chanalet S, Malandain G, Bedoucha P, Lebrun C. Early detection of

cognitive impairment in relapsing-remitting multiple sclerosis: functional-anatomical correlations and longitudinal follow-up. Rev Neurol (Paris) 2006; 162: 1221-31.

Bergendal G, Fredrikson S, Almkvist O. Selective decline in information processing in subgroups of multiple sclerosis: an 8-year longitudinal study. Eur Neurol 2007; 57: 193-202.

166

Boringa JB, Lazeron RH, Reuling IE, et al. The brief repeatable battery of neuropsychological tests: normative values allow application in multiple sclerosis clinical practice. Mult Scler 2001; 7: 263-267.

Brochet B, Deloire MS, Bonnet M, et al. Should SDMT substitute for PASAT in MSFC? A 5-year longitudinal study. Mult Scler 2008; 14: 1242-9.

Brochet B. Assessing incapacity at early stages of Multiple sclerosis using the EDSS. Rev Neurol (Paris) 2009; 165: S173-S179.

Brochet B. Prevalence, profile and functional impact of cognitive impairment in multiple sclerosis. In Cognitive Impairment in Multiple Sclerosis, Amato MP (Elsevier) Milano, 2011; pp. 1-8.

Deloire MS, Bonnet MC, Salort E, Arimone Y, Boudineau M, Petry KG, et al. How to detect cognitive dysfunction at early stages of multiple sclerosis? Mult Scler 2006, 12. 445-52.

Deloire M, Ruet A, Hamel D, Bonnet M, Brochet B. Early cognitive impairment in multiple sclerosis predicts disability outcome several years later. Mult Scler 2010; 16: 581-7.

Drake AS, Weinstock-Guttman B, Morrow SA, Hojnacki D, Munschauer FE, Benedict RH. Psychometrics and normative data for the Multiple Sclerosis Functional Composite: replacing the PASAT with the Symbol Digit Modalities Test. Mult Scler 2010; 16: 228-37.

Dujardin K, Sockeel P, Cabaret M, De Sèze J, Vermersch P. BCcogSEP: a French test battery evaluating cognitive functions in multiple sclerosis. Rev Neurol (Paris) 2004; 160: 51-62.

Fischer JS, Rudick RA, Cutter GR, Reingold SC. The Multiple Sclerosis Functional Composite Measure (MSFC): an integrated approach to MS clinical outcome assessment. National MS Society Clinical Outcomes Assessment Task Force. Mult Scler 1999; 5: 244-50.

Jennekens-Schinkel A, Sanders EA, Lanser JB, Van der Velde EA. Reaction time in ambulant multiple sclerosis patients. Part I. Influence of prolonged cognitive effort. J Neurol Sci 1988; 85: 173-86.

Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444–1452.

Langdon DW, Amato MP, Boringa J, et al. Recommendations for a Brief International

Cognitive Assessment for Multiple Sclerosis (BICAMS). Mult Scler 2012; 18: 891-8.

Morrow SA, Drake A, Zivadinov R, Munschauer F, Weinstock-Guttman B, Benedict RH. Predicting loss of employment over three years in multiple sclerosis: clinically meaningful cognitive decline. Clin Neuropsychol 2010a, 24. 1131-45.

Morrow SA, O'Connor PW, Polman CH, et al. Evaluation of the symbol digit modalities test (SDMT) and MS neuropsychological screening questionnaire (MSNQ) in natalizumab-treated MS patients over 48 weeks. Mult Scler 2010b; 16: 1385-92.

167

Parmenter BA, Weinstock-Guttman B, Garg N, Munschauer F, Benedict RH. Screening for cognitive impairment in multiple sclerosis using the Symbol digit Modalities Test. Mult Scler 2007; 13: 52-7.

Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the ‘McDonald Criteria’’. Ann Neurol 2005; 58: 840–846.

Rao SM, the Cognitive Function Study Group of the National Multiple Sclerosis Society. A manual for the Brief Repeatable Battery of Neuropsychological Tests in multiple sclerosis. Milwaukee, WI: Medical College of Wisconsin; 1990.

Rao SM, Leo GJ, Bernardin L, Unverzagt F. Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction. Neurology 1991; 41: 685-691.

Ruet A, Charré-Morin J, Brochet B, Hamel D, Deloire M. How to detect cognitive impairment in relapsing-remitting and primary-progressive multiple sclerosis. Mult Scler 2011; 17 Suppl 10: S277-S505.

Schulz D, Kopp B, Kunkel A, Faiss JH. Cognition in the early stage of multiple sclerosis. J Neurol 2006; 253: 1002-10.

Sharrack B, Hughes RA. The Guy's Neurological Disability Scale (GNDS): a new disability measure for multiple sclerosis. Mult Scler 1999; 5: 223-233.

Smith A. Symbol Digit Modalities Test (SDMT) Manual (revised). Los Angeles: Western Psychological Services; 1982.

Spielberg CD. Manual for the state-trait anxiety inventory. Palo Alto: Consulting Psychologists Press; 1983.

Tinnefeld M, Treitz FH, Haase CG, Wilhelm H, Daum I, Faustmann PM. Attention and memory dysfunctions in mild multiple sclerosis. Eur Arch Psychiatry Clin Neurosci 2005; 255: 319-26.

Wechsler D. Wechsler Adult Intelligence Scale-III administration and scoring manual. San Antonio, TX: The Psychological Incorporation 1997

Zakzanis KK. Distinct neurocognitive profiles in multiple sclerosis subtypes. Arch Clin Neuropsychol 2000; 15: 115-36.

Zimmermann P, Fimm B. Tests d’évaluation de l’attention (TEA). Würselen: Psytest 1994.

Zimmermann P, Fimm B. TAP 2.1 Tests d’évaluation de l’attention Würzelen: Psytest 2009.

168

Discussion de cet article

1. Résumé des résultats de cet article

Cette étude décrit les caractéristiques psychométriques d'un nouveau test cognitif informatisé appelé le CSCT (Ruet et al., 2011b) à partir d’un échantillon de 101 patients ayant une SEP et 415 sujets sains contrôles appariés sur les données démographiques. Il a été retrouvé une excellente fiabilité pour ce test dans les deux groupes de sujets, ainsi qu’un faible effet d’apprentissage («practice effect») lors de la passation à 6 mois. Ces résultats étaient cliniquement pertinents et le CSCT était associé à une bonne validité écologique pour les patients ayant une SEP. Il existait une forte corrélation entre le CSCT et le SDMT et les autres tests de VTI pour les sujets du groupe SEP. Parmi les tests de la batterie NP utilisés pour estimer la VTI, le CSCT avait la meilleure sensibilité, et était l'un des tests les plus précis pour détecter un ralentissement de la VTI dans cet échantillon de patients. Ces résultats permettent de proposer le CSCT comme un bon candidat parmi les tests NP utiles pour l'évaluation de la VTI dans la SEP.

2. Intégration des résultats aux données actuelles de la littérature

Les troubles cognitifs sont fréquents dans la SEP (Brochet, dans Amato, 2011), et leur détection précoce apparaît aussi nécessaire que l'évaluation des déficits moteurs ou des autres déficiences chez les patients ayant un diagnostic de SEP. Cette détection peut fournir des informations importantes pour l'évaluation de la maladie, et pourrait être très utile dans la gestion des patients atteints de SEP pour adapter les stratégies de prise en charge de la maladie, la réadaptation cognitive et l'orientation professionnelle de ces patients. Malheureusement, la réalisation d’une évaluation NP complète n'est pas disponible partout, est chronophage, et la détection de l’atteinte cognitive n'est pas facile en pratique clinique quotidienne lors des visites de suivi neurologique. Par conséquent, le défi consiste à identifier les tests NP pouvant prédire les troubles cognitifs dans la SEP. Dans une précédente étude, la capacité d’un test NP à prédire des troubles cognitifs mis en évidence à partir d’une large batterie NP a été testée sur un échantillon de patients atteints de SEP-RR récemment diagnostiqués (Deloire et al., 2006). Le SDMT était le test NP associé à la meilleure exactitude pour prédire l’atteinte cognitive. Ce test permet d’explorer principalement la VTI et correspond à une tâche de substitution symboles/chiffres. Ces résultats ont été confirmé à partir d’un échantillon mixte de patients ayant une SEP-RR et une SEP-SP (Parmenter et al., 2007). Le SDMT était le test le plus sensible parmi la batterie NP de Rao (Rao Brief Repeatable Neuropsychological Battery: BRB-N) et «the Minimal Assessment of Cognitive Function in MS» (MACFIMS) dans un échantillon de 65 patients ayant une SEP (Strober et al., 2009). Le test de dépistage n'est pas destiné à se substituer à la passation de l'ensemble de l’évaluation NP qui reste essentielle pour caractériser précisément les déficits cognitifs, mais peut aider à identifier les patients avec un ralentissement de la VTI qui nécessitent une évaluation NP plus complète.

Récemment, un comité international d'experts en SEP a recommandé l’utilisation d’une brève batterie NP pour évaluer les troubles cognitifs des patients ayant une SEP

169

appelée «the Brief International Cognitive Assessment for MS» (BICAMS) (Langdon et al., 2012). Cette batterie comprend un test de VTI (le SDMT) (Rao et al., 1990) et deux tests de mémoire épisodique (la deuxième édition du «California Verbal Learning Test» (CVLT2) (Delis et al., 2000), et «the revised Brief Visuospatial Memory Test»(BVMTR) (Benedict et al., 1997)). Le comité a également proposé des recommandations pour la validation internationale d'un test NP (Benedict et al., 2012b). Cinq étapes ont été décrites par ce groupe d'experts. L'étape 1 et l’étape 2 requièrent respectivement la standardisation et la traduction des stimuli du test et de ses instructions. L'étape 3 consiste en la normalisation du test qui doit être fournie par l'étude d’au moins 65 sujets sains appariés aux patients sur l'âge, le sexe et le niveau d'éducation. L'étape 4 est la démonstration de la fiabilité lors du test-retest, qui peut être étudiée dans un petit échantillon de patients et/ou de sujets sains avec un intervalle d’une à trois semaines. L'étape 5 correspond au critère de validité du test en comparant les patients ayant une SEP à des sujets sains contrôles.

L'approche actuelle met l'accent sur des tests NP évaluant la VTI, qui est le domaine cognitif le plus souvent altéré chez les patients ayant une SEP (Zakzanis et al., 2000). Comme mentionné précédemment, le SDMT a été proposé comme un bon candidat pour détecter un ralentissement de la VTI. Les critères psychométriques de ce test ont été soulignés avec une bonne fiabilité lors du test-retest pour les patients ayant une SEP (Langdon et al., 2012). Néanmoins, ce test est associé à un effet d’apprentissage, et l’équivalence des différentes formes du SDMT issues de la BRB-N (Rao et al, 1990) n'a pas été démontrée (Boringa et al., 2001; Benedict et al., 2012a). Le CSCT qui est basé sur le même principe que d'autres tests de substitution de symboles/chiffres comme le WAIS-R Digit Symbol (Harris et al., 2007) ou le SDMT (Smith, 1982) comporte beaucoup de formes alternatives générées automatiquement par un logiciel informatique. Par conséquent, l’effet d’apprentissage devrait être moindre avec le CSCT qu'avec le SDMT.

3. Défis et difficultés

Le défi consistera à déterminer la valeur minimale du changement du score du CSCT cliniquement significative, et à appliquer ces informations en vue de l'amélioration de la gestion des patients ayant une SEP. Nos résultats doivent être validés dans d'autres études avec un plus long suivi, et le faible effet d’apprentissage du CSCT devra être confirmé avant de le proposer comme un outil pertinent pour le suivi de la progression du ralentissement de la VTI pour les patients ayant une SEP.

4. Conclusion et perspectives

L'enjeu sera d'identifier des tests psychométriques brefs permettant d’évaluer les principaux domaines cognitifs altérés dans la SEP: la VTI et la mémoire épisodique. En plus des techniques dites classiques, il est envisageable que des tests informatisés joueront prochainement un plus grand rôle pour la sélection des patients qui ont besoin d'une évaluation NP complète effectuée par un neuropsychologue qualifié. Comme le CSCT comporte de nombreux critères psychométriques pertinents, il est souhaitable d'utiliser ce test dans d'autres études longitudinales et internationales de patients ayant une SEP.

170

171

Chapitre 4:

Discussion générale

et perpectives

172

La SEP est la pathologie neurologique chronique invalidante d’origine non-traumatique la plus fréquente parmi les jeunes adultes en France. Son diagnostic repose sur la mise en évidence dans le système nerveux central de lésions disséminées dans l'espace et le temps. Cette pathologie aux multiples facettes est caractérisée par des lésions multiples focales (appelées plaques) dans le cerveau et la moelle épinière et d’une atteinte cérébrale diffuse. Le pronostic de la SEP est dominé par le développement d'une invalidité à long terme. De nombreux arguments illustrent que les premiers stades de la SEP sont cruciaux dans l'évolution ultérieure de la maladie et de son pronostic. La façon dont ces stades ont été considérés a considérablement changé au cours de ces dernières années en raison de plusieurs avancées importantes dans la connaissance de la maladie. L’atteinte inflammatoire démyélinisante a été traditionnellement considérée comme le processus principal de la maladie. Néanmoins, la survenue de lésions et/ou de pertes axonales est de plus en plus documentée au début de la maladie, et il a été rapporté des corrélations entre l'étendue de l’atteinte axonale et le degré de l'inflammation (Trapp et al., 1998). Plusieurs études neuropathologiques ont montré la relation entre l'inflammation et les lésions axonales, renforçant l'idée que le processus inflammatoire précoce est très important dans la détermination de la neurodégénérescence secondaire (Kutzelnigg et al., 2005; Lassmann et al., 2007). Il a été reconnu le rôle majeur de la perte axonale dans l’instauration de l'incapacité permanente des patients ayant une SEP (Bjartmar et al., 2000). Plusieurs études utilisant l’ITM, la spectroscopie par résonance magnétique ou des mesures d’atrophie cérébrale ont également montré de façon constante que la perte axonale peut apparaître tôt au cours de la maladie (Agosta et al., 2006; De Stefano et al., 2003; Horakova et al., 2009; Lukas et al., 2010;. Filippi et al., 2012). Des corrélations entre l'atrophie du cerveau et l’invalidité (Fisher et al., 2002) ont été rapportées, ce qui suggère que la progression de l'atrophie du cerveau peut être un marqueur utile de la progression de la maladie. Ces résultats illustrent aussi la pertinence clinique des lésions cérébrales diffuses, même aux stades précoces de la maladie (Fisher et al., 2002; Kalkers et al., 2002), car ils reflètent les processus pathologiques probablement les plus destructeurs et irréversibles. Néanmoins, il a été constaté que le développement de l'atrophie cérébrale n'est pas indépendant des lésions tissulaires focales présentes aux stades précoces de la SEP (Fisher et al., 2002; Kalkers et al., 2002) et est influencé par l’activité inflammatoire de la maladie, illustré en partie par la présence de lésions rehaussées après injection de produit de contraste (Simon et al., 1999). Récemment, le dépistage effectué à l’échelle du génome entier de la plupart des polymorphismes alléliques associés aux gènes qui contrôlent le système immunitaire (Gourraud et al., 2012) a renforcé l'idée que le dysfonctionnement du système immunitaire était essentiel dans le développement de la SEP. Cependant, le fait que l'inflammation soit la cause de la dégénérescence axonale et neuronale dans la SEP a été contesté. L'hypothèse selon laquelle un processus neurodégénératif primaire pourrait se produire indépendamment du processus inflammatoire a été suggérée à partir d’études de cohorte (Confavreux et al., 2003), et est encore un sujet à discussion. On ignore toujours si la progression de la maladie et de la neurodégénérescence dans la SEP peut se produire en l'absence même d'inflammation.

Ces résultats ont contribué à l'évolution de la conception des stades précoces de la maladie. Selon les données de cohorte, la progression du handicap au début de la SEP-RR pourrait être un marqueur utile pour la prédiction de l'incapacité à long-terme. Par conséquent, le diagnostic précoce de la SEP a suscité un intérêt croissant afin de pouvoir débuter un traitement de fond dès les premiers stades de la maladie. Les essais cliniques

173

pivots ont mis en évidence que le traitement par immunomodulateurs était efficace en cas d’utilisation dès les stades précoces de la maladie (Jacobs et al., 2000; Kappos et al., 2006;. Comi et al., 2009; Comi et al., 2012). La meilleure connaissance de l'évolution clinique de la maladie acquise à partir des études de cohortes et la meilleure utilisation des outils paracliniques (principalement l'IRM et, à un moindre degré, la ponction lombaire) ont fourni d'importants progrès pour porter le diagnostic de la SEP depuis quelques années. Ces progrès ont conduit à diverses propositions de critères diagnostic appelés «critères de McDonald» qui ont été discutés en détail dans la première partie de cette thèse. La mise en application des ces critères et leur large utilisation dans la communauté neurologique ont radicalement changé la façon dont la maladie est diagnostiquée. Cependant, plusieurs points importants au sujet de ces critères doivent être discutés. La structure générale de ces critères est issue des règles fondamentales du diagnostic de la SEP établies par Schumacher (Schumacher et al., 1965.), c'est à dire la démonstration de la dissémination des lésions dans l'espace et dans le temps après la survenue d’un d'épisode neurologique typique d’un premier événement démyélinisant et l’absence de meilleure explication. Cependant, l'évolution des critères a abouti à de fortes modifications concernant ces critères de DIS et de DIT. Dans les études de recherche et dans la pratique clinique, les critères de Barkhof-Tintoré sont largement acceptés pour démontrer la DIS à l’aide de l'IRM (Barkhof et al., 1997; Tintoré et al., 2000). En fait, ces critères sont issus d'études par IRM visant à déterminer après un SCI le risque de «conversion» en SEP-CD (Barkhof et al., 1997; Tintoré et al., 2000). Plus récemment, un comité international d’experts a proposé de nouvelles révisions des critères d’imagerie utilisés pour démontrer la DIS et la DIT après un SCI typique (Polman et al., 2011). Nous n'avons pas discuté le critère de DIT en détail dans la première partie de ce travail. Il est intéressant de noter que la possibilité d'établir une DIT à partir d'une seule IRM (Rovira et al., 2009) en cas de coexistence de lésions asymptomatiques gadolinium-positive et -négative acceptée dans la dernière version révisée des critères de McDonald (Polman et al., 2011) a été fondée sur l'hypothèse que cette caractéristique IRM avait la même exactitude dans la prédiction d’une «conversion» vers une SEP-CD. Ces critères simplifiés proposés en 2011 peuvent permettre l'établissement d'un diagnostic de SEP après une seule IRM au décours d’un SCI typique. Ils ont été issus des résultats de précédentes études européennes multicentriques d'IRM dans le cadre d’un réseau collaboratif de recherche appelé «MAGNIMS» (Swanton et al., 2007; Rovira et al., 2009; Montalban et al., 2010) utilisés pour valider ce concept sur une durée de suivi relativement courte. Cependant, ces nouveaux critères de DIT ne sont pas une réelle démonstration d’une DIT, comme dans les versions antérieures des critères de 2001 et de 2005, mais seulement une prédiction de la DIT. Puisque les BOC dans le LCR n'ont pas été évaluées dans ces études IRM en raison de données manquantes, les résultats du LCR n’ont pas été pris en compte dans les nouveaux critères. Dans la première partie de ce travail, nous avons identifié sans sélection au préalable des facteurs prédictifs du diagnostic de SEP par l'étude de paramètres à la fois cliniques, biologiques, et d’imagerie par IRM. A partir de deux cohortes distinctes, nous avons appliqué très facilement ces facteurs pour prédire le diagnostic de SEP, et il a été retrouvé une exactitude au moins similaire après l’utilisation de ces facteurs et celle des critères de DIS de 2005. La nécessité d'une véritable DIT, par la survenue d’un deuxième épisode clinique ou par l’IRM, avant de débuter un traitement de fond pourrait être débattue, mais ce n'est pas ici le sujet. Dans l'avenir, le bilan diagnostique pourrait changer radicalement en raison du développement d'un biomarqueur sensible et spécifique de la maladie. La SEP est reconnue comme étant une maladie auto-immune, mais son antigène cible reste à prouver. La publication récente de la fréquence des

174

anticorps anti-KIR4.1 dans la SEP (Srivastava et al., 2012) est prometteuse. Il est encourageant de constater qu'un biomarqueur pourrait être utilisé pour étayer le diagnostic de SEP, comme c'est déjà le cas dans la neuromyélite optique (NMO) avec l’utilisation des anticorps anti-NMO (Lennon et al., 2004). La présence d'anticorps anti-NMO a été incluse dans les critères diagnostiques de la NMO (Wingerchuk et al., 2006). Il serait très intéressant d'évaluer la valeur diagnostique d'un tel biomarqueur chez les patients considérés à haut risque de SEP.

À l'avenir, l'un des principaux défis dans la prise en charge de la SEP serait de pouvoir soit limiter soit réparer l’atteinte cérébrale diffuse. Dans la deuxième partie de cette thèse, nous avons confirmé que cette atteinte cérébrale diffuse peut se produire dès le début de la maladie et que l’atteinte cognitive, qui a été associée à cette atteinte, pourrait être utilisée comme un marqueur pronostique intéressant. Plus précisément, nous avons rapporté la relation entre l’atteinte de la VTI aux stades précoces de SEP et des marqueurs IRM connus pour être prédictifs de l’'incapacité physique dans la SEP. En outre, la valeur pronostique de l'atteinte de la VTI a été illustrée par sa capacité à prédire l'incapacité physique et le statut professionnel des patients ayant une SEP. Par conséquent, la détection d’un ralentissement de la VTI dans la SEP apparaît primordiale au vu de ses implications cliniques. Ainsi, un nouveau test cognitif informatisé, le CSCT, a été développé dans notre unité de recherche pour détecter une atteinte de la VTI. Nous avons démontré ses nombreuses caractéristiques psychométriques souhaitables pour un outil de dépistage. La prochaine étape sera la détermination de la valeur minimale du changement du score du CSCT cliniquement significative, et l'application de cette information pour améliorer la prise en charge des patients ayant une SEP. Par ailleurs, il faut souligner la démonstration récente qu'un traitement précoce par interféron β peut contribuer à préserver les capacités cognitives des patients SEP (Penner et al., 2012) potentiellement dû en partie à la préservation des mécanismes de compensation au niveau cérébral. Une hypothèse intéressante est que la relation entre le fonctionnement cognitif et l’atteinte cérébrale diffuse est du à la relation entre cette atteinte cérébrale diffuse et les capacités des mécanismes de compensation au niveau cérébral. Chez certains patients atteints de SEP, l’atteinte lésionnelle sévère visible à l’IRM cérébrale ne se traduit pas par des troubles cognitifs détectables aux stades précoces de la maladie (Bonnet et al., 2006), ce qui suggère la présence de mécanismes de compensation cérébrale. Dans les premiers stades de SEP-RR, les capacités compensatoires sont relativement maintenues, en particulier chez les patients hautement qualifiés (Bonnet et al., 2006). Il a été discuté l'influence protectrice d’une plus grande réserve cognitive contre les déficits cognitifs liés à la maladie (Sumowski et al., 2009; 2010). L’implication de mécanismes compensatoires pour l'exécution normale des tâches cognitives chez des patients ayant une SEP a été illustrée à partir d’études en IRM fonctionnelle (IRMf) (Filippi et Rocca, dans Amato, 2011). Ces études ont démontré que les patients atteints de SEP devaient recruter de plus grands réseaux cérébraux que les sujets sains pour effectuer une tâche cognitive donnée avec des niveaux de performance similaires (Audoin et al., 2005). L'efficacité de la réorganisation corticale des patients ayant une SEP peut jouer un rôle majeur dans l'explication de l'hétérogénéité interindividuelle des manifestations cliniques liées à la maladie. Le degré d'atteinte cérébrale diffuse peut limiter les possibilités de compensation cérébrale en déconnectant les réseaux impliqués dans la compensation ou en affectant des aires corticales d’intérêt (voies fronto-cérébelleuses, cortex préfrontal dorsolatéral, lobe pariétal médial, et cingulum) et ainsi contribuer à l’atteinte cognitive des patients. L'utilisation d'une tâche de go/no/go de complexité croissante dans une étude

175

d’IRMf chez des patients atteints de SEP-RR a permis d’observer une diminution du recrutement cérébral lorsque les patients doivent accomplir une tâche plus complexe exigeant plus de ressources attentionnelles (Bonnet et al., 2009; 2010). Ce résultat suggère une défaillance des phénomènes de compensation cérébrale pouvant conduire au ralentissement de la VTI.

Plusieurs questions demeurent quant à la physiopathologie des troubles cognitifs dans la SEP. L'interruption de la communication au sein des grands réseaux corticaux, survenant à la suite de l’atteinte diffuse de la SB, peut constituer l'un des substrats anatomiques des troubles cognitifs lors des stades précoces de la maladie. L’atteinte de la SG profonde et les lésions corticales peuvent également contribuer aux déficits cognitifs, soit directement en affectant l'intégrité des réseaux cognitifs spécialisés, soit indirectement en affectant les réseaux impliqués dans la compensation cérébrale. La mise en évidence de ces anomalies dès les premières manifestations cliniques de la maladie (SCI) pourrait fournir des marqueurs utiles et intéressants pour évaluer le pronostic à plus long terme des patients ayant une SEP.

En conclusion, nous avons confirmé notre hypothèse principale selon laquelle l’atteinte cognitive présente au début de la SEP a une valeur pronostique chez ces patients. Nous recommandons la détection précoce des troubles cognitifs dans la SEP afin de pouvoir adapter les stratégies thérapeutiques dès les premiers stades de la maladie. Puisque l’atteinte cérébrale diffuse reflète étroitement le processus neurodégénératif qui est responsable de l'accumulation ultérieure de l’invalidité irréversible dans la SEP, sa détection précoce est utile pour prédire le pronostic de la maladie à long terme. Un important défi pour les chercheurs réside dans la prévention de l’atteinte cérébrale diffuse et la limitation de sa progression dès les stades précoces de la maladie.

176

PERSPECTIVES

La détection des troubles cognitifs dans les mois suivant un SCI pourrait être un marqueur diagnostic utile de SEP, puisque l’atteinte cognitive a été associée à un taux de élevé de SEP-CD après un SCI typique (Zipoli et al., 2010). En outre, les troubles cognitifs au stade de SCI pourraient être utilisés comme marqueurs pronostiques intéressants, et leur étude dès ce stade à partir duquel ils deviennent détectables pourrait fournir des informations pertinentes concernant l’atteinte lésionnelle cérébrale sous-jacente. Il est connu que les lésions cérébrales focales et diffuses coexistent, même à des stades précoces de la maladie, mais il reste à déterminer la contribution respective de l’atteinte de la SG et de la SB au déclin cognitif dans la SEP. Nous supposons que les premiers dommages au niveau de la SB soient associés à une atteinte de la VTI, de la mémoire de travail, et des troubles résultant du ralentissement cognitif, pouvant affecter secondairement l'attention et la mémoire épisodique, tandis que, les troubles les plus graves de la mémoire épisodique, de l'attention, et des fonctions exécutives pourraient être liés à des atteintes plus tardives de la SG, en particulier au niveau du cortex. Le ralentissement de la VTI est supposé être le premier déficit des fonctions cognitives dans la SEP, et cette altération pourrait être la conséquence d'une défaillance des mécanismes de compensation plutôt qu'une simple déconnexion au sein des réseaux neuronaux impliqués dans cette fonction. Pour explorer cette question, nous prévoyons de réaliser une étude prospective longitudinale de 60 patients inclus dans un délai de 4 mois après un SCI typique évocateur de SEP et 60 sujets sains. Les principaux objectifs seront de mieux comprendre la physiopathologie des déficits cognitifs, qui sont détectables dès le premier événement clinique démyélinisant évocateur de SEP, et d’explorer la relation entre l’atteinte cérébrale diffuse, présente dès ce stade précoce, et l'échec des mécanismes de compensation au niveau cérébral et ses implications cliniques. Cette étude pourrait aider à déterminer quels sont les paramètres d’imagerie par RMN utiles comme marqueurs de l’atteinte cognitive après un SCI. La technique d’ITD sera utilisée pour nous aider à comprendre les mécanismes sous-jacents des troubles cognitifs à un stade précoce. Cette technique est particulièrement adaptée pour l'étude des déconnexions entre les différentes zones cérébrales au niveau des faisceaux de SB. La technique appelée «Tract-based spatial statistics» (TBSS) sera également utile pour fournir des informations pertinentes sur la localisation des dommages sans a priori, et permettra de générer des hypothèses spécifiques concernant la localisation des processus lésionnels en lien avec les troubles cognitifs aux stades précoces de la SEP (Smith et al., 2006). En outre, les mesures d’atrophie, qui représentent la méthode quantitative la plus robuste pour estimer les dommages structuraux de la SG, seront réalisées d’une part par la détermination de l’épaisseur corticale (pour évaluer l'atrophie corticale), et d’autre part par le calcul du volume des noyaux gris centraux (pour évaluer l'atrophie profonde de la SG) à partir d’une séquence pondérée en 3D-T1 et en utilisant le logiciel "Freesurfer". Enfin, nous allons effectuer une séquence d’IRM fonctionnelle au repos («resting state») pour évaluer la connectivité fonctionnelle cérébrale. Au cours de l’état dit «de repos», il a été rapporté que le cerveau pourrait être très actif, même en l'absence de toute tâche explicite. L’étude du réseau à l'état de repos pourrait donner un meilleur aperçu de l'architecture intrinsèque fonctionnelle du cerveau, et les mécanismes de fonctionnement cognitif dans la SEP. Une étude récente a indiqué que l'état de repos était fortement associé au degré d’invalidité, évalué à l’aide du score MSFC, chez 13 patients atteints de SEP aux stades précoces (Faivre et al., 2012).

177

Ces résultats suggèrent que l'étude de l’état de repos en IRM fonctionnelle constitue une approche encourageante pour caractériser la maladie. De plus, les marqueurs d’IRM pourraient aussi être utiles pour prédire l'efficacité de la réadaptation cognitive ou des programmes de remédiation cognitive des patients ayant une SEP et des déficits cognitifs. Enfin, il est important de souligner que nous prévoyons un suivi longitudinal des patients inclus dans cette étude après un SCI typique afin d’obtenir des informations supplémentaires pour étayer la valeur pronostique des troubles cognitifs détectés aux stades précoces pour prédire l’invalidité à long terme des patients ayant un diagnostic de SEP.

178

REFERENCES

179

A

Achiron A, Barak Y. Cognitive impairment in probable multiple sclerosis. J Neurol Neurosurg Psychiatry 2003; 74: 443-6.

Agosta F, Rovaris M, Pagani E, Sormani MP, Comi G, Filippi M. Magnetization transfer MRI metrics predict the accumulation of disability 8 years later in patients with multiple sclerosis. Brain 2006; 129: 2620-2627.

Allen IV, McQuaid S, Mirakhur M, Nevin G. Pathological abnormalities in the normal-appearing white matter in multiple sclerosis. Neurol Sci 2001; 22: 141-144.

Amato MP, Ponziani G, Bartolozzi ML, Siracusa G. A prospective study on the natural history of multiple sclerosis: clues to the conduct and interpretation of clinical trials. J Neurol Sci 1999; 168: 96-106.

Amato MP, Ponziani G. A prospective study on the prognosis of multiple sclerosis. Neurol Sci 2000; 21: S831-8.

Amato MP, Ponziani G, Siracusa G, Sorbi S. Cognitive dysfunction in early-onset multiple sclerosis: a reappraisal after 10 years. Arch Neurol 2001a; 58, 1602-6.

Amato MP, Ponziani G, Rossi F, Liedl CL, Stefanile C and Rossi L. Quality of life in multiple sclerosis: the impact of depression, fatigue and disability. Mult Scler 2001b; 7: 340–344.

Amato MP, Bartolozzi ML, Zipoli V, et al. Neocortical volume decrease in relapsing–remitting MS patients with mild cognitive impairment. Neurology 2004; 63: 89–93.

Amato MP, Zipoli V, Goretti B, et al. Benign multiple sclerosis: cognitive, psychological and social aspects in a clinical cohort. J Neurol 2006; 253: 1054-9.

Amato MP, Portaccio E, Goretti B, et al. Association of neocortical volume changes with cognitive deterioration in relapsing-remitting multiple sclerosis. Arch Neurol 2007; 64: 1157-1161.

Amato MP, Portaccio E, Stromillo ML, et al. Cognitive assessment and quantitative magnetic resonance metrics can help to identify benign multiple sclerosis. Neurology 2008; 71: 632–638.

Amato MP, Portaccio E, Goretti B, et al. Relevance of cognitive deterioration in early relapsing–remitting MS: a 3-year follow-up study. Mult Scler 2010; 16: 1474–1482.

Amato MP, Hakiki B, Goretti B, et al. Association of MRI metrics and cognitive impairment in radiologically isolated syndromes. Neurology 2012; 78: 309–314.

Audoin B, Ranjeva JP, van Au Dong M, et al. Voxel-based analysis of MTR images: a method to locate gray matter abnormalities in patients at the earliest stage of multiple sclerosis. J Magn Reson Imaging 2004; 20: 765–771.

180

Audoin B, Au Duong MV, Ranjeva JP, et al. Magnetic resonance study of the influence of tissue damage and cortical reorganization on PASAT performance at the earliest stage of multiple sclerosis. Hum Brain Mapp 2005; 24: 216-28.

Audoin B, Davies GR, Finisku L, Chard DT, Thompson AJ, Miller DH. Localization of grey matter atrophy in early RRMS: A longitudinal study. J Neurol 2006; 253: 1495-1501.

Audoin B, Zaaraoui W, Reuter F, et al. Atrophy mainly affects the limbic system and the deep grey matter at the first stage of multiple sclerosis. J Neurol Neurosurg Psychiatry 2010; 81: 690-695.

B

Bakshi R, NeemaM, Healy BC, et al. Predicting clinical progression in multiple sclerosis with the magnetic resonance disease severity scale. Arch Neurol 2008; 65: 1449–53.

Barkhof F, Filippi M, Miller DH, et al. Comparison of MRI criteria at first presentation to predict conversion to clinically definite multiple sclerosis. Brain 1997; 120: 2059–69.

Barkhof F. The clinico-radiological paradox in multiple sclerosis revisited. Curr Opin Neurol. 2002; 15: 239-245.

Bashir K, Whitaker JN. Importance of paraclinical and CSF studies in the diagnosis of MS in patients presenting with partial cervical transverse myelopathy and negative cranial MRI. Mult Scler 2000; 6: 312-6.

Baumstarck-Barrau K, Pelletier J, Simeoni MC, Auquier P; MusiQol Study Group. French validation of the Multiple Sclerosis International Quality of Life Questionnaire. Rev Neurol (Paris) 2011a; 167: 511-21.

Baumstarck-Barrau K, Simeoni MC, Reuter F, et al. Cognitive function and quality of life in multiple sclerosis patients: a cross-sectional study. BMC Neurol 2011b; 2; 11:17.

Baumstarck K, Pelletier J, Aghababian V, et al. Is the concept of quality of life relevant for multiple sclerosis patients with cognitive impairment? Preliminary results of a cross-sectional study. PLoS One 2012; 7:e30627.

Beatty WW, Goodkin DE, Hertsgaard D, Monson N. Clinical and demographic predictors of cognitive performance in multiple sclerosis: do diagnostic type, disease duration and disability matter? Arch Neurol 1990; 47: 305-308.

Bellmann-Strobl J, Wuerfel J, Aktas O, et al. Poor PASAT performance correlates with MRI contrast enhancement in multiple sclerosis. Neurology 2009; 73: 1624-7.

Benedict RH. Brief visuospatial memory test-revised. Professional manual. Odessa, Floriday: psychological assessment resources, Inc; 1997.

Benedict RH, Fischer JS, Archibald C, et al. Minimal neuropsychological assessment of MS patients: a consensus approach. Clin Neuropsychol 2002; 16: 381-97.

181

Benedict RH, Munschauer F, Linn R, et al. Screening for multiple sclerosis cognitive impairment using a self-administered 15-item questionnaire. Mult Scler 2003; 9: 95-101.

Benedict RH, Cox D, Thompson LL, Foley F, Weinstock-Guttman B, Munschauer F. Reliable screening for neuropsychological impairment in multiple sclerosis. Mult Scler 2004a; 10: 675-8.

Benedict RH, Weinstock-Guttman B, Fishman I, Sharma J, Tjoa CW, Bakshi R. Prediction of neuropsychological impairment in multiple sclerosis: comparison of conventional magnetic resonance imaging measures of atrophy and lesion burden. Arch Neurol 2004b; 61: 226-30.

Benedict RH, Wahlig E, Bakshi R, et al. Predicting quality of life in multiple sclerosis: accounting for physical disability, fatigue, cognition, mood disorder, personality, and behavior change. J Neurol Sci 2005; 231: 29-34.

Benedict RH, Cookfair D, Gavett R, et al. Validity of the minimal assessment of cognitive function in multiple sclerosis (MACFIMS). J Int Neuropsychol Soc 2006a; 12: 549-58.

Benedict RH, Bruce JM, Dwyer MG, et al. Neocortical atrophy, third ventricular width, and cognitive dysfunction in multiple sclerosis. Arch Neurol 2006b; 63: 1301-6.

Benedict RH, Bruce J, Dwyer MG et al. Diffusion-weighted imaging predicts cognitive impairment in multiple sclerosis. Mult Scler 2007; 13: 722-30.

Benedict RH, Duquin JA, Jurgensen S, et al. Repeated assessment of neuropsychological deficits in multiple sclerosis using the Symbol Digit Modalities Test and the MS Neuropsychological Screening Questionnaire. Mult Scler 2008; 14: 940-6.

Benedict RH, Smerbeck A, Parikh R, Rodgers J, Cadavid D, Erlanger D. Reliability and equivalence of alternate forms for the Symbol Digit Modalities Test: implications for multiple sclerosis clinical trials. Mult Scler 2012a; 18: 1320-5.

Benedict R, Amato MP, Boringa J, et al. Brief International Cognitive Assessment for MS (BICAMS): international standards for validation. BMC Neurol 2012b, 12: 55. [Epub ahead of print].

Benito-León J, Morales JM, Rivera-Navarro J. Health-related quality of life and its relationship to cognitive and emotional functioning in multiple sclerosis patients. Eur J Neurol 2002; 9: 497-502.

Berman M, Feldman S, Alter M, Zilber N, Kahana E. Acute transverse myelitis: incidence and etiologic considerations. Neurology 1981; 31: 966-71.

Bermel RA, Bakshi R, Tjoa C, Puli SR, Jacobs L. Bicaudate ratio as a magnetic resonance imaging marker of brain atrophy in multiple sclerosis. Arch Neurol 2002; 59:275-280.

Bjartmar C, Kinkel RP, Kidd G, Rudick RA, Trapp BD. Axonal loss in normal-appearing white matter in a patient with acute MS. Neurology 2001; 57: 1248-1252.

182

Blinkenberg M, Mathiesen HK, Tscherning T, et al. Cerebral metabolism, magnetic resonance spectroscopy and cognitive dysfunction in early multiple sclerosis: an exploratory study. Neurol Res 2012; 34: 52-8.

Bo L, Vedeler CA, Nyland HI, Trapp BD, Mork SJ. Subpial demyelination in the cerebral cortex of multiple sclerosis patients. J Neuropathol Exp Neurol 2003; 62: 723-732.

Bobholz JA, Rao SM. Cognitive dysfunction in multiple sclerosis: a review of recent developments. Curr Opin Neurol. 2003; 16: 283-8.

Bonnet MC, Deloire MS, Salort E, Dousset V, Petry KG, Brochet B. Evidence of cognitive compensation associated with educational level in early relapsing-remitting multiple sclerosis. J Neurol Sci 2006; 251: 23-8.

Bonnet MC, Dilharreguy B, Allard M, Deloire MS, Petry KG, Brochet B. Differential cerebellar and cortical involvement according to various attentional load: role of educational level. Hum Brain Mapp 2009; 30: 1133-43.

Bonnet MC, Allard M, Dilharreguy B, Deloire M, Petry KG, Brochet B. Cognitive compensation failure in multiple sclerosis. Neurology 2010; 75: 1241-8.

Boringa JB, Lazeron RH, Reuling IE, et al. The brief repeatable battery of neuropsychological tests: normative values allow application in multiple sclerosis clinical practice. Mult Scler 2001;7: 263-267.

Bourre B, Zéphir H, Ongagna JC, et al. Long-term follow-up of acute partial transverse myelitis. Arch Neurol 2012; 69: 357-62.

Brex PA, Ciccarelli O, O’Riordan JI, Sailer M, Thompson AJ, Miller DH. A longitudinal study of abnormalities on MRI and disability from multiple sclerosis. N Engl J Med 2002; 346: 158-164.

Brochet B, Dousset V. Pathological correlates of magnetization transfer imaging abnormalities in animal models and humans with multiple sclerosis. Neurology 1999; 53:S12-7.

Brochet B. Prevalence, profile and functional impact of cognitive impairment in multiple sclerosis. In Amato MP (Elsevier) Cognitive Impairment in Multiple Sclerosis, Milano, 2011; pp. 1-8.

Broman T, Andersen O, Bergmann L. Clinical studies on multiple sclerosis. I. Presentation of an incidence material from Gothenburg. Acta Neurol Scand 1981; 63: 6-33.

Brozek JL, Guyatt GH, Schünemann HJ. How a well-grounded minimal important difference can enhance transparency of labelling claims and improve interpretation of a patient reported outcome measure. Health Qual Life Outcomes 2006; 4: 69.

Bruna J, Martínez-Yélamos S, Martínez-Yélamos A, et al. Idiopathic acute transverse myelitis: a clinical study and prognostic markers in 45 cases. Mult Scler 2006; 12: 169-73.

183

C

Calabrese M, Agosta F, Rinaldi F et al. Cortical lesions and atrophy associated with cognitive impairment in relapsing-remitting multiple sclerosis. Arch Neurol 2009; 66: 1144-50.

Calabrese M, Rocca M, Atzori M, et al. A three-year MRI study of cortical lesions in relapse-onset multiple sclerosis. Ann Neurol 2010a; 67: 376–83.

Calabrese M, Rinaldi F, Mattisi I, et al. Widespread cortical thinning characterizes patients with MS with mild cognitive impairment. Neurology 2010b; 74: 321-8.

Calabrese M, Rinaldi F, Sappi D, et al. Cortical diffusion-tensor imaging abnormalities in multiple sclerosis: a 3-year longitudinal study. Radiology 2011; 261: 891–898.

Calabrese M, Federle L, Bernardi V, et al. The association of intrathecal immunoglobulin synthesis and cortical lesions predicts disease activity in clinically isolated syndrome and early relapsing-remitting multiple sclerosis. Mult Scler 2012; 18: 174-80.

Callanan MM, Logsdail SJ, Ron MA, Warrington EK. Cognitive impairment in patients with clinically isolated lesions of the type seen in multiple sclerosis. A psychometric and MRI study. Brain. 1989; 112: 361-74.

Campi A, Filippi M, Comi G, et al. Acute transverse myelopathy: spinal and cranial MR study with clinical follow-up. AJNR Am J Neuroradiol 1995;16: 115-23.

Ceccarelli A, Bakshi R, Neema M. MRI in multiple sclerosis: a review of the current literature. Curr Opin Neurol 2012; 25: 402–409.

Cercignani M, Bozzali M, Iannucci G, Comi G, Filippi M. Magnetisation transfer ratio and mean diffusivity of normal appearing white and grey matter from patients with multiple sclerosis. J Neurol Neurosurg Psychiatry 2001; 70: 311-317.

Chahraoui K, Bonin B, Couvreur G, Fromont A, Viegas N, Moreau T.Subjective quality of life profile in patients with multiple sclerosis. Rev Neurol (Paris) 2010; 166:745-9.

Charcot J.M, 1877. Lectures on the diseases of the nervous system. New Sydenham Society, London.

Chard DT, Griffin CM, Parker GJ, Kapoor R, Thompson AJ, Miller DH. Brain atrophy in clinically early relapsing–remitting multiple sclerosis. Brain 2002; 125: 327–37.

Chard DT. MRI only conversion to multiple sclerosis following a clinically isolated syndrome. J Neurol Neurosurg Psychiatry 2011; 82: 176-9.

Chiaravalloti ND, DeLuca J. Cognitive impairment in multiple sclerosis. Lancet Neurol 2008; 7: 1139–1151.

Choi S.R, Howell O.W, Carassiti D, et al. Meningeal inflammation plays a role in the pathology of primary progressive multiple sclerosis. Brain 2012; 1-13.

Christodoulou C, Krupp LB, Liang Z et al. Cognitive performance and MR markers of cerebral injury in cognitively impaired MS patients. Neurology 2003; 60: 1793-1798.

184

Clavelou P, Auclair C, Taithe F, Gerbaud L. Quality of life in multiple sclerosis. Rev Neurol (Paris) 2009;165 Suppl 4:S123-8.

Comi G, Filippi M, Martinelli V, et al. Brain magnetic resonance imaging correlates of cognitive impairment in primary and secondary progressive multiple sclerosis. J Neurol Sci 1995; 132: 222-227.

Comi G, Martinelli V, Rodegher M, et al. Effect of glatiramer acetate on conversion to clinically definite multiple sclerosis in patients with clinically isolated syndrome (PreCISe study): a randomized, double-blind, placebo-controlled trial. Lancet 2009; 374: 1503-11.

Comi G, De Stefano N, Freedman MS, et al. Comparison of two dosing frequencies of subcutaneous interferon beta-1a in patients with a first clinical demyelinating event suggestive of multiple sclerosis (REFLEX): a phase 3 randomised controlled trial. Lancet Neurol 2012; 11: 33–41.

Confavreux C, Aimard G, Devic M. Course and prognosis of multiple sclerosis assessed by the computerized data processing of 349 patients. Brain 1980; 103: 281-300.

Confavreux C, Vukusic S, Moreau T, Adeleine P. Relapses and progression of disability in multiple sclerosis. N Engl J Med 2000; 343: 1430-1438.

Confavreux C, Vukusic S, Adeleine P. Early clinical predictors and progression of irreversible disability in multiple sclerosis: an amnesic process. Brain 2003; 126: 770–82.

Confavreux C, Vukusic S. Age at disability milestones in multiple sclerosis. Brain 2006; 129: 595-605.

Cordonnier C, de Seze J, Breteau G, et al. Prospective study of patients presenting with acute partial transverse myelopathy. J Neurol 2003; 250: 1447-52.

D

Dalton CM, Chard DT, Davies GR, et al. Early development of multiple sclerosis is associated with progressive grey matter atrophy in patients presenting with clinically isolated syndromes. Brain 2004; 127: 1101–7.

Dalton C, Bodini B, Samson R, et al. Brain lesion location and clinical status 20 years after a diagnosis of clinically isolated syndrome suggestive of multiple sclerosis. Mult Scler 2012; 18: 322–328.

Debette S, de Seze J, Pruvo JP, et al. Long-term outcome of acute and subacute myelopathies. J Neurol 2009; 256: 980-8.

Debouverie M, Pittion-Vouyovitch S, Louis S, et al. LORSEP Group. Natural history of multiple sclerosis in a population-based cohort. Eur J Neurol 2008; 15: 916-921.

Delis DC, Kramer JH, Kaplan E, Ober BA. California verbal learning test manual: second edition, adult version. San Antonio, TX : Psychological Corporation ; 2000.

185

Deloire-Grassin MS, Brochet B, Quesson B, et al. In vivo evaluation of remyelination in rat brain by magnetization transfer imaging. J Neurol Sci 2000; 178:10-16.

Deloire MS, Salort E, Bonnet M, et al. Cognitive impairment as marker of diffuse brain abnormalities in early relapsing-remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2005; 76: 519-526.

Deloire MS, Bonnet MC, Salort E, et al. How to detect cognitive dysfunction at early stages of multiple sclerosis? Mult Scler 2006; 12: 445-52.

Deloire M, Ruet A, Hamel D, Bonnet M, Brochet B. Early cognitive impairment in multiple sclerosis predicts disability outcome several years later. Mult Scler 2010; 16, 581-7.

Deloire MS, Ruet A, Hamel D, Bonnet M, Dousset V, Brochet B. MRI predictors of cognitive outcome in early multiple sclerosis. Neurology 2011; 76, 1161-7.

DeLuca J, Chelune GJ, Tulsky DS, Lengenfelder J, Chiaravalloti ND. Is speed of processing or working memory the primary information processing deficit in multiple sclerosis? J Clin Exp Neuropsychol 2004; 26: 550–562.

Denney DR, Lynch SG, Parmenter BA, Horne N. Cognitive impairment in relapsing and primary progressive multiple sclerosis: mostly a matter of speed. J Int Neuropsychol Soc 2004; 10: 948-56.

Denney DR, Lynch SG, Parmenter BA. A 3-year longitudinal study of cognitive impairment in patients with primary progressive multiple sclerosis: speed matters. J Neurol Sci 2008; 267: 129-36.

Detels R, Clark VA, Valdiviezo NL, Visscher BR, Malmgren RM, Dudley JP. Factors associated with a rapid course of multiple sclerosis. Arch Neurol 1982; 39: 337-41.

de Seze J, Stojkovic T, Breteau G, et al, Acute myelopathies: Clinical, laboratory and outcome profiles in 79 cases. Brain 2001; 124: 1509-21.

de Seze J, Delalande S, Michelin E, et al. Brain MRI in late-onset multiple sclerosis. Eur J Neurol 2005 Apr; 12: 241-4.

De Sonneville LM, Boringa JB, Reuling IE, Lazeron RH, Adèr HJ, Polman CH. Information processing characteristics in subtypes of multiple sclerosis. Neuropsychologia 2002; 40: 1751-65.

De Stefano N, Matthews PM, Filippi M, et al. Evidence of early cortical atrophy in MS: relevance to white matter changes and disability. Neurology 2003; 60: 1157–1162.

De Stefano N, Battaglini M, Stromillo ML, et al. Brain damage as detected by magnetization transfer imaging is less pronounced in benign than in early relapsing multiple sclerosis. Brain 2006; 129: 2008-2016.

De Stefano N, Filippi M. MR spectroscopy in multiple sclerosis. J Neuroimaging 2007; 17 Suppl 1: 31S-35S.

186

De Stefano N, Stromillo ML, Rossi F, et al. Improving the characterization of radiologically isolated syndrome suggestive of multiple sclerosis. PLoS One 2011; 6: e19452.

Dineen RA, Vilisaar J, Hlinka J, et al. Disconnection as a mechanism for cognitive dysfunction in multiple sclerosis. Brain 2009; 132: 239–249.

Dousset V, Grossman RI, Ramer KN, et al. Experimental allergic encephalomyelitis and multiple sclerosis: lesion characterization with magnetization transfer imaging Radiology 1992; 183: 483-491

Dousset V, Gayou A, Brochet B, Caillé JM. Early structural changes in acute nascent MS lesions suggesting demyelination and remyelination assessed by in vivo serial quantitative magnetization transfer studies. Neurology 1998; 51: 1150-1155.

Duque B, Sepulcre J, Bejarano B, Samaranch L, Pastor P, Villoslada P. Memory decline evolves independently of disease activity in MS. Mult Scler 2008; 14: 947-53.

E

Ebers GC, Heigenhauser L, Daumer M, Lederer C, Noseworthy JH. Disability as an outcome in MS clinical trials. Neurology 2008; 71: 624-631.

Eriksson M, Andersen O, Runmarker B. Long-term follow up of patients with clinically isolated syndromes, relapsing-remitting and secondary progressive multiple sclerosis. Mult Scler 2003; 9: 260-74.

Ellison GW, Myers LW, Leake BD, et al. Design strategies in multiple sclerosis clinical trials. Cyclosporine Multiple Sclerosis Study Group. Ann Neurol 1994; 36: S108-S112.

Evangelou N, Esiri MM, Smith S, Palace J, Matthews PM. Quantitative pathological evidence for axonal loss in normal appearing white matter in multiple sclerosis. Ann Neurol 2000; 47: 391-395.

F

Faivre A, Rico A, Zaaraoui W, et al. Assessing brain connectivity at rest is clinically relevant in early multiple sclerosis. Mult Scler 2012; 18: 1251-8.

Fazekas F, Offenbacher H, Fuchs S, et al. Criteria for an increased specificity of MRI interpretation in elderly subjects with suspected multiple sclerosis. Neurology 1988; 38: 1822-1825.

Feinstein A, Kartsounis LD, Miller DH, Youl BD, Ron MA. Clinically isolated lesions of the type seen in multiple sclerosis: a cognitive, psychiatric, and MRI follow up study. J Neurol Neurosurg Psychiatry 1992; 55: 869-76.

187

Fernández O, Baumstarck-Barrau K, Simeoni MC, Auquier P; MusiQoL study group. Patient characteristics and determinants of quality of life in an international population with multiple sclerosis: assessment using the MusiQoL and SF-36 questionnaires. Mult Scler 2011; 17: 1238-49.

Feuillet L, Reuter F, Audoin B, et al. Early cognitive impairment in patients with clinically isolated syndrome suggestive of multiple sclerosis. Mult Scler 2007; 13: 124-7.

Filippi M, Paty DW, Kappos L, et al. Correlations between changes in disability and T2-weighted brain MRI activity in multiple sclerosis: a follow-up study. Neurology 1995; 45: 255-60.

Filippi M, Iannucci G, Tortorella C, et al. Comparison of MS clinical phenotypes using conventional and magnetization transfer MRI. Neurology 1999; 52: 588-594.

Filippi M, Tortorella C, Rovaris M, et al. Changes in the normal appearing brain tissue and cognitive impairment in multiple sclerosis. J Neurol Neurosurg Psychiatry 2000; 68: 157-161.

Filippi M, Rocca MA, Benedict RH, et al. The contribution of MRI in assessing cognitive impairment in multiple sclerosis. Neurology 2010a; 75: 2121–2128.

Filippi M, Rocca MA, Calabrese M, et al. Intracortical lesions: relevance for new MRI diagnostic criteria for multiple sclerosis. Neurology 2010b; 75: 1988–1994.

Filippi M., Rocca MA. MRI correlates. In Amato MP (Elsevier) Cognitive Impairment in Multiple Sclerosis, Milano, 2011; pp 53-60.

Filippi M, Rocca MA, Barkhof F, et al. Association between pathological and MRI findings in multiple sclerosis. Lancet Neurol 2012; 11: 349–60.

Filli L, Hofstetter L, Kuster P, et al. Spatiotemporal distribution of white matter lesions in relapsing-remitting and secondary progressive multiple sclerosis. Mult Scler 2012; 11. [Epub ahead of print]

Fischer JS, Rudick RA, Cutter GR, Reingold SC. The Multiple Sclerosis Functional Composite Measure (MSFC): an integrated approach to MS clinical outcome assessment. National MS Society Clinical Outcomes Assessment Task Force. Mult Scler 1999; 5: 244-50.

Fisher E, Rudick R.A; Simon J.H, et al. Eight-year follow-up study of brain atrophy in patients with MS. Neurology 2002; 59:1412–1420.

Fisniku LK, Brex PA, Altmann DR, et al. Disability and T2 MRI lesions: a 20-year follow-up of patients with relapse onset of multiple sclerosis. Brain 2008; 131: 808-17.

Foong J, Rozewicz L, Chong WK, Thompson AJ, Miller DH, Ron MA. A comparison of neuropsychological deficits in primary and secondary progressive multiple sclerosis. J Neurol 2000 ;247: 97–101.

Ford B, Tampieri D, Francis G. Long-term follow-up of acute partial transverse myelopathy. Neurology 1992; 42: 250-2.

188

Forn C, Belenguer A, Parcet-Ibars MA, Avila C. Information-processing speed is the primary deficit underlying the poor performance of multiple sclerosis patients in the Paced Auditory Serial Addition Test (PASAT). J Clin Exp Neuropsychol 2008; 13:1-8.

Freedman MS, Thompson EJ, Deisenhammer F, et al. Recommended standard of cerebrospinal fluid analysis in the diagnosis of multiple sclerosis: a consensus statement. Arch Neurol 2005; 62: 865-70.

Fromont A, Binquet C, Sauleau EA, et al. Geographic variations of multiple sclerosis in France. Brain 2010; 133: 1889-99.

G

Gadea M, Martínez-Bisbal MC, Marti-Bonmatí L, et al. Spectroscopic axonal damage of the right locus coeruleus relates to selective attention impairment in early stage relapsing-remitting multiple sclerosis. Brain 2004; 127: 89-98.

Gass A, Barker GJ, Kidd D, et al. Correlation of magnetization transfer ratio with disability in multiple sclerosis. Ann Neurol 1994; 36: 62-67.

Gaudino EA, Chiaravalloti ND, DeLuca J, Diamond BJ. A comparison of memory performance in relapsing-remitting, primary progressive and secondary progressive, multiple sclerosis. Neuropsychiatry Neuropsychol Behav Neurol 2001;14 : 32-44.

Gerbaud L, Deffond D, Mulliez A, Benausse F, Vernay D, Clavelou P. Cognitive impairment and quality of life in multiple sclerosis patients. Rev Neurol (Paris) 2006; 162: 970-9.

Geurts JJ, Bo L, Pouwels PJ, Castelijns JA, Polman CH, Barkhof F. Cortical lesions in multiple sclerosis: combined post-mortem MR imaging and histopathology. AJNR Am J Neuroradiol 2005a; 26: 572–577.

Geurts JJ, Pouwels PJ, Uitdehaag BM, Polman CH, Barkhof F, Castelijns JA. Intracortical lesions in multiple sclerosis: improved detection with 3D double inversion-recovery MR imaging. Radiology 2005b; 236: 254–260.

Geurts JJG, Barkhof F. Grey matter pathology in multiple sclerosis. Lancet Neurol 2008; 7: 841–51.

Giorgio A, Stromillo ML, Rossi F, et al. Cortical lesions in radiologically isolated syndrome. Neurology 2011; 77: 1896–1899.

Giorgio A, De Stefano N. MRI correlates. In Amato MP (Elsevier) Cognitive Impairment in Multiple Sclerosis, Milano, 2011; pp. 45-52.

Goretti B, Portaccio E, Zipoli V, Razzolini L, Amato MP. Coping strategies, cognitive impairment, psychological variables and their relationship with quality of life in multiple sclerosis. Neurol Sci 2010; 31 (Suppl 2): S227-30.

189

Gourraud PA, Harbo HF, Hauser SL, Baranzini SE. The genetics of multiple sclerosis: an up-to-date review. Immunol Rev 2012; 248: 87-103.

Grassiot B, Desgranges B, Eustache F, Defer G. Quantification and clinical relevance of brain atrophy in multiple sclerosis: a review. J Neurol 2009; 256: 1397–1412.

Grønning M, Hannisdal E, Mellgren SI. Multivariate analyses of factors associated with unemployment in people with multiple sclerosis. J Neurol Neurosurg Psychiatry 1990; 53: 388-90.

H

Hakiki B, Goretti B, Portaccio E, Zipoli V, Amato MP. ‘Subclinical MS’: follow-up of four cases. Eur J Neurol 2008;15: 858–861.

Hannoun S, Bagory M, Durand-Dubief F, et al. Correlation of Diffusion and Metabolic Alterations in Different Clinical Forms of Multiple Sclerosis. PLoS ONE 2012; 7: e32525. doi:10.1371/journal.pone.0032525

Harris JG, Wagner B, Cullum CM. Symbol vs. Digit Substitution Task Performance in Diverse Cultural and Linguistic Groups. Clin Neuropsychol 2007; 21: 800-810.

Harzheim M, Schlegel U, Urbach H, Klockgether T, Schmidt S. Discriminatory features of acute transverse myelitis: a retrospective analysis of 45 patients. J Neurol Sci 2004; 217: 217-23.

Hasan K.M, Walimuni IS, Abid H, Datta S, Wolinsky JS, Narayana PA. Human brain atlas-based multimodal MRI analysis of volumetry, diffusimetry, relaxometry and lesion distribution in multiple sclerosis patients and healthy adult controls: Implications for understanding the pathogenesis of multiple sclerosis and consolidation of quantitative MRI results in MS. J Neurol Sci 2012; 313: 99–109.

Heesen C, Kolbeck J, Gold SM, Schulz H, Schulz KH. Delivering the diagnosis of MS-results of a survey among patients and neurologists. Acta Neurol Scand 2003; 107: 363-8.

Henry RG, Shieh M, Okuda DT, et al. Regional grey matter atrophy in clinically isolated syndromes at presentation. J Neurol Neurosurg Psychiatry 2008; 79: 1236–44.

Hohol MJ, Guttmann CR, Orav J, et al. Serial neuropsychological assessment and magnetic resonance imaging analysis in multiple sclerosis. Arch Neurol 1997; 54: 1018-25.

Horakova D, Cox JL, Havrdova E, et al. Evolution of different MRI measures in patients with active relapsing-remitting multiple sclerosis over 2 and 5 years: a case-control study. J Neurol Neurosurg Psychiatry 2008;79: 407-14.

Horakova D, Dwyer MG, Havrdova E, et al. Gray matter atrophy and disability progression in patients with early relapsing-remitting multiple sclerosis: a 5-year longitudinal study. J Neurol Sci 2009; 282: 112–119.

190

Houtchens MK, Benedict RH, Killiany R, et al. Thalamic atrophy and cognition in multiple sclerosis. Neurology 2007; 69: 1213-23.

Hu B, Ye B, Yang Y, et al. Quantitative diffusion tensor deterministic and probabilistic fiber tractography in relapsing–remitting multiple sclerosis. Eur J Radiol 2011; 79: 101–107.

Huijbregts SC, Kalkers NF, de Sonneville LM, de Groot V, Reuling IE, Polman CH. Differences in cognitive impairment of relapsing remitting, secondary, and primary progressive MS. Neurology 2004; 63: 335-339.

Hulst HE, Geurts JJ. Gray matter imaging in multiple sclerosis: what have we learned? BMC Neurol 2011; 11: 153.

I

Iaffaldano P, Viterbo RG, Paolicelli D, et al. Impact of natalizumab on cognitive performances and fatigue in relapsing multiple sclerosis: a prospective, open-label, two years observational study. PLoS One 2012; 7: e35843.

Iannucci G, Tortorella C, Rovaris M, Sormani MP, Comi G, Filippi M. Prognostic value of MR and magnetization transfer imaging findings at presentation in patients with clinically isolated syndromes suggestive of multiple sclerosis at presentation. AJNR Am J Neuroradiol 2000; 21: 1034-1038.

Inglese M, Horsfield MA, Filippi M. Scan-rescan variation of measures derived from brain magnetization transfer ratio histograms obtained in healthy volunteers by use of a semi-interleaved magnetization transfer sequence. AJNR Am J Neuroradiol 2001; 22: 681-4.

J

Jacobs LD, Beck RW, Simon JH, et al. Intramuscular interferon beta-1a therapy initiated during a first demyelinating event in multiple sclerosis. CHAMPS Study Group. N Engl J Med 2000; 343: 898–904.

Jeffery DR, Mandler RN, Davis LE. Transverse myelitis. Retrospective analysis of 33 cases, with differentiation of cases associated with multiple sclerosis and parainfectious events. Arch Neurol 1993; 50: 532-5.

Jennekens-Schinkel A, Laboyrie PM, Lanser JB, van der Velde EA. Cognition in patients with multiple sclerosis after four years. J Neurol Sci 1990; 99: 229–247.

Julian LJ, Vella L, Vollmer T, Hadjimichael O, Mohr DC. Employment in multiple sclerosis. Exiting and re-entering the work force. J Neurol 2008; 255: 1354-60.

K

191

Kalkers NF, Vrenken H, Uitdehaag BM, Polman CH, Barkhof F. Brain atrophy in multiple sclerosis: impact of lesions and of damage of whole brain tissue. Mult Scler 2002; 8: 410-4.

Kantarci O, Siva A, Eraksoy M., et al. Survival and predictors of disability in Turkish MS patients. Turkish Multiple Sclerosis Study Group (TUMSSG). Neurology 1998; 51: 765-72.

Kappos L, Polman CH, Freedman MS, et al. Treatment with interferon beta-1b delays conversion to clinically definite and McDonald MS in patients with clinically isolated syndromes. Neurology 2006; 67: 1242–49.

Kappos L, Freedman MS, Polman CH, et al, for the BENEFIT Study Group. Effect of early versus delayed interferon beta-1b treatment on disability after a first clinical event suggestive of multiple sclerosis: a 3-year follow-up analysis of the BENEFIT study. Lancet 2007; 370: 389–97.

Kappos L, Freedman MS, Polman CH, et al, for the BENEFIT Study Group. Long-tem effect of early treatment with interferon beta-1b after a first clinical event suggestive of multiple sclerosis: a 5-year active treatment extension of the phase 3 BENEFIT trial. Lancet 2009; 8: 987-97.

Kern K.C., Sarcona J., Montag M., Giesser B.S., Sicotte N.L. Corpus callosal diffusivity predicts motor impairment in relapsing–remitting multiple sclerosis: A TBSS and tractography study. NeuroImage 2011; 55: 1169–1177.

Khoury SJ, Guttmann CR, Orav EJ, et al. Longitudinal MRI in multiple sclerosis: correlation between disability and lesion burden. Neurology 1994; 44: 2120-4.

Kornblith AB, La Rocca NG, Baum HM. Employment in individuals with multiple sclerosis. Int J Rehabil Res 1986; 9: 155-65.

Korteweg T, Tintoré M, Uitdehaag B, et al. MRI criteria for dissemination in space in patients with clinically isolated syndromes: a multicentre follow-up study. Lancet Neurol 2006; 5:221-7.

Korteweg T, Tintore M, Uitdehaag BM, et al. A search for new MRI criteria for dissemination in space in subjects with a clinically isolated syndrome. Eur Radiol 2009; 19: 2244-8.

Kujala P, Portin R, Ruutiainen J. The progress of cognitive decline in multiple sclerosis A controlled 3-year follow-up. Brain 1997; 120: 289–297.

Kürtzke JF, Beebe GW, Nagler B, Kurland LT, Auth TL. Studies on the natural history of multiple sclerosis: Early prognostic features of the later course of the illness. J Chronic Dis 1977; 30: 819–830.

Kürtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444-52.

Kutzelnigg A, Lassmann H Cortical lesions and brain atrophy in MS. J Neurol Sci 2005; 233: 55–59.

192

L

Langdon DW. Cognition in multiple sclerosis. Curr Opin Neurol 2011; 24: 244–249.

Langdon DW, Amato MP, Boringa J, et al. Recommendations for a Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS). Mult Scler 2012; 18: 891-8.

Langer-Gould A, Popat RA, Huang SM, et al. Clinical and demographic predictors of long-term disability in patients with relapsing-remitting multiple sclerosis. Arch Neurol 2006; 63: 1686-1691.

LaRocca N, Kalb R, Kendall P, Scheinberg L. The role of disease and demographic factors in the employment of patients with multiple sclerosis. Arch Neurol 1982; 39: 256.

LaRocca N, Kalb R, Scheinberg L, Kendall P. Factors associated with unemployment of patients with multiple sclerosis. J Chronic Dis 1985; 38: 203-10.

Lassmann H, Brück W, Lucchinetti CF. The immunopathology of multiple sclerosis: an overview. Brain Pathol 2007; 17: 210-8.

Lebrun C, Bensa C, Debouverie M, et al. Association between clinical conversion to multiple sclerosis in radiologically isolated syndrome and magnetic resonance imaging, cerebrospinal fluid, and visual evoked potential: follow-up of 70 patients. Arch Neurol 2009; 66: 841-6.

Lebrun C, Blanc F, Brassat D, Zephir H, de Seze J; CFSEP. Cognitive function in radiologically isolated syndrome. Mult Scler 2010; 16: 919-25.

Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet 2004; 364: 2106-12.

Leray E, Yaouanq J, Le Page E., et al Evidence for a two-stage disability progression in multiple sclerosis. Brain 2010: 133; 1900–1913.

Lin X, Tench CR, Morgan PS, et al. ‘Importance sampling’ in MS: use of diffusion tensor tractography to quantify pathology related to specific impairment. J Neurol Sci 2005; 237: 13–19.

Lin X, Tench CR, Morgan PS, Constantinescu CS. Use of combined conventional and quantitative MRI to quantify pathology related to cognitive impairment in multiple sclerosis. J Neurol Neurosurg Psychiatry 2008; 79: 437-441.

Liu Y, Duan Y, He Y, et al. Whole brain white matter changes revealed by multiple diffusion metrics in multiple sclerosis: A TBSS study. Eur Journ Radiol 2012; 81: 2826–2832.

Llufriu S, Blanco Y, Martinez-Heras E, et al. Influence of Corpus Callosum Damage on Cognition and Physical Disability in Multiple Sclerosis: A Multimodal Study. PLoS ONE 2012; 7: e37167. doi:10.1371/journal.pone.0037167

193

Lublin FD, Reingold SC. Defining the clinical course of multiple sclerosis: results of an international survey. National Multiple Sclerosis Society (USA) Advisory Committee on Clinical Trials of New Agents in Multiple Sclerosis. Neurology 1996; 46: 907-11.

Lucchinetti CF, Popescu BF, Bunyan RF, et al. Inflammatory cortical demyelination in early multiple sclerosis. N Engl J Med 2011; 365: 2188-97.

Lukas C, Minneboo A, de Groot V, et al. Early central atrophy rate predicts 5 year clinical outcome in multiple sclerosis. J Neurol Neurosurg Psychiatry 2010; 81: 1351-6.

Lynch SG, Parmenter BA, Denney DR. The association between cognitive impairment and physical disability in multiple sclerosis. Mult Scler 2005; 11: 469-76.

Lyon-Caen O, Jouvent R, Hauser S, et al. Cognitive function in recent-onset demyelinating diseases. Arch Neurol 1986; 43: 1138-41.

M

Magliozzi R, Howell O, Vora A, et al. Meningeal B-cell follicles in secondary progressive multiple sclerosis associate with early onset of disease and severe cortical pathology. Brain 2007; 130: 1089–104.

Masjuan J, Alvarez-Cermeño JC, García-Barragán N, et al. Clinically isolated syndromes: a new oligoclonal band test accurately predicts conversion to MS. Neurology 2006; 66: 576-8.

McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001; 50: 121–27.

Mesaros S, Rocca MA, Sormani MP, Charil A, Comi G, Filippi M. Clinical and conventional MRI predictors of disability and brain atrophy accumulation in RRMS. A large scale, short-term follow-up study. J Neurol 2008; 255: 1378-83.

Mesaros S, Rocca MA, Kacar K, et al. Diffusion tensor MRI tractography and cognitive impairment in multiple sclerosis. Neurology 2012; 78: 969–975.

Miller DH, Hornabrook RW, Purdie G. The natural history of multiple sclerosis: a regional study with some longitudinal data. J Neurol Neurosurg Psychiatry 1992; 55: 341-6.

Miller DH, Weinshenker BG, Filippi M, et al. Differential diagnosis of suspected multiple sclerosis: a consensus approach. Mult Scler 2008; 14: 1157-74.

Miller DH, Chard DT, Ciccarelli O. Clinically isolated syndromes. Lancet Neurol 2012; 11: 157-69.

Minneboo A, Barkhof F, Polman CH, et al. Infratentorial lesions predict long-term disability in patients with initial findings suggestive of multiple sclerosis. Arch Neurol 2004; 61: 217-221.

194

Mitchell AJ, Benito-León J, Morales González JM, Rivera-Navarro J. Quality of life and its assessment in multiple sclerosis: integrating physical and psychological components of wellbeing. Lancet Neurol 2005; 4, 556-66.

Montalban X, Tintoré M, Swanton J, et al. MRI criteria for MS in patients with clinically isolated syndromes. Neurology 2010; 74: 427-34.

Moodie J, Healy BC, Buckle GJ, et al. Magnetic resonance disease severity scale (MRDSS) for patients with multiple sclerosis: A longitudinal study. J Neurol Sci 2012; 315: 49–54.

Moraal B, Pohl C, Uitdehaag BM, et al. Magnetic resonance imaging predictors of conversion to multiple sclerosis in the BENEFIT study. Arch Neurol 2009; 66: 1345-52.

Morgen K, Sammer G, Courtney SM, et al. Evidence for a direct association between cortical atrophy and cognitive impairment in relapsing-remitting MS. Neuroimage 2006; 30: 891-898.

Morrow SA, Drake A, Zivadinov R, Munschauer F, Weinstock-Guttman B, Benedict RH. Predicting loss of employment over three years in multiple sclerosis: clinically meaningful cognitive decline. Clin Neuropsychol 2010; 24, 1131-45.

Myhr KM, Riise T, Vedeler C, et al. Disability and prognosis in multiple sclerosis: demographic and clinical variables important for the ability to walk and awarding of disability pension. Mult Scler 2001; 7: 59-65.

N

Nocentini U, Pasqualetti P, Bonavita S, et al. Cognitive dysfunction in patients with relapsing-remitting multiple sclerosis. Mult Scler 2006; 12: 77-87.

O

O'Brien AR, Chiaravalloti N, Goverover Y, Deluca J. Evidenced-based cognitive rehabilitation for persons with multiple sclerosis: a review of the literature. Arch Phys Med Rehabil 2008; 89: 761-9.

O'Connor P; Canadian Multiple Sclerosis Working Group. Key issues in the diagnosis and treatment of multiple sclerosis. An overview. Neurology 2002; 59: S1-33.

Okuda DT, Mowry EM, Beheshtian A, et al. Incidental MRI anomalies suggestive of multiple sclerosis: the radiologically isolated syndrome. Neurology 2009; 72: 800-5.

Okuda DT, Mowry EM, Cree BA, et al. Asymptomatic spinal cord lesions predict disease progression in radiologically isolated syndrome. Neurology 2011; 76: 686–692.

Optic Neuritis Study Group. Visual function five years after optic neuritis: experience of the Optic Neuritis Treatment Trial. Arch Ophthalmol 1997; 115: 1545–1552.

195

Optic Neuritis Study Group. High and Low Risk Profiles for the Development of Multiple Sclerosis Within Ten Years After Optic Neuritis. Experience of the Optic Neuritis Treatment Trial. Arch Ophthalmol 2003; 121: 944–949.

Optic Neuritis Study Group. Multiple sclerosis risk after optic neuritis: final optic neuritis treatment trial follow-up. Arch Neurol 2008; 65: 727-32.

Oreja-Guevara C, Charil A, Caputo D, Cavarretta R, Sormani M.P, Filippi M. Magnetization Transfer Magnetic Resonance Imaging and Clinical Changes in Patients With Relapsing-Remitting Multiple Sclerosis. Arch Neurol 2006; 63: 736-740.

P

Paolillo A, Pozzilli C, Giugni E, et al. A 6-year clinical and MRI follow-up study of patients with relapsing-remitting multiple sclerosis treated with Interferon-beta. Eur J Neurol. 2002; 9: 645-55.

Parmenter BA, Weinstock-Guttman B, Garg N, Munschauer F, Benedict RH. Screening for cognitive impairment in multiple sclerosis using the Symbol digit Modalities Test. Mult Scler 2007; 13: 52-7.

Patrucco L, Rojas JI, Cristiano E. Assessing the value of spinal cord lesions in predicting development of multiple sclerosis in patients with clinically isolated syndromes. J Neurol 2012; 259: 1317-20.

Patti F, Amato M, Trojano M, et al. Cognitive impairment and its relation with disease measures in mildly disabled patients with relapsing–remitting multiple sclerosis: baseline results from the Cognitive Impairment in Multiple Sclerosis (COGIMUS) study. Mult Scler 2009; 15: 779–788.

Patti F, Amato MP, Trojano M, et al. Quality of life, depression and fatigue in mildly disabled patients with relapsing–remitting multiple sclerosis receiving subcutaneous interferon beta-1a: 3-year results from the COGIMUS (COGnitive Impairment in MUltiple Sclerosis) study. Mult Scler Journal 2011; 17: 991–1001.

Paty DW, Oger JJ, Kastrukoff LF, et al. MRI in the diagnosis of MS: a prospective study with comparison of clinical evaluation, evoked potentials, oligoclonal banding, and CT. Neurology 1988; 38: 180-185.

Pelletier J, Suchet L, Witjas T, et al. A longitudinal study of callosal atrophy and interhemispheric dysfunction in relapsing-remitting multiple sclerosis. Arch Neurol 2001; 58: 105-11.

Pelosi L, Geesken JM, Holly M, Hayward M, Blumhardt LD. Working memory impairment in early multiple sclerosis. Evidence from an event-related potential study of patients with clinically isolated myelopathy. Brain 1997; 120: 2039-58.

Penner IK, Kappos L, Rausch M, Opwis K, Radü EW. Therapy-induced plasticity of cognitive functions in MS patients: insights from fMRI. J Physiol Paris 2006; 99: 455-62.

196

Penner IK, Bechtel N, Raselli C, et al. Fatigue in multiple sclerosis: relation to depression, physical impairment, personality and action control. Mult Scler 2007; 13: 1161–1167.

Penner IK, Stemper B, Calabrese P, et al. Effects of interferon beta-1b on cognitive performance in patients with a first event suggestive of multiple sclerosis. Mult Scler 2012; 4 [Epub ahead of print].

Penny S, Khaleeli Z, Cipolotti L, Thompson A, Ron M. Early imaging predicts later cognitive impairment in primary progressive multiple sclerosis. Neurology 2010; 74: 545-552.

Peyser JM, Edwards KR, Poser CM, Filskov SB. Cognitive function in patients with multiple sclerosis. Arch Neurol 1980; 37: 577-9.

Phadke JG. Clinical aspects of multiple sclerosis in north-east Scotland with particular references to its course and prognosis. Brain 1990; 113: 1597–628.

Pirko I, Lucchinetti CF, Sriram S, Bakshi R. Gray matter involvement in multiple sclerosis. Neurology 2007; 68: 634–64.

Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald Criteria”. Ann Neurol 2005; 58: 840–46.

Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: 2010 Revisions to the McDonald Criteria. Ann Neurol 2011; 69: 292-302.

Pompeii LA, Moon SD, McCrory DC. Measures of physical and cognitive function and work status among individuals with multiple sclerosis: a review of the literature. J Occup Rehabil 2005; 15: 69-84.

Portaccio E, Stromillo ML, Goretti B, et al. Neuropsychological and MRI measures predict short-term evolution in benign multiple sclerosis. Neurology 2009; 73: 498-503.

Portaccio E, Amato MP. Natural history In Amato MP (Elsevier) Cognitive Impairment in Multiple Sclerosis, Milano, 2011: pp 29-36.

Poser S, Bauer HJ, Poser W. Prognosis of multiple sclerosis. Results from an epidemiological area in Germany. Acta Neurol Scand 1982; 65: 347-54.

Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research proposals. Ann Neurol 1983; 13: 227–231.

Prakash RS, Snook EM, Lewis JM, Motl RW, Kramer AF. Cognitive impairments in relapsing–remitting multiple sclerosis: a meta-analysis. Mult Scler 2008; 14: 1250–1261.

Preziosa P, Rocca MA, Mesaros S, et al. Intrinsic damage to the major white matter tracts in patients with different clinical phenotypes of multiple sclerosis: a voxelwise diffusion-tensor MR study. Radiology 2011; 260: 541–550.

R

197

Ramasamy DP, Benedict RHB, Cox JL, et al. Extent of cerebellum, subcortical and cortical atrophy in patients with MS: a case-control study. J Neurol Sci 2009; 282: 47–54.

Rao SM, the Cognitive Function Study Group of the National Multiple Sclerosis Society. A manual for the Brief Repeatable Battery of Neuropsychological Tests in multiple sclerosis. Milwaukee, WI: Medical College of Wisconsin; 1990.

Rao SM, Leo GJ, Bernardin L, Unverzagt F. Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction. Neurology 1991a; 41: 685-691.

Rao SM, Leo GJ, Ellington L, Nauertz T, Bernardin L, Unverzagt F Cognitive dysfunction in multiple sclerosis. II. Impact on employment and social functioning. Neurology 1991b; 41, 692-6.

Reuter F, Zaaraoui W, Crespy L, et al. Frequency of cognitive impairment dramatically increases during the first 5 years of multiple sclerosis. J Neurol Neurosurg Psychiatry. 2011; 82: 1157-9.

Riccitelli G, Rocca MA, Pagani E, et al. Cognitive impairment in multiple sclerosis is associated to different patterns of gray matter atrophy according to clinical phenotype. Hum Brain Mapp 2011; 32: 1535–1543.

Riise T, Grønning M, Fernández O, et al. Early prognostic factors for disability in multiple sclerosis, a European multicenter study. Acta Neurol Scand 1992; 85: 212-8.

Rinaldi F, Calabrese M, Grossi P, Puthenparampil M, Perini P, Gallo P. Cortical lesions and cognitive impairment in multiple sclerosis. Neurol Sci 2010; 31: S235-7.

Rocca MA, Mastronardo G, Rodegher M, Comi G, Filippi M. Long-term changes of magnetization transfer-derived measures from patients with relapsing-remitting and secondary-progressive multiple sclerosis. AJNR Am J Neuroradiol 1999; 20: 821-827.

Rocca MA, Falini A, Colombo B, Scotti G, Comi G, Filippi M. Adaptive functional changes in the cerebral cortex of patients with nondisabling multiple sclerosis correlate with the extent of brain structural damage. Ann Neurol 2002; 51: 330–339.

Rocca MA, Colombo B, Falini A, et al. Cortical adaptation in patients with MS: a cross-sectional functional MRI study of disease phenotypes. Lancet Neurol 2005; 4: 618-26.

Rocca MA, Mesaros S, Pagani E, et al. Thalamic damage and long-term progression of disability in multiple sclerosis. Radiology 2010; 257: 463–469.

Roosendaal SD, Moraal B, Pouwels PJ, et al. Accumulation of cortical lesions in MS: relation with cognitive impairment. Mult Scler 2009; 15: 708–14.

Roosendaal SD, Bendfeldt K, Vrenken H, et al. Grey matter volume in a large cohort of MS patients: relation to MRI parameters and disability. Mult Scler 2011; 17: 1098–1106.

Rosti E, Hämäläinen P, Koivisto K, Hokkanen L. One-year follow-up study of relapsing-remitting MS patients' cognitive performances: Paced Auditory Serial Addition Test's susceptibility to change. J Int Neuropsychol Soc 2007; 13: 791-8.

198

Rosti-Otajärvi EM, Hämäläinen PI. Neuropsychological rehabilitation for multiple sclerosis. Cochrane Database Syst Rev 2011; 9: CD009131.

Rovaris M, Iannucci G, Falautano M, at al. Cognitive dysfunction in patients with mildly disabling relapsing-remitting multiple sclerosis: an exploratory study with diffusion tensor MR imaging. J Neurol Sci 2002; 195: 103-109.

Rovaris M, Agosta F, Sormani MP, et al. Conventional and magnetization transfer MRI predictors of clinical multiple sclerosis evolution: a medium-term follow-up study. Brain 2003; 126: 2323-2332.

Rovaris M, Comi G, Filippi M. MRI markers of destructive pathology in multiple sclerosis-related cognitive dysfunction. J Neurol Sci 2006; 245: 111–116.

Rovaris M, Barkhof F, Calabrese M, et al. MRI features of benign multiple sclerosis: toward a new definition of this disease phenotype. Neurology 2009; 72: 1693-701.

Rovira A, Swanton J, Tintoré M, et al. A single, early magnetic resonance imaging study in the diagnosis of multiple sclerosis. Arch Neurol 2009; 66: 587-92.

Rudick RA, Fisher E, Lee JC, Simon J, Jacobs L. Use of the brain parenchymal fraction to measure whole brain atrophy in relapsing-remitting MS. Multiple Sclerosis Collaborative Research Group. Neurology 1999; 53: 1698-1704.

Rudick RA, Cutter G, Baier M, Fisher E, Dougherty D, Weinstock-Guttman B, Mass MK, Miller D, Simonian NA. Use of the multiple sclerosis functional composite to predict disability in relapsing MS. Neurology 2001; 56: 1324–1330.

Ruet A, Deloire MS, Ouallet JC, et al. Predictive factors for multiple sclerosis in patients with clinically isolated spinal cord syndrome. Mult Scler 2011a; 17: 312-318.

Ruet A, Charré-Morin J, Brochet B, Hamel D, Deloire M. How to detect cognitive impairment in relapsing-remitting and primary-progressive multiple sclerosis. Mult Scler 2011b; 17: S277-S505.

Runmarker B, Andersen O. Prognostic factors in a multiple sclerosis incidence cohort with twenty-five years of follow-up. Brain 1993; 116: 117-34.

S

Santos AC, Narayanan S, de Stefano N, et al. Magnetization transfer can predict clinical evolution in patients with multiple sclerosis. J Neurol 2002; 249: 662-668.

Sastre-Garriga J, Alonso J, Renom M, et al. A functional magnetic resonance proof of concept pilot trial of cognitive rehabilitation in multiple sclerosis. Mult Scler 2011; 17: 457-67.

Sbardella E, Tomassini V, Stromillo ML, et al. Pronounced focal and diffuse brain damage predicts short-term disease evolution in patients with clinically isolated syndrome suggestive of multiple sclerosis. Mult Scler 2011; 17: 1432-40.

199

Scalfari A, Neuhaus A, Degenhardt A, et al. The natural history of multiple sclerosis, a geographically based study 10: relapses and long-term disability. Brain 2010; 133: 1914-29.

Schmahmann JD. Disorders of the Cerebellum: Ataxia, Dysmetria of Thought, and the Cerebellar Cognitive Affective Syndrome. J Neuropsychiatry Clin Neurosci Summer 2004; 16: 367–378.

Schmahmann JD. The role of the cerebellum in cognition and emotion: personal reflections since 1982 on the dysmetria of thought hypothesis, and its historical evolution from theory to therapy. Neuropsychol Rev 2010; 20: 236-60.

Schmierer K, Scaravilli F, Altmann DR, Barker GJ, Miller DH. Magnetization transfer ratio and myelin in postmortem multiple sclerosis brain. Ann Neurol 2004; 56: 407-415.

Schumacher GA, Beebe G, Kibler RF, et al. Problems of experimental trials of therapy in multiple sclerosis report by the panel on the evaluation of experimental trials of therapy in multiple sclerosis. Ann N Y Acad Sci 1965; 122: 552-68.

Scott TF, Bhagavatula K, Snyder PJ, Chieffe C. Transverse myelitis. Comparison with spinal cord presentations of multiple sclerosis. Neurology 1998; 50: 429-33.

Scott TF, Kassab SL, Singh S. Acute partial transverse myelitis with normal cerebral magnetic resonance imaging: transition rate to clinically definite multiple sclerosis. Mult Scler 2005; 11: 373-7.

Scott TF. Nosology of idiopathic transverse myelitis syndromes. Acta Neurol Scand 2007; 115: 371-6.

Sellner J, Lüthi N, Bühler R, et al. Acute partial transverse myelitis: risk factors for conversion to multiple sclerosis. Eur J Neurol 2008; 15: 398-405.

Serafini B, Rosicarelli B, Magliozzi R, Stigliano E, Aloisi F. Detection of ectopic B-cell follicles with germinal centers in the meninges of patients with secondary progressive multiple sclerosis. Brain Pathol 2004; 14: 164-74.

Shiee N, Bazin P-L, Zackowski KM, Farrell SK, Harrison DM, et al. Revisiting Brain Atrophy and Its Relationship to Disability in Multiple Sclerosis. PLoS ONE 2012; 7: e37049. doi:10.1371/journal.pone.0037049.

Shu N, Liu Y, Li K, et al. Diffusion tensor tractography reveals disrupted topological efficiency in white matter structural networks in multiple sclerosis. Cereb Cortex 2011; 21: 2565–2577.

Sicotte NL, Kern KC, Giesser BS, et al. Regional hippocampal atrophy in multiple sclerosis. Brain 2008; 131: 1134–1141.

Simeoni M, Auquier P, Fernandez O, et al. Validation of the Multiple Sclerosis International Quality of Life questionnaire. Mult Scler 2008; 14: 219–230.

200

Simmons RD, Tribe KL, McDonald EA. Living with multiple sclerosis: longitudinal changes in employment and the importance of symptom management. J Neurol 2010; 257: 926-36.

Simon JH, Jacobs LD, Campion MK, et al. for the Multiple Sclerosis Collaborative Research Group (MSCRG). A longitudinal study of brain atrophy in relapsing multiple sclerosis. Neurology 1999; 53: 139-148.

Smith A. Symbol Digit Modalities Test (SDMT) Manual (revised). Los Angeles: Western Psychological Services; 1982.

Smith MM, Arnett PA. Factors related to employment status changes in individuals with multiple sclerosis. Mult Scler 2005; 11: 602-9.

Smith SM, Jenkinson M, Johansen-Berg H, et al. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. NeuroImage 2006; 31: 1487–1505.

Sormani MP, Iannucci G, Rocca MA, et al. Reproducibility of magnetization transfer ratio histogram-derived measures of the brain in healthy volunteers. AJNR Am J Neuroradiol 2000; 21: 133-6.

Srivastava R, Aslam M, Kalluri SR, et al. Potassium Channel KIR4.1 as an Immune Target in Multiple Sclerosis. N Engl J Med 2012; 367: 115-23.

Staffen W, Zauner H, Mair A, et al. Magnetic resonance spectroscopy of memory and frontal brain region in early multiple sclerosis. J Neuropsychiatry Clin Neurosci 2005; 17: 357-63.

Stoodley CJ, Schmahmann JD. Evidence for topographic organization in the cerebellum of motor control versus cognitive and affective processing. Cortex 2010; 46: 831-44.

Strober L, Englert J, Munschauer F, Weinstock-Guttman B, Rao S, Benedict RH. Sensitivity of conventional memory tests in multiple sclerosis: comparing the Rao Brief Repeatable Neuropsychological Battery and the Minimal Assessment of Cognitive Function in MS. Mult Scler 2009; 15: 1077-84.

Strober LB, Christodoulou C, Benedict RH, et al. Unemployment in multiple sclerosis: the contribution of personality and disease. Mult Scler 2012; 18: 647-53.

Summers M, Fisniku L, Anderson V, Miller D, Cipolotti L, Ron M. Cognitive impairment in relapsing-remitting multiple sclerosis can be predicted by imaging performed several years earlier. Mult Scler 2008;14: 197-204.

Sumowski JF, Chiaravalloti N, DeLuca J. Cognitive reserve protects against cognitive dysfunction in multiple sclerosis. J Clin Exp Neuropsychol 2009; 31: 913-26.

Sumowski JF, Wylie GR, Deluca J, Chiaravalloti N. Intellectual enrichment is linked to cerebral efficiency in multiple sclerosis: functional magnetic resonance imaging evidence for cognitive reserve. Brain 2010; 133: 362-74.

201

Swanton JK, Fernando KT, Dalton CM, et al. Modification of MRI criteria for multiple sclerosis in patients with clinically isolated syndromes. J Neurol Neurosurg Psychiatry 2006; 77: 830–33.

Swanton JK, Rovira A, Tintoré M, et al. MRI criteria for multiple sclerosis in patients presenting with clinically isolated syndromes: a multicentre retrospective study. Lancet Neurol 2007; 6: 677-86.

Swanton JK, Fernando KT, Dalton CM, et al. Early MRI in optic neuritis: the risk for disability. Neurology 2009; 72: 542-50.

Swanton JK, Fernando KT, Dalton CM, et al. Early MRI in optic neuritis: the risk for clinically definite multiple sclerosis. Mult Scler 2010; 16: 156-165.

T

Tiberio M, Chard DT, Altmann DR, et al. Gray and white matter volume changes in early RRMS: a 2-year longitudinal study. Neurology 2005; 64: 1001–1007.

Tiemann L, Penner IK, Haupts M, Schlegel U, Calabrese P. Cognitive decline in multiple sclerosis: impact of topographic lesion distribution on differential cognitive deficit patterns. Mult Scler 2009;15: 1164-74.

Tintoré M, Rovira A, Martinez M, et al. Isolated demyelinating syndromes: comparison of different MR imaging criteria to predict conversion to clinically definite multiple sclerosis. Am J Neuroradiol 2000; 21: 702–06.

Tintoré M, Rovira A, Rio J, et al. Is optic neuritis more benign than other first attacks in multiple sclerosis? Ann Neurol 2005; 57: 210-5.

Tintoré M, Rovira A, Rio J, et al. Baseline MRI predicts future attacks and disability in clinically isolated syndromes. Neurology 2006; 67: 968–72.

Tintoré M, Rovira A, Río J, et al. Do oligoclonal bands add information to MRI in first attacks of multiple sclerosis? Neurology 2008; 70:1079-83.

Tintoré M, Rovira A, Arrambide G, et al. Brainstem lesions in clinically isolated syndromes. Neurology 2010; 75: 1933–1938.

Transverse Myelitis Consortium Working Group. Proposed diagnostic criteria and nosology of acute transverse myelitis. Neurology 2002; 59: 499-505.

Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mork S, Bo L. Axonal transection in the lesions of multiple sclerosis. N Engl J Med 1998; 338: 278-285.

Trapp BD, Nave KA. Multiple sclerosis: an immune or neurodegenerative disorder? Annu Rev Neurosci 2008; 31: 247-69.

Trojano M, Liguori M, Bosco Zimatore G, et al. Age-related disability in multiple sclerosis. Ann Neurol 2002; 51: 475-80.

202

Truelle JL, Palisson E, Le Gall D, Stip E, Derouesne C. Intellectual and mood disorders in multiple sclerosis. Rev Neurol (Paris) 1987; 143: 595-601.

Truyen L, van Waesberghe JH, van Walderveen MA, et al. Accumulation of hypointense lesions ("black holes") on T1 spin-echo MRI correlates with disease progression in multiple sclerosis. Neurology 1996; 47: 1469-76.

V

van Buchem MA, McGowan JC, Kolson DL, Polansky M, Grossman RI. Quantitative volumetric magnetization transfer analysis in multiple sclerosis: estimation of macroscopic and microscopic disease burden. Magn Reson Med 1996; 36: 632–636.

van den Burg W, van Zomeren AH, Minderhoud JM, Prange AJA, Meijer NSA. Cognitive impairment in patients with multiple sclerosis and mild physical disability. Arch Neurol 1987; 44: 494-501.

Vaneckova M, Kalincik T, Krasensky J, et al. Corpus Callosum Atrophy – A Simple Predictor of Multiple Sclerosis Progression: A Longitudinal 9-Year Study. Eur Neurol 2012; 68: 23–27.

van Waesberghe JH, Kamphorst W, De Groot CJ, et al. Axonal loss in multiple sclerosis lesions: magnetic resonance imaging insights into substrates of disability. Ann Neurol 1999; 46: 747-754.

van Walderveen MA, Barkhof F, Hommes OR, et al. Correlating MRI and clinical disease activity in multiple sclerosis: relevance of hypointense lesions on short-TR/short-TE (T1-weighted) spin-echo images. Neurology 1995; 45: 1684-90.

Van Walderveen MA, Lycklama A Nijeholt GJ, et al. Hypointense lesions on T 1 - weighted spin-echo magnetic resonance imaging: relation to clinical characteristics in subgroups of patients with multiple sclerosis. Arch Neurol 2001; 58: 76–81.

Vernay D, Gerbaud L, Biolay S, et al. Quality of life and multiple sclerosis: validation of the French version of the self-questionnaire (SEP-59) Rev Neurol (Paris) 2000; 156, 247-263.

Vickrey BG. Multiple Sclerosis Quality of life (MSQOL)-54 Instrument. University of California, Los Angeles 1995.

Vickrey BG, Hays RD, Genovese BJ, Myers LW, Ellison GW. Comparison of a generic to disease-targeted health-related quality-of-life measures for multiple sclerosis. J Clin Epidemiol 1997; 50: 557-69.

Villar LM, García-Barragán N, Sádaba MC, et al. Accuracy of CSF and MRI criteria for dissemination in space in the diagnosis of multiple sclerosis. J Neurol Sci 2008; 266: 34-7.

203

Visscher BR, Liu KS, Clark VA, Detels R, Malmgren RM, Dudley JP. Onset symptoms as predictors of mortality and disability in multiple sclerosis. Acta Neurol Scand 1984; 70: 321-8.

W

Wachowius U, Talley M, Silver N, Heinze HJ, Sailer M. Cognitive impairment in primary and secondary progressive multiple sclerosis. J Clin Exp Neuropsychol 2005; 27: 65–77.

Wang JL, Reimer MA, Metz LM, Patten SB. Major depression and quality of life in individuals with multiple sclerosis. Int J Psychiatry Med 2000; 30: 309-17.

Weinshenker BG, Bass B, Rice GP, et al. The natural history of multiple sclerosis: a geographically based study. I. Clinical course and disability. Brain 1989a; 112: 133–46.

Weinshenker BG, Bass B, Rice GP, et al. The natural history of multiple sclerosis: a geographically based study. 2. Predictive value of the early clinical course. Brain 1989b; 112: 1419–28.

Weinshenker BG, Rice GP, Noseworthy JH, Carriere W, Baskerville J, Ebers GC. The natural history of multiple sclerosis: a geographically based study. 3. Multivariate analysis of predictive factors and models of outcome. Brain 1991; 114: 1045-1056.

Wingerchuk DM, Lennon VA, Pittock SJ, Lucchinetti CF, Weinshenker BG. Revised diagnostic criteria for neuromyelitis optica. Neurology 2006; 66: 1485–89.

Wybrecht D, Reuter F, Zaaraoui W, et al. Voxelwise analysis of conventional magnetic resonance imaging to predict future disability in early relapsing-remitting multiple sclerosis. Mult Scler 2012; 28. [Epub ahead of print].

Y

Young IR, Hall AS, Pallis CA, Legg NJ, Bydder GM, Steiner RE. Nuclear magnetic resonance imaging of the brain in multiple sclerosis. Lancet 1981; 14: 1063-6.

Yu C.S., Lin F.C., Liu Y., Duan Y., Lei H., Li K.C. Histogram analysis of diffusion measures in clinically isolated syndromes and relapsing-remitting multiple sclerosis. Eur J Radiol 2008; 68: 328–334.

Z

Zaaraoui W, Deloire M, Merle M, et al. Monitoring demyelination and remyelination by magnetization transfer imaging in the mouse brain at 9.4 T. MAGMA 2008; 21: 357-62.

Zaaraoui W, Konstandin S, Audoin B, et al. Distribution of Brain Sodium Accumulation Correlates with Disability in Multiple Sclerosis: A Cross-sectional 23Na MR Imaging Study. Radiology 2012; 264: 859-67.

204

Zakzanis KK. Distinct neurocognitive profiles in multiple sclerosis subtypes. Arch Clin Neuropsychol 2000; 15: 115-36.

Zipoli V, Hakiki B, Portaccio E, et al. The contribution of cerebrospinal fluid oligoclonal bands to the early diagnosis of multiple sclerosis. Mult Scler 2009; 15: 472-8.

Zipoli V, Goretti B, Hakiki B, et al. Cognitive impairment predicts conversion to multiple sclerosis in clinically isolated syndromes. Mult Scler 2010; 16: 62-7.

Zivadinov R, Sepcic J, Nasuelli D et al. A longitudinal study of brain atrophy and cognitive disturbances in the early phase of relapsing-remitting multiple sclerosis. J Neurol Neurosurg Psychiatry 2001; 70: 773-780.

Zivadinov R, Leist TP. Clinical–MRI correlations in multiple sclerosis. J Neuroimaging 2005; 15: 10S–21S.