Multiple Sclerosis and Related Disorders 48 (2021) 102673 Contents lists available at ScienceDirect Multiple Sclerosis and Related Disorders journal homepage: www.elsevier.com/locate/msard Original article Effect of Ozanimod on Symbol Digit Modalities Test Performance in Relapsing MS John DeLuca a, *, Sven Schippling b, Xavier Montalban c, Ludwig Kappos d, Bruce A.C. Cree e, Giancarlo Comi f, Douglas L. Arnold g, Hans-Peter Hartung h, James K. Sheffield i, Hongjuan Liu i, Diego Silva i, Jeffrey A. Cohen j a Kessler Foundation, 1199 Pleasant Valley Way, West Orange, NJ 07052 USA and Departments of Physical Medicine and Rehabilitation, and Neurology, Rutgers - New Jersey Medical School, Newark 07103, NJ, USA b Neuroimmunology and Multiple Sclerosis Research, Department of Neurology, University Hospital and University of Zürich and Neuroscience Center Zürich, University of Zürich, Frauenklinikstrasse 26, 8091 Zurich, Switzerland and Federal Institute of Technology (ETH) Zürich, Rämistrasse 101, 8092 Zürich, Switzerland c Department of Neurology-Neuroimmunology, Centre d’Esclerosi Múltiple de Catalunya (Cemcat), Hospital Universitario Vall d’Hebron, Pg. Vall d’Hebron, 119-129, 08035 Barcelona, Spain d Research Center for Clinical Neuroimmunology and Neuroscience Basel (RC2NB), Departments of Medicine, Clinical Research, Biomedicine, and Biomedical Engineering, University Hospital and University of Basel, Petersgraben 4, CH-4031 Basel, Switzerland e Weill Institute for Neurosciences, Department of Neurology, University of California San Francisco, 675 Nelson Rising Lane, San Francisco, CA 94158 USA f Department of Neurology, San Raffaele Scientific Institute, Vita-Salute San Raffaele University, via Olgettina 48, 20132 Milan, Italy g NeuroRx Research and Montréal Neurological Institute, McGill University, 3801 University Street, Montreal, QC H3A 2B4, Canada h Department of Neurology, Medical Faculty, Heinrich-Heine University, University Hospital Dusseldorf, Moorenstr. 5 40225 Dusseldorf, Germany i Bristol Myers Squibb, 3401 Princeton Pike, Lawrenceville, NJ 08648, USA j Department of Neurology, Mellen Center for MS Treatment and Research, Neurological Institute, Cleveland Clinic, 9500 Euclid Avenue, Cleveland, OH 44195 USA A R T I C L E I N F O A B S T R A C T Keywords: Background: Cognitive dysfunction, including slowed cognitive processing speed (CPS), is one of the most Multiple sclerosis disabling symptoms of multiple sclerosis (MS). The Symbol Digit Modalities Test (SDMT) is a preferred measure Cognition of CPS for MS trials and routine screening. Based on encouraging SDMT results in the phase 3 SUNBEAM trial, Neuropsychological tests these post hoc, exploratory analyses were conducted to further compare effects of the sphingosine 1-phosphate Ozanimod Brain volume receptor modulator ozanimod versus intramuscular interferon β-1a on CPS in participants with relapsing mul­ tiple sclerosis (RMS). Methods: In the phase 3, double-blind, double-dummy, SUNBEAM study, adults (aged 18‒55 years) with RMS (N=1,346) were randomized to once-daily oral ozanimod 0.92 or 0.46 mg, or weekly intramuscular interferon β-1a 30 µg. The study continued until the last participant was treated for 12 months. CPS was measured as part of a secondary endpoint using the SDMT. Exploratory, post hoc analyses evaluated SDMT change and percentages of participants with clinically meaningful (≥4-point) SDMT improvement or worsening at months 6 and 12, and relationship between SDMT and brain volume on magnetic resonance imaging. Results: Ozanimod improved SDMT scores compared with interferon β-1a at months 6 and 12. At month 12, least squares mean difference in SDMT z-scores for ozanimod 0.92 mg versus interferon β-1a was 0.102 (95% CI, 0.031‒0.174, nominal p = 0.0051; standardized mean difference = 0.1376). A greater percentage of ozanimod 0.92 mg‒treated participants had clinically meaningful improvements in SDMT scores versus interferon β-1a at month 6 (30.0% versus 22.2%) and month 12 (35.6% versus 27.9%). Of those with SDMT improvement at month 6, 66.4% of those treated with ozanimod 0.92 mg and 55.9% of those treated with interferon β-1a had sustained Statistician: Hongjuan Liu, Bristol Myers Squibb, Princeton, New Jersey. Nonstandard abbreviations: CPS = cognitive processing speed; PwMS = persons with multiple sclerosis; SMD = standardized mean differences. * Corresponding author at: Senior Vice President for Research and Training, Kessler Foundation, 1199 Pleasant Valley Way, West Orange, NJ 07052. E-mail addresses: jdeluca@kesslerfoundation.org (J. DeLuca), sven.schippling@usz.ch (S. Schippling), xavier.montalban@cem-cat.org (X. Montalban), ludwig. kappos@usb.ch (L. Kappos), bruce.cree@ucsf.edu (B.A.C. Cree), comi.giancarlo@hsr.it (G. Comi), douglas.arnold@mcgill.ca (D.L. Arnold), hans-peter.hartung@ uni-duesseldorf.de (H.-P. Hartung), James.Sheffield@bms.com (J.K. Sheffield), Hongjuan.Liu@bms.com (H. Liu), Diego.Silva@bms.com (D. Silva), COHENJ@ccf. org (J.A. Cohen). https://doi.org/10.1016/j.msard.2020.102673 Received 5 August 2020; Received in revised form 18 November 2020; Accepted 2 December 2020 Available online 10 December 2020 2211-0348/© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 improvement at month 12. Brain volume loss was similar for those with SDMT improvement versus worsening at month 12. Conclusions: In these exploratory analyses, ozanimod had modestly beneficial effects on CPS in RMS participants. The effects of ozanimod on SDMT are being further evaluated in an ongoing 3-year clinical trial. SUNBEAM is registered on clinicaltrials.gov (NCT02294058) and the European Clinical Trials Database (EudraCT 2014- 002320-27). 1. Introduction Scans included 3 scout images (fast 3-plane localizer scan, true mid-line sagittal scan, and fast axial scan) followed by 7 sequences (PD-weighted, Cognitive dysfunction, one of the most disabling symptoms of mul­ T2-weighted, MT-off, MT-on, T1-weighted, FLAIR, and GdE T1- tiple sclerosis (MS), can include diminished cognitive processing speed weighted scans). Images were acquired using a standardized process (CPS), memory, learning, and executive ability (Benedict et al., 2017). and analyzed by a blinded central imaging facility (NeuroRx, Montréal, Slowed CPS, generally the first cognitive deficit to emerge in persons Quebec, Canada). Percentage change from baseline to month 12 in with MS (PwMS) (Van Schependom et al., 2015), may underlie other whole brain volume was a secondary endpoint; changes in cortical grey cognitive dysfunctions (Costa et al., 2017). It is associated with matter (CGM) and thalamic volumes were prespecified exploratory decreased quality of life, interference with everyday tasks, depression, endpoints. Whole brain volume and CGM were measured using SienaX, and unemployment (Barker-Collo, 2006; Clemens and Langdon, 2018; which provides an automated analysis of brain volume normalized for Eizaguirre et al., 2018). The Symbol Digit Modalities Test (SDMT) is a head size. Thalamic volume was measured using ThalamicVolume preferred measure of CPS for MS trials and routine screening (Drake software. Loss of whole brain volume, CGM volume, and thalamic vol­ et al., 2010; Strober et al., 2019; Kalb et al., 2018). ume was measured using paired Jacobian integration via JacobianA­ Ozanimod, an oral sphingosine 1-phosphate receptor 1 and 5 trophy software (Nakamura et al., 2014). modulator, causes lymphocyte retention in lymphoid tissues (Scott et al., 2016). In phase 3 trials (RADIANCE, SUNBEAM), ozanimod resulted in lower annualized relapse rates (ARR), fewer gadolinium-enhancing 2.2. Study population (GdE) and new/enlarging T2 lesions on brain magnetic resonance im­ aging (MRI), and less loss of whole brain, cortical grey matter, and SUNBEAM participants were aged 18 to 55 years inclusive with a thalamic volume than intramuscular interferon β-1a 30 μg weekly in diagnosis of MS by 2010 McDonald criteria (Polman et al., 2011), a participants with relapsing MS (RMS) (Cohen et al., 2019; Comi et al., clinical course and history of brain MRI lesions consistent with RMS, and 2019). Ozanimod 0.92 mg is approved by regulatory agencies in mul­ baseline EDSS score of ≤5.0. Participants had to have ≥1 documented tiple countries for the treatment of RMS. relapse within 12 months before screening, or ≥1 relapse in the 24 SUNBEAM included SDMT in the Multiple Sclerosis Functional months before screening plus ≥1 GdE lesion on brain MRI in the 12 Composite (MSFC), a secondary endpoint. Difference versus interferon months before randomization but no relapses from 30 days before β-1a in mean change in MSFC score (average of component z-scores) screening until randomization. Key exclusion criteria included primary from baseline to month 12 was 0.040 (95% CI − 0.009‒0.090, nominal p progressive MS, illness duration >15 years with an EDSS score ≤2.0, = 0.1091) for ozanimod 0.92 mg. Difference for mean change in SDMT presence of >20 GdE lesions at baseline, current signs of major z-score was 0.111 (0.039‒0.182, nominal p = 0.0024) (Comi et al., depression or history of suicide attempts, and history of alcohol or drug 2019). Based on the encouraging SDMT results, we undertook additional abuse in the year before randomization. Persons with any condition (eg, post hoc, exploratory analyses to further evaluate ozanimod’s effects on neurological, ophthalmological) that would make implementation of CPS in PwMS compared with interferon β-1a. study procedures or interpretation of the study results difficult, in the opinion of the treating investigator, were also excluded. Complete 2. Methods inclusion/exclusion criteria are available in the primary SUNBEAM publication appendix (Comi et al., 2019). The institutional review 2.1. Study design and procedures board/ethics committee at each site approved the study protocol, which conformed to Good Clinical Practice guidelines and Declaration of SUNBEAM was a randomized, double-blind, active-controlled, Helsinki principles. All participants provided informed consent before parallel-group, phase 3 study conducted in 20 countries from December enrollment. 2014 through December 2016 (Comi et al., 2019). Participants with RMS were randomized (1:1:1) to once-daily oral ozanimod 0.92 mg (equivalent to ozanimod HCl 1 mg) or 0.46 mg (equivalent to ozanimod 2.3. Assessment of CPS with SDMT HCl 0.5 mg), or weekly intramuscular interferon β-1a 30 µg. To maintain blinding, participants in the ozanimod groups also received weekly The same person, either the blinded evaluator or another designated placebo intramuscular injections, and participants in the interferon β-1a team member trained in conducting MSFC assessments, administered group received once-daily oral placebo tablets. Randomization was the SDMT throughout the study. The SDMT pairs symbols with numer­ carried out according to a blocked algorithm stratified by country and ical digits, and participants must match the appropriate symbol to the baseline Expanded Disability Status Scale (EDSS) score (≤3.5 or >3.5). paired digit using a key (Benedict et al., 2017). Scoring is based on the Ozanimod was dose-escalated as follows: ozanimod 0.23 mg correct number of responses given within 90 seconds; higher scores (equivalent to ozanimod HCl 0.25 mg) on days 1‒4, 0.46 mg on days 5‒ indicate faster processing (Benedict et al., 2017). Normative scores in a 7, and the assigned dose starting on day 8. Treatment continued until the (non-MS) general community population <40 years of age ranged from last participant randomized was treated for 12 months. The MSFC, 50‒58 for written responses and 59‒66 for oral responses (Sheridan which included the written SDMT, was a secondary outcome performed et al., 2006). Scores tend to decrease with age and are influenced by at screening, baseline, month 6, month 12, end of treatment, and at the education and gender (Sheridan et al., 2006; Strober et al., 2020). time of suspected relapse. Updated normative values for the oral SDMT range from 55‒68 in men Brain MRI for assessment of brain volume and lesions was performed aged 18‒54 years, and 60‒70 in women in the same age range (Strober at screening, month 6, and month 12 using 1.5 or 3.0 Tesla scanners. et al., 2020). A score change of ≥4 points is considered clinically meaningful (Benedict et al., 2017). 2 J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 2.4. Analyses of cognitive outcomes Table 1 LS mean changes from baseline Post hoc analyses of cognitive outcomes compared ozanimod 0.92 in SDMT scores and z-scores for ozanimod and interferon β-1a based on mixed mg with interferon β-1a 30 µg in the intent-to-treat (ITT) population (all model for repeated measures regression modeling.a randomized participants who received ≥1 dose of assigned study drug). Interferon Ozanimod All reported p-values are nominal and not controlled for type I error. β-1a 30 µg 0.92 mg Least squares (LS) mean change in SDMT score and z-score from Month 6, N 446 444 baseline to months 6 and 12 and between-treatment differences in LS LS mean (SE) change from baseline − 0.5 (0.48) 0.5 (0.47) means were estimated using a mixed model for repeated measures in SDMT score LS mean difference (95% CI) vs 1.0 (0.08‒1.93) (MMRM), a method that implicitly imputes missing data (Siddiqui et al., – interferon β-1a 2009). The ITT study population served as the reference population for Nominal p value – 0.0336 the z-scores. The model included change from baseline in SDMT score or SMD vs interferon β-1a 0.10 z-score as the dependent variable, stratification factors (region [Eastern LS mean (SE) change from baseline − 0.034 (0.036) 0.041 (0.035) in SDMT z-score Europe versus rest of world] and baseline EDSS category [≤3.5 versus LS mean difference (95% CI) vs – 0.074 (0.006‒0.143) >3.5]), baseline SDMT score or z-score, age at baseline, brain volume at interferon β-1a baseline, and the interaction between treatment and time point as fixed Nominal p value – 0.0336 effects, and subject as a random effect. An unstructured covariance was SMD vs interferon β-1a – 0.10 used to model within-subject errors. This analysis differs from the pre­ Month 12, N 426 427 LS mean (SE) change from baseline − 0.1 (0.49) 1.3 (0.48) viously reported analysis in which change in SDMT z-scores were re­ in SDMT score ported descriptively, and comparisons of ozanimod versus interferon LS mean difference – 1.4 (0.42‒2.34) β-1a were made using an analysis of covariance (ANCOVA) model with (95% CI) vs adjustments for region, baseline EDSS score, and baseline SDMT z-score; interferon β-1a Nominal p value 0.0051 data were imputed by the last observation carried forward (LOCF) – SMD vs interferon β-1a – 0.14 method (Comi et al., 2019). Compared with the ANCOVA and LOCF LS mean (SE) change from baseline − 0.009 (0.036) 0.093 (0.036) approach, MMRM provides a less biased approach to minimizing type I in SDMT z-score error and handling missing data (Siddiqui et al., 2009). Standardized LS mean difference (95% CI) vs – 0.102 (0.031‒0.174) mean differences (SMD, also called Cohen’s d) between ozanimod and interferon β-1a Nominal p value 0.0051 interferon β-1a were calculated as a measure of effect size (Faraone, – SMD vs interferon β-1a – 0.14 2008). Per established interpretations, SMD 0.2 was considered small; 0.5, medium; and 0.8, large (Faraone, 2008). CI = confidence interval; EDSS = Expanded Disability Status Scale; LS = least SDMT response was analyzed categorically as percentage of partic­ squares; SDMT = Symbol Digit Modalities Test; SE = standard error; SMD = standardized mean difference (Cohen’s d), in which 0.2 = small effect, 0.5 = ipants at months 6 and 12 with clinically meaningful improvement (≥4- medium effect, 0.8 = large effect (Faraone, 2008). point increase), stability (<4-point [±] change), or worsening (≥4-point a Adjusted by baseline SDMT score (or z-score), region (Eastern Europe vs rest decrease) in SDMT scores relative to baseline. Rate ratios with 95% of world), baseline EDSS category (≤3.5 vs >3.5), age at baseline, brain volume confidence intervals (CIs), and nominal p-values for comparison of at baseline, and interaction between treatment and time point as fixed effects, ozanimod with interferon β-1a were calculated. Cohen’s d was calcu­ assuming unstructured within-subject covariate structure. lated based on the corresponding odds ratios and interpreted using the same cutoffs for small, medium, and large effect size as for the SMDs. improvement (≥4-point increase) and those within SDMT worsening Rate ratios of SDMT improvement relative to baseline with ozanimod (≥4-point decrease) at month 12, and the other to compare SDMT 0.92 mg versus interferon β-1a were analyzed using a generalized esti­ improvement versus SDMT worsening within each treatment group. mating equation model adjusted by baseline SDMT score, age at base­ SMD was calculated for each analysis, and effect size was interpreted as line, brain volume at baseline, stratification factors, and the interaction described above. between treatment and time point as fixed effects, assuming unstruc­ tured within-subject covariate structure. The generalized estimating 3. Results equation model implicitly imputes missing data. A shift table was produced for number and percentage of participants 3.1. Subject disposition and demographics in each response category (improved/stable/worsened) at month 12 based on month 6 category. The improvement rate (≥4-point increase) In SUNBEAM, 447 participants received ≥1 dose of ozanimod 0.92 maintained at months 6 and 12 was analyzed using a generalized esti­ mg and 448 received ≥1 dose of interferon β-1a; 418 (93.5%) and 412 mating equation model adjusted by baseline SDMT score, age at base­ (92.0%) of those participants, respectively, completed the study. Mean line, and brain volume at baseline, stratification factors, and treatment. (standard deviation [SD]) duration of treatment exposure in these 2 Rate ratios with 95% CIs, and nominal p-values for comparison of oza­ groups, respectively, was 13.6 (2.7) and 13.5 (2.9) months. nimod with interferon β-1a were calculated for each shift, and Cohen’s As previously reported, baseline demographics and disease charac­ d effect sizes were calculated based on corresponding odds ratios. teristics in the overall population were similar across treatment groups Demographics and baseline disease characteristics were assessed (Comi et al., 2019). Mean age was 34.8 years in the ozanimod 0.92 mg among those with SDMT improvement (≥4-point increase) and wors­ group and 35.9 years in the interferon β-1a group, 63.3% and 67.0% ening (≥4-point decrease) at month 12 in the interferon β-1a and oza­ were female, 99.8% in each group were white, and 92.8% and 93.5% nimod 0.92 mg groups. For each treatment group, the nominal p value were from Eastern Europe. Mean time since first MS symptoms was 6.9 and SMD were calculated for those with SDMT improvement versus years and mean EDSS was 2.6 in both groups, and 28.6% of the ozani­ worsening at month 12. mod 0.92 mg group and 33.7% of the interferon β-1a group had previ­ Percentage change from baseline (screening MRI) to month 6 and ously used an MS disease-modifying therapy (DMT) (Comi et al., 2019). month 12 in whole brain volume, thalamic volume, and CGM volume Baseline mean (SD) SDMT score was 47.7 (13.7) in the ozanimod 0.92 were calculated using an ANCOVA model adjusted for stratification mg group and 47.1 (13.5) in the interferon β-1a group; SDMT z-scores factors (region and baseline EDSS category), age at baseline, and base­ were 0.045 (1.02) and 0.002 (1.00), respectively. At baseline, the oza­ line brain volume value of interest. Two analyses were performed, one to nimod 0.92 mg group and interferon β-1a group had similar whole brain test ozanimod 0.92 mg versus interferon β-1a in those within SDMT volume (mean 1456 and 1443 cm3, respectively), thalamic volume 3 J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 Fig. 1. (A) Categorical analysis of clinically meaningful change in SDMT. (B) SDMT clinically meaningful improvement relative to baseline with ozanimod 0.92 mg vs interferon β-1a (A) Percentage of participants treated with interferon β-1a or ozanimod who experienced clinically meaningful changes in SDMT scores at months 6 and 12. (B) Rate ratios (ozanimod vs interferon β-1a) for clinically meaningful SDMT improvement (≥ 4-point increase) relative to baseline at months 6 and 12 were analyzed based on a generalized estimating equation model adjusted by baseline SDMT score, stratification factors, age at baseline, brain volume at baseline, and treatment. Cohen’s d effect size was calculated based on odds ratios; interpretation: 0.2 = small effect, 0.5 = medium effect, 0.8 = large effect (Faraone, 2008). CI = confidence interval; SDMT = Symbol Digit Modalities Test. (mean 15.4 and 15.1 cm3), and CGM volume (mean 527 and 521 cm3). SDMT scores and z-scores between ozanimod and interferon β-1a were all 0.10‒0.14 at months 6 and 12 (trivial effect size) (Table 1). 3.2. LS mean change in SDMT score and z-scores 3.3. SDMT response Ozanimod nominally significantly improved SDMT scores compared with interferon β-1a at months 6 and 12 (Table 1). At month 12, LS mean A greater proportion of participants in the ozanimod 0.92 mg group change from baseline in z-scores was 0.093 in the ozanimod 0.92 mg achieved clinically meaningful improvement (≥4-point increase) in group and -0.009 in the interferon β-1a group (LS mean difference: SDMT score relative to interferon β-1a at months 6 and 12 (Fig. 1A). The 0.102 [95% CI 0.031‒0.174], nominal p = 0.0051). SMDs in LS mean related rate ratios favored ozanimod over interferon β-1a at both time 4 J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 Table 2 improvement (49/82 [59.8%]) versus worsening (33/82 [40.2%]), Shift table: SDMT response at months 12 relative to baseline based on month 6 whereas a larger percentage of women experienced SDMT worsening response categories. (87/157 [55.4%]) versus improvement (70/157 [44.5%]) at month 12; Interferon Ozanimod the effect of gender was small (SMD 0.29) but nominally significant (p = β-1a 30 µg 0.92 mg 0.0260) (Table 3). A similar trend was seen in the ozanimod 0.92 mg n/N (%) n/N (%) group for men (61/93 [65.6%] improved, 32/93 [34.4%] worsened), Month 6 Month 12 RR (95% CI),a and Cohen’s db vs but women were also more likely to be improved (91/153 [59.5%]) than Response Response Interferon β-1a worsened (62/153 [40.5%]); gender differences did not achieve nomi­ nal significance and had a trivial effect size (SMD 0.13). To explore this Improved 52/93 (55.9) 85/128 (66.4) 1.1 (0.95‒1.27), 0.21 finding further, we analyzed baseline SDMT by gender and found that Stable 33/93 (35.5) 39/128 (30.5) mean SDMT scores and z-scores were nominally significantly (p < Improved 0.9 (0.63‒1.33), − 0.07 0.0005) lower at baseline in men (score: 45.3 [SD 14.15]; z-score: − 0.14 Worsened 8/93 (8.6) 4/128 (3.1) [SD 1.05]) than women (score: 48.1 [SD 13.07]; z-score: 0.07 [SD NE, -0.58 Improved 56/228 (24.6) 55/204 (27.0) 0.97]). When baseline SDMT was added to the MMRM model, gender 1.1 (0.78‒1.47), 0.06 was not significant (p = 0.66). Stable 123/228 (53.9) 115/204 (56.4) Stable 1.0 (0.88‒1.25), 0.06 3.6. Change in brain volume based on SDMT response category at month 12 Worsened 49/228 (21.5) 34/204 (16.7) 0.8 (0.53‒1.15), − 0.18 Improved 11/105 (10.5) 12/95 (12.6) Ozanimod 0.92 mg was associated with a slower rate of whole brain 1.2 (0.58‒2.65), 0.13 volume loss over 12 months than interferon β-1a among participants Worsened Stable 31/105 (29.5) 27/95 (28.4) who showed clinically meaningful SDMT improvement at month 12 0.9 (0.59‒1.43), − 0.08 (difference 0.27% [95% CI 0.09‒0.44]; nominal p = 0.0030; SMD = 0.26 Worsened 63/105 (60.0) 56/95 (58.9) 1.0 (0.81‒1.24), 0.01 [small effect]) (Fig. 2A). Brain volume differences at month 12 in par­ ticipants with ≥4-point worsened SDMT were similar between treat­ NE = not evaluated due to small n; SDMT = Symbol Digit Modalities Test; n/N = ment groups (p = 0.1368). number of participants with corresponding response category at months 6 and Ozanimod was associated with a slower rate of thalamic volume loss 12/number of participants in response category at month 6. than interferon β-1a in those with clinically meaningful SDMT SDMT response is defined as follows: Improvement: ≥4-point increase; wors­ ened: ≥4-point decrease; stable: <4-point change in either direction. A 4-point improvement (difference 0.68% [95% CI 0.25‒1.10]; nominal p = change is considered clinically meaningful. 0.0019; SMD = 0.28 [small effect]) and those with clinically meaningful a All nominal p > 0.05. SDMT worsening (difference 1.02 [95% CI 0.40‒1.64]; nominal p = b Interpretation: 0.2 = small effect, 0.5 = medium effect, 0.8 = large effect 0.0013; SMD = 0.30 [small effect]) (Fig. 2B). Ozanimod also was (Faraone, 2008). associated with a slower rate of CGM loss than interferon β-1a in those with SDMT improvement (difference 1.06% [95% CI 0.83‒1.30; nomi­ points, being nominally significant but of small effect size (Fig. 1B). nal p < 0.0001; SMD = 0.78 [medium effect]) and those with SDMT worsening (difference 0.80% [95% CI 0.51‒1.09]; nominal p < 0.0001; SMD = 0.50 [medium effect]) at month 12 (Fig. 2C). 3.4. Maintenance of early SDMT response Irrespective of treatment assignment, rates of whole brain volume loss (Fig. 2D), thalamic volume loss (Fig. 2E), and CGM loss (Fig. 2F) In the subgroup of participants with clinically meaningful improve­ were similar among those with ≥4-point SDMT improvement versus ≥4- ment in SDMT at month 6 relative to baseline, 66.4% of those receiving point SDMT worsening. ozanimod 0.92 mg versus 55.9% of those receiving interferon β-1a showed sustained improvement at month 12 (Table 2). Differences be­ 4. Discussion tween ozanimod 0.92 mg and interferon β-1a did not achieve nominal significance, and the effect size for sustained improvement was small The current secondary analyses further support the original SDMT (Cohen’s d = 0.21). analyses from SUNBEAM, demonstrating that ozanimod 0.92 mg was Among all participants with SDMT assessments at months 6 and 12, associated with greater improvements in CPS at months 6 and 12 the overall rate of sustained ≥4-point improvement at both time points compared with interferon β-1a. These differences were nominally sig­ was 19.9% with ozanimod and 12.2% with interferon β-1a. Rate ratios nificant, although effect size was trivial. At month 6 and month 12, a for sustained improvement at months 6 and 12 favored ozanimod over greater proportion of ozanimod 0.92 mg–treated participants exhibited interferon β-1a (1.6 [95% CI 1.18‒2.18], nominal p = 0.0028, Cohen’s d clinically meaningful (≥4-point) improvements on the SDMT compared = 0.31 [small effect]). with those who received interferon β-1a, and the participants treated Some participants with stable or worsened SDMT at month 6 showed with ozanimod 0.92 mg more commonly had clinically meaningful clinically meaningful improvement relative to baseline at month 12 improvement at both those time points. (Table 2). These shifts did not achieve nominal significance, and effect The SDMT was administered as a component of the MSFC, which is sizes were trivial. reported in the primary publication of SUNBEAM (Comi et al., 2019). In contrast to the SDMT results, mean change from baseline to month 12 in 3.5. Baseline characteristics by SDMT response category at month 12 overall composite MSFC score (average of component z-scores) in the ozanimod 0.92 mg group was not significantly different from interferon Most baseline and disease characteristics were not significantly β-1a (difference: 0.040 [95% CI − 0.009‒0.090], nominal p = 0.1091). different in those who had clinically meaningful SDMT improvement The MSFC, as originally designed, included the Paced Auditory Serial versus worsening at month 12 with either ozanimod 0.92 mg or inter­ Addition Test (PASAT) as an assessment of CPS (Drake et al., 2010); feron β-1a (Table 3). One exception was that in both treatment groups, however, the SDMT is increasingly preferred because it is a more sen­ baseline SDMT scores were nominally significantly lower in those who sitive and more reliable measure of CPS in MS and its administration is subsequently showed ≥4-point SDMT improvement than those who had easier for examiners and participants (Drake et al., 2010; Strober et al., ≥4-point worsening; the SMD (0.53) was medium in both groups. In the 2019; Langdon et al., 2012). A recent pooled analysis of 14 clinical trials interferon β-1a group, a larger percentage of men experienced SDMT by the Multiple Sclerosis Outcome Assessments Consortium found that 5 J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 Table 3 Baseline demographics and disease characteristics in those with clinically meaningful SDMT improvement (≥4-point increase) or worsening (≥4-point decrease) at month 12. Data are mean (SD) unless otherwise noted Interferon β-1a 30 µg Ozanimod 0.92 mg SDMT SDMT SMDa SDMT SDMT SMDa Improved Worsened Improved Worsened (n = 119) (n = 120) (n = 152) (n = 94) Age, y 35.8 (9.0) 35.7 (9.2) − 0.01 34.2 (8.9) 35.2 (8.5) 0.11 Gender, n (%) 0.29b,* 0.13b Female 70 (58.8) 87 (72.5) 91 (59.9) 62 (66.0) Male 49 (41.2) 33 (27.5) 61 (40.1) 32 (34.0) Time since MS symptom onset, y 6.4 (5.1) 6.9 (6.2) 0.08 6.7 (5.7) 6.7 (6.4) 0.01 Time since MS diagnosis, y 3.1 (3.7) 3.8 (4.9) 0.16 3.5 (3.9) 3.4 (3.9) − 0.03 Type of MS, n (%) 0.06b − 0.01b RRMS 116 (97.5) 118 (98.3) 149 (98.0) 92 (97.9) PRMS 2 (1.7) 1 (0.8) 3 (2.0) 2 (2.1) SPMS 1 (0.8) 1 (0.8) 0 0 EDSS score 2.5 (1.0) 2.5 (1.2) 0.04 2.5 (1.2) 2.7 (1.2) 0.21 MSFC score − 0.08 (0.58) 0.09 (0.85) 0.23 0.04 (0.62) 0.13 (0.61) 0.13 SDMT score 44.7 (13.6) 52.1 (14.5) 0.53† 45.6 (12.0) 52.5 (14.9) 0.53‡ SDMT z-score − 0.18 (1.01) 0.37 (1.08) 0.53† − 0.12 (0.89) 0.40 (1.10) 0.53‡ MSQOL score 70.8 (15.1)c 71.4 (16.5) 0.04 70.8 (17.0) 67.2 (17.1)d − 0.21 No. of relapses in 12 mo prior to baseline 1.2 (0.52) 1.3 (0.56) 0.07 1.2 (0.50) 1.3 (0.63) 0.17 Prior DMT, n (%)e 38 (31.9) 37 (30.8) − 0.02 50 (32.9) 24 (25.5) − 0.16 No. of GdE lesions 1.8 (3.5) 2.0 (4.0) 0.05 1.8 (3.3) 1.6 (2.8) − 0.08 No. of T2 lesions 52.3 (37.1) 59.1 (39.1) 0.18 52.5 (37.0) 58.4 (40.6) 0.15 T2 lesion volume, cm3 12.3 (13.8) 14.3 (14.7) 0.14 11.2 (12.9) 12.7 (17.5) 0.10 WBV, cm3 1439.6 (78.3) 1447.4 (79.7) 0.10 1463.3 (73.7) 1451.4 (75.7) − 0.16 CGM volume, cm3 519.2 (44.9)f 519.9 (49.6) 0.01 532.2 (45.1) 521.9 (44.0) − 0.23 TV, cm3 15.1 (1.8) 15.1 (2.0) − 0.01 15.6 (1.9) 15.3 (2.0) − 0.15 CGM = cortical grey matter; EDSS = Expanded Disability Status Scale; DMT = disease-modifying therapy; GdE = gadolinium enhancing; MS = multiple sclerosis; MSFC = Multiple Sclerosis Functional Composite; MSQOL = Multiple Sclerosis Quality of Life; PRMS = progressive-relapsing multiple sclerosis; RRMS = relapsing- remitting multiple sclerosis; SD = standard deviation; SDMT = Symbol Digit Modalities Test; SMD, standardized mean difference; SPMS = secondary progressive multiple sclerosis; TV = thalamic volume; WBV = whole brain volume. * Nominal p < 0.05 for comparison of SDMT improved vs SDMT worsened † Nominal p < 0.0001 for comparison of SDMT improved vs SDMT worsened ‡ Nominal p < 0.001 for comparison of SDMT improved vs SDMT worsened. a SMD was calculated by using the adjusted mean difference of worsened minus improved. Interpretation: 0.2 = small effect, 0.5 = medium effect, 0.8 = large effect (Faraone, 2008). b For categorical variables, the SMD and chi-square test nominal p value were based on the proportion of the majority category to a combination of all minority categories c n = 117 d n = 93 e Prior DMTs used by any SUNBEAM participants included interferon β-1a, pegylated interferon β-1a, interferon β-1b, glatiramer acetate, daclizumab, dimethyl fumarate, teriflunomide, and mitoxantrone f n = 118. SDMT does not have floor or ceiling effects, and that a 4-point change is trial is daclizumab (Benedict et al., 2018), which was withdrawn from clinically meaningful (Goldman et al., 2019). There are no published the global market in 2018. Some DMTs (fingolimod, natalizumab, data on clinically meaningful change on the PASAT. interferon β-1b) showed benefit on the PASAT in published, Participants whose SDMT performance improved during treatment peer-reviewed phase 3 trials (Weinstock-Guttman et al., 2012; Kappos had lower baseline SDMT scores, and therefore may have had a greater et al., 2016; Penner et al., 2012), although a second phase 3 trial of opportunity to show a ≥4-point improvement than those with higher natalizumab failed to find benefit versus placebo (Weinstock-Guttman baseline SDMT scores. In both treatment groups, men more commonly et al., 2012). A recent meta-analysis of 41 studies of the effects of DMTs experienced clinically meaningful SDMT improvement than worsening. on CPS (SDMT or PASAT) reported a small to moderate benefit overall, Women more commonly had clinically meaningful SDMT improvement with no advantage identified for any DMT over another (Landmeyer versus worsening if treated with ozanimod 0.92 mg whereas they more et al., 2020). A separate systematic review of 87 studies concluded that commonly worsened if treated with interferon β-1a. SDMT scores were good-quality evidence was still lacking and insufficient to recommend lower at baseline among men. Furthermore, normative scores on the oral use of DMTs or other pharmacologic therapies to improve cognitive SDMT in healthy volunteers are significantly (p < 0.001) lower for men function in PwMS (Chen et al., 2020). than women (Strober et al., 2020). The previously published primary Relative to interferon β-1a, ozanimod was associated with reduced analysis of SUNBEAM did not show a significant variation in ozanimod loss of thalamic and CGM volume over 12 months in those with either efficacy by gender, based on ARR, new/enlarging T2 lesion count, or clinically meaningful SDMT improvement or worsening at month 12. GdE lesion count (Comi et al., 2019). Ozanimod was also associated with reduced whole brain volume loss in There are no approved treatments available for slowed CPS in PwMS. those with SDMT improvement, but not SDMT worsening. Whole brain, Current research includes cognitive rehabilitation (DeLuca et al., 2020) CGM, and thalamic volume loss were similar (nominal p > 0.05, trivial and explorations of whether MS DMTs preserve or improve CPS effect sizes) for those with SDMT improvement and those with SDMT (Weinstock-Guttman et al., 2012; Benedict et al., 2018; Kappos et al., worsening at month 12 in either treatment group, which contrasts with 2016; Penner et al., 2012). To date, the only other DMT with published, previous reports that indicated that SDMT scores correlate with peer-reviewed results of cognitive analyses using the SDMT in a phase 3 neocortical and all deep grey matter volumes, particularly thalamic 6 J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 Fig. 2. Change in (A) whole brain volume, (B) thalamic volume, and (C) cortical grey matter volume over 12 months (solid bars) and over 6 months (dotted bars) comparing treatment difference between ozanimod 0.92 mg or interferon β-1a among those with clinically meaningful SDMT improvement and among those with clinically meaningful SDMT worsening at month 12. Change in (D) whole brain volume, (E) thalamic volume, and (F) cortical grey matter volume over 12 months (solid bars) and over 6 months (dotted bars) comparing clinically meaningful SDMT improvement vs SDMT worsening within each treatment group. All comparisons conducted using analysis of covariance model adjusted for region (Eastern Europe vs rest of world), baseline EDSS category (≤3.5 vs >3.5), age at baseline, and baseline brain volume value of interest. CI = confidence interval; EDSS = Expanded Disability Status Scale; LS = least squares; SDMT = Symbol Digit Modalities Test; SE = standard error; SMD = standardized mean difference (Cohen’s d); interpretation: 0.2 = small effect, 0.5 = medium effect, 0.8 = large effect (Faraone, 2008). 7 J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 volume (Batista et al., 2012; Bergsland et al., 2016; Bisecco et al., 2018, innovative treatment approaches. Data requests may be submitted to 2019). It is unclear why results from SUNBEAM, which prospectively Celgene, a Bristol Myers Squibb company, at https://vivli.org/our­ evaluated both SDMT and brain volume, failed to corroborate those member/celgene/ and must include a description of the research previous findings. The short (≥12 months) study duration may have proposal. hindered the ability to detect an association. Brain atrophy alone might be limited in its capacity to reflect the entire degenerative component of Declaration of Competing Interest MS. Other factors that may be involved in SDMT change are diffuse white and grey matter damage and dendritic spine loss, which have been J. DeLuca reports personal compensation for consulting from Cel­ described in pathology studies (Kutzelnigg et al., 2005; Bo et al., 2003; gene, Biogen IDEC, Novartis, Consortium of MS Centers; speaker for Jurgens et al., 2016), but may not be detected by conventional MRI. Sanofi-Genzyme, Canadian MS Society, EXCEMED; grant funding from SDMT results are affected not only by CPS but also visual acuity and Biogen IDEC, EMD Serono, Canadian MS Society, National MS Society, ocular motor function, and over repeated assessments patients may and Consortium of MS Centers. S. Schippling reports personal compen­ learn the symbol-digit associations, facilitating response (Benedict et al., sation for employment, consulting, serving on a scientific advisory 2017). Finally, negative change or no change in SDMT may reflect board, speaking, or other activities from Biogen, Celgene, Merck, Sanofi, limited cognitive reserve/learning capacity in addition to, or even and Teva. X. Montalban has received speaking honoraria and travel independently from, a poorer performance due to tissue damage. expenses for scientific meetings, has been a steering committee member Further studies with CPS as a primary outcome are needed to better of clinical trials or participated in advisory boards of clinical trials in the understand the relationship between CPS and thalamic volume loss. past 3 years with Actelion, Alexion, Bayer, Biogen, Celgene, EMD Strengths of this analysis include use of a well-established, validated Serono, EXCEMED, Genzyme, MedDay, Merck, Multiple Sclerosis In­ tool (SDMT) for assessment of a clinically relevant cognitive deficit ternational Federation, National Multiple Sclerosis Society, Nervgen, (CPS) in PwMS and use of SDMT and brain volume data derived from a Novartis, Roche, Sanofi Genzyme, Teva Pharmaceuticals, and TG large, well-controlled phase 3 study. It is also notable that ozanimod Therapeutics. L. Kappos’s institution (University Hospital Basel) has showed benefit in SDMT results in the prespecified SDMT analysis in received in the last 3 years the following, which was used exclusively for SUNBEAM against an active comparator (interferon β-1a) that is known research support: steering committee, advisory board, consultancy fees, to have some potential benefits with regard to cognition (Mokhber et al., and support of educational activities from Actelion, Allergan, Almirall, 2014). However, these SDMT exploratory results should be interpreted Baxalta, Bayer, Biogen, Celgene, CSL Behring, Desitin, EXCEMED, Eisai, cautiously given that they are post hoc analyses. The SUNBEAM study Genzyme, Japan Tobacco, Merck, Minoryx, Novartis, Pfizer, F. sample size determination was based on the primary endpoint (ARR), Hoffmann-La Roche Ltd, Sanofi, Santhera, Teva, and license fees for and the study was not specifically designed to evaluate change in SDMT. Neurostatus-UHB products; the research of the MS Center in Basel has Although physical or cognitive disability could potentially interfere with been supported by grants from Bayer, Biogen, European Union, Inno­ performance on the written SDMT, people with conditions that might suisse, Novartis, Roche Research Foundations, Swiss MS Society, and hinder implementation of study procedures or interpretation of the Swiss National Research Foundation. B.A.C. Cree reports personal study results were excluded, and the mean baseline EDSS indicated compensation for consulting for Akili, Alexion, Biogen, EMD Serono, limited disability in the study population. It is unknown whether de­ Novartis, Sanofi, and TG Therapeutics. G. Comi reports compensation creases in SDMT or worsening by ≥4 points were related to clinical or for consulting and/or speaking activities from Almirall, Biogen, Celgene, cognitive relapse (Giedraitiene et al., 2018; Meli et al., 2020). SDMT EXCEMED, Forward Pharma, Genzyme, Merck, Novartis, Roche, Sanofi, data from SUNBEAM were limited to 12 months, as the study ended once and Teva. D.L. Arnold reports personal fees for consulting and/or grants the last enrolled participant completed 12 months of treatment, and a from Albert Charitable Trust, Biogen, Celgene, F. Hoffmann-La Roche, limited number of participants had SDMT assessments at subsequent Frequency Therapeutics, MedDay, Merck Serono, Novartis, and Sanofi time points. Longer studies are needed to better elucidate the long-term Aventis, and equity interest in NeuroRx Research. H-P Hartung reports effects of ozanimod on cognition. While the current analyses do not personal fees for consulting, serving on steering committees, and control for type I error, nominally significant differences in least squares speaking from Bayer Healthcare, Biogen, Celgene, GeNeuro, Genzyme, mean changes from baseline in SDMT scores and z-scores between Merck, MedImmune, Novartis, Octapharma, Roche, Sanofi, and Teva. J. ozanimod and interferon β-1a were observed at multiple time points, K. Sheffield is an employee of Bristol Myers Squibb. H. Liu is an suggesting that the observations are attributed to a true treatment effect employee of Bristol Myers Squibb. D. Silva is an employee of Bristol and are not due to chance. Myers Squibb. J.A. Cohen reports personal compensation for consulting for Adamas, Atara, Bristol Myers Squibb, Convelo, MedDay, and Mylan; 5. Conclusions and serving as an Editor of Multiple Sclerosis Journal. These post hoc, exploratory results suggest that ozanimod has Funding disclosure modestly beneficial effects on CPS in PwMS. The sustained effect of these encouraging findings needs to be confirmed in prospective, long- SUNBEAM was sponsored by Celgene Corporation. The sponsor was term studies focusing on cognitive endpoints. The SDMT is being used involved in data analysis and interpretation, and manuscript prepara­ as a measure of CPS in both a long-term open-label extension study tion, review, and approval. All authors vouch for data accuracy, (study RPC01-3001, DAYBREAK; NCT02576717), as well as in a phase reviewed all drafts, and approved the final manuscript. 3b trial (ENLIGHTEN; NCT04140305) where clinically meaningful improvement in CPS is the primary endpoint. The relationship between Acknowledgment brain volume loss and changes in SDMT will also be further character­ ized in these studies. Support for third-party writing assistance for this manuscript was provided by Peloton Advantage, LLC, an OPEN Health company, and 5.1. Data availability statement was funded by Bristol Myers Squibb. Celgene, a Bristol Myers Squibb company, is committed to respon­ References sible and transparent sharing of clinical trial data with patients, healthcare practitioners, and independent researchers for the purpose of Barker-Collo, SL., 2006. Quality of life in multiple sclerosis: does information-processing improving scientific and medical knowledge as well as fostering speed have an independent effect. Arch Clin Neuropsychol 21 (2), 167–174. 8 J. DeLuca et al. Multiple Sclerosis and Related Disorders 48 (2021) 102673 Batista, S, Zivadinov, R, Hoogs, M, et al., 2012. Basal ganglia, thalamus and neocortical Goldman, MD, LaRocca, NG, Rudick, RA, et al., 2019. Evaluation of multiple sclerosis atrophy predicting slowed cognitive processing in multiple sclerosis. J Neurol 259 disability outcome measures using pooled clinical trial data. Neurology 93, (1), 139–146. e1921–e1931. Benedict, RHB, DeLuca, J, Enzinger, C, Geurts, JJG, Krupp, LB, Rao, SM, 2017. Jurgens, T, Jafari, M, Kreutzfeldt, M, et al., 2016. Reconstruction of single cortical Neuropsychology of multiple sclerosis: looking back and moving forward. J Int projection neurons reveals primary spine loss in multiple sclerosis. Brain 139 (Pt 1), Neuropsychol Soc 23 (9-10), 832–842. 39–46. Benedict, RH, DeLuca, J, Phillips, G, LaRocca, N, Hudson, LD, Rudick, R, 2017. Validity Kalb, R, Beier, M, Benedict, RH, et al., 2018. Recommendations for cognitive screening of the symbol digit modalities test as a cognition performance outcome measure for and management in multiple sclerosis care. Mult Scler 24 (13), 1665–1680. multiple sclerosis. Mult Scler 23 (5), 721–733. Kappos, L, Radue, EW, Chin, P, Ritter, S, Tomic, D, Lublin, F, 2016. Onset of clinical and Benedict, RH, Cohan, S, Lynch, SG, et al., 2018. Improved cognitive outcomes in patients MRI efficacy occurs early after fingolimod treatment initiation in relapsing multiple with relapsing-remitting multiple sclerosis treated with daclizumab beta: results sclerosis. J Neurol 263 (2), 354–360. from the DECIDE study. Mult Scler 24 (6), 795–804. Kutzelnigg, A, Lucchinetti, CF, Stadelmann, C, et al., 2005. Cortical demyelination and Bergsland, N, Zivadinov, R, Dwyer, MG, Weinstock-Guttman, B, Benedict, RH, 2016. diffuse white matter injury in multiple sclerosis. Brain 128 (Pt 11), 2705–2712. Localized atrophy of the thalamus and slowed cognitive processing speed in MS Landmeyer, NC, Bürkner, PC, Wiendl, H, et al., 2020. Disease-modifying treatments and patients. Mult Scler 22 (10), 1327–1336. cognition in relapsing-remitting multiple sclerosis: A meta-analysis. Neurology 94 Bisecco, A, Stamenova, S, Caiazzo, G, et al., 2018. Attention and processing speed (22), e2373–e2383. performance in multiple sclerosis is mostly related to thalamic volume. Brain Langdon, DW, Amato, MP, Boringa, J, et al., 2012. Recommendations for a Brief Imaging Behav 12 (1), 20–28. International Cognitive Assessment for Multiple Sclerosis (BICAMS). Mult Scler 18 Bisecco, A, Capuano, R, Caiazzo, G, et al., 2019. Regional changes in thalamic shape and (6), 891–898. volume are related to cognitive performance in multiple sclerosis. Mult Scler. Meli, R, Roccatagliata, L, Capello, E, et al., 2020. Ecological impact of isolated cognitive https://doi.org/10.1177/1352458519892552 [Epub ahead of print]. relapses in MS. Mult Scler 26 (1), 114–117. Bo, L, Vedeler, CA, Nyland, HI, Trapp, BD, Mork, SJ, 2003. Subpial demyelination in the Mokhber, N, Azarpazhooh, A, Orouji, E, et al., 2014. Cognitive dysfunction in patients cerebral cortex of multiple sclerosis patients. J Neuropathol Exp Neurol 62 (7), with multiple sclerosis treated with different types of interferon beta: a randomized 723–732. clinical trial. J Neurol Sci 342 (1-2), 16–20. Chen, MH, Goverover, Y, Genova, HM, DeLuca, J, 2020. Cognitive efficacy of Nakamura, K, Guizard, N, Fonov, VS, Narayanan, S, Collins, DL, Arnold, DL, 2014. pharmacologic treatments in multiple sclerosis: a systematic review. CNS Drugs 34 Jacobian integration method increases the statistical power to measure gray matter (6), 599–628. atrophy in multiple sclerosis. NeuroImage Clinical 4, 10–17. Clemens, L, Langdon, D., 2018. How does cognition relate to employment in multiple Penner, IK, Stemper, B, Calabrese, P, et al., 2012. Effects of interferon beta-1b on sclerosis? A systematic review. Mult Scler Relat Disord 26, 183–191. cognitive performance in patients with a first event suggestive of multiple sclerosis. Cohen, JA, Comi, G, Selmaj, KW, et al., 2019. Safety and efficacy of ozanimod versus Mult Scler 18 (10), 1466–1471. interferon beta-1a in relapsing multiple sclerosis (RADIANCE): a multicentre, Polman, CH, Reingold, SC, Banwell, B, et al., 2011. Diagnostic criteria for multiple randomised, 24-month, phase 3 trial. Lancet Neurol 18 (11), 1021–1033. sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol 69 (2), 292–302. Comi, G, Kappos, L, Selmaj, KW, et al., 2019. Safety and efficacy of ozanimod versus Scott, FL, Clemons, B, Brooks, J, et al., 2016. Ozanimod (RPC1063) is a potent interferon beta-1a in relapsing multiple sclerosis (SUNBEAM): a multicentre, sphingosine-1-phosphate receptor-1 (S1P1) and receptor-5 (S1P5) agonist with randomised, minimum 12-month, phase 3 trial. Lancet Neurol 18 (11), 1009–1020. autoimmune disease-modifying activity. Br J Pharmacol 173 (11), 1778–1792. Costa, SL, Genova, HM, DeLuca, J, Chiaravalloti, ND, 2017. Information processing Sheridan, LK, Fitzgerald, HE, Adams, KM, et al., 2006. Normative Symbol Digit speed in multiple sclerosis: Past, present, and future. Mult Scler 23 (6), 772–789. Modalities Test performance in a community-based sample. Arch Clin Neuropsychol DeLuca, J, Chiaravalloti, ND, Sandroff, BM, 2020. Treatment and management of 21 (1), 23–28. cognitive dysfunction in patients with multiple sclerosis. Nat Rev Neurol 16 (6), Siddiqui, O, Hung, HM, O’Neill, R, 2009. MMRM vs. LOCF: a comprehensive comparison 319–332. based on simulation study and 25 NDA datasets. J Biopharm Stat 19 (2), 227–246. Drake, AS, Weinstock-Guttman, B, Morrow, SA, Hojnacki, D, Munschauer, FE, Strober, L, DeLuca, J, Benedict, RH, et al., 2019. Symbol Digit Modalities Test: A valid Benedict, RH, 2010. Psychometrics and normative data for the Multiple Sclerosis clinical trial endpoint for measuring cognition in multiple sclerosis. Mult Scler 25 Functional Composite: replacing the PASAT with the Symbol Digit Modalities Test. (13), 1781–1790. Mult Scler 16 (2), 228–237. Strober, LB, Bruce, JM, Arnett, PA, et al., 2020. A new look at an old test: Normative data Eizaguirre, MB, Vanotti, S, Merino, A, et al., 2018. The role of information processing of the symbol digit modalities test -oral version. Mult Scler Relat Disord 43. https:// speed in clinical and social support variables of patients with multiple sclerosis. doi.org/10.1016/j.msard.2020.102154. Epub 2020 May 3. J Clin Neurol 14 (4), 472–477. Van Schependom, J, D’Hooghe M, B, Cleynhens, K, et al., 2015. Reduced information Faraone, SV., 2008. Interpreting estimates of treatment effects: implications for managed processing speed as primum movens for cognitive decline in MS. Mult Scler 21 (1), care. P & T 33 (12), 700–711. 83–91. Giedraitiene, N, Kaubrys, G, Kizlaitiene, R, 2018. Cognition during and after multiple Weinstock-Guttman, B, Galetta, SL, Giovannoni, G, et al., 2012. Additional efficacy sclerosis relapse as assessed with the Brief International Cognitive Assessment for endpoints from pivotal natalizumab trials in relapsing-remitting MS. J Neurol 259 Multiple Sclerosis. Sci Rep 8 (1), 8169. (5), 898–905. 9