Highlights
- •Ozanimod and ponesimod have not been directly compared in multiple sclerosis.
- •A matching-adjusted indirect comparison of both treatments was conducted.
- •Ozanimod was associated with nonsignificant reductions in annualized relapse rates.
- •Reductions in brain volume loss were significant after treatment with ozanimod.
- •Ozanimod had a favorable safety profile.
Abstract
Background
Ozanimod and ponesimod are sphingosine 1-phosphate receptor modulators approved by the U.S. Food and Drug Administration for treatment of relapsing forms of multiple sclerosis (MS). Given that no head-to-head trials have assessed these two treatments, we performed a matching-adjusted indirect comparison (MAIC) to compare efficacy and safety outcomes between ozanimod and ponesimod for MS.
Methods
A MAIC compared efficacy and safety of ozanimod and ponesimod at 2 years. Outcomes included annualized relapse rate (ARR) and percentage change from baseline in brain volume loss (BVL) as well as rates of any treatment-emergent adverse events (TEAEs), serious adverse events (AEs), AEs leading to discontinuation, and other safety outcomes. Individual patient-level data were obtained for ozanimod from the RADIANCE-B trial, while aggregate-level patient data were obtained for ponesimod from the OPTIMUM trial. The MAIC was not anchored owing to lack of a common comparator across the two trials. The following characteristics were matched between the trials’ populations: age, sex, time since MS symptom onset, relapses in prior year, Expanded Disability Status Scale score, disease-modifying therapies received in the prior 2 years, absence of gadolinium-enhancing T1 lesions, and percentage of patients from Eastern Europe.
Results
After matching, key baseline characteristics were balanced between patients receiving ozanimod and ponesimod. Compared with ponesimod, ozanimod had a numerically lower ARR (rate ratio: 0.80 [95% CI: 0.57, 1.10]) and was associated with a significant reduction in BVL (% change difference: 0.20 [95% CI: 0.05, 0.36]). Additionally, ozanimod was associated with a significantly lower risk of TEAEs (risk difference: −11.9% [95% CI: –16.8%, –7.0%]), AEs leading to discontinuation (−6.1% [95% CI: −8.9%, −3.4%]), and lymphocyte count <0.2 K/μL (−2.3% [95% CI: −4.2%, −0.5%]). There were no statistically significant differences in the other safety outcomes.
Conclusion
The MAIC results suggest that, compared with ponesimod, ozanimod is more effective in preserving brain volume, is comparable in terms of reducing relapse rates, and has a favorable safety profile.
Keywords
1. Introduction
Multiple sclerosis (MS) is a chronic autoimmune and neurodegenerative disease of the central nervous system, characterized by inflammation, demyelination, neuronal and oligodendrocyte loss, and disruption of the blood-brain barrier. Brain volume loss (BVL), the result of white and gray matter focal lesions, as well as diffuse damage in normal-appearing white and gray matter, correlates with long-term disability progression and cognitive impairment in people with MS (
Kappos et al., 2016
; Radue et al., 2015
; Vollmer et al., 2015
). MS may present as clinically isolated syndrome, relapsing-remitting MS (RRMS), secondary progressive MS, or primary progressive MS (Lublin et al., 2014
; - Lublin F.D.
- Reingold S.C.
- Cohen J.A.
- Cutter G.R.
- Sorensen P.S.
- Thompson A.J.
- Wolinsky J.S.
- Balcer L.J.
- Banwell B.
- Barkhof F.
- Bebo Jr., B.
- Calabresi P.A.
- Clanet M.
- Comi G.
- Fox R.J.
- Freedman M.S.
- Goodman A.D.
- Inglese M.
- Kappos L.
- Kieseier B.C.
- Lincoln J.A.
- Lubetzki C.
- Miller A.E.
- Montalban X.
- O'Connor P.W.
- Petkau J.
- Pozzilli C.
- Rudick R.A.
- Sormani M.P.
- Stuve O.
- Waubant E.
- Polman C.H.
Defining the clinical course of multiple sclerosis: the 2013 revisions.
Olek and Howard, 2022
). RRMS, the most prevalent type of MS, is characterized by episodes of relapses followed by a period of remission. Approximately 85% of individuals with MS present with RRMS (European Medicines Agency (EMA) 2015
). Complete physical recovery from relapse can occur, but approximately 50% of relapses are associated with residual neurologic deficits (Lublin et al., 2003
) resulting in sustained increase in disability, typically measured in phase 3 MS clinical trials using the Expanded Disability Status Scale (EDSS) (Giovannoni et al., 2016
; Ziemssen et al., 2016
). The disease often has a highly debilitating impact on quality of life for people with MS and their families (Aronson, 1997
) and is associated with considerable economic burden (Tremlett et al., 2001
).Numerous disease-modifying therapies (DMTs) are available for the treatment of MS (
Doshi and Chataway, 2017
; Finkelsztejn, 2014
; Garg and Smith, 2015
; Mayo Clinic 2022
; Straus Farber et al., 2016
). These include injectable treatments, such as interferon beta medications (IFN ß-1a) and glatiramer acetate, oral treatments, such as ozanimod, ponesimod, fingolimod, and teriflunomide, among others, and infusion treatments, such as ocrelizumab, natalizumab and alemtuzumab.Owing to the lack of head-to-head randomized controlled trials comparing these DMTs, pairwise comparisons between treatments in RRMS have been conducted using matching-adjusted indirect comparisons (MAICs). Currently, published comparisons include dimethyl fumarate (DMF) versus fingolimod (
Fox et al., 2015
), delayed-release DMF versus fingolimod (Fox et al., 2017
), cladribine versus fingolimod, natalizumab, alemtuzumab, and ocrelizumab in adults (Berardi et al., 2019
), ozanimod versus DMF (- Berardi A.
- Siddiqui M.K.
- Treharne C.
- Harty G.
- Wong S.L.
Estimating the comparative efficacy of cladribine tablets versus alternative disease modifying treatments in active relapsing-remitting multiple sclerosis: adjusting for patient characteristics using meta-regression and matching-adjusted indirect treatment comparison approaches.
Cohan et al., 2021a
), ozanimod versus teriflunomide (Cohan et al., 2021b
), and ozanimod versus fingolimod (- Cohan S.
- Tencer T.
- Arndorfer S.
- Zhu X.
- Zivkovic M.
- Kumar J.
Matching-adjusted indirect treatment comparison of ozanimod versus teriflunomide for relapsing multiple sclerosis.
Mult. Scler. Relat. Disord. 2021; 52102972https://doi.org/10.1016/j.msard.2021.102972
Swallow et al., 2020
).Comparisons between ozanimod, ponesimod, and fingolimod are of particular interest, given that these DMTs share a similar mechanism of action as selective sphingosine 1-phosphate (S1P) receptor modulators. All 3 have been approved by the U.S. Food and Drug Administration for treatment of RRMS based on efficacy demonstrated in phase 3 trials: RADIANCE-B (
Cohen et al., 2019
) and SUNBEAM (- Cohen J.A.
- Comi G.
- Selmaj K.W.
- Bar-Or A.
- Arnold D.L.
- Steinman L.
- Hartung H.P.
- Montalban X.
- Kubala Havrdová E.
- Cree B.A.C.
- Sheffield J.K.
- Minton N.
- Raghupathi K.
- Huang V.
- Kappos L.
Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (RADIANCE): a multicentre, randomised, 24-month, phase 3 trial.
Comi et al., 2017
) for ozanimod, OPTIMUM (- Comi G.
- Kappos L.
- Selmaj K.W.
- Bar-Or A.
- Arnold D.L.
- Steinman L.
- Hartung H.P.
- Montalban X.
- Havrdova E.K.
- Cree B.A.C.
- Sheffield J.K.
- Raghupathi K.
- Cohen J.A.
Ozanimod demonstrates efficacy and safety in a phase 3 trial of relapsing multiple sclerosis (SUNBEAM) [abstract 232].
Mult. Scler. J. 2017; 23: 73-74
Kappos et al., 2021
) for ponesimod, TRANSFORMS (- Kappos L.
- Fox R.J.
- Burcklen M.
- Freedman M.S.
- Havrdova E.K.
- Hennessy B.
- Hohlfeld R.
- Lublin F.
- Montalban X.
- Pozzilli C.
- Scherz T.
- D'Ambrosio D.
- Linscheid P.
- Vaclavkova A.
- Pirozek-Lawniczek M.
- Kracker H.
- Sprenger T.
Ponesimod compared with teriflunomide in patients with relapsing multiple sclerosis in the active-comparator phase 3 OPTIMUM study: a randomized clinical trial.
Cohen et al., 2010
) for fingolimod, but the comparative safety and efficacy of these S1P receptor-modulating agents has not yet been fully characterized. A previous MAIC was recently used to assess the comparative safety and efficacy of ozanimod versus fingolimod (Swallow et al., 2020
).Ozanimod's efficacy and safety for the treatment of RRMS was evaluated in the RADIANCE-B and SUNBEAM phase 3 clinical trials; these trials enrolled 1320 and 1346 participants for 24 and 12 months, respectively. In the ozanimod trials, patients were randomly assigned 1:1:1 to receive daily oral ozanimod 1.0 mg, daily oral ozanimod 0.5 mg, or weekly intramuscular IFN ß-1a 30 μg. Both studies established favorable efficacy and safety profiles for ozanimod compared with IFN ß-1a. In particular, the annualized relapse rate (ARR) for patients treated with ozanimod 1.0 mg was significantly lower than for those treated with IFN ß-1a (0.17 vs. 0.28 [P<0.0001] in RADIANCE and 0.18 vs. 0.35 [P<0.0001] in SUNBEAM). The incidence of adverse events (AEs) was either lower or similar in the ozanimod versus the IFN ß-1a arm.
Results from the phase 3 trial of ponesimod, the OPTIMUM study, showed favorable results for the treatment of RRMS. A total of 1133 participants were enrolled, randomly assigned 1:1 to receive ponesimod 20 mg or teriflunomide 14 mg once daily oral tablets for 108 weeks. The ARR in the ponesimod arm was significantly lower than it was in the teriflunomide arm (0.20 vs. 0.29 [P<0.001]). The safety profile of ponesimod was assessed as being similar to that of other S1P modulators.
We used data from these phase 3 trials and conducted indirect comparisons between ozanimod and ponesimod, adjusting for cross-trial differences in patient populations to account for the absence of a head-to-head randomized trial.
2. Methods
2.1 Data sources
Data from three phase 3 clinical trials were assessed for suitability for use in an indirect treatment comparison: individual patient data (IPD) from the RADIANCE-B (
Cohen et al., 2019
) (NCT02047734, RPC01–201B) and SUNBEAM (- Cohen J.A.
- Comi G.
- Selmaj K.W.
- Bar-Or A.
- Arnold D.L.
- Steinman L.
- Hartung H.P.
- Montalban X.
- Kubala Havrdová E.
- Cree B.A.C.
- Sheffield J.K.
- Minton N.
- Raghupathi K.
- Huang V.
- Kappos L.
Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (RADIANCE): a multicentre, randomised, 24-month, phase 3 trial.
Comi et al., 2019
) (NCT02294058, RPC01–301) clinical trials of ozanimod and summary-level data from the OPTIMUM pivotal trial of ponesimod (- Comi G.
- Kappos L.
- Selmaj K.W.
- Bar-Or A.
- Arnold D.L.
- Steinman L.
- Hartung H.P.
- Montalban X.
- Kubala Havrdova E.
- Cree B.A.C.
- Sheffield J.K.
- Minton N.
- Raghupathi K.
- Ding N.
- Cohen J.A.
for the SUNBEAM Study Investigators
Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (SUNBEAM): a multicentre, randomised, minimum 12-month, phase 3 trial.
Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (SUNBEAM): a multicentre, randomised, minimum 12-month, phase 3 trial.
Kappos et al., 2021
) (NCT02425644, AC–058B301). The feasibility assessment revealed that only 2-year outcomes could be compared. Consequently, SUNBEAM trial data were not included in the analysis.- Kappos L.
- Fox R.J.
- Burcklen M.
- Freedman M.S.
- Havrdova E.K.
- Hennessy B.
- Hohlfeld R.
- Lublin F.
- Montalban X.
- Pozzilli C.
- Scherz T.
- D'Ambrosio D.
- Linscheid P.
- Vaclavkova A.
- Pirozek-Lawniczek M.
- Kracker H.
- Sprenger T.
Ponesimod compared with teriflunomide in patients with relapsing multiple sclerosis in the active-comparator phase 3 OPTIMUM study: a randomized clinical trial.
2.2 Study outcomes
Based on data availability and clinical expert input, 2-year outcomes were selected for comparison between ozanimod and ponesimod. Efficacy outcomes in this analysis were ARR and % change from baseline in BVL. Safety outcomes were AEs leading to discontinuation, death, any treatment-emergent AE (TEAE), macular edema, absolute lymphocyte count <0.2 K/μL, and any serious AE (SAE), which included serious infections and infestations (specifically, appendicitis and herpesvirus infection), serious neoplasms (basal cell carcinoma and melanoma, including melanoma in situ), and serious cardiac events (atrioventricular block 1st and 2nd degree).
2.3 Statistical methods
To assess the comparative safety and efficacy of ozanimod versus ponesimod, a MAIC (
Signorovitch et al., 2012
, - Signorovitch J.
- Erder M.H.
- Xie J.
- Sikirica V.
- Lu M.
- Hodgkins P.S.
- Wu E.Q.
Comparative effectiveness research using matching-adjusted indirect comparison: an application to treatment with guanfacine extended release or atomoxetine in children with attention-deficit/hyperactivity disorder and comorbid oppositional defiant disorder.
Signorovitch et al., 2010
) was conducted, adjusting for baseline differences between treatment groups. Thus, individual patients in the ozanimod trial were re-weighted so that their mean baseline characteristics exactly matched those reported for the ponesimod trial. The weight of each patient was equal to his or her estimated odds of enrollment in the ozanimod trial versus ponesimod trial, conditional on enrollment within either of the trial populations. A logistic regression model using the method of moments (Signorovitch et al., 2010
) was applied to estimate the weights for the propensity of enrollment in the ponesimod trial versus the ozanimod trial. Based on data availability and clinical considerations, cross-trial differences for the following baseline characteristics were adjusted, using the mean for continuous variables and percentages for categorical variables: age, sex (female), duration of MS since first symptom, relapses within previous 12 months before study entry, EDSS score, prior use of DMTs received within 2 years before randomization, absence of gadolinium-enhancing (GdE) T1 lesions, and region Eastern European (%).Before and after matching, the baseline characteristics and selected outcomes were described and compared between the ozanimod 1 mg arm (RADIANCE-B, n = 433 before matching) versus the ponesimod 20 mg arm (OPTIMUM, n = 567). Means and standard deviations were computed for continuous variables, while frequencies and percentages were computed for categorical variables. Differences in ARRs were presented as risk ratios (RRs) based on a negative binomial model. Safety outcomes were summarized as proportions and compared using risk differences (RDs). The standard error, 95% confidence interval (CI), and P value for the indirect comparisons were based on a robust estimate of the variance, based on a sandwich estimator, which accounts for the variability in the propensity score weights. Chi-square tests were used to compare binary variables before matching, and Wald tests were used after matching. Wald tests were used to compare continuous variables before and after matching. A P value of 0.05 was used to determine statistical significance. Because of the lack of a common comparator arm between both trials, the comparisons of 2-year outcomes were non-anchored.
3. Results
3.1 Assessment of cross-trial similarities and differences
All patients in the intention-to-treat population from RADIANCE-B (n = 433) were included in the analysis (see Appendix for details on RADIANCE-B patient eligibility criteria). Patients had largely comparable baseline characteristics. However, in addition to slight variations in baseline characteristics, there were differences in trial design, such as the trials’ active comparators, treatment duration, inclusion/exclusion criteria, and outcomes, which are discussed below.
3.1.1 Active comparators
The RADIANCE-B and SUNBEAM trials compared ozanimod with IFN ß-1a, while OPTIMUM compared ponesimod with teriflunomide.
3.1.2 Treatment duration
Duration of treatment was 24 months in the RADIANCE-B trial and 108 weeks (24 months and 21 days) in the OPTIMUM trial. Duration of ozanimod treatment was 12 months in SUNBEAM and 24 months in RADIANCE-B. Ponesimod treatment duration was 108 weeks (i.e., 24 months and 21 days). No 1-year outcomes were reported in the OPTIMUM trial. Therefore, comparisons that included SUNBEAM data were not possible, and only a 2-year comparison between RADIANCE-B and OPTIMUM (2-year) was feasible.
3.1.3 Inclusion/exclusion criteria
The OPTIMUM trial included patients with RRMS or secondary progressive MS with superimposed relapses (2.6% of patients). The RADIANCE-B trial included patients with MS who had baseline EDSS scores between 0 and 5.0 (vs. ≤5.5 in OPTIMUM). There were differences in exclusion criteria related to prior treatment and the time periods over which the treatments could be administered. Additionally, differences between exclusion criteria in terms of MS phenotype, contraindications, and comorbidities were observed.
3.1.4 Baseline characteristics
Duration of MS since first symptom was slightly longer for patients in the OPTIMUM trial than for patients in the RADIANCE-B trial (7.6 vs. 6.9 years, respectively); the difference was not statistically significant. A higher percentage of patients in OPTIMUM than in RADIANCE-B had previously received a DMT (37.6% vs. 28.4%, respectively), and a greater proportion of patients in RADIANCE-B than in OPTIMUM was from Eastern Europe (86.4% vs. 73.5%, respectively) (Table 1).
Table 1Non–Anchor-Based Comparison of 2-Year Safety Outcomes for Ozanimod 1 mg and Ponesimod 20 mg: Baseline Characteristics.
| Before Matching | After Matching | |||||
|---|---|---|---|---|---|---|
| RADIANCE-B | OPTIMUM | P value [A] vs. [B] | RADIANCE-B | OPTIMUM | P value [A] vs. [B] | |
| Baseline Characteristics | Ozanimod 1 mg n = 433 [A] | Ponesimod 20 mg n = 567 [B] | Ozanimod 1 mg ESS = 359 [A] | Ponesimod 20 mg n = 567 [B] | ||
| Age, years | 36.0 ± 8.9 | 36.7 ± 8.7 | 0.2 | 36.7 ± 9.0 | 36.7 ± 8.7 | 1.0 |
| Sex, female,% | 67.20 | 64.0 | 0.3 | 64.0 | 64.0 | 1.0 |
| Time since MS symptom onset, years | 6.9 ± 6.2 | 7.6 ± 6.8 | 0.1 | 7.6 ± 6.7 | 7.6 ± 6.8 | 1.0 |
| Patients relapsing in 12 months before study entry, mean ± SD | 1.3 ± 0.6 | 1.2 ± 0.6 | 0.1 | 1.2 ± 0.5 | 1.2 ± 0.6 | 1.0 |
| EDSS score, mean ± SD | 2.6 ± 1.1 | 2.6 ± 1.2 | 0.8 | 2.6 ± 1.2 | 2.6 ± 1.2 | 1.0 |
| Patients receiving DMTs in 2 years before randomization,% | 28.4 | 37.6 | <0.01* | 37.6 | 37.6 | 1.0 |
| Absence of GdE T1 lesions,% patients | 58.9 | 60.1 | 0.01 | 60.1 | 60.1 | 1.0 |
| Region Eastern Europe,% patients 1 Countries were classified as Eastern European based on the classification used in the RADIANCE-B trial (Cohen et al., 2019 (Cohen et al., 2019) [see Appendix for details]). Countries in the OPTIMUM trial not included in RADIANCE-B (e.g., Czech Republic) were classified as Eastern European based on the United Nations delineation of Geographic Regions (https://unstats.un.org/unsd/methodology/m49/#geo-regions). | 86.4 | 73.5 | <0.001* | 73.5 | 73.5 | 1.0 |
Abbreviations: DMT, disease-modifying therapy; EDSS, Expanded Disability Status Scale; ESS, effective sample size; GdE, gadolinium-enhancing; MS, multiple sclerosis; SD, standard deviation.
1 Countries were classified as Eastern European based on the classification used in the RADIANCE-B trial (
Cohen et al., 2019
(- Cohen J.A.
- Comi G.
- Selmaj K.W.
- Bar-Or A.
- Arnold D.L.
- Steinman L.
- Hartung H.P.
- Montalban X.
- Kubala Havrdová E.
- Cree B.A.C.
- Sheffield J.K.
- Minton N.
- Raghupathi K.
- Huang V.
- Kappos L.
Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (RADIANCE): a multicentre, randomised, 24-month, phase 3 trial.
Cohen et al., 2019
) [see Appendix for details]). Countries in the OPTIMUM trial not included in RADIANCE-B (e.g., Czech Republic) were classified as Eastern European based on the United Nations delineation of Geographic Regions (https://unstats.un.org/unsd/methodology/m49/#geo-regions).- Cohen J.A.
- Comi G.
- Selmaj K.W.
- Bar-Or A.
- Arnold D.L.
- Steinman L.
- Hartung H.P.
- Montalban X.
- Kubala Havrdová E.
- Cree B.A.C.
- Sheffield J.K.
- Minton N.
- Raghupathi K.
- Huang V.
- Kappos L.
Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (RADIANCE): a multicentre, randomised, 24-month, phase 3 trial.
3.1.5 Outcomes
In the ARR analysis, RADIANCE-B used a negative binomial regression model, (sensitivity analyses) adjusting for region (Eastern Europe vs rest of world), baseline age, and baseline number of GdE T1 lesions. OPTIMUM used a negative binomial regression model adjusting for prior use of DMTs, EDSS category, and number of relapses in the year before study entry.
In the BVL analysis, in RADIANCE-B, change from baseline brain volume was calculated using descriptive statistics. In OPTIMUM it was calculated as least squares mean change using a mixed model with linear time effect and covariates of EDSS strata (≤3.5 or >3.5), DMTs within the 2 years before randomization strata, GdE T1 lesions at baseline, and baseline brain volume.
In the RADIANCE-B trial, confirmed disability progression (CDP) was defined as a sustained worsening (increase) in EDSS score of at least 1.0. In the OPTIMUM trial, CDP was defined as an increase of at least 1.5 from a baseline EDSS score of 0, an increase of at least 1.0 from a baseline EDSS score of 1.0 to 5.0, or an increase of at least 0.5 from a baseline EDSS score of at least 5.5. Because these differences in definitions could introduce bias, this outcome was not included in the analysis.
The proportion of patients free of GdE T1 lesions was an outcome in the RADIANCE-B trial, and the cumulative number of new GdE T1 lesions was an outcome in the OPTIMUM trial. Also, there were no data reported for T2 lesions in OPTIMUM trial (T2 lesions are considered only as part of the composite outcome of cumulative number of combined unique active lesions). Therefore, lesion-related outcomes were not included in the analyses.
Despite the differences noted above, the trial designs were deemed to be highly similar and suitable for an adjusted cross-trial comparison.
3.2 Adjusted analysis
3.2.1 Baseline characteristics before and after matching
Before adjustment for cross-trial differences, demographics and disease severity were generally similar between trial populations (Table 1). However, two clinically relevant differences between patient populations receiving ozanimod and those receiving ponesimod were identified (Table 1): patients in RADIANCE-B, compared with those in OPTIMUM, were less likely to have received DMTs within the 2 years before randomization (28.4% vs. 37.6% [P<0.01]) and were more likely to be from Eastern Europe (86.4% vs. 73.5% [P<0.001]). After adjustment, baseline averages and percentages for all included patient characteristics were balanced between trials.
3.2.2 Adjusted analyses of 2-year outcomes
Overall, numerical differences in measurements of the efficacy outcomes favored ozanimod, and the matching adjustment had little impact on the estimated differences (Table 2). Compared with ponesimod treatment, ozanimod treatment was associated with a 19% and 20% decrease in ARR at 2 years before and after matching, respectively, although these differences in treatment effects were not significant. However, reductions in BVL were significant among patients treated with ozanimod before and after matching (RR: 0.20 [95% CI: 0.05, 0.35] before matching [P<0.01] and 0.20 [95% CI: 0.05, 0.36] after matching [P<0.01]) compared with patients treated with ponesimod. See additional details in Table A1 in the Appendix.
Table 2Comparison of 2-Year Efficacy Outcomes: Ozanimod 1 mg vs. Ponesimod 20 mg: Assessment of Risk Differences.
| Before Matching | After Matching | |||||
|---|---|---|---|---|---|---|
| Outcome | Δ | 95% CI | P value | Δ | 95% CI | P value |
| Annualized relapse rate, risk ratio 1 Ratio of ARRs for ozanimod vs. ponesimod. A value <1 indicates that the ozanimod group had fewer relapses than the ponesimod group. ARR estimates for ozanimod are based on a model that takes input data from the ozanimod arm only, whereas the ARR estimates for ponesimod are based on a model that takes input data from ponesimod and teriflunomide arms. RADIANCE-B used a negative binomial regression model adjusted for region (Eastern Europe vs rest of world), baseline age, and baseline number of GdE lesions (sensitivity analyses in the trial). OPTIMUM used a negative binomial regression model adjusted for prior use of DMTs, EDSS category, and the number of relapses in the year before study entry. | 0.81 | (0.57, 1.15) | 0.24 | 0.80 | (0.57, 1.10) | 0.17 |
| Percent change of brain volume from baseline to year 2 2 A positive value indicates that ozanimod group had less BVL than ponesimod group. In the RADIANCE-B trial, change from baseline was calculated using descriptive statistics. In the OPTIMUM trial, change from baseline was calculated as least squares mean change using a mixed model with linear time effect and covariates of EDSS strata (≤3.5 or >3.5), DMTs within the 2 years before randomization strata, GdE T1 lesions at baseline, and baseline brain volume. | 0.20 | (0.06, 0.35) | <0.01* | 0.20 | (0.05, 0.36) | < 0.01* |
Abbreviations: ARR, annualized relapse rate; BVL, brain volume loss; CI, confidence interval; DMT, disease-modifying therapy; EDSS, Expanded Disability Status Scale; GdE, gadolinium-enhancing.
1 Ratio of ARRs for ozanimod vs. ponesimod. A value <1 indicates that the ozanimod group had fewer relapses than the ponesimod group. ARR estimates for ozanimod are based on a model that takes input data from the ozanimod arm only, whereas the ARR estimates for ponesimod are based on a model that takes input data from ponesimod and teriflunomide arms. RADIANCE-B used a negative binomial regression model adjusted for region (Eastern Europe vs rest of world), baseline age, and baseline number of GdE lesions (sensitivity analyses in the trial). OPTIMUM used a negative binomial regression model adjusted for prior use of DMTs, EDSS category, and the number of relapses in the year before study entry.
2 A positive value indicates that ozanimod group had less BVL than ponesimod group. In the RADIANCE-B trial, change from baseline was calculated using descriptive statistics. In the OPTIMUM trial, change from baseline was calculated as least squares mean change using a mixed model with linear time effect and covariates of EDSS strata (≤3.5 or >3.5), DMTs within the 2 years before randomization strata, GdE T1 lesions at baseline, and baseline brain volume.
Safety outcomes were overall statistically comparable between both treatments (Table 3), but numerical differences tended to favor ozanimod. After adjustment, treatment with ozanimod was associated with a statistically significant lower risk of AE leading to discontinuation (RD: −6.1% [95% CI: −8.9%, −3.4%]), a statistically significant lower risk of any TEAE (−11.9% [95% CI: −16.8%, −7.0%]), and a statistically significant lower risk of absolute lymphocyte count <0.2 K/μL (−2.3% [95% CI: −4.2%, −0.5%]). Herpesvirus infections were less common among patients treated with ozanimod compared with ponesimod before matching, but the difference became statistically insignificant after matching. One safety outcome (incidence of basal cell carcinoma) had a change in sign before and after the adjustment, but neither value was statistically significantly different between treatments. See additional details in Table A1 in Appendix.
Table 3Comparison of 2-Year Safety Outcomes: Ozanimod 1 mg vs. Ponesimod 20 mg: Assessment of Risk Differences.
| Before Matching | After Matching | |||||
|---|---|---|---|---|---|---|
| Outcome | Δ | 95% CI | P value | Δ | 95% CI | P value |
| Death | 0 | – | – | 0 | – | – |
| AE leading to discontinuation | −5.7 | (−8.5, −2.9) | <0.001* | −6.1 | (−8.9, −3.4) | <0.001* |
| Any TEAE | −14.0 | (−18.8, −9.1) | <0.001* | −11.9 | (−16.8, −7.0) | <0.001* |
| Any SAE | −2.2 | (−5.5, 1.0) | 0.18 | −3.0 | (−6.2, 0.2) | 0.06 |
| Infections and infestations SAE | −0.3 | (−1.5, 1.0) | 0.67 | −0.4 | (−1.6, 0.9) | 0.58 |
| Appendicitis | −0.1 | (−0.9, 0.8) | 0.88 | −0.1 | (−0.9, 0.8) | 0.9 |
| Herpesvirus infection | −2.7 | (−4.9, −0.5) | <0.05* | −1.2 | (−4.2, 1.8) | 0.42 |
| Neoplasm SAE | ||||||
| Basal-cell carcinoma, SAE | −0.1 | (−0.8, 0.5) | 0.73 | 0.2 | (−1.0, 1.3) | 0.76 |
| Melanoma (including in situ), SAE | −0.2 | (−0.5, 0.2) | 0.31 | −0.2 | (−0.5, 0.2) | 0.31 |
| Cardiac SAE | 0 | – | – | 0 | – | – |
| Atrioventricular block 1st degree | −0.5 | (−1.1, 0.1) | 0.09 | −0.5 | (−1.1, 0.1) | 0.09 |
| Atrioventricular block 2nd degree | 0 | − | – | 0 | – | – |
| Macular edema | −0.9 | (−1.8, 0.1) | 0.08 | −0.9 | (−1.8, 0.1) | 0.09 |
| Absolute lymphocyte count <0.2 K/μL | −2.4 | (−4.1, −0.8) | <0.01* | −2.3 | (−4.2, −0.5) | <0.05* |
Abbreviations: AE, adverse event; CI, confidence interval; SAE, serious adverse event; TEAE, treatment-emergent adverse event.
1 A negative value indicates that ozanimod group had lower rates of safety events than ponesimod group.
2 In OPTIMUM trial, herpesvirus infection, basal cell carcinoma, melanoma, and macular edema were reported as adverse events of special interest.
3 In RADIANCE-B, absolute lymphocyte count <0.2 K/μL for patients with missing lymphocyte counts were assumed to be not available.
4. Discussion
Evidence of the comparative effectiveness of ponesimod and ozanimod in the treatment of RRMS is important for clinicians, payers, regulators, and other decision-makers to assess the relative value of these therapies. However, there are no head-to-head randomized data available that compare the efficacy and safety of outcomes with ponesimod and ozanimod. In the absence of these data, this study conducted indirect comparisons across clinical trials of ponesimod and ozanimod, adjusting for cross-trial differences between patient populations. This analysis is based on a comprehensive evaluation of cross-trial heterogeneity and potential sources of bias to ensure robust comparisons of treatment effects across trials and used MAIC methodology to control for important imbalances in the observed prognostic factors across the trials.
The use of a MAIC was supported by an assessment of the differences and similarities between trial designs and patient populations. The trials’ size, duration, inclusion/exclusion criteria, and mode of treatment administration were all found to be similar (ozanimod as a daily oral capsule and ponesimod as a daily oral tablet). Some differences were noted with respect to the methodologies used to estimate the selected outcomes (ARR and BVL), but the main differences identified in the trials were in the patient characteristics measured, specifically the percentage of patients receiving DMTs within 2 years before randomization and patients’ geographic representation. The latter differences were accounted for by the MAIC methodology, by weighting patients in the ozanimod arm to match the average baseline characteristics in the ponesimod arm.
The results of this MAIC analysis indicate that ozanimod and ponesimod are comparable in that the difference in ARR was nonsignificant. However, ozanimod is associated with numerical reductions in ARR at 2 years compared with ponesimod and with statistically significantly better prevention of BVL at 2 years than ponesimod. Ozanimod also has a favorable safety profile, as it is associated with significantly lower risks of key safety outcomes of interest compared with ponesimod, including AEs leading to discontinuation, any AEs, herpetic infections, bradycardia, and abnormal liver enzymes. All the numerical differences in safety outcomes were in the direction of greater safety associated with ozanimod versus ponesimod, except for basal cell carcinoma (but this difference was not statistically significant). These findings suggest that ozanimod has an overall improved benefit-risk profile compared with ponesimod.
This study was subject to some limitations. First, cross-trial differences were identified in trial design that cannot be fully controlled for in the MAIC (e.g., in exclusion criteria related to prior treatment and the time periods over which the treatments could be administered [see Appendix for details]).
Second, despite the MAIC adjusting for differences in observed baseline patient characteristics (e.g., ozanimod trial included a larger proportion of patients from the Eastern European region compared with the ponesimod trial), multiple unobserved differences may have confounded the analysis. Only a well-conducted, head-to-head randomized trial can avoid this limitation.
Third, there were also cross-trial differences in the definition of key efficacy outcomes (e.g., ARR, BVL, CDP). In particular, the differences in CDP imply a higher threshold to observe a disease progression in OPTIMUM, potentially leading to fewer events, which may introduce a bias in the analysis. Therefore, this outcome was not included in the analysis. Furthermore, each trial adjusted for different variables to derive the respective ARR. In RADIANCE-B, BVL was calculated using descriptive statistics, whereas in OPTIMUM, BVL was determined with least squares mean change in brain volume, using a mixed model with linear time effect and covariates of EDSS strata (≤3.5 or >3.5), DMTs within the 2 years before randomization strata, GdE T1 lesions at baseline, and baseline brain volume. These discrepant methodologic approaches may limit the comparability of some of the outcomes included in the analysis.
Fourth, the MAIC was based on a non-anchored comparison because of the lack of common comparator across trials, which lessens the opportunity to assess the presence of confounding bias (which would have been feasible by inspecting the differences between patient populations before and after matching in the common comparator arms) and limits our ability to address possible confounding bias via an anchor-based comparison.
Finally, the data used in these studies comes from clinical trial populations with specific patient selection criteria, which may not be representative of the broader RRMS population. As such, the results may not be generalizable to the real-world population. Despite these limitations, this real-world comparative effectiveness study provides decision-makers and clinicians further evidence on the clinical value of ozanimod and ponesimod for the treatment of RRMS.
5. Conclusions
This MAIC of clinical trial data suggests that ozanimod is associated with a favorable benefit-risk profile compared with ponesimod when considering key efficacy and safety outcomes over 2 years: patients treated with ozanimod had a numerically lower ARR and a significantly lower risk of BVL, as well as lower risk of TEAEs, AEs leading to treatment discontinuation, or absolute lymphocyte count <0.2 K/μL.
Author credits statement
Elyse Swallow: Conceptualization, Data acquisition, and Data interpretation. Timothy Pham: Conceptualization and Data interpretation. Oscar Patterson-Lomba: Conceptualization, Data analysis, and Data interpretation. Lei Yin: Data acquisition and Data analysis. Andres Gomez-Lievano: Data acquisition and Data analysis. Jingyi Liu: Data acquisition and Data analysis. Tom Tencer: Conceptualization and Data interpretation. Komal Gupte-Singh: Conceptualization and Data interpretation.
Disclosures
Role of the funding source
The study was supported by Bristol Myers Squibb (Princeton, NJ, USA). Bristol Myers Squibb had a role in the design of the study; analysis and interpretation of the data; review and approval of the manuscript; and decision to submit the manuscript for publication.
Data sharing statement
Bristol Myers Squibb policy on data sharing may be found at https://www.bms.com/researchers-and-partners/independent-research/data-sharing-request-process.html.
Declaration of Competing Interest
Elyse Swallow, Oscar Patterson-Lomba, Lei Yin, Andres Gomez-Lievano, and Jingyi Liu are employees of Analysis Group, which received funding from Bristol Myers Squibb for the conduct of this study.
Timothy Pham and Tom Tencer were employees of Bristol Myers Squibb at the time of this study and may be shareholders of the company.
Komal Gupte-Singh is an employee of Bristol Myers Squibb and may be a shareholder of the company.
Acknowledgements
Medical writing support provided by Loraine Georgy, PhD, of Analysis Group, Inc., and editorial support provided by Eleanor Bush, MA, of Peloton Advantage, LLC (Parsippany, NJ, USA), an OPEN Health company, were funded by Bristol Myers Squibb.
Appendix
Details on eligibility criteria in RADIANCE-B trial
- IInclusion criteria
- APatients in the RADIANCE-B trial who met the criteria for the OPTIMUM trial will be included in the comparison of ozanimod vs. ponesimod, including:
- 1Male or female patients aged 18 to 55 years
- 2Diagnosis of relapsing-remitting multiple sclerosis (RRMS), as defined by the 2010 revised McDonald criteria (Polman et al., 2011)
- 3At least (a) one documented relapse in the previous year before screening or (b) one documented relapse within the 24 months before screening and evidence of at least one gadolinium-enhancing (GdE) lesion on brain magnetic resonance imaging (MRI) in the 12 months before randomization
- 4Score between 0.0 and 5.0 on the Expanded Disability Status Scale (EDSS)
- 1
- A
- IIExclusion criteria
- APrior treatments, such as use of any investigational agent within 6 months before enrollment or receipt of a vaccine 4 weeks before randomization:
- 1Prior use of any investigational agent within 6 months of enrollment
- 2Receipt of a live vaccine within 4 weeks before randomization
- 3Systemic corticosteroid therapy or adrenocorticotropic hormone use within 30 days before screening
- 4Intolerance of or contraindication to oral or intravenous corticosteroids
- 5Previous treatment with lymphocyte-depleting therapies
- 6Non–lymphocyte-depleting disease-modifying MS agents must be discontinued from signing of informed consent until randomization
- 7Prior treatment with lymphocyte trafficking blockers (e.g., natalizumab, ponesimod, other S1P1 receptor modulators)
- 8Treatment with other disease-modifying therapies (e.g., dimethyl fumarate, teriflunomide, daclizumab, laquinimod) within 3 months before randomization
- 9Treatment with other immunosuppressant agents such as azathioprine, cyclosporine, methotrexate, or mycophenolate within 6 months before randomization
- 10Treatment with intravenous immune globulin, plasmapheresis, within 3 months before randomization
- 11Use of therapies that are not allowed based on metabolism by the cytochrome P450 3A4 isoenzyme within 4 weeks before randomization
- 12Treatment with medications that have a known impact on the cardiac conduction system are excluded
- 1
- BPrimary progressive MS at screening or disease duration of more than 15 years in patients with an EDSS ≤2.0
- 1Contraindications to MRI or gadolinium contrast, such as known allergy to gadolinium contrast dyes, renal insufficiency, claustrophobia, body size incompatible with the scanner, pacemaker, cochlear implants, or intracranial vascular clips
- 1
- CPregnancy and breastfeeding, and clinically relevant cardiovascular conditions or other relevant diseases that could affect the implementation or interpretation of the trial or put the patient at risk
- A
Treatment-arm and Comparison of Efficacy and Safety Outcomes for Ozanimod (weighted) and Ponesimod
Table A12-Year Efficacy and Safety Outcomes: Ozanimod 1 mg vs. Ponesimod 20 mg.
| Outcome | Measure | Weighted Ozanimod 1 mg n = 433 [A] | Ponesimod 20 mg n = 567 [B] | MAIC [A] vs. [B] | MAIC 95% CI | |
|---|---|---|---|---|---|---|
| Efficacy Outcomes | ||||||
| ARR 1 ARR estimates for ozanimod are based on a model that takes as input data from the ozanimod arm only, whereas the ARR estimates for ponesimod are based on a model that takes as input data from ponesimod and teriflunomide arms. The RADIANCE-B trial used a negative binomial regression model adjusted for region (Eastern Europe vs rest of world), baseline age, and baseline number of GdE lesions (sensitivity analyses in the trial). The OPTIMUM trial used a negative binomial regression model adjusted for prior use of DMTs, EDSS category and the number of relapses in the year before study entry. | RR | 0.16 | 0.20 | 0.80 | (0.57, 1.10) | |
| Percent change of brain volume from baseline to year 2 2 In the RADIANCE-B trial, change from baseline in brain volume was calculated using descriptive statistics. In the OPTIMUM trial, change from baseline was calculated as least squares mean change in brain volume using a mixed model with linear time effect and covariates of EDSS strata (≤3.5 or >3.5), DMTs within the 2 years before randomization strata, GdE T1 lesions at baseline, and baseline brain volume. | RD | −0.71 | −0.91 | 0.20 | (0.05, 0.36) | |
| Safety Outcomes, % | ||||||
| Death | RD | 0.00 | 0.00 | 0.0 | – | |
| AE leading to discontinuation | RD | 2.58 | 8.70 | −6.1 | (–8.9, –3.4) | |
| Any TEAE | RD | 76.89 | 88.80 | −11.9 | (–16.8, –7.0) | |
| Any SAE | RD | 5.68 | 8.70 | −3.0 | (–6.2, 0.2) | |
| Infections and infestations SAE | RD | 0.85 | 1.20 | −0.4 | (–1.6, 0.9) | |
| Appendicitis | RD | 0.47 | 0.53 | −0.1 | (–0.9, 0.8) | |
| Herpesvirus infection | RD | 3.56 | 4.80 | −1.2 | (–4.2, 1.8) | |
| Neoplasm SAE | ||||||
| Basal cell carcinoma, SAE | RD | 0.52 | 0.35 | 0.2 | (–1.0, 1.3) | |
| Melanoma (including in situ), SAE | RD | 0.00 | 0.18 | −0.2 | (–0.5, 0.2) | |
| Cardiac SAE | RD | 0.00 | 0.00 | 0 | – | |
| Atrioventricular block 1st degree | RD | 0.00 | 0.50 | −0.5 | (–1.1, 0.1) | |
| Atrioventricular block 2nd degree | RD | 0.00 | 0.00 | 0 | – | |
| Macular edema | RD | 0.25 | 1.10 | −0.9 | (−1.8, 0.1) | |
| Absolute lymphocyte count <0.2 K/uL | RD | 0.87 | 3.20 | −2.3 | (−4.2, –0.5) | |
Abbreviations: ARR, annualized relapse rate; CI, confidence interval; DMT, disease-modifying therapy; EDSS, Expanded Disability Status Scale; GdE, gadolinium-enhancing; RD, risk difference; RR, risk ratio; MAIC, matching-adjusted indirect comparison; SAE, serious adverse events; TEAE, treatment-emergent adverse events.
1 ARR estimates for ozanimod are based on a model that takes as input data from the ozanimod arm only, whereas the ARR estimates for ponesimod are based on a model that takes as input data from ponesimod and teriflunomide arms. The RADIANCE-B trial used a negative binomial regression model adjusted for region (Eastern Europe vs rest of world), baseline age, and baseline number of GdE lesions (sensitivity analyses in the trial). The OPTIMUM trial used a negative binomial regression model adjusted for prior use of DMTs, EDSS category and the number of relapses in the year before study entry.
2 In the RADIANCE-B trial, change from baseline in brain volume was calculated using descriptive statistics. In the OPTIMUM trial, change from baseline was calculated as least squares mean change in brain volume using a mixed model with linear time effect and covariates of EDSS strata (≤3.5 or >3.5), DMTs within the 2 years before randomization strata, GdE T1 lesions at baseline, and baseline brain volume.
3 In the OPTIMUM trial, herpesvirus infection, basal cell carcinoma, melanoma, and macular edema were reported as adverse events of special interest.
4 In the RADIANCE-B trial, absolute lymphocyte count <0.2 K/uL for patients with missing lymphocyte counts were assumed to be not available.
References
- Quality of life among persons with multiple sclerosis and their caregivers.Neurology. 1997; 48: 74-80https://doi.org/10.1212/wnl.48.1.74
- Estimating the comparative efficacy of cladribine tablets versus alternative disease modifying treatments in active relapsing-remitting multiple sclerosis: adjusting for patient characteristics using meta-regression and matching-adjusted indirect treatment comparison approaches.Curr. Med. Res. Opin. 2019; 35: 1371-1378https://doi.org/10.1080/03007995.2019.1585779
- Comparative efficacy and safety of ozanimod and dimethyl fumarate for relapsing-remitting multiple sclerosis using matching-adjusted indirect comparison.CNS Drugs. 2021; 35: 795-804https://doi.org/10.1007/s40263-021-00805-0
- Matching-adjusted indirect treatment comparison of ozanimod versus teriflunomide for relapsing multiple sclerosis.Mult. Scler. Relat. Disord. 2021; 52102972https://doi.org/10.1016/j.msard.2021.102972
- Oral fingolimod or intramuscular interferon for relapsing multiple sclerosis.N. Engl. J. Med. 2010; 362: 402-415https://doi.org/10.1056/NEJMoa0907839
- Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (RADIANCE): a multicentre, randomised, 24-month, phase 3 trial.Lancet Neurol. 2019; 18: 1021-1033https://doi.org/10.1016/S1474-4422(19)30238-8
- Ozanimod demonstrates efficacy and safety in a phase 3 trial of relapsing multiple sclerosis (SUNBEAM) [abstract 232].Mult. Scler. J. 2017; 23: 73-74
- Safety and efficacy of ozanimod versus interferon beta-1a in relapsing multiple sclerosis (SUNBEAM): a multicentre, randomised, minimum 12-month, phase 3 trial.Lancet Neurol. 2019; 18: 1009-1020https://doi.org/10.1016/S1474-4422(19)30239-X
- Multiple sclerosis, a treatable disease.Clin. Med. (Lond). 2017; 17: 530-536https://doi.org/10.7861/clinmedicine.17-6-530
- Guideline On Clinical Investigation of Medicinal Products For the Treatment of Multiple Sclerosis.2015 (<) (>accessed July 5, 2021)
- Multiple sclerosis: overview of disease-modifying agents.Perspect. Medicin. Chem. 2014; 6: 65-72https://doi.org/10.4137/PMC.S13213
- Comparative effectiveness using a matching-adjusted indirect comparison between delayed-release dimethyl fumarate and fingolimod for the treatment of multiple sclerosis.Curr. Med. Res. Opin. 2017; 33: 175-183https://doi.org/10.1080/03007995.2016.1248380
- Comparative effectiveness using a matching-adjusted indirect comparison between delayed-release dimethyl fumarate and fingolimod for the treatment of relapsing-remitting multiple sclerosis.Value Health. 2015; 18: A750https://doi.org/10.1016/j.jval.2015.09.2902
- An update on immunopathogenesis, diagnosis, and treatment of multiple sclerosis.Brain Behavior. 2015; 5: e00362https://doi.org/10.1002/brb3.362
- Brain health: time matters in multiple sclerosis.Mult. Scler. Relat. Disord. 2016; 9: S5-S48https://doi.org/10.1016/j.msard.2016.07.003
- Inclusion of brain volume loss in a revised measure of 'no evidence of disease activity' (NEDA-4) in relapsing-remitting multiple sclerosis.Mult. Scler. 2016; 22: 1297-1305https://doi.org/10.1177/1352458515616701
- Ponesimod compared with teriflunomide in patients with relapsing multiple sclerosis in the active-comparator phase 3 OPTIMUM study: a randomized clinical trial.JAMA Neurol. 2021; 78: 558-567https://doi.org/10.1001/jamaneurol.2021.0405
- Effect of relapses on development of residual deficit in multiple sclerosis.Neurology. 2003; 61: 1528-1532https://doi.org/10.1212/01.wnl.0000096175.39831.21
- Defining the clinical course of multiple sclerosis: the 2013 revisions.Neurology. 2014; 83: 278-286https://doi.org/10.1212/WNL.0000000000000560
- Multiple Sclerosis.2022 (<) (>accessed January 31, 2022)
- Clinical presentation, course, and prognosis of multiple sclerosis in adults.2022 (Accessed July 9, 2019)
- Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria.Ann. Neurol. 2011; 69: 292-302
- Correlation between brain volume loss and clinical and MRI outcomes in multiple sclerosis.Neurology. 2015; 84: 784-793https://doi.org/10.1212/WNL.0000000000001281
- Comparative effectiveness research using matching-adjusted indirect comparison: an application to treatment with guanfacine extended release or atomoxetine in children with attention-deficit/hyperactivity disorder and comorbid oppositional defiant disorder.Pharmacoepidemiol Drug Saf. 2012; 21: 130-137https://doi.org/10.1002/pds.3246
- Comparative effectiveness without head-to-head trials: a method for matching-adjusted indirect comparisons applied to psoriasis treatment with adalimumab or etanercept.Pharmacoeconomics. 2010; 28: 935-945https://doi.org/10.2165/11538370-000000000-00000
- Novel agents for relapsing forms of multiple sclerosis.Annu. Rev. Med. 2016; 67: 309-321https://doi.org/10.1146/annurev-med-052814-023415
- Comparative safety and efficacy of ozanimod versus fingolimod for relapsing multiple sclerosis.J. Comp. Eff. Res. 2020; 9: 275-285https://doi.org/10.2217/cer-2019-0169
- Prescribing for multiple sclerosis patients in general practice: a case-control study.J. Clin. Pharm. Ther. 2001; 26: 437-444https://doi.org/10.1046/j.1365-2710.2001.00377.x
- The natural history of brain volume loss among patients with multiple sclerosis: a systematic literature review and meta-analysis.J. Neurol. Sci. 2015; 357: 8-18https://doi.org/10.1016/j.jns.2015.07.014
- Optimizing treatment success in multiple sclerosis.J. Neurol. 2016; 263: 1053-1065https://doi.org/10.1007/s00415-015-7986-y
Article info
Publication history
Published online: February 06, 2023
Accepted:
February 3,
2023
Received in revised form:
January 10,
2023
Received:
September 23,
2022
Identification
Copyright
© 2023 The Authors. Published by Elsevier B.V.
User license
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
