Elicit: Ocrelizumab vs Ofatumumab: Relapse and Disability Outcomes (public)
Ocrelizumab vs Ofatumumab: Relapse and Disability Outcomes (public)
Research reportView only
Create alertChat
September 4, 2025
What is the annualized relapse rate and confirmed disability progression with ocrelizumab vs ofatumumab in registry data?
In registry studies, ocrelizumab and ofatumumab show comparable annualized relapse rates (0.059 vs 0.038, p = 0.185) and both achieve 0% confirmed disability progression at 12 months.
Abstract
Registry studies in multiple sclerosis report that ocrelizumab and ofatumumab yield comparable outcomes. In one large, multicenter, propensity‐matched study (Zanghì et al., 2024), the annualized relapse rate was 0.059 with ocrelizumab versus 0.038 with ofatumumab (p = 0.185), and both treatments showed 0% confirmed disability progression over 12 months. A small post–natalizumab cohort (Cunha et al., 2025) likewise observed no significant change in relapse rate or in Expanded Disability Status Scale scores between ocrelizumab and ofatumumab. Other registry reports of ocrelizumab document relapse rates ranging from 0.014 to 0.11, with very low rates of confirmed disability progression over follow-up intervals from 6 months to several years.
These findings indicate that, based on available registry data, ocrelizumab and ofatumumab produce similar outcomes in annualized relapse rate and short-term disability progression among patients with multiple sclerosis.
Methods
We analyzed 40 sources from an initial pool of 999, using 8 screening criteria. Each paper was reviewed for 7 key aspects that mattered most to the research question. More on methods
Papers identified with Elicit search
n = 999
Papers screened using: Multiple Sclerosis Population, Target Medications, Real-World Data Source, Relevant Outcomes, Appropriate Study Design, MS-Focused Study, Adequate Study Type, Adult or Mixed Population
n = 999
Papers screened out
n = 959
Papers included for extraction
n = 40
Press enter or space to select a node.You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel.
Press enter or space to select an edge. You can then press delete to remove it or escape to cancel.
Paper search
Using your research question “What is the annualized relapse rate and confirmed disability progression with ocrelizumab vs ofatumumab in registry data?”, we searched across over 126 million academic papers from the Semantic Scholar corpus. We retrieved the 999 papers most relevant to the query.
Screening
We screened in sources based on their abstracts that met these criteria:
- Multiple Sclerosis Population: Does this study include patients with multiple sclerosis (any subtype)?
- Target Medications: Does this study report outcomes for patients treated with ocrelizumab and/or ofatumumab, allowing for direct or indirect comparison between these treatments?
- Real-World Data Source: Does this study utilize registry data, real-world databases, or observational cohorts (rather than randomized controlled trials)?
- Relevant Outcomes: Does this study report annualized relapse rate and/or confirmed disability progression?
- Appropriate Study Design: Is this study an observational study (cohort study, case-control study), systematic review, or meta-analysis?
- MS-Focused Study: Is this study NOT focusing exclusively on conditions other than multiple sclerosis?
- Adequate Study Type: Is this study NOT a case report, case series, editorial, letter, or conference abstract?
- Adult or Mixed Population: Is this study NOT focusing exclusively on pediatric populations?
We considered all screening questions together and made a holistic judgement about whether to screen in each paper.
Data extraction
We asked a large language model to extract each data column below from each paper. We gave the model the extraction instructions shown below for each column.
- Study Design:
Identify and extract the specific type of study design. Look in the methods section for details such as:
- Retrospective or prospective
- Observational (registry, cohort) or interventional
- Multicenter or single-center
- Matching method used (e.g., propensity score matching)
If multiple design characteristics are present, list all. If unclear, note “design not clearly specified”. Prioritize the most specific description of the study design.
- Patient Population Characteristics:
Extract key patient demographic and clinical characteristics:
- Total number of patients
- Mean/median age
- Gender distribution
- Multiple sclerosis type (RRMS, PPMS, aSPMS)
- Mean disease duration
- Baseline Expanded Disability Status Scale (EDSS) score
- Treatment-naive status vs. previous treatment history
If ranges or confidence intervals are provided, include those. If data is missing for any characteristic, note “not reported”.
- Interventions Compared:
List the specific treatments being compared:
- Drug names
- Dosage (if specified)
- Treatment duration
- Number of patients in each treatment group
If multiple comparisons were made, list all. Ensure the interventions match the research question (ocrelizumab vs ofatumumab).
- Annualized Relapse Rate (ARR):
Extract ARR data:
- Baseline ARR
- On-treatment ARR
- Confidence intervals
- Statistical significance of differences between groups
Record the exact numerical values and confidence intervals. If ARR is reported differently (e.g., percentage of relapse-free patients), note this explicitly.
- Disability Progression:
Extract confirmed disability progression data:
- Method of measuring disability progression (e.g., 12-week confirmed disability progression)
- Percentage of patients with disability progression
- Changes in EDSS score
- Statistical significance of differences between groups
If multiple time points are reported, prioritize the primary follow-up period specified in the study.
- Follow-up Duration:
Record:
- Total follow-up period
- Median follow-up time
- Range of follow-up times
If multiple follow-up periods are reported, note the primary analysis period. If not clearly specified, note “follow-up duration not clearly reported”.
- Safety Outcomes:
Extract information on:
- Treatment-related side effects
- Infection rates
- Infusion-related reactions
- Any serious adverse events
If no safety issues are reported, explicitly note “no significant safety concerns reported”. Quantify side effects if possible (percentage of patients affected).
Results
Characteristics of Included Studies
Study
Study Design
Registry Source
Population Size
Follow-up Duration
Full text retrieved
Zanghì et al., 2024
Retrospective cohort, multicenter, propensity-matched
Italian multiple sclerosis centers
396 (ocrelizumab: 216, ofatumumab: 180)
Mean 13.2 months
Yes
Zhu et al., 2022
Retrospective, multicenter, inverse probability of treatment weighting, registry
MSBase
1045
ocrelizumab: median 1.8 years; cladribine: 1.1 years; natalizumab: 3.6 years
Yes
Ghajarzadeh et al., 2025
Systematic review/meta-analysis
Multiple
No mention found
No mention found
No
Cunha et al., 2025
Retrospective cohort
No mention found
59
No mention found
No
Epstein et al., 2021
Retrospective chart review
No mention found
56
2 years
No
Ellwardt et al., 2020
Retrospective, multicenter, observational
3 neurology centers
210
Median 200 days (range 30–1674)
No
Cellerino et al., 2021
Prospective, single-center, observational
No mention found
153
Median 1.9 years (1.3–2.7)
Yes
Muros-Le Rouzic et al., 2024
Retrospective, registry, multicenter, propensity score matching
OPERA, NTD
611 (OPERA), 7141 (NTD)
5.5 years
No
Zhu et al., 2023
Retrospective, multicenter, inverse probability of treatment weighting, registry
MSBase
1386
Median 2.7 years
Yes
Schuldesz et al., 2025
Prospective, single-center, cohort
No mention found
98
2 years
No
toof
Pageof
View 30 more rows
Summary of study characteristics:
- Study design:
- 28 studies were retrospective.
- 7 studies were prospective.
- 2 studies were systematic reviews or meta-analyses.
- 19 studies were multicenter; 11 were single-center or monocentric.
- 14 studies were described as observational; 9 as cohort studies.
- 34 studies used registry-based data.
- 9 studies used propensity score matching or similar methods; 3 used inverse probability of treatment weighting.
- Other designs included randomized controlled trial/open-label extension (1), chart review (2), cross-sectional (1), post-authorization (1), and Bayesian propensity score matching (1).
- Registry source:
- 24 studies used a named registry, most commonly MSBase (8 studies), NTD (2), CONFIDENCE (2), OPERA (2), MENACTRIMS (1), DMSR (1), Kuwait (1), Danish (1), and Bari (1).
- 4 studies used multiple or unspecified registries.
- No registry source was found for 12 studies.
- Countries/regions represented include Germany (at least 5 studies), Italy (at least 3), United Kingdom (at least 2), France, Denmark, Turkey, Australia, and Kuwait.
- Population size:
- 8 studies included fewer than 100 participants.
- 14 studies included 100–499 participants.
- 6 studies included 500–999 participants.
- 10 studies included 1000 or more participants.
- No mention of population size was found for 5 studies.
- Follow-up duration:
- 6 studies had follow-up durations of less than 1 year.
- 8 studies had follow-up durations of 1–2 years.
- 15 studies had follow-up durations of 2–5 years.
- 4 studies had follow-up durations longer than 5 years.
- No mention of follow-up duration was found for 7 studies.
Effects
Annualized Relapse Rate
Study
Treatment
Annualized Relapse Rate (95% CI)
Study Population
Effect Size
Zanghì et al., 2024
ocrelizumab: 0.059; ofatumumab: 0.038 (p=0.185)
Relapsing multiple sclerosis
No significant difference
Zhu et al., 2022
ocrelizumab: 0.07 (0.04–0.13); natalizumab: 0.11 (0.09–0.14); cladribine: 0.25 (0.12–0.57)
Relapsing-remitting multiple sclerosis post-fingolimod
ocrelizumab superior to cladribine, lower than natalizumab
Cunha et al., 2025
rituximab: 0.08 (post-switch); ocrelizumab/ofatumumab: no significant change
Post-natalizumab
No significant change for ocrelizumab/ofatumumab
Boz et al., 2023
ocrelizumab: 0.08 (0.06–0.11); natalizumab: 0.09 (0.07–0.12); fingolimod: 0.17 (0.15–0.19)
Relapsing-remitting multiple sclerosis
ocrelizumab/natalizumab similar, both superior to fingolimod
Zhu et al., 2023
ocrelizumab: 0.06 (0.04–0.08); fingolimod: 0.26 (0.12–0.48); dimethyl fumarate: 0.27 (0.12–0.56)
Relapsing-remitting multiple sclerosis post-natalizumab
ocrelizumab superior to both
Roos et al., 2024
ocrelizumab: 0.05; cladribine: 0.09; natalizumab: 0.06; fingolimod: 0.12; alemtuzumab: 0.04
Relapsing-remitting multiple sclerosis
ocrelizumab superior to cladribine/fingolimod, similar to natalizumab/alemtuzumab
Roos et al., 2022
ocrelizumab: 0.08 vs interferon: 0.27; ocrelizumab: 0.03 vs fingolimod: 0.14; ocrelizumab: 0.06 vs natalizumab: 0.08
Relapsing-remitting multiple sclerosis
ocrelizumab superior to interferon/fingolimod, similar to natalizumab
Axhausen et al., 2025
ocrelizumab/ofatumumab: no mention found
Active multiple sclerosis
Effectiveness comparable
Buttmann et al., 2025
ocrelizumab: 0.11 (SD 0.31)
Relapsing multiple sclerosis
Annualized relapse rate declines over 5 years
Guerra et al., 2025
ocrelizumab: 0.02 (0.004–0.062) (year 2)
Relapsing-remitting/primary progressive/secondary progressive multiple sclerosis
Annualized relapse rate drops from 0.61 pre-treatment
toof
Pageof
View 7 more rows
Summary of annualized relapse rate findings:
- Range of annualized relapse rate for ocrelizumab:0.014 to 0.11 across 12 studies reporting this outcome.
- Direct head-to-head comparisons:9 studies included direct comparisons between ocrelizumab, ofatumumab, or rituximab and other disease-modifying therapies.
- Within-group changes:5 studies reported a drop in annualized relapse rate after switching to ocrelizumab, with pre-treatment rates as high as 0.8 and post-treatment rates as low as 0.014.
- Comparative effectiveness:
- Ocrelizumab was reported as superior to at least one comparator in 5 studies (most often fingolimod or cladribine).
- Ocrelizumab was reported as similar in effectiveness to comparators (most often natalizumab or alemtuzumab) in 7 studies.
- Ocrelizumab was reported as inferior to a comparator in 1 study (lower than natalizumab in Zhu et al., 2022).
- Ofatumumab data:Only 1 study (Zanghì et al., 2024) reported annualized relapse rate for ofatumumab, finding no significant difference compared to ocrelizumab.
- Rituximab data:1 study (Cunha et al., 2025) reported annualized relapse rate for rituximab post-switch.
- Comparators:Natalizumab and fingolimod were the most common comparators, each included in 5–6 studies.
- Reporting gaps:5 studies did not mention annualized relapse rate or comparative effectiveness for ocrelizumab, ofatumumab, or rituximab.
Confirmed Disability Progression
Study
Treatment
Confirmed Disability Progression Rate
Time to Progression
Statistical Significance
Zanghì et al., 2024
ocrelizumab/ofatumumab: 0%
12 months
No difference
Zhu et al., 2022
ocrelizumab/natalizumab: no mention found
No mention found
No significant difference
Cunha et al., 2025
ocrelizumab/ofatumumab: no significant change; rituximab: Expanded Disability Status Scale increased (p=0.022)
No mention found
No significant difference for ocrelizumab/ofatumumab
Boz et al., 2023
ocrelizumab/natalizumab/fingolimod: no mention found
1–2 years
No significant difference
Zhu et al., 2023
ocrelizumab: 48 events; fingolimod: 184 events
No mention found
Fingolimod higher risk than ocrelizumab (hazard ratio 1.49, p=0.02)
Roos et al., 2024
ocrelizumab: hazard ratio 0.45 (0.26–0.78) vs cladribine
No mention found
ocrelizumab lower risk than cladribine
Axhausen et al., 2025
ocrelizumab/ofatumumab: no mention found
No mention found
Comparable effectiveness
Buttmann et al., 2025
ocrelizumab: no mention found
5 years
No mention found
Guerra et al., 2025
ocrelizumab: 26.4% (year 2) → 7.85% (year 4)
4 years
Significant only in secondary progressive multiple sclerosis
Santiago-Setien et al., 2023
ocrelizumab: 0%
96 weeks
No difference
toof
Pageof
View 5 more rows
Summary of confirmed disability progression findings:
- Studies with ocrelizumab as a treatment group:7 studies.
- Studies with ofatumumab alone:1 study.
- Studies with ocrelizumab/ofatumumab combined:3 studies.
- Other treatments compared:rituximab (1 study), natalizumab (2 studies, plus 1 ocrelizumab/natalizumab), fingolimod (2 studies), cladribine (1 study).
- Numerical confirmed disability progression rates or event counts:Found in 6 studies, with reported values including 0% (2 studies), 48 events (ocrelizumab), 184 events (fingolimod), 26.4% (year 2) → 7.85% (year 4), standard interval dosing 8.3%, extended interval dosing 10.9%, and for subtypes: relapsing-remitting multiple sclerosis >97% confirmed disability progression-free, secondary progressive multiple sclerosis 48.9%, primary progressive multiple sclerosis 57.2%.
- Time to progression or follow-up duration:Reported in 9 studies, ranging from 6 months to 5 years. The most common time points were 12 months (2 studies) and 5 years (2 studies).
- Statistical significance of confirmed disability progression or progression comparisons:
- 8 studies reported no significant difference or comparable effectiveness between groups.
- 3 studies reported a significant difference, with ocrelizumab associated with lower risk than fingolimod (1 study), lower risk than cladribine (1 study), and a significant effect only in secondary progressive multiple sclerosis (1 study).
- No mention of statistical significance was found in 4 studies.
Comparative Effectiveness
Direct registry-based comparisons of ocrelizumab and ofatumumab:
- Zanghì et al., 2024 (large, multicenter, propensity-matched study): No significant difference in annualized relapse rate or confirmed disability progression between ocrelizumab and ofatumumab over 12 months.
- Axhausen et al., 2025: Lateral switching between ocrelizumab and ofatumumab did not reduce effectiveness.
- Cunha et al., 2025 (small cohort post-natalizumab): No significant difference in annualized relapse rate or Expanded Disability Status Scale change for ocrelizumab and ofatumumab.
- The evidence base for ofatumumab is much smaller and less mature than for ocrelizumab, with limited sample sizes and shorter follow-up.
Comparative findings for ocrelizumab:
- In these studies, ocrelizumab was reported to be superior to fingolimod and cladribine for annualized relapse rate, and similar to natalizumab and alemtuzumab.
- Confirmed disability progression rates were low and similar between ocrelizumab and natalizumab, and lower than fingolimod.
- Safety profiles were not systematically reported; we did not find mention of new safety signals in the available studies.
Registry Data Considerations
- Several studies of ocrelizumab included large sample sizes and long follow-up (for example, CONFIDENCE and MSBase registries).
- The evidence base for ofatumumab is limited, with only a few small, short-term studies.
- Observational design, potential confounding, and differences in patient populations, prior treatment history, and follow-up duration may affect comparability across studies.
- The diversity of registry cohorts supports generalizability of findings for ocrelizumab, but the limited data for ofatumumab restricts conclusions about its comparative effectiveness in broader populations.
References
No authors found (2022).ePosters. European Journal of Neurology
Download BIBDownload RISDownload TXT
Report
Status
Gather sources
999 sources found
Details
Screen sources
40 sources included
Details
Extract data
280 data points extracted
Details
Generate report
Save PDF
BIBLaTeX, ZoteroRISZotero, MendeleyTXTAPA BibliographyPDFPDFDOCXMicrosoft Word
Chat
Got some follow-up questions?
Sign up or sign in to chat with this report.
Back
Comparing switch to ocrelizumab, cladribine or natalizumab after fingolimod treatment cessation in multiple sclerosis
Chao Zhu, Zhen Zhou, I. Roos, Daniele Merlo, T. Kalincik, S. Ozakbas, O. Skibina, J. Kuhle, S. Hodgkinson, C. Boz, R. Alroughani, J. Lechner-Scott, M. Barnett, G. Izquierdo, A. Prat, D. Horáková, E. Kubala Havrdová, R. Macdonell, F. Patti, S. Khoury, M. Slee, R. Karabudak, M. Onofrj, V. van Pesch, J. Prevost, M. Monif, V. Jokubaitis, A. van der Walt, H. Butzkueven
Journal of Neurology Neurosurgery & Psychiatry·
2022·
17 citations
SourceDOI
Plain textPDFFigures
Searching for PDF
Unable to find PDF from source
Back
Study Design
Retrospective, observational, multicenter cohort study using inverse-probability-treatment-weighting (IPTW) and propensity scores (PS)
Patient Population Characteristics
- Total number of patients: 1045 - Mean age: 40.6 years (SD: 10.4) - Gender distribution: Female: 780 (74.6%) - Multiple sclerosis type: RRMS - Mean disease duration: 10.2 years (IQR: 5.7-15.4) - Baseline EDSS score: Median: 3 (IQR: 2-5) - Treatment-naive status vs. previous treatment history: Median previous therapies (excluding fingolimod): 1 (IQR: 1-2)
Interventions Compared
- Ocrelizumab: 600 mg every 24 weeks, n=445, median follow-up period: 1.8 years - Cladribine: 3.5 mg/kg total dose orally in two courses, n=76, median follow-up period: 1.1 years - Natalizumab: 300 mg intravenously every 4 weeks, n=524, median follow-up period: 3.6 years
Annualized Relapse Rate (ARR)
- Baseline ARR: Not mentioned - On-treatment ARR: - Ocrelizumab: 0.07 (95% CI, 0.04 to 0.13) - Natalizumab: 0.11 (95% CI, 0.09 to 0.14) - Cladribine: 0.25 (95% CI, 0.12 to 0.57) - Statistical significance of differences between groups: - Ocrelizumab vs. Natalizumab: ARR ratio 0.67 (95% CI, 0.47 to 0.96), statistically significant reduction - Cladribine vs. Natalizumab: ARR ratio 2.31 (95% CI, 1.30 to 4.10), statistically significant increase
Disability Progression
- Method of measuring disability progression: Confirmed increase in EDSS of ≥0.5 steps for patients with a baseline score >5.5, or ≥1.0 step for those with a baseline score between 1.0 and 5.5, and ≥1.5 steps if the baseline score was 0. - Percentage of patients with disability progression: Not mentioned. - Changes in EDSS score: Not mentioned. - Statistical significance of differences between groups: No significant difference in disability progression between ocrelizumab and natalizumab users.
Follow-up Duration
- Total follow-up period: Not explicitly stated - Median follow-up time: - Ocrelizumab: 1.8 years (IQR: 1.2-2.4 years) - Cladribine: 1.1 years (IQR: 0.9-1.8 years) - Natalizumab: 3.6 years (IQR: 1.9-5.3 years) - Range of follow-up times: - Ocrelizumab: 1.2-2.4 years - Cladribine: 0.9-1.8 years - Natalizumab: 1.9-5.3 years
Safety Outcomes
no significant safety concerns reported (specific details on treatment-related side effects, infection rates, infusion-related reactions, and serious adverse events not provided)
Background To compare the effectiveness and treatment persistence of ocrelizumab, cladribine and natalizumab in patients with relapsing–remitting multiple sclerosis switching from fingolimod. Methods Using data from MSBase registry, this multicentre cohort study included subjects who had used fingolimod for ≥6 months and then switched to ocrelizumab, cladribine or natalizumab within 3 months after fingolimod discontinuation. We analysed relapse and disability outcomes after balancing covariates using an inverse-probability-treatment-weighting method. Propensity scores for the three treatments were obtained using multinomial-logistic regression. Due to the smaller number of cladribine users, comparisons of disability outcomes were limited to natalizumab and ocrelizumab. Results Overall, 1045 patients switched to ocrelizumab (n=445), cladribine (n=76) or natalizumab (n=524) after fingolimod. The annualised relapse rate (ARR) for ocrelizumab was 0.07, natalizumab 0.11 and cladribine 0.25. Compared with natalizumab, the ARR ratio (95% confidence interval [CI]) was 0.67 (0.47 to 0.96) for ocrelizumab and 2.31 (1.30 to 4.10) for cladribine; the hazard ratio (95% CI) for time to first relapse was 0.57 (0.40 to 0.83) for ocrelizumab and 1.18 (0.47 to 2.93) for cladribine. Ocrelizumab users had an 89% lower discontinuation rate (95% CI, 0.07 to 0.20) than natalizumab, but also a 51% lower probability of confirmed disability improvement (95% CI, 0.32 to 0.73). There was no difference in disability accumulation. Conclusion After fingolimod cessation, ocrelizumab and natalizumab were more effective in reducing relapses than cladribine. Due to the low ARRs in all three treatment groups, additional observation time is required to determine if statistical difference in ARRs results in long-term disability differences.
INTRODUCTION
2][3][4] It is approved as a secondline therapy in Europe and first-line therapy in Australia, the USA, Canada and other countries. 5 owever, it is common that patients discontinue fingolimod after a few years of treatment, mainly due to adverse events, relapse, MRI activity, disease progression or pregnancy planning. 6 Fingolimod cessation is associated with severe relapses resembling immune reconstitution inflammatory syndrome, therefore, the standard practice is to switch to another high-efficacy DMT with a short treatment gap, including natalizumab, ocrelizumab and cladribine. 7 8 Evidence is lacking for the relative
WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ There is no clear consensus on the optimum treatment in the context of patients switching from fingolimod. Is there any difference in the effectiveness and treatment persistence between ocrelizumab, cladribine and natalizumab among patients with relapsingremitting multiple sclerosis (RRMS) who switched from fingolimod?
WHAT THIS STUDY ADDS
⇒ In this observational study including 1045 patients with RRMS who ceased fingolimod, switch to ocrelizumab was associated with reduced annualised relapse rate compared with natalizumab and cladribine switches.
Ocrelizumab was associated with a lower rate of disability improvement than natalizumab.
Multiple sclerosis
net clinical benefit between these treatment choices after fingolimod cessation.
Ocrelizumab is a humanised monoclonal antibody 9 that depletes CD20-expressing B cells with minimal impact on preexisting humoral immunity. 10 11 Cladribine is a semi-selective immune reconstitution pulse therapy that temporarily depletes 80% of peripheral B cells and 50% of T cells. 12 ][15][16] Natalizumab is the first monoclonal antibody used for multiple sclerosis (MS). 1][22] However, no study to date has directly compared the relative efficacy of ocrelizumab, natalizumab and cladribine.
We, therefore, performed a retrospective analysis to compare treatment outcomes with ocrelizumab, cladribine and natalizumab after fingolimod cessation using the MSBase registry data set. 23 Outcomes of interest included time to first relapse, relapse rate, confirmed disability progression or improvement and DMT discontinuation.
Standard protocol approvals, registrations and patient consents
The MSBase registry is an international observational cohort study of MS. It was established in 2004. 23 Written informed consent was obtained from all enrolled registry participants.
Study population
Patients continuously treated with fingolimod monotherapy ≥6 months before discontinuation were included if they had definite relapse-onset MS 24 25 and continued using ocrelizumab, natalizumab or cladribine ≥6 months after switching. They were also required to have at least two visits with a minimum 6-month gap and complete Expanded Disability Status Scale (EDSS, a non-linear ordinal disability scale with a range 0-10) assessment during follow-up to allow for the ascertainment of disability progression. Study inclusion also required complete information for sex, age, the date of starting and stopping fingolimod, the starting and stop date of the new treatments after fingolimod, EDSS assessments, and dates of relapses. To reflect current switch practice and minimise complete loss of treatment efficacy, including rebound risk, we only included patients with a treatment gap less than 3 months. 26
Study inclusion criteria and definitions
Included patients were treated with fingolimod (0.5 mg orally daily) for at least 6 months and were categorised into three switch groups: ocrelizumab (600 mg, every 24 weeks), 11 cladribine (3.5 mg/kg total dose orally, initial treatment consisting of two courses completed at week 1 and week 5) 16 and natalizumab (300 mg intravenously every 4 weeks). 19 Patients were censored at either treatment discontinuation or the last recorded EDSS visit, whichever occurred first. The discontinuation was defined as starting a new treatment. In addition, the closest EDSS visit within 6 months before or after the switching date was chosen as the baseline EDSS visit.
Patient data were recorded as part of routine clinical visits at participating centres via the locally installed iMed or MDS MSBase data entry systems and monitored through a series of procedures to maintain quality. 26
Study endpoints
The primary study outcomes were annualised relapse rate (ARR, calculated by dividing the total number of relapses by the total number of person-years at risk) and time to the first relapse. Secondary outcomes were disability accumulation events, disability improvement events and treatment discontinuation.
Relapse was defined as new symptom occurrence or exacerbation of existing symptoms for at least 24 hours, in the absence of concurrent illness or fever and occurring at least 30 days after a previous relapse. 27 Disability progression was defined as the confirmed increase in EDSS of ≥0.5 steps for patients with a baseline score >5.5, or ≥1.0 step for those with a baseline score between 1.0 and 5.5, and ≥1.5 steps if the baseline score was 0. Similarly, the confirmed disability improvement was defined as a decrease in EDSS by one step (1.5 steps if baseline EDSS was 0 and 0.5 steps if baseline EDSS was >5.5). Scores obtained <30 days after relapse were excluded. Six-month confirmed disability progression was defined as disability progression sustained over two subsequent visits relative to the baseline EDSS score, with a minimum of 6 months between each assessment. Treatment discontinuation event dates were always recorded. The treating neurologist recorded the primary reason for discontinuation using predefined terms, but these data are not part of the MSBase minimum data set, with relatively high missingness.
Statistical analysis
The demographic information and the baseline characteristics were reported as number and percentage for discrete variables and as mean (standard deviation [SD]) or median (interquartile range [IQR]) for continuous variables, as appropriate. To mitigate baseline differences between the groups and selection bias, we applied an inverse-probability-treatment-weighting (IPTW) approach based on propensity scores (PS) for three treatments groups to achieve covariate balance. 28 29 The PS represents the probability of receiving a treatment conditional on observed covariates. We calculated the PS using a multinomial-logistic regression model with treatment groups as a dependent variable and the baseline covariates listed in table 1 as independent variables.
The weights were obtained by taking the inverse of the probability of receiving the corresponding treatment. We truncated the weights at the 1st and 99th percentiles 30 and used the stabilised weights approach 31 to mitigate the influence of the extreme weights on the variability of the estimated treatment effect. 29 he covariate balance was evaluated using absolute standardised difference measure (ASD), where ASD≥0.1 indicates an imbalance. 32 33 n IPTW-weighted negative binomial model was used to compare ARRs, with the relapse count as a dependent variable, the treatment group as an independent variable, and the natural logarithm of the follow-up time of the corresponding treatment after fingolimod discontinuation as an offset term. We used the IPTW-weighted Cox proportional-hazard regression model with robust standard errors and Kaplan-Meier cumulative hazard curves to assess and visualise treatment effect differences for the time-to-event outcomes. The proportionalhazard assumption was checked by the Schoenfeld global test, and no violation was detected. 34 All statistical tests were twosided with a statistical significance defined as p≤0.05. All analyses were performed in R, V. 4
Sensitivity analyses
We performed seven sensitivity analyses to assess the robustness of our primary outcome results. First, the analyses were rerun using the doubly-robust approach, 35 adjusting the baseline covariates in the IPTW-weighted models to eliminate the potential risk of insufficient covariate balance. Second, we introduced the cerebral MRI information (number of hyperintense T2 lesions) at baseline to the PS model and repeated the analysis. MRI information was reported by the practicing neurologists according to the local MRI protocols and policies. Due to the sparse frequency of MRI, the baseline MRI was defined as the closest brain MRI taken within 12 months before, or 6 months after, the start date of treatment. The missing values were handled with multiple imputations. Overall, 20 imputed data sets were generated, and the estimates were computed separately and then combined using Rubin's rules. 36 Third, an intention-totreat analysis was conducted with the start of study therapies and censoring at either event occurrence or the end of the follow-up. Fourth, the intention-to-treat analyses were repeated using patients with any duration of on-treatment follow-up (ie, not excluding patients who discontinued the switch therapy before 6 months.) Fifth, to eliminate possible confounding by country, we added country (Australia/others) to the PS model and reran the analysis with the newly generated weights. Sixth, to test the consistency of results across different baseline EDSS, we redefined baseline EDSS as obtained 6 months before or 1 month after the new treatment. Seventh, to mitigate the possible influence of different reasons for fingolimod discontinuation, we added it as a categorical variable to the PS model and repeated the analysis. Eighth, we also conducted an analysis with all patients censored at 1 year of follow-up to eliminate the possible effect in three treatment groups due to the different length of the follow-up. Ninth, to eliminate the potential effect of different approved dates, we conducted an analysis limiting to patients who started the study therapies after 1 January 2018.
Study population
We assessed the eligibility of a total of 62 100 patients with MS from MSBase (figure 1 ). Following the selection criteria fulfilment, we included 1045 patients from 26 countries and 64 centres who had discontinued fingolimod. Among them, 445 patients were treated with ocrelizumab (median [IQR] follow-up period: 1.8 [1.2-2.4] years), 76 were treated with cladribine (1.1 [0.9-1.8] years) and 524 were treated with natalizumab (3.6 [1.9-5.3] years) between 24 May 2010 and 13 November 2020.
Baseline characteristics of the original (non-weighted) study population are presented in table 1 and of the weighted study population were presented in online supplemental eTable 1. Before weighting, natalizumab users were more likely to be younger, have a shorter disease duration and a shorter period of fingolimod use compared with cladribine or ocrelizumab users. The baseline EDSS score and the number of previous treatments used were similar between the three groups. The main difference between the three treatment groups was the number of relapses in the 1 and 2 years prior to the treatment switch, with the highest relapse rate in the natalizumab group and the lowest rate in the cladribine group (table 1 ). After weighting, all baseline covariates were balanced (ASD<0.1).
Effectiveness
The mean ARR for ocrelizumab (IPTW-weighted ARR, 0.07; 95% CI, 0.04 to 0.13), cladribine (0.25; 95% CI, 0.12 to 0.57) and natalizumab (0.11; 95% CI, 0.09 to 0.14) were obtained using the negative binomial model. The IPTW-weighted ARR ratios between treatment groups were reported in figure 2 . Ocrelizumab use was associated with a statistically significant reduction in ARR by 33% when compared with natalizumab; cladribine was less effective than natalizumab, with an ARR ratio of 2.31 (95% CI, 1.30 to 4.10). Consistent results were observed in the cumulative hazards of time to first relapse. The cumulative hazard of experiencing the first relapse in the ocrelizumab users was 43% lower than that in the natalizumab users over time (IPTW-weighted HR, 0.57; 95% CI, 0.40 to 0.83). In addition, the cumulative hazard curves for ocrelizumab and natalizumab were similar until about 5 months, after which the hazard of the first relapse was significantly increased in the natalizumab group compared with ocrelizumab (figure 3 ). There were no significant differences in the cumulative hazards of time to first relapse between cladribine and natalizumab users (IPTW-weighted HR, 1.18; 95% CI, 0.47 to 2.93).
Because the number of patients with cladribine was insufficient in the EDSS progression analyses, we only compared ocrelizumab with natalizumab. There was no significant difference in the cumulative hazard of disability However, the probability of disability improvement in the ocrelizumab users was 51% lower than natalizumab users (IPTW-weighted HR, 0.49; 95% CI, 0.32 to 0.73) (figure 2 ). Of those who improved, 79% of patients in the natalizumab group maintained improvement until the end of follow-up, with a median follow-up of 1.6 years and 83% of patients in the ocrelizumab group maintained improvement, with a median follow-up of 1.1 years (online supplemental eTable 2).
Treatment discontinuation
There was no significant difference in treatment discontinuation between cladribine and natalizumab users (IPTWweighted HR, 0.74; 95% CI, 0.26 to 2.10). In contrast, the hazard of treatment discontinuation was 89% lower in the ocrelizumab users compared with natalizumab users (IPTWweighted HR, 0.11; 95% CI, 0.07 to 0.20). The Kaplan-Meier cumulative probability of discontinuation curves were shown in figure 3 . The most reported reason for treatment discontinuation in the natalizumab and cladribine users was scheduled stop (36% and 50%) and in the ocrelizumab users was adverse events (18%) (online supplemental eTable 3).
Of the three treatment groups, the most reported reason for fingolimod discontinuation was lack of efficacy (cladribine 35%, ocrelizumab 56% and natalizumab 64%) (online supplemental eTable 4).
Sensitivity analyses
We assessed the association between the primary outcome and the treatment groups using a doubly robust approach, and there was no evidence of substantial change in outcome (online supplemental eFigure 1). Consistency of the primary results was maintained in the remainder of the sensitivity analyses (online supplemental eFigure 2-8), except for a loss of statistical significance observed in the switch-date restricted analysis after 1 January 2018 due to the major loss of patients in the natalizumab group and, therefore, lack of power (online supplemental eFigure 9). Compared with natalizumab users, the ARR was statistically significantly lower in ocrelizumab users but higher in cladribine users. Consistently, the incidence of the time to first relapse and discontinuation rate were significantly lower in the ocrelizumab users compared with natalizumab users. Cladribine and natalizumab use were not significantly different for time to first relapse and discontinuation, acknowledging that this analysis was likely underpowered due to the relatively small sample size of cladribine users. Natalizumab was associated with a higher probability of sustained disability improvement than ocrelizumab. No significant difference in disability accumulation was found between the two groups. In this cohort, natalizumab was associated with a higher ARR and shorter time to first relapse than ocrelizumab, yet natalizumab has a higher rate of sustained disability improvement than ocrelizumab. Our results comparing ocrelizumab to natalizumab appear to be inconsistent. However, we do not consider these results to be contradictory as EDSS improvement likely relates to the cessation of ongoing lesional central nervous system inflammation and enhanced remyelination whereas prevention of new relapses relates to new lesion formation. These mechanisms could easily be differentially affected by the different mechanisms of action of natalizumab and ocrelizumab.
It might also be unexpected that, in relation to relapse prevention, ocrelizumab was found to be more effective than natalizumab. This could relate to the particular patient population being studied, namely patients ceasing fingolimod due to recurrent disease activity. Perhaps in these patients, a drug mechanism interfering with lymphocyte migration (ie, fingolimod, natalizumab) could be less effective than one targeting B cell functions (ie, ocrelizumab). Our results are consistent with a prior comparative study of rituximab versus natalizumab. Like ocrelizumab, rituximab is also a monoclonal antibody targeting the B cell surface antigen CD20. A retrospective cohort study found that natalizumab use was associated with an increased relapse rate (HR 5.1; 95% CI, 1.20 to 22.20) and increased treatment discontinuation rate (HR, 13.90; 95% CI, 3.80 to 50.60) compared with rituximab. 37 Another retrospective observational study found that rituximab was superior to natalizumab in medication persistence (HR, 0.05; 95% CI, 0.01 to 0.38) but not in control of disease activity. 38 The authors stated that the latter result was possibly due to the small number of participants (n=48) in the rituximab group. An observational study of 740 patients showed no difference in the time to first relapse and EDSS worsening between anti-CD20 use (including ocrelizumab [n=59] and rituximab [n=278]) and natalizumab use (n=403), with an HR of 1.04 (95% CI, 0.74 to 1.46) and 1.13 (95% CI, 0.75 to 1.70) respectively. 39 Overall, their result of EDSS progression were consistent with ours. The result of time to first relapse in their study is inconsistent with ours and this may be explained by the different descriptions of relapse rate and analysis methodology used in the two studies.
Previous studies have compared the efficacy of ocrelizumab, cladribine and natalizumab separately. A PS-matched study comparing cladribine with natalizumab showed that the cladribine users were more likely to experience a relapse than the natalizumab users (HR, 1.80; 95% CI, 1.08 to 2.97). 40 A cohort study using data from two MS data sets reported that patients treated with cladribine had a significantly higher ARR than those treated with natalizumab (ARR ratio, 2.13; 95% CI, 1.17 to 3.88). 41 Both studies support our results. Two network meta-analyses reported no significant difference in the efficacy between cladribine, ocrelizumab and natalizumab (cladribine vs ocrelizumab ARR ratio 1.06, 95% CI, 0.78 to 1.45, and cladribine vs natalizumab 1.16, 95% CI, 0.89 to 1.53 21 ; cladribine vs ocrelizumab ARR ratio 1.14, 95% CI, 0.81 to 1.60, and cladribine vs natalizumab 1.22, 95% CI, 0.89 to 1.68 42 ), but the trend of these results was comparable with ours. However, the sample size of the cladribine group is small in our study, limiting the conclusions for this treatment. Future studies will be needed to confirm our findings.
Multiple sclerosis
In addition, to our knowledge, a three-way comparison of efficacy and persistence among the three treatments is novel and necessary, as all are reasonable choices for subsequent DMT after fingolimod cessation.
Limitations
First, this study is an observational study and treatment indication bias is inevitable. To address this question, We used an IPTW approach that has been shown to be superior to the conventional adjustment approach in mitigating treatment indication bias. It allows observational studies to be designed similar to randomised experiments and estimates the causal effect based on all eligible participants. 29 Second, the sample size of cladribine users is still relatively small, as cladribine has only been approved in the last 2-3 years in most parts of the world. However, it is important to perform this analysis as the relevant treatment decisions are current. Overall, 70% of the cladribine users were from Australian centres, where cladribine is a popular treatment choice. The potential confounding influence due to country was assessed in a sensitivity analysis, and the results were consistent. Furthermore, we did not incorporate the MRI data in the primary PS model in our main analyses as only 38% of study participants had this information. However, we conducted a sensitivity analysis using MRI information, in combination with the data obtained by multiple imputations, and the results were consistent. Drug safety is an important part for comparing the net clinical benefit of different DMTs, especially in long-term. The collection of safety data from patients involved in the MSBase is currently underway.
CONCLUSION
In patients with RRMS who discontinue fingolimod and commence another disease-modifying treatment within 3 months, ocrelizumab was associated with a significant reduction in ARRs, longer time to first relapse and a lower discontinuation rate compared with cladribine and natalizumab. Natalizumab was associated with better sustained disability improvement compared with ocrelizumab. The relapse rates in all treatment groups were relatively low, suggesting that all these three DMTs are feasible and reasonable treatment options.
Ethics approval
This study involves human participants and was approved by The Alfred Health Human Research and Ethic Committee No 528/12, Version: 7, Dated 7 May 2020. Participants gave informed consent to participate in the study before taking part.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. In principle, patient-level data sharing is possible. However, permission from each contributing data controller is required.
Supplemental material
This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Acknowledgements
AcknowledgementsWe wish to thank all of the patients and their caregivers who participated in this study and contributed data to the MSBase registry.Contributors Guarantor: CZ.Concept and design: CZ, AvdW and HB.Acquisition, analysis or interpretation of data: CZ, ZZ, AvdW and HB.Drafting of the manuscript: CZ.Critical revision of the manuscript for important intellectual content: All authors.Statistical analysis: CZ and ZZ.Administrative, technical or material support: AvdW and HB.Supervision: AvdW and HB. FundingThe authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Competing interestsCZ, OS, RM, AP, JK, MO, RK, SO, DM, IR and ZZ report no disclosures.VJ received conference travel support from Merck and Roche and speaker's honoraria from Biogen and Roche outside of the submitted work.She receives research support from the Australian National Health and Medical Research Grant and MS Research Australia.MB received conference travel support from Biogen and Novartis.His institution has received research support from Biogen, Merck and Novartis.TK received conference travel support and/or speaker honoraria from WebMD Global, Eisai, Novartis, Biogen, Sanofi-Genzyme, Teva, BioCSL and Merck and received research or educational event support from Biogen, Novartis, Genzyme, Roche, Celgene and Merck.JL-S received travel compensation from Novartis, Biogen, Roche and Merck.Her institution receives the honoraria for talks and advisory board commitment and research grants from Biogen, Merck, Roche, TEVA and Novartis.SJK received compensation for scientific advisory board activity from Merck and Roche.SH received honoraria and consulting fees from Novartis, Bayer Schering and Sanofi, and travel grants from Novartis, Biogen Idec and Bayer Schering.MS participated in, but not received honoraria for, advisory board activity for Biogen, Merck, Bayer Schering, Sanofi Aventis and Novartis.AvdW served on advisory boards and receives unrestricted research grants from Novartis, Biogen, Merck and Roche.She has received speaker's honoraria and travel support from Novartis, Roche and Merck.She also received grant support from the National Health and Medical Research Council of Australia and MS Research Australia.VVP has received travel grants from Merck Healthcare KGaA (Darmstadt, Germany), Biogen, Sanofi, Bristol Meyer Squibb, Almirall and Roche.His institution has received research grants and consultancy fees from Roche, Biogen, Sanofi, Merck Healthcare KGaA (Darmstadt, Germany), Bristol Meyer Squibb, Janssen, Almirall and Novartis Pharma.JP received travel compensation from Novartis, Biogen, Genzyme and Teva, and speaking honoraria from Biogen, Novartis, Genzyme and Teva.DH received speaker honoraria and consulting fees from Biogen, Merck, Teva, Roche, Sanofi Genzyme and Novartis, as well as support for research activities from Biogen and Czech Ministry of Education.EKH received honoraria/research support from Biogen, Merck Serono, Novartis, Roche and Teva.GI received speaking honoraria from Biogen, Novartis, Sanofi, Merck, Roche, Almirall and Teva.FP received speaker honoraria and advisory board fees from Almirall, Bayer, Biogen, Celgene, Merck, Novartis, Roche, Sanofi-Genzyme and TEVA.He received research funding from Biogen, Merck, FISM (Fondazione Italiana Sclerosi Multipla), Reload Onlus Association and University of Catania.RA received honoraria as a speaker and for serving on scientific advisory boards from Bayer, Biogen, GSK, Merck, Novartis, Roche and Sanofi-Genzyme.CB received conference travel support from Biogen, Novartis, Bayer-Schering, Merck and Teva; he participated in Sanofi Aventis, Roche and Novartis clinical trials.MM has served on advisory board for Merck, has received speaker honoraria from Merck and Biogen.Her institution receives funding from Merck, Australian National Health Medical Research Council, Brain Foundation, Charles and Sylvia Viertel Foundation, Bethlehem Griffith Foundation and MS Research Australia.HB has received institutional (Monash University) funding from Biogen, F. Hoffmann-La Roche Ltd, Merck, Alexion, CSL and Novartis; has carried out contracted research for Novartis, Merck, F. Hoffmann-La Roche Ltd and Biogen; has taken part in speakers' bureaus for Biogen, Genzyme, UCB, Novartis, F. Hoffmann-La Roche Ltd and Merck; has received personal compensation from Oxford Health Policy Forum for the Brain Health Steering Committee.Patient consent for publication Not applicable.
Failed to load PDF:
StripeM-Inner
Paper sources
Abstract screening pilot
Abstract screening results
Extraction pilot
Extraction results
Research report
Modify setup