Multiple sclerosis

J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
Original research

Comparing ocrelizumab to interferon/glatiramer

acetate in people with multiple sclerosis over age 60
Yi Chao Foong ‍ ‍,1,2,3 Daniel Merlo,1,4 Melissa Gresle,1,2,5 Katherine Buzzard,4,6
Michael Zhong,1,2 Wei Zhen Yeh,1,2 Vilija Jokubaitis ‍ ‍,1,2 Mastura Monif,1,2
Olga Skibina,2,4 Serkan Ozakbas,,7 Francesco Patti ‍ ‍,8,9 Pierre Grammond,10
Maria Pia Amato,11 Tomas Kalincik ‍ ‍,6,12 Dana Horakova,13 Eva Kubala Havrdova,13
Bianca Weinstock-­Guttman,14 Jeanette Lechner Scott,15,16 Cavit Boz,17
Maria Jose Sa,18,19 Helmut Butzkueven,1,2 Anneke van der Walt ‍ ‍,1,2 Chao Zhu ‍ ‍,1
on behalf of MSBASE study group

For numbered affiliations see ABSTRACT

end of article.
Correspondence to
Dr Anneke van der Walt,
Monash University Central
Clinical School, Melbourne,
Background Ongoing controversy exists regarding
optimal management of disease modifying therapy
(DMT) in older people with multiple sclerosis (pwMS).
There is concern that the lower relapse rate, combined
with a higher risk of DMT-­related infections and side
WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ In older people with multiple sclerosis (pwMS),
relapse rates decrease while risk of adverse
events related to disease modifying therapy
(DMT) increases. The comparative efficacy of

(HINARI) - Group A. Protected by copyright.

VIC 3000, Australia; ​anneke.​
vanderwalt@​monash.​edu
HB, AvdW and CZ are joint
senior authors.
Received 27 October 2023
Accepted 11 February 2024
© Author(s) (or their
employer(s)) 2024. No
commercial re-­use. See rights
and permissions. Published
by BMJ.
To cite: Foong YC, Merlo D,
Gresle M, et al. J Neurol
Neurosurg Psychiatry Epub
ahead of print: [please
include Day Month Year].
doi:10.1136/jnnp-2023-
332883
effects, may alter the risk-­benefit balance in older pwMS.
Given the lack of pwMS above age 60 in randomised
controlled trials, the comparative efficacy of high-­efficacy
DMTs such as ocrelizumab has not been shown in
older pwMS. We aimed to evaluate the comparative
effectiveness of ocrelizumab, a high-­efficacy DMT, versus
interferon/glatiramer acetate (IFN/GA) in pwMS over the
age of 60.
Methods Using data from MSBase registry, this
multicentre cohort study included pwMS above 60
who switched to or started on ocrelizumab or IFN/
GA. We analysed relapse and disability outcomes
after balancing covariates using an inverse probability
treatment weighting (IPTW) method. Propensity scores
were obtained based on age, country, disease duration,
sex, baseline Expanded Disability Status Scale, prior
relapses (all-­time, 12 months and 24 months) and prior
DMT exposure (overall number and high-­efficacy DMTs).
After weighting, all covariates were balanced. Primary
outcomes were time to first relapse and annualised
relapse rate (ARR). Secondary outcomes were 6-­month
confirmed disability progression (CDP) and confirmed
disability improvement (CDI).
Results A total of 248 participants received
ocrelizumab, while 427 received IFN/GA. The IPTW-­
weighted ARR for ocrelizumab was 0.01 and 0.08 for
IFN/GA. The IPTW-­weighted ARR ratio was 0.15 (95% CI
0.06 to 0.33, p<0.001) for ocrelizumab compared with
IFN/GA. On IPTW-­weighted Cox regression models, HR
for time to first relapse was 0.13 (95% CI 0.05 to 0.26,
p<0.001). The hazard of first relapse was significantly
reduced in ocrelizumab users after 5 months compared
with IFN/GA users. However, the two groups did not
differ in CDP or CDI over 3.57 years.
Conclusion In older pwMS, ocrelizumab effectively
reduced relapses compared with IFN/GA. Overall relapse
activity was low. This study adds valuable real-­world data
for informed DMT decision making with older pwMS.
Our study also confirms that there is a treatment benefit
Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883
DMT has not been shown in older pwMS.
WHAT THIS STUDY ADDS
⇒ This study demonstrates a clear treatment
benefit in older people with multiple sclerosis
with regard to relapse activity.
HOW THIS STUDY MIGHT AFFECT RESEARCH,
PRACTICE OR POLICY
⇒ Our findings add valuable real-­world data for
informed DMT discussions with older pwMS.
in older people with MS, given the existence of a clear
differential treatment effect between ocrelizumab and
IFN/GA in the over 60 age group.
INTRODUCTION
Ongoing controversy exists regarding the optimal
management of disease modifying therapy (DMT) in
older people with multiple sclerosis (pwMS). With
increasing age, the relapse rate in MS falls, with a
concomitant marked reduction in MRI activity.1–4
There is concern that reduced potential benefits of
immune modulation, combined with a higher risk
of DMT-­ related infections and side effects, may
alter the risk-­benefit balance in older pwMS.5–7
This has led to limited studies of DMT discontinua-
tion in pwMS over 60 which have shown no differ-
ence in patient-­reported outcomes.8 9 However, the
comparative efficacy of DMTs in an elderly MS
population has not been shown, with pivotal DMT
randomised controlled trials excluding pwMS
above the age of 60.10–12
Ocrelizumab is a humanised monoclonal anti-
body that selectively depletes CD20+ B cells. It is
highly effective in relapsing-­remitting MS (RRMS)
and primary progressive MS (PPMS) in reducing
relapse activity and clinical progression.11 13–16
1
 Multiple sclerosis
We aimed to address this gap in knowledge by comparing the
outcomes of pwMS above the age of 60 who had switched to or
started on ocrelizumab, a high-­efficacy DMT, versus interferon/
glatiramer acetate (IFN/GA), considered low-­efficacy DMTs.17 18
METHODS
Standard protocol approvals and patient consent
The MSBase registry is an international observational cohort
study of pwMS established in 2004.19 This study followed the
Strengthening the Reporting of Observational Studies in Epide-
miology reporting guidelines.
Study population
Participants were required to have at least two follow-­up outpa-
tient visits (minimum 6 months between visits) with complete
Expanded Disability Status Scale (EDSS, a non-­linear ordinal
disability scale with a range 0–10) assessments to allow for the
ascertainment of confirmed disability progression (CDP).
Study inclusion also required complete information for sex,
age, duration of disease, the date of starting and/or stopping
DMTs and dates of relapses.
Study inclusion criteria and definitions
We included patients who were started on or switched to ocre-
lizumab or IFN/GA above the age of 60. For those who were
started on ocrelizumab and one of either IFN/GA after age
60, we grouped them based on the DMT started soonest after
age 60. Patients were required to be on therapy for at least 6
months. Patients were censored at the last recorded visit with
EDSS score recorded. The closest EDSS score assessed within
6 months before or after the initiation date was chosen as the
baseline EDSS score. All EDSS scores recorded within 1 month
of a relapse were excluded from baseline EDSS.
Patient data was recorded during routine clinical visits at
participating centres via the locally installed MSBase data entry
systems and monitored through a series of procedures to main-
tain quality.
High-­efficacy DMT in this study was defined as: natalizumab,
ocrelizumab, alemtuzumab, mitoxantrone, rituximab and autol-
ogous stem cell transplantation.
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 first relapse (after
the age of 60). For ARR, we examined both ARR over 2 years of
follow-­up and total follow-­up time.20 21
Secondary outcomes were CDP and confirmed disability
improvement (CDI). CDP was defined as an increase of
≥1.5 EDSS steps from a baseline score of 0, 1 step from baseline
scores 1.0–5.5, or 0.5 step from a baseline score ≥6.0, sustained
at two or more consecutive visits separated by ≥150 days. CDI
was defined as a decrease of 1 or 0.5 EDSS steps if baseline EDSS
was 2–6 or >6, respectively, sustained at two or more consecu-
tive visits separated by >150 days.22
Sensitivity analyses
We performed the following sensitivity analyses for our primary
outcomes: (1) We used a 1:1 nearest neighbour matching based
on the propensity score. (2) We limited the analysis to those
with a baseline EDSS within 1 month of starting or switching
DMT. (3) To minimise the effect of differing treatment dates,
we conducted an analysis limited to patients starting DMTs after
2 Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883
J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
January 201614 23 as well as using the doubly robust standardi-
sation method, which combines IPTW with G-­computation to
achieve double robustness.24–26 (4) We added MS phenotypes (PP,
RR or secondary progressive (SP) MS) to the propensity scores
(PS) model and reran the primary outcome analysis. (5) We
examined excluding PPMS phenotypes from the analysis.22 (6)
We examined hazard of time to unconfirmed disability progres-
sion as a marker of disability. (7) We examined hazard of time to
EDSS of 6 as a marker of disability—this analysis was restricted
to those with a baseline EDSS of less than 6 (362 participants in
total).
Statistical analysis
The demographic information and the baseline characteristics
were reported as frequencies and proportions for discrete vari-
ables and as mean (SD) or median (IQR) for continuous vari-
ables, as appropriate. To mitigate baseline differences between
the groups and selection bias, we applied an inverse probability
treatment weighting (IPTW) approach based on PS for two treat-
ments groups to achieve covariate balance.27 28 The PS represents
the probability of receiving a treatment based on observed
covariates. We calculated the PS using a logistic regression model
with treatment groups as a dependent variable and the baseline
covariates listed in as independent variables.
(HINARI) - Group A. Protected by copyright.
The weights were obtained by taking the inverse of the prob-
ability of receiving the corresponding treatment. The covariate
balance was evaluated using the absolute standardised difference
measure (ASD), where ASD >0.1 indicates an imbalance.29
An 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 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 proportional hazard assumption
was checked with the Schoenfeld global test, and no violation
was detected. All statistical tests were two-­sided with a statistical
significance defined as p≤0.05. All analyses were performed in
R, V.4.1.1. (R Foundation for Statistical Computing).
RESULTS
Participant characteristics
We assessed the eligibility of 81 769 patients from MSBase
(figure 1). Following the selection criteria we identified 675
patients, of which 248 were started on ocrelizumab and 427
were started on IFN/GA.
Baseline characteristics of the original (non-­weighted) study
population are presented in table 1. Before weighting, ocreli-
zumab users were more likely to be older, live in Australia, have a
longer disease duration, higher baseline EDSS, less relapses in the
1 and 2 years prior to the treatment, used more previous DMTs
and have a higher chance of prior high efficacy DMTs. After
weighting, all baseline covariates were balanced (ASD<0.1).
Effectiveness
The total number of relapses was 8 and 182 for ocrelizumab
and IFN/GA, respectively. Follow-­ up person-­ years were 571
(with median follow-­up time of 2.47 years) and 2238 (with
median follow-­up time of 4.48 years) for ocrelizumab and IFN/
GA respectively. The crude ARR in the original (non-­weighted)
study population was 0.01 for ocrelizumab and 0.08 for IFN/
 Multiple sclerosis

J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
Figure 1 Flow chart of the patient inclusion/exclusion criteria. EDSS, Expanded Disability Status Scale; GA, glatiramer acetate; IFN, interferon; OCR,
ocrelizumab.

(HINARI) - Group A. Protected by copyright.

GA. The crude 2-­year ARR was 0.01 for ocrelizumab and 0.11
for IFN/GA.
The mean IPTW-­ weighted ARR for ocrelizumab (IPTW-­
weighted ARR, 0.01; 95% CI 0.00 to 0.03) and IFN/GA (IPTW-­
weighted ARR, 0.08; 95% CI 0.06 to 0.10) were obtained using
the negative binomial model. Ocrelizumab use was associated
with a statistically significant reduction in ARR by 85% when
compared with IFN/GA (ARR ratio 0.15, 95% CI 0.06 to 0.33,
p<0.001). Similar results were found when the follow-­up dura-
tion was limited to 2 years (ARR ratio 0.10, 95% CI 0.03 to
0.26, p<0.001).
We evaluated the cumulative hazards of time to first relapse
and observed consistent results. The risk of experiencing the
first relapse in the ocrelizumab users was 89% lower than
that in the IFN/GA users over time (IPTW-­weighted HR 0.11,
95% CI 0.05 to 0.26, p<0.001). There was a significantly
increased risk of experiencing first relapse in the IFN/GA
group compared with ocrelizumab after 5 months of treatment
initiation (figure 2).
Confirmed disability progression/confirmed disability
improvement
The weights were regenerated for the smaller CDP cohort.
Covariate balance was achieved for all variables.
Of the 167 participants on ocrelizumab with sufficient
follow-­up, 37 experienced a CDP event, while 146 out of 340
IFN/GA users experienced a CDP event. Follow-­up person-­years
were 410 and 1401 for ocrelizumab and IFN/GA, respectively.
There was no significant difference in the cumulative hazard of
CDP between ocrelizumab and IFN/GA users (IPTW-­weighted
HR 0.78, 95% CI 0.48 to 1.26, p=0.307) (figure 3). Similarly,
there was no significant difference between cumulative hazard
of CDI between ocrelizumab and IFN/GA users (IPTW-­weighted
HR 0.97, 95% CI 0.52 to 1.80, p=0.92).
Sensitivity analysis
We performed further sensitivity analyses by using 1:1 nearest
neighbour matching based on the propensity score for the
primary outcomes of time to first relapse and ARR. The results

Table 1 Baseline characteristics of study population by ocrelizumab versus interferon/glatiramer acetate

ASD
Ocrelizumab (n=248) Interferon/glatiramer (n=427) Before IPTW After IPTW
Age, mean (SD), years 64.53±3.98 63.75±3.49 0.21 0.01
Female, n (%) 159 (64.11) 303 (70.96) 0.07 0.01
Country, n (%)* 95 (38.31) 67 (15.69) 0.23 0.01
Disease duration, mean (SD), years 17.7±11.43 16.1±12.00 0.14 0.06
Baseline EDSS, mean (SD) 5.22±1.58 3.89±1.96 0.75 0.05
Relapses 1 year prior to baseline, mean (SD) 0.21±0.46 0.39±0.63 0.32 0.01
Relapses 2 years prior to baseline, mean (SD) 0.33±0.65 0.59±0.78 0.36 0.02
Previous DMTs, median (IQR) 1(1-­2) 1 (0–1) 0.49 0.05
Previous high-­efficacy DMTs, n (%) 68 (27.42) 49 (11.48) 0.16 0.00
Values are median (IQR), n (%) or mean±SD and rounded to two decimal points. ASD is the absolute difference in means or proportions divided by the SE. An imbalance was
defined as an ASD >0.10, indicated in bold.
*This was a binary variable with Australia versus other countries, with n representing the number of participants from Australia.
ASD, absolute standardised difference; DMT, disease modifying therapy; EDSS, Expanded Disability Status Scale; IPTW, inverse probability treatment weighting.

Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883 3

 Multiple sclerosis
Figure 2 IPTW-­weighted cumulative hazard of time to first relapse. Weighted Kaplan-­Meier curves were applied to show the cumulative hazard of time
to first relapse in ocrelizumab users (blue line) and interferon/glatiramer acetate users (red line). GA, glatiramer acetate; IFN, interferon; IPTW, inverse
probability treatment weighting; OCR, Ocrelizumab.
were consistent with the main analysis results. The HR of time to
first relapse was 0.16 (95% CI 0.08 to 0.32, p<0.001), and the
ARR ratio was 0.18 (95% CI 0.08 to 0.36, p<0.001).
We performed sensitivity analysis restricting the inclusion
criteria to those having a baseline EDSS 1 month post-­ DMT
initiation. For this we had 585 patients meeting criteria, 367 on
IFN/GA and 218 on ocrelizumab. There were five relapses in the
ocrelizumab group over 491 follow-­up years (crude ARR 0.01)
and 140 relapses in the IFN/GA group over 1902 follow-­up
years (crude ARR 0.07). On propensity score-­ weighted Cox
regression models, HR for time to first relapse was 0.10 (95%
CI 0.03 to 0.31, p<0.001). The IPTW-­weighted ARR ratio was
0.17 (95% CI 0.06 to 0.40, p<0.001).
The use of retrospective cohorts raises the possibility of treat-
ment indication bias (given the decline in relapse rates over
time) and violation of the positivity assumption. To address
this, we conducted two sensitivity analysis.14 22 23 First, we
restricted the analysis to those who initiated one of the study
DMTs after 2014. For this analysis we had 196 in the IFN/GA
group with 31 relapses in total, while the ocrelizumab group
was unchanged. Covariate balance was achieved. Similar results
were found compared with the primary analysis, although with
slightly smaller magnitudes of difference. The HR for time to
first relapse was 0.20 (95% CI 0.08 to 0.48, p<0.001), while
the IPTW-­weighted ARR ratio was 0.25 (95% CI 0.08 to 0.76,
p=0.02). Second, we used the doubly robust standardisation
method as the doubly robust estimators are consistent when
the positivity assumption fails.24–26 For the final model, gender
4 Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883
J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
(HINARI) - Group A. Protected by copyright.
and disease duration was not included as the negative binomial
model ran into divergence issues. The results were consistent
with our primary analysis (ARR ratio 0.16, 95% CI 0.03 to
0.35).
We then added MS phenotype into the PS model. As MS
phenotypes were not updated in some patients, we only had a
total of 641 participants (395 on IFN/GA, 246 on ocrelizumab)
for this sensitivity analysis. Of the 395 patients on IFN/GA, 36,
280 and 79 were PP, RR and SPMS, respectively. Of the 246
patients on ocrelizumab, 83, 106 and 58 were PP, RR and SPMS,
respectively. Covariate balance was achieved for all coviariates.
The HR of time to first relapse was 0.14 (95% CI 0.06 to 0.33,
p<0.001), and the ARR ratio was 0.20 (95% CI 0.08 to 0.40,
p<0.001).
Furthermore, we ran another sensitivity analysis restricting the
analysis to SPMS and RRMS. This sensitivity analysis has shown
consistent results with the primary analysis. As expected, the
magnitude of the difference between ocrelizumab and IFN/GA
was greater for the relapse analysis. We did not achieve covariate
balance for baseline EDSS (ASD 0.13) and prior DMT numbers
(0.13). We had 522 patients meeting criteria, 359 on IFN/GA
and 163 on ocrelizumab. There were seven relapses in the ocre-
lizumab group over 385 follow-­up years (crude ARR 0.02) and
152 relapses in the IFN/GA group over 1895 follow-­up years
(crude ARR 0.08). On IPTW-­weighted Cox regression models,
HR for time to first relapse was 0.09 (95% CI 0.04 to 0.23,
p<0.001). The IPTW-­weighted ARR ratio was 0.13 (95% CI
0.04 to 0.34, p<0.001).

Figure 3 IPTW-­weighted cumulative hazard of time to CDP. Weighted Kaplan-­Meier curves were applied to show cumulative hazard of time to first
relapse in ocrelizumab users (blue line) and interferon/glatiramer acetate users (red line). GA, glatiramer acetate; IFN, interferon; IPTW, inverse probability
treatment weighting; OCR, ocrelizumab.
When we analysed unconfirmed disability progression, similar
results were found (HR 0.95, 95% CI 0.68 to 1.32, p=0.76). We
performed further sensitivity analysis by analysing time to EDSS
of 6 among those with a baseline EDSS less than 6. We had a
total of 314 participants, of which 67 were on ocrelizumab and
247 were on IFN/GA. We failed to achieve covariate balance for
disease duration (ASD=0.15), likely due to the smaller number
of participants. Given the minor deviation from accepted ASD
cut-­offs, we proceeded with the analysis. There was no differ-
ence between the two groups regarding the hazard of time to
EDSS of 6 (HR=1.07, 95% CI 0.56 to 2.04, p=0.836).
DISCUSSION
This retrospective analysis compared the effectiveness of ocreli-
zumab against IFN/GA in pwMS above the age of 60. Compared
with IFN/GA users, the ARR was significantly lower in ocrel-
izumab users. Consistently, the hazard of time to first relapse
was significantly lower in the ocrelizumab users compared with
IFN/GA users. However, there was no difference in CDP and
CDI between the compared interventions. We believe that this
finding is novel and supports the comparative higher efficacy of
ocrelizumab in relapse prevention for older pwMS. Logically, our
analysis also demonstrates the presence of an ongoing treatment
effect for DMTs in people over 60. However, given the overall
low relapse rate and lack of efficacy in disability accumulation,
nuanced shared decision-­making with older pwMS regarding the
risk-­benefit ratio of high efficacy DMTs is required.
Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883
Multiple sclerosis
J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
(HINARI) - Group A. Protected by copyright.
In this cohort, IFN/GA was associated with a higher ARR and
shorter time to first relapse than ocrelizumab. This is consistent
with prior comparative studies in younger pwMS.11 This study is
one of the first to demonstrate significant comparative efficacy of
ocrelizumab in older pwMS. Our study population differs from
previous treatment de-­escalation or cessation studies, which have
generally been conducted in pwMS with minimal disease activity
for 3–5 years.30–33 For example, a recent discontinuation study,
the DISCOMS trial, included pwMS over 55 who were free of
clinical relapses for 5 years and radiological relapse for 3 years.
In our study, mean relapses in the ocrelizumab group in the year
prior to treatment initiation were 0.20, and 0.40 in the IFN/GA
group. Furthermore, many of the de-­escalation/cessation studies
were performed on pwMS on low-­efficacy treatments such as
IFN/GA. Their results may not be generalisable to contemporary
MS cohorts. Even among these older pwMS who had relatively
stable disease, treatment cessation was found to have shorter
time to clinical relapse or a new brain MRI lesion over 2 years
in the DISCOMS study.30 This supports our findings of compar-
ative effectiveness of DMTs in older pwMS.
We did not find a difference between the two groups with
regards to CDP or CDI. The landmark OPERA and ORATORIO
trials showed the comparative effectiveness of ocrelizumab in
preventing disability progression in both RRMS and PPMS,
respectively.11 16 However, these trials excluded older pwMS (age
>55) and those with high EDSS scores (5.5 for OPERA and 6.5
for ORATORIO). It should be noted that this differs from our
5
 Multiple sclerosis
cohort, who were much older, and had a relatively low relapse
rate, high baseline EDSS and a higher rate of SPMS. To further
support our findings, the subgroup analysis for ORATORIO for
those over 45 showed a significant drop in efficacy from 24%
to 12% and the result was no longer statistically significant.34
While there is a dearth of literature on the effect of ocrelizumab
on SPMS, small-­scale studies have suggested the limited effect of
ocrelizumab on older progressive pwMS.35 Our study supports
this hypothesis. Furthermore, studies of other DMTs (including
IFN/GA) have shown limited efficacy in slowing disease progres-
sion in SPMS.36–38 This is thought to be due to DMTs’ inability
to stop the neurodegenerative processes that predominate MS
activity with more advanced age.39
Limitations
Given the observational nature of MSBase, there will be treat-
ment indication bias. Relapses were reported by physicians from
the individual participating sites with no additional adjudica-
tion, giving rise to the possibility of overdiagnosis. For patients
who had DMTs switched to either ocrelizumab or IFN/GA, the
reason was not recorded. In response, the IPTW approach was
used to minimise this source of bias, and a number of sensitivity
analyses were included to examine the robustness of the primary
outcomes. This allows us to infer with greater confidence that
the outcome differences between groups are likely treatment
related instead of differences in baseline characteristics between
groups. Ideally, we would have liked to perform the sensitivity
analysis for those initiating one of the study DMTs after 2016.
However, this led to a significantly smaller IFN/GA cohort,
unbalanced covariates and loss of statistical power. Therefore,
we used a 2014 cut-­off instead. By demonstrating similar find-
ings to our primary analysis with a doubly robust method, time
to first relapse survival models and post-­2014 analysis, we can be
confident that our findings are not the result of indication bias or
violation of the positivity assumption.
Significant limitations should be noted in our CDP/CDI anal-
ysis. We suffered a significant loss of power in the CDP/CDI
analysis, and mean follow-­up duration was only 3.57 years. We
also did not have data on comorbidities which can increase risk
of disease progression, particularly in older pwMS.40–42 Future
studies with larger cohorts and longer follow-­up periods are
required to confirm our findings.
We did not examine safety data for ocrelizumab compared with
IFN/GA. Given the greater propensity for infections and malig-
nancies with increasing age, this will be an important outcome
in future studies with longer follow-­ up, including systematic
serious adverse event follow-­up. These are now being conducted
in MSBase and other registries, in the context of POST-­Approval
Safety Studies.43 We were also unable to incorporate MRI data in
the primary PS model due to the limited number of participants
(n=133) with available MRI data. MRI lesions may present a
more sensitive marker of disease activity, and future studies will
examine any difference between ocrelizumab and IFN/GA on
MRI outcomes in this older cohort, as MSBase and other regis-
tries are now systematically collecting MRI scans from patients.
CONCLUSION
In pwMS above the age of 60 who either start or switch to ocreli-
zumab or IFN/GA, ocrelizumab was associated with a significant
reduction in ARRs and longer time to first relapse compared
with IFN/GA. The relapse rates in both treatment groups were
relatively low. This study adds novel information for clinicians
counselling older pwMS on the benefits and risks of low vs high
6 Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883
J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
efficacy DMTs, and highlights the presence of ongoing beneficial
treatment effects in pwMS over 60.
Author affiliations
1
Department of Neuroscience, Central Clinical School, Monash University, Melbourne,
Victoria, Australia
2
Alfred Health, Melbourne, Victoria, Australia
3
Department of Neurology, Tasmanian Health Service, Hobart, Tasmania, Australia
4
Eastern Health, Box Hill, Victoria, Australia
5
6
Melbourne Health, Melbourne, Victoria, Australia
Department of Neurology, The Royal Melbourne Hospital, Parkville, Victoria, Australia
7
Medical Point Hospital, Izmir, Turkey
8
Neuroscience, University of Catania Department of Surgical and Medical Sciences
and Advanced Technologies ’G.F. Ingrassia’, Catania, Italy
9
University of Catania, Catania, Italy
10
Hotel-­Dieu de Levis, Levis, Quebec, Canada
11
Department of Neurological Siences, University of Florence, Florence, Italy
12
CORe, Department of Medicine, University of Melbourne, Melbourne, Victoria,
Australia
13
Department of Neurology and Center of Clinical Neuroscience, Charles University
in Prague, 1st Faculty of Medicine and General University Hospital in Prague, Prague,
Czech Republic
14
Neurology, Jacobs Comprehensive MS Treatment and Research Center, Buffalo,
New York, USA
15
Hunter New England Health, New Lambton, New South Wales, Australia
16
The University of Newcastle, Newcastle, New South Wales, Australia
17
Karadeniz Technical University, Trabzon, Turkey
18
Neurology, Centro Hospitalar de São João, Porto, Portugal
19
Faculty of Health Sciences University Fernando Pessoa, Porto, Portugal
(HINARI) - Group A. Protected by copyright.
Twitter Yi Chao Foong @foongahh
Acknowledgements We thank the MSBase Operations team Ms Charlotte Sartori,
Ms Eloise Hinson, Ms Pamela Farr, Ms Rein Moore, Mr Dusko Stupar, Ms Cynthia
Tang and Ms Sonya Smirnova for maintaining the MSBase Registry that enabled
these analyses.
Collaborators MSBASE study group.
Contributors YCF performed the statistical analyses, interpreted the results,
drafted, revised the manuscript and acts as the study guarantor. HB, CZ and AvdW
contributed to study design, data collection, designed the statistical analyses,
interpreted the results, edited and reviewed the manuscript. MG, CZ and DM
contributed to the study design, interpreted the results and edited and reviewed
the manuscript. KB, MZ, WZY, VJ, MM, OS, SO, FP, PG, MPA, TK, DH, EKH, BW-­G, JLS,
CB and MJS contributed data, interpreted the results and edited and reviewed the
manuscript.
Funding The authors have not declared a specific grant for this research from any
funding agency in the public, commercial or not-­for-­profit sectors.
Competing interests YCF received travel compensation from Biogen, National
Health and Medical Research Council, Multiple Sclerosis Research Australia and
Australian and New Zealand Association of Neurologists. MG is currently working on
observational studies funded by Biogen and Roche. DM has received honoraria from
Novartis. CZ has nothing to disclose. KB has received honoraria for presentations
and/or educational support from Biogen, Sanofi Genzyme, Merck, Roche, Alexion and
Teva and serves on medical advisory boards for Merck and Biogen. MZ has received
conference travel support from Novartis and Roche, and research support from the
Australian Government Research Training Program and MS Research Australia. WZY:
received honoraria from Merck and Novartis. VJ receives research grant support form
F.Hoffmann La-­Roche, MS Research Australia and the National Health and Medical
Research Council of Australia (NHMRC 1156519). MM has received research support
from National Health and Medical Research Council, Medical Research Future Fund,
Brain Foundation, Charles and Sylvia Viertel Foundation, MS Research Australia,
Merck and Ku Leuven University, served on advisory board for Merck, has received
speaker honoraria from Merck and Biogen. OS received honoraria and consulting
fees from Bayer Schering, Novartis, Merck, Biogen and Genzyme. SO has nothing
to disclose. FP received personal compensation for serving on advisory board by
Almirall, Alexion, Biogen, Bristol, Merck, Novartis and Roche and received research
grant by Biogen, Merck and Roche and by FISM, Reload Association (Onlus),
Italian Health Minister, University of Catania. PG has served in advisory boards for
Novartis, EMD Serono, Roche, Biogen idec, Sanofi Genzyme, Pendopharm and has
received grant support from Genzyme and Roche, has received research grants
for his institution from Biogen idec, Sanofi Genzyme, EMD Serono. MPA received
honoraria as consultant on scientific advisory boards by Biogen, Bayer-­Schering,
Merck, Teva and Sanofi-­Aventis and has received research grants by Biogen,
Bayer-­Schering, Merck, Teva and Novartis. TK served on scientific advisory boards
for MS International Federation and WHO, BMS, Roche, Janssen, Sanofi Genzyme,

Novartis, Merck and Biogen, steering committee for Brain Atrophy Initiative by Sanofi
Genzyme, received conference travel support and/or speaker honoraria from WebMD
Global, Eisai, Novartis, Biogen, Roche, Sanofi-­Genzyme, Teva, BioCSL and Merck and
received research or educational event support from Biogen, Novartis, Genzyme,
Roche, Celgene and Merck. DH supported by the Charles University: Cooperatio
Programme in Neuroscience, by the project National Institute for Neurological
Research (Programme EXCELES, ID Project No. LX22NPO5107)—funded by the
European Union – Next Generation EU, General University Hospital in Prague project
MH CZ-­DRO-­VFN64165 and received compensation for travel, speaker honoraria and
consultant fees from Biogen Idec, Novartis, Merck, Bayer, Sanofi Genzyme, Roche,
and Teva, as well as support for research activities from Biogen Inc. EKH received
honoraria/research support from Biogen, Merck Serono, Novars, Roche, and Teva;
has been member of advisory boards for Actelion, Biogen, Celgene, Merck Serono,
Novars, and Sanofi Genzyme; received honoraria/research support from Biogen,
Merck Serono, Novars, Roche, and Teva; has been member of advisory boards for
Actelion, Biogen, Celgene, Merck Serono, Novars and Sanofi Genzyme; and has been
supported by the Czech Ministry of Education—project Cooperatio LF1, research
area Neuroscience, and the project National Institute for Neurological Research
(Programme EXCELES, ID project No LX22NPO5107)—funded by the European
Union-­Next Generation EU. BW-­G participated in speaker’s bureaus and/or served
as a consultant and received grants from Biogen, EMD Serono, Novartis, Genentech,
Celgene/Bristol Meyers Squibb, Sanofi Genzyme, Bayer, Janssen and Horizon; she
also serves in the editorial board for BMJ Neurology, Children, CNS Drugs, MS
International and Frontiers Epidemiology. JLS received travel compensation from
Novartis, Biogen, Roche and Merck and her institution receives the honoraria for
talks and advisory board commitment as well as research grants from Biogen, Merck,
Roche, TEVA and Novartis. CB received conference travel support from Biogen,
Novartis, Bayer-­Schering, Merck and Teva; has participated in clinical trials by Sanofi
Aventis, Roche and Novartis. MJS: Received consulting fees, speaker honoraria, and/
or travel expenses for scientific meetings from Alexion, Bayer Healthcare, Biogen,
Bristol Myers Squibb, Celgene, Janssen, Merck-­Serono, Novartis, Roche, Sanofi and
Teva. AvdW served on advisory boards for Novartis, Biogen, Merck and Roche and
NervGen, received research grants from National Health and Medical Research
Council of Australia, MS Research Australia, Novartis, Biogen, Merck and Roche,
received speaker’s honoraria and travel support from Novartis, Roche, Biogen and
Merck and is currently a coprincipal investigator on a cosponsored observational
study with Roche. HB’s Institution has received compensation for advisory boards
or lecture fees from Novartis, Biogen, Merck, UCB Pharma and Roche and receives
research funding from Novartis, Biogen, Merck, Roche, The National Health and
Medical Research Council of Australia, The Medical Research Future Fund (Australia),
Monash Partners, the Trish MS Foundation, The Pennycook Foundation, and MS
Australia; he also receives personal compensation as the Managing Director of the
MSBase Foundation and from the Oxford Health Policy Forum Brain Health Initiative.
Patient consent for publication Consent obtained directly from patient(s).
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.
ORCID iDs
Yi Chao Foong http://orcid.org/0000-0001-6447-078X
Vilija Jokubaitis http://orcid.org/0000-0002-3942-4340
Francesco Patti http://orcid.org/0000-0002-6923-0846
Tomas Kalincik http://orcid.org/0000-0003-3778-1376
Anneke van der Walt http://orcid.org/0000-0002-4278-7003
Chao Zhu http://orcid.org/0000-0003-3951-7501
REFERENCES
1 Schwehr NA, Kuntz KM, Butler M, et al. Age-­related decreases in relapses among
adults with relapsing-­onset multiple sclerosis. Mult Scler 2020;26:1510–8.
2 Sanai SA, Saini V, Benedict RH, et al. Aging and multiple sclerosis. Mult Scler
2016;22:717–25.
3 Mahad DH, Trapp BD, Lassmann H. Pathological mechanisms in progressive multiple
sclerosis. Lancet Neurol 2015;14:183–93.
4 Zeydan B, Kantarci OH. Impact of age on multiple sclerosis disease activity and
progression. Curr Neurol Neurosci Rep 2020;20:1–7.
5 Schweitzer F, Laurent S, Fink GR, et al. Age and the risks of high-­efficacy disease
modifying drugs in multiple sclerosis. Curr Opin Neurol 2019;32:305–12.
6 Prosperini L, Haggiag S, Tortorella C, et al. Age-­related adverse events of disease-­
modifying treatments for multiple sclerosis: a meta-­regression. Mult Scler
2021;27:1391–402.
Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883
Multiple sclerosis
J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
7 Vollmer BL, Wolf AB, Sillau S, et al. Evolution of disease modifying therapy benefits
and risks: an argument for de-­escalation as a treatment paradigm for patients with
multiple sclerosis. Front Neurol 2022;12:2646.
8 Hua LH, Fan TH, Conway D, et al. Discontinuation of disease-­modifying therapy in
patients with multiple sclerosis over age 60. Mult Scler 2019;25:699–708.
9 Hua LH, Harris H, Conway D, et al. Changes in patient-­reported outcomes between
continuers and discontinuers of disease modifying therapy in patients with multiple
sclerosis over age 60. Mult Scler Relat Disord 2019;30:252–6.
10 Kappos L, Bar-­Or A, Cree BAC, et al. Siponimod versus placebo in secondary
progressive multiple sclerosis (EXPAND): a double-­blind, randomised, phase 3 study.
Lancet 2018;391:1263–73.
11 Hauser SL, Bar-­Or A, Comi G, et al. Ocrelizumab versus interferon Beta-­1A in relapsing
multiple sclerosis. N Engl J Med 2017;376:221–34.
12 O’Connor P, Wolinsky JS, Confavreux C, et al. Randomized trial of oral teriflunomide
for relapsing multiple sclerosis. N Engl J Med 2011;365:1293–303.
13 Turner B, Cree BAC, Kappos L, et al. Ocrelizumab efficacy in subgroups of patients
with relapsing multiple sclerosis. J Neurol 2019;266:1182–93.
14 Roos I, Hughes S, McDonnell G, et al. Rituximab vs ocrelizumab in relapsing-­remitting
multiple sclerosis. JAMA Neurol 2023;80:789–97.
15 Hauser SL, Kappos L, Arnold DL, et al. Five years of ocrelizumab in relapsing multiple
sclerosis: OPERA studies open-­label extension. Neurology 2020;95:e1854–67.
16 Montalban X, Hauser SL, Kappos L, et al. Ocrelizumab versus placebo in primary
progressive multiple sclerosis. N Engl J Med 2017;376:209–20.
17 He A, Spelman T, Jokubaitis V, et al. Comparison of switch to fingolimod or interferon
beta/glatiramer acetate in active multiple sclerosis. JAMA Neurol 2015;72:405–13.
18 D’Agostino RB. Propensity score methods for bias reduction in the comparison of a
treatment to a non-­randomized control group. Statist Med 1998;17:2265–81.
19 Butzkueven H, Chapman J, Cristiano E, et al. Msbase: an international, online registry
and platform for collaborative outcomes research in multiple sclerosis. Mult Scler
2006;12:769–74.
(HINARI) - Group A. Protected by copyright.
20 Fox RJ, Miller DH, Phillips JT, et al. Placebo-­controlled phase 3 study of oral BG-­12 or
glatiramer in multiple sclerosis. N Engl J Med 2012;367:1087–97.
21 Newsome SD, Kieseier BC, Arnold DL, et al. Subgroup and sensitivity analyses
of annualized relapse rate over 2 years in the ADVANCE trial of peginterferon
Beta-­1A in patients with relapsing-­remitting multiple sclerosis. J Neurol
2016;263:1778–87.
22 Zhu C, Zhou Z, Roos I, et al. Comparing switch to ocrelizumab, cladribine or
natalizumab after fingolimod treatment cessation in multiple sclerosis. J Neurol
Neurosurg Psychiatry 2022;93:1330–7.
23 Zhu C, Kalincik T, Horakova D, et al. Comparison between dimethyl fumarate,
fingolimod, and ocrelizumab after natalizumab cessation. JAMA Neurol
2023;80:739–48.
24 Chatton A, Borgne FL, Leyrat C, et al. G-­computation and doubly robust
standardisation for continuous-­time data: a comparison with inverse probability
weighting. Stat Methods Med Res 2022;31:706–18.
25 Neugebauer R, van der Laan M. Why prefer double robust estimators in causal
inference. J Statist Plann Inference 2005;129:405–26.
26 Petersen ML, Porter KE, Gruber S, et al. Diagnosing and responding to violations in
the positivity assumption. Stat Methods Med Res 2012;21:31–54.
27 Olmos A, Govindasamy P. A practical guide for using propensity score weighting in R.
Pract Assess Res Eval 2015;20:13.
28 Austin PC, Stuart EA. Moving towards best practice when using inverse probability of
treatment weighting (IPTW) using the propensity score to estimate causal treatment
effects in observational studies. Stat Med 2015;34:3661–79.
29 Austin PC. Balance diagnostics for comparing the distribution of baseline covariates
between treatment groups in propensity-­score matched samples. Stat Med
2009;28:3083–107.
30 Corboy JR, Fox RJ, Kister I, et al. Risk of new disease activity in patients with multiple
sclerosis who continue or discontinue disease-­modifying therapies (DISCOMS): a
multicentre, randomised, single-­blind, phase 4, non-­inferiority trial. Lancet Neurol
2023;22:568–77.
31 Bonenfant J, Bajeux E, Deburghgraeve V, et al. Can we stop immunomodulatory
treatments in secondary progressive multiple sclerosis. Eur J Neurol 2017;24:237–44.
32 Kister I, Spelman T, Alroughani R, et al. Discontinuing disease-­modifying therapy in
MS after a prolonged relapse-­free period: a propensity score-­matched study. J Neurol
Neurosurg Psychiatry 2016;87:1133–7.
33 Kaminsky A-­L, Omorou AY, Soudant M, et al. Discontinuation of disease-­modifying
treatments for multiple sclerosis in patients aged over 50 with disease inactivity.
J Neurol 2020;267:3518–27.
34 Wolinsky J, Montalban X, Hauser S, et al. Prespecified subgroup analyses of
ocrelizumab efficacy in patients with primary progressive multiple sclerosis from the
phase III ORATORIO study. Int J MS Care 2018;20:1–128.
35 Epstein S, Fong KT, De Jager PL, et al. Evaluation of ocrelizumab in older progressive
multiple sclerosis patients. Mult Scler Relat Disord 2021;55:103171.
36 Andersen O, Elovaara I, Färkkilä M, et al. Multicentre, randomised, double blind,
placebo controlled, phase III study of weekly, low dose, subcutaneous interferon
Beta-­1A in secondary progressive multiple sclerosis. J Neurol Neurosurg Psychiatry
2004;75:706–10.
7
 Multiple sclerosis
37 McAdams M, Stankiewicz JM, Weiner HL, et al. Review of phase III clinical trials
outcomes in patients with secondary progressive multiple sclerosis. Mult Scler Relat
Disord 2021;54:103086.
38 Kapoor R, Ho P-­R, Campbell N, et al. Effect of natalizumab on disease progression
in secondary progressive multiple sclerosis (ASCEND): a phase 3, randomised,
double-­blind, placebo-­controlled trial with an open-­label extension. Lancet Neurol
2018;17:405–15.
39 Friese MA, Schattling B, Fugger L. Mechanisms of neurodegeneration and axonal
dysfunction in multiple sclerosis. Nat Rev Neurol 2014;10:225–38.
8 Foong YC, et al. J Neurol Neurosurg Psychiatry 2024;0:1–8. doi:10.1136/jnnp-2023-332883
J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2023-332883 on 7 March 2024. Downloaded from http://jnnp.bmj.com/ on May 13, 2024 at World Health Organisation
40 Marrie RA, Rudick R, Horwitz R, et al. Vascular comorbidity is associated
with more rapid disability progression in multiple sclerosis. Neurology
2010;74:1041–7.
41 McKay KA, Tremlett H, Fisk JD, et al. Psychiatric comorbidity is associated with
disability progression in multiple sclerosis. Neurology 2018;90:e1316–23.
42 Zhang T, Tremlett H, Zhu F, et al. Effects of physical comorbidities on disability
progression in multiple sclerosis. Neurology 2018;90:e419–27.
43 Ng HS, Rosenbult CL, Tremlett H. Safety profile of ocrelizumab for the treatment of
multiple sclerosis: a systematic review. Expert Opin Drug Saf 2020;19:1069–94.
(HINARI) - Group A. Protected by copyright.