One of the co‐authors of the protocol, for unforeseen reasons, could no longer collaborate on the review and asked to be excluded from authoring.

We planned the following subgroup analyses at the protocol stage but did not perform them due to lack of sufficient data: different doses and regimens, disease type (relapsing‐remitting, primary‐progressive, secondary‐progressive or progressive‐relapsing), disability at baseline (EDSS scores ≤ 5.0 or ≥ 5.5).

We did not plan the subgroup analysis taking into account previous treatment (naive versus pre‐treated patients) at the protocol stage but as it seemed to be clinically relevant, we performed it when there were sufficient data.

In the protocol, we proposed to assess 'number of participants without relapse' and 'number of progression‐free participants' as primary outcomes. However, we found available data only for 'relapse‐free survival' and 'sustained disease progression‐free survival'. We decided to use both of these measures in the final review, since they are the opposite of the same measurement and this change would not influence the effect of the intervention nor its interpretation.

In the protocol, we proposed to assess the 'Changes in the number of MRI T2‐ and T1‐weighted lesions after treatment'. However, we discussed among the review authors and peer reviewers whether 'Patients with new or enlarging T2‐hyperintense lesions' could more relevant for decision‐making, since it comprises the total area of all MS lesions. Moreover, since this is a dichotomous measure, it could be easier for physicians and consumers to interpret. We considered this outcome relevant enough (despite the fact that it is a secondary outcome) to warrant a change in its presentation regardless of the tool/measurement used to assess it. Therefore we decided to change the way it was measured, leading to the presentation of the existing results.

We planned the following outcomes for the 'Summary of findings' table at the protocol stage but we did not include them due to lack of data: one‐point EDSS score increase confirmed at three months' follow‐up and quality of life. We included the outcome dropout rate in the 'Summary of findings' table.

In the protocol we did not pre‐define the use of the hazard ratio to estimate the effect size of time‐to‐event outcomes. We added this information in the final version of the review because we used this measure for the outcomes relapse‐free survival and sustained disease progression.