Assessment

Problem

Is the problem a priority?

Judgement

Research evidence

Additional considerations

The Atlas of MS estimates there are 2.8 million people living with MS (pwMS). 85% of these are initially diagnosed with relapsing forms of MS (RMS). There is unequal access to disease-modifying therapies (DMTs) globally, with 14% of countries not having access to any on-label DMTs. Low income countries (LICs), lower middle income countries (LMICs) and upper middle income countries (UMICs) are affected more than high income countries (HICs) by lack of access to DMTs . Evidence for both on-label and off-label DMTs should be considered when considering essential medicines for MS. PICO 1: The Panel decided to review DMTs for active and/or worsening forms of relapsing MS to consider the most appropriate treatment approach. PICO 2: The Panel decided to review DMTs for not active and not worsening or indeterminate forms of relapsing MS to consider the most appropriate treatment approach. PICO 3: The Panel decided to review DMTs for active and/or worsening forms of relapsing MS when there is a lack of treatment response to consider the most appropriate treatment approach. Panel members with COI for DMTs reviewed for RMS: Anthony Traboulsee, Jagannadha Avasarala, Carlos Navas, Maya Zeineddine, Riley Bove, Dina Jacobs, Shanthi Viswanathan, Bassem Yamout, Kathy Costello. Undetermined COI: Hans-Peter Hartung.

Desirable Effects

How substantial are the desirable anticipated effects for each intervention?

Judgement

Research evidence

Additional considerations

The evidence base on disease-modifying treatments (DMTs) for relapsing multiple sclerosis (RMS) was retrieved through a systematic review of the biomedical literature developed according to the Cochrane methodology. The search was performed on February 11, 2022. Included studies were randomised-controlled trials (RCTs).

Thirty treatments (with registered indications for MS, as well as non-licensed but used off-label in clinical practice), compared vs placebo or vs any other DMT, were included in a network metanalysis (NMA).

Direct, pairwise comparisons were assessed assuming placebo as the common comparator, a choice that inevitably resulted in not including in the analysis comparisons with active comparator. However, in the NMA, estimates from indirect comparisons included also such evidence, provided that a comparison with placebo was included in the loop. An alternative NMA featuring interferon beta 1a as the common comparator, given its higher relevance than placebo in current clinical practice, was performed by the evidence review team upon request by MEMP. However, the panel concluded that choosing placebo as the common comparator allowed a more comprehensive assessment.

Among people with RMS, three populations were identified by MEMP: with active RMS, with non-active RMS and with active RMS when there is a lack of treatment response (switching).

We retrieved 50 RCTs (36,541 participants in total) eligible for analysis. Twenty studies included only people with active RMS. Twenty-six studies included a mixed population of people with active RMS and lack of treatment response together with treatment-naive people. The proportion of people with previous lack of treatment response in these studies varied from 3% to 75% (median 33%). Separate results for people with previous lack of treatment response were not reported in studies and the inclusion criteria featured a number of different definitions for “allowed previous treatments” (more or less drug-specific and with different washout time windows, depending on the treatment). Such heterogeneity did not allow a meaningful data pooling of the population with previous lack of treatment response.

Two small studies (88 participants in total) included people with non-active RMS and in two other studies (240 participants in total) the RMS phenotype was not reported.
The panel agreed in considering as the evidence base the analysis including all retrieved RCT as representative of people with active RMS.

Among the desirable effects, disability worsening and frequency of relapse were assessed for most DMTs.
Disability at 24 months assessed by means of the EDSS is the desirable effect on which most data were available, when considering placebo as the common comparator. All 18 DMTs with disability at 24 months data reported an absolute difference in favour of the intervention, with two notable exceptions: ozanimod and interferon beta products (beta 1a and 1b considered together), showing values in favour of placebo. However, such estimates need to be interpreted with caution, since both show a very low certainty due to imprecision (and also risk of bias for interferon beta products). In particular, the point estimate for interferon beta products, showing very wide CIs, came from only indirect comparisons in the network evidence (see network plot), referring to two small studies (less than 250 participants in total) comparing beta interferons with azathioprine. Point estimates from studies directly comparing interferon beta 1a or beta 1b vs placebo, showed values in favour of the intervention.
No study of DMTs vs placebo assessed disability at 36 months.
Relapse was assessed at 12 and 24 months for most DMTs, showing values in favour of the intervention. Considerations mentioned above on disability and the certainty of point estimates of beta interferon products, compared together vs. placebo, can be made about relapses (see "Additional Considerations"). Direct evidence about the frequency of relapse at 36 months vs. placebo was available only for interferon beta 1b, with values favouring the intervention.
Data on MRI outcomes (new or enlarging T2-weigthed lesions and new gadolinium-enhancing positive T-1 weighted lesions) were available at 12 and 24 months. The majority of MRI estimates were available for DMTs compared to placebo relative to gadolinium-enhancing positive T1-weigthed lesions at 24 months. Most absolute point estimates were in favour of the intervention with some exceptions: for T2-weighted MRI lesions at 12 months most estimates came only from indirect evidence and wide loops in the network plot, with resulting very wide CIs and very low certainty mostly due to imprecision. Therefore such values should be interpreted with caution (see "Additional Considerations").
Quality of life was assessed, by means of several different scales, for cladribine, teriflunomide, daclizumab, ozanimod and interferons beta 1b and 1a vs placebo, showing values in favour of the intervention.
Cognitive decline was assessed in no study comparing a DMT vs placebo, therefore no estimates on this outcome were available in the NMA.