Most RRMS patients starting treatment with a disease-modifying therapy (DMT) will necessarily begin with a front-line injectable (interferon-beta, glatiramer acetate) or oral therapy (teriflunomide, dimethyl fumarate; fingolimod in some countries). As previously noted (see Part 2 in this series), clinicians opt for a treatment with a favourable risk-benefit profile, which is generally defined as low-risk and low-benefit.
The expectation is that many patients may respond adequately to a first-choice agent, and those with “breakthrough” disease activity can be switched at a later visit. So risk can be titrated up according to the clinical need.
“Breakthrough disease” is poorly characterized and a number of attempts have been made to identify patients at risk of progression while on treatment. The Barcelona group developed the Rio criteria, which scored risk from 0 to 3 after one year on interferon-beta treatment according to MRI (>2 active T2 lesions), relapse (>1 relapse), and EDSS (>1 point) criteria (Rio et al. Mult Scler 2009;15:848-853). A change in only one domain (MRI or relapses or EDSS) was not predictive of new clinical activity or disease worsening in the following two years. Interestingly, isolated relapses were associated with a non-significant decrease in the risk of progression (odds ratio 0.5). However, the positive predictive value of a Rio score >2 was only 47.8%, which was less than the PPV of new/enlarging T2 lesions alone (Rio & Ruiz-Pena. J Neurol Sci 2016;361:158-167).
The modified Rio criteria simplified patient assessment by using only MRI and relapse criteria (Sormani et al. Mult Scler 2013;19:605-612). Risk was stratified as low (score 0: < 4 new T2 lesions + no relapses), intermediate (score 1: < 4 new T2 lesions + 1 relapse; or >4 new T2 lesions + no relapses) or high (score 2 or 3: < 4 new T2 lesions + >2 relapses; >4 new T2 lesions + 1 relapse; or both >4 new T2 lesions + >2 relapses). So patients could be considered low-risk even if they had ongoing inflammatory activity. This contrasts with the treatment optimization recommendations developed by the Canadian MS Working Group (Freedman et al. Can J Neurol Sci 2013;40:307-323). With this approach, a patient with no relapses and 3 new lesions (mRio score of 0) would be classified as a High level of concern and would be a candidate for switching therapies.
While the mRio scoring system provides some indication of relative risks, it should be noted that the risk of progression was high in all groups. In the low-risk group (score 0), the risk of progression at 3-year follow-up was 32% (vs. 50% for score 2-3) in interferon-treated patients. Similar results were seen in an analysis of teriflunomide-treated patients in the TEMSO trial (Sormani et al. ECTRIMS 2015; abstract P1131): an estimated one-third of patients with an mRio score of 0 would progress at 7-year follow-up. In addition, in a MAGNIMS Study Group analysis, the risk of treatment failure was 27% for patients with 1 relapse or >3 new T2 lesions in the first year of treatment; and 48% for those with >1 relapse, or 1 relapse and >3 new T2 lesions (Sormani et al. Neurology 2016;87:134-140).
Thus, one-third of patients who are seemingly stable – relapse-free and a low number of new lesions – during the first year on treatment would be expected to experience worsening disability over the next few years.
A more stringent benchmark of treatment response is no evidence of disease activity (NEDA), defined as no relapses, no MRI activity and no EDSS change. Perhaps the most common objection is that this measure is not sustainable in clinical practice. For example, in the CLIMB (Comprehensive Longitudinal Investigation of Multiple Sclerosis at Brigham and Women’s Hospital) cohort study, 46% achieved NEDA at 1 year, but only 7.9% maintained their NEDA status at 7 years (Chitnis et al. JAMA Neurol 2015;72:152-158). As the Canadian MS Working Group noted, attaining NEDA may be optimal, but may not be feasible for most patients (Freedman 2013).
However, it is unclear if even NEDA is adequate as a benchmark, or whether the metrics of relapses/MRI + EDSS are useful as a guide. For example, in the EPIC (Expression/genomics, proteomics, imaging, and clinical) longitudinal study, patients with NEDA at year 2 did not have better outcomes at 10-year follow-up (Cree et al. Ann Neurol 2016;80:499-510). NEDA does not appear to indicate remission, as the authors noted, and many patients who meet NEDA criteria will develop significant disability, which may be the result of ongoing axonal degeneration, spinal cord lesions, or other factors. This is the rationale for assessing NEDA-4, which incorporates brain atrophy rates, to provide some indication of the rate of tissue loss on therapy (Rojas et al. Neurol Res 2014;36:615-618). While there are numerous methodological and logistical challenges to using brain atrophy measures in clinical practice, brain atrophy data from phase III trials are useful as a proof-of-principle to provide some indication of the comparative benefits that may be achieved with higher-efficacy agents.
A further consideration is that NEDA status and “breakthrough disease” do not take into consideration the idea that some degree of ongoing inflammation may be beneficial. An observational study of natalizumab-treated patients found that relapses and mRio scores in the first year of therapy were predictive of EDSS change at year two, however, this effect disappeared with longer follow-up (Raffel et al. PLoS One 2017;12:e0169546). In fact, ongoing inflammatory activity in the first year of treatment was predictive of a lower risk of EDSS progression in years 3-7. These findings bring to mind the concept of good versus bad inflammation that was bruited a decade ago (Martino et al. Lancet Neurology 2002;1:499-509). Thus, some degree of “breakthrough” inflammatory activity may represent a re-regulation of the aberrant immune response in MS.
Accordingly, recalibrating different components of the immune system – by sequestering pathogenic T cells (e.g. fingolimod) or reconstituting T and B cell subsets (e.g. alemtuzumab, BMT, cladribine) – may provide greater long-term benefits. While this remains to be determined, treatment escalation may be appropriate even in patients who appear stable according to conventional measures in the first few years of frontline treatment, but who are in reality at high risk of significant disability progression over the next 5-10 years.
Dr. Paul Giacomini: The advent of numerous new disease modifying therapies in recent years has greatly expanded the therapeutic options available for patients with MS. However, this has also presented new challenges to patients and physicians alike, specifically with regard to when to switch therapies, and what are the optimal agents and sequences to use. Significant variability persists among neurologists regarding therapeutic preferences, even among second- and third-line choices. Consensus among MS experts has remained elusive.
Many other factors also influence the choice of therapy, including many patient-related factors (risk tolerance, patient preferences, etc.), as detailed by Dr. Yeung in the commentary on part 2 of this series. Another consideration should be the risk of compound toxicities – the risk of exposing patients to several different therapies that may all alter immune function in different and lasting ways. Although the focus among neurologists thus far has largely been on achieving as much suppression of disease activity as possible, a lasting remission is very difficult to achieve even with the most potent agents available, as indicated above. Less consideration has been given to the potential long-term risks of sequencing between numerous potent agents, each of which may each be altering the immune system in lasting ways, even after the patient has been moved to a newer drug.
Many of the newer therapies have not been on the market long enough to have convincing long-term safety data, and we should not assume that there are no long-term risks. We are already seeing that some of the newer oral and monoclonal therapies may have lasting effects on lymphocyte subsets, even years after being discontinued. This should give us pause, and remind us that although therapeutic efficacy should be our prime consideration, we should also remain vigilant about the potential long-term risks of therapies we prescribe. Furthermore, we should be thoughtful about how, and when we sequence between medications. The therapeutic goal should remain to try and achieve adequate control of disease activity, but ideally with the fewest number of agents possible, and in the safest possible order.
In part 4 we will discuss the need to develop a long-term treatment plan.
NeuroSens survey on sequencing – Part 3