Point/Counterpoint: Food and Drug Administration Multiple Sclerosis Categorization Changes

A Brief History of Multiple Sclerosis Classification

An accurate description of the clinical course of multiple sclerosis (MS) patients is helpful for clinical trial design, providing study population homogeneity. It also helps prognosticate and guide everyday clinical decision making, including the selection of disease modifying therapy (DMT).

The1996 classification of MS disease course was developed by an international survey of clinicians and included 4 categories or phenotypes of MS. These categories (or phenotypes) were relapsing-remitting MS (RRMS) secondary progressive MS (SPMS), primary progressive MS (PPMS) and progressive relapsing MS (PRMS).¹ Key to this classification is conceptualizing of MS as a disease with 2 distinct courses: 1) PPMS, in which insidious progression in the absence of a history of acute relapse(s) occurs, and 2) RRMS, which is progressive disability related to relapses and MRI measures of disease activity (new T2 lesions or T1 gadolinium-enhancing lesions), followed by SPMS, in which disability progression is independent of relapses and MRI disease activity. These fundamental phenotypes served as the foundation for designing pivotal trials for potential DMTs for MS over 2 decades and have been the basis for approving DMTs as being principally for RRMS or “relapsing forms” of the disease.

Although the 1996 classification was rapidly adopted, advances in MRI measures and emergence of new therapies identified limitations, particularly in relation to response to treatment and defining the transition from RRMS to SPMS. In 2013, updates to the classification were made.² These updates added the categories clinically isolated syndrome (CIS) and radiologically isolated syndrome (RIS), potential predecessors to the ultimate diagnosis of MS, with CIS considered as a first presentation of MS that does not yet meet the MS diagnostic criteria of dissemination in time. The MS clinical phenotypes were also updated to include disease-activity assessments—the presence of clinical relapse or new MRI lesions being termed active disease and used in combination with RRMS, SPMS, and PPMS. The word progression was defined to be used in combination with SPMS and PPMS to denote disability progression independent of relapse and/or MRI disease activity. The intention was to ensure that clinicians consistently and repeatedly evaluate their patients with MS patients to enhance vigilance for the presence of active disease.

Classifications and Treatment Indications

Because most of the previous pivotal studies for DMTs were done in populations with RRMS who had relapses and/or MRI activity, the Food and Drug Administration (FDA) approval remained consistent with the 2013 terminology and was for people with RRMS. In clinical practice, this meant the status of many individuals who progressed from RRMS to SPMS was not updated because of fear that continued use of DMTs might be denied by insurance companies or pharmacy benefit managers. This discrepancy in what patients and clinicians experienced and what was documented caused confusion for patients, families, and clinicians.

In 2017, the first treatment for PPMS, ocrelizumab, was approved,³ and in March 2019 (3 days apart), siponimod and oral cladribine were both approved for treating RRMS and active SPMS (aSPMS).^4,5 The incorporation of the term active in the FDA’s approval subtly—but significantly—changed MS categories used for defining treatment indications.

The indication for treatment of RRMS and aSPMS was also retroactively applied (with certain nuances) to already available DMTs.

A Step Forward!

By Svetlana Primma Eckert and Bianca Weinstock-Guttman

Changing the indications for DMTs for MS to include both RRMS and aSPMS is significant because a large majority of people with aSPMS are simply people with SPMS with ongoing underlying inflammatory activity and a level of disability. It appears that the clinical course of RRMS conversion to SPMS occurs over a long and variable period of time rather than at a defined specific time point. Therefore, many people with early SPMS may still experience relapses MRI disease activity. The SPMS definition previously precluded the use of many DMTs in more advanced stages of MS. It is also important to mention that for some people with SPMS taking DMTs that were initiated in the RRMS phase, relapse and MRI disease activity can both be masked by the DMT.

The FDA approval of siponimod was based largely on a phase 3 study with 1,645 participants with SPMS.⁶ This was the first large-scale study of a DMT in a population with SPMS, providing the highest level of evidence for a positive therapeutic outcome to date for SPMS. The study population included participants with high disability (Expanded Disability Status Scale [EDSS] median score of 6) and a median 2.6 years since conversion to SPMS. A high proportion of participants had no relapses in the year (79%) or 2 years (64%) before enrollment and 75% had no gadolinium-enhancing lesions. Despite this, the FDA did not provide approval for all people with SPMS based on their own assessment of data from the study. Nevertheless, with this broader indication, individuals with SPMS who experience a relapse or disease activity on MRI can now potentially qualify for the different available therapies.

Although individuals with SPMS who have not had a relapse or MRI disease activity are excluded from the indications for DMTs, the inclusion of those with aSPMS is a big step forward for managing SPMS for several reasons.

First, more DMTs for MS will be available to patients and potentially covered by health insurance. Many insurance companies follow recommendations from the FDA when updating their formularies with lists of specifically approved medications. The new definitions and criteria may provide physicians with more leverage to prescribe DMTs to patients who have SPMS. Additionally, specific DMT initiation will be more freely approved based on clinical features of MS rather than consideration of a multitude of financial factors.

Second, the inclusion of CIS as an indication for DMTs initiation can allow for earlier initiation of highly effective first-line therapies thereby increasing relapse-free periods and significantly slowing down disability progression, accumulation of new or enhancing MRI lesions, brain-volume loss, and cognitive changes. This has the potential for increasing the number of years patients live with a good quality of life with less disability, and for decreasing the healthcare costs associated with active inflammatory pathology.

Furthermore, the FDA update may allow greater transparency and specificity of MS course and promote better communication regarding prognostication and appropriate goals of care in discussions with people with MS and their loved ones.

We therefore posit that the updated FDA recommendations on the use of MS DMTs in aSPMS and CIS are indeed a large positive step in making treatment more easily available to MS patients and will likely lead to better more efficient and cost-effective care.

A Step Backward or a Well-Intentioned Stumble!

By Stephen Krieger

Changing indications retroactively has meant that treatment indications for a DMT do not always align with the populations studied in the clinical trials of that agent. This has created inconsistencies in the approved indications and dissonance between the approval language and the data from the pivotal trials. I argue that this is a step backward—or at least, a well-intentioned stumble—that is nonetheless out of step with the direction that the field is heading.

Approaches to categorizing and phenotyping a disease must be made with clear goals and intentions. In my opinion, the FDA MS disease categorizations ideally need to bridge 3 different approaches, including 1) how MS should be assessed over time, 2) how to apply clinical trial data in MS, and 3) how we currently understand MS disease course and pathophysiology.

Assessing Over Time

The 1996 classification and 2013 revisions specify that the assessment of disease activity is a time-locked phenomenon; disease activity must be considered present or absent over a specific time interval. The new FDA categorization of aSPMS incorporates the concept of disease activity from most recent classification, but it does not specify a time frame for that disease activity, other than a vague, and certainly not universal, “in the first few years” of SPMS. This leaves great uncertainty in how the new categorizations should be applied in practice. When disease activity ceases in SPMS is not merely an academic or semantic question. There is no consensus in the field, and by not specifying a time frame within which disease activity ought to be assessed (perhaps every year, or even every 2 or 3 years), the FDA’s categorizations fail to make good on the intent of the 2013 classification update.

Applying Clinical Trial Data

The FDA’s approval indication language needs to be based on the inclusion criteria and outcome measures used in the pivotal clinical trials so that clinicians can determine which patients are best suited to specific treatments. Drug approvals and indications for MS have become a contortionist act, in an attempt to fit trial outcomes to new disease category definitions that do not line up with how the trials were designed. This has yielded a series of unfortunate inconsistencies that clinicians are left to navigate. In the case of siponimod, the FDA approval was not based on the clinical trial as designed or the success in meeting the primary endpoint of preventing disability in SPMS,⁶ but rather was based on a subgroup with aSPMS and the category of RRMS, which wasn’t studied in a phase 3 trial. In contrast, FDA approval of ocrelizumab was for all PPMS, not just active PPMS, even though the efficacy data is similarly more impressive in this subgroup.³

There are more inconsistencies because the current FDA categories are retrofitted onto subtypes of MS for which the treatments were not studied or were not successful. Because siponimod was approved as a first-line agent for RRMS, this approval also includes CIS, even though the siponimod pivotal trial enrolled only people with SPMS—the opposite end of the disease spectrum. Natalizumab received approval for active SPMS, even though the ASCEND SPMS trial^a failed to show benefits of natalizumab for SPMS. In other cases, new agents were not approved for subtypes for which they have been studied. For example, oral cladribine was approved for aSPMS, for which it was not studied, but not for CIS, although there was a successful trial of oral cladribine for CIS trial a decade ago.⁷ Here the FDA is conflating a risk-benefit assessment (whether an agent ought to be first-line or not) with an assessment of the appropriate clinical course categorization that is best justified by clinical trial populations studied.

Clinicians are left with conundrums, increased complexity, and lack of actionable guidance—and thus insurance providers and third-party payers are given opportunities to further restrict approvals. Complexity, vagueness, and lack of consensus are to the insurance industry’s advantage, wielded to dictate how neurologists can practice in a way that serves their bottom lines. That is a serious real-world problem.

Multiple Sclerosis Course and Pathophysiology

Let’s also consider the third point on the above list of what FDA categories need to accomplish, and that is to move closer to the ground truth of MS. There is an emerging understanding that MS is, from the outset, comprised of an admixture of disease activity and progression, a view encapsulated by the topographic model of MS⁸ that dynamically depicts disease course in a unified way without clear boundaries between the categories. The topographical model of MS posits that although MS often begins with demonstrable relapses, there is an accelerated loss of reserve, beginning at the outset of the disease. Thus, what we think of as clinical progression is actually a late manifestation of lost compensatory reserve.⁹ This model is both unifying, in that it represents the dynamic blurring of distinctions between phenotype categories, and personalized, in that it encapsulates an individual’s disease course as being defined by their specific disease topography. It suggests there is not a definitive difference between PPMS and SPMS, which cannot be prognostically distinguished,¹⁰ and that progressive disease ought to be considered a single entity.

A Concluding Example

An example of a recent FDA approval highlights the way their categorizations force distinctions that may not be biologically justified. Ocrelizumab is approved for RRMS, aSPMS, and PPMS, but not SPMS without activity. So, if a person has progressive MS but never had a relapse, they can receive ocrelizumab —for this we call PPMS. But if a person has progressive MS and did have a relapse 10 years ago, they cannot receive ocrelizumab—for this we call nonactive SPMS. Is ocrelizumab less likely effective in the person who had the relapse 10 years ago? Even though this same agent is also approved for RRMS? This example speaks to all 3 issues outlined. The new FDA categories impose arbitrary distinctions that are not true to the biology of the disease; they yield inconsistencies in how clinical trial data were interpreted, and they do not provide time-based guidance on how to assess and optimize treatment for our patients in practice.

Rebuttal: Potentially Better Access to Treatment is a Step Forward

By Svetlana Primma Eckert and Bianca Weinstock-Guttman

As elegantly described by Dr. Kreiger, MS disease progression is well known as a continuum, related initially to occurrence of relapses and the gradual loss of compensatory reserve. Nevertheless, the compensatory reserve is linked to many factors including age at disease onset, disease duration, and aging. The underlying pathophysiologic loss of compensatory reserve is associated primarily with the loss of neural tissue and functional connectivity for which most of approved DMTs showed certain benefit. Furthermore, factors such as education or leisure activity are also strongly involved in maintaining the functional reserve.¹¹ These additional factors are currently considered only within post-hoc analyses of pivotal studies, or during phase 4 trials.

Most of previous studies of DMTs in MS were launched before the 2013 new disease course criteria became available and included participants with higher EDSS scores of 3 to 5.5, who most probably had already begun the conversion to a secondary progressive disease course, but still had superimposed relapses or disease activity on MRI as per study inclusion criteria. Therefore, retroactive FDA approval of most DMTs that include the subset of patients with aSPMS is, in our opinion, a step forward, allowing more evidence-based and perhaps more appropriate application of DMT studies’ data. As more information becomes available on the underlying pathophysiology of MS, and the benefits of DMTs with different mechanisms of action on controlling the CNS inflammatory milieu, providing the new classification of aSPMS versus not active SPMS is beneficial.

Furthermore, the consideration of aging not only as related to chronologic definition of higher age but also to biologic changes and functional neurologic reserve is also important. The repair processes and the capacity for remyelination is clearly important in maintaining the functional reserve and unfortunately decreases with aging.¹²

Studies suggest that certain biomarkers associated with biologic aging can be seen earlier in people with MS. For example, shorter telomere length is associated with faster disability progression and brain atrophy.¹³ Therefore, disease duration and age should be considered when a specific DMT is recommended. The recent 2017 McDonald criteria that require inclusion of clinical disease course status support the need for more exact definitions to help guide clinicians, which is included in the new FDA approval updates. With the greater number of therapies to choose from, DMT selection by clinicians may be more appropriate when trying to simultaneously account for disease course, disease duration, and a patient’s age.

In conclusion, with the new FDA updated definitions, both physicians and patients will have more leeway in choosing an appropriate DMTs, as subsequent changes are expected within insurance formularies. However, the safety and benefit of different DMTs available today for use in individual patients will remain an important deliberation, as part of the decision-making discussion between physicians and their patients.

Rebuttal: Indvidualization of Care on a Continuum Is Preferable

By Stephen Krieger

The essential problem is that the new FDA categories import the concepts of disease activity and progression from the 2013 clinical course phenotypes in a way that does not adhere to their time-locked intent, which defines MS as a dynamic disease. The concepts of active or nonactive MS are meaningless without a specified time frame. The new nomenclature therefore fails to provide actionable guidance to clinicians with regard to when, how, and how often, to best surveil for both disease activity and progression. That is the real missed opportunity of the new FDA categories as they stand now. This may give clinicians less leverage, rather than more, over third-party payers, who may take advantage of the lack of consensus in our field to effectively make clinical decisions for us.

Drs. Eckert and Weinstock-Guttman take an optimistic view that the revised FDA categories will enable easier access to medications across the spectrum of MS disease. Although I agree that would be a terrific outcome of this new nomenclature, I am afraid there is little evidence to suggest this will be the case. Adding a CIS indication for agents approved as first-line treatments for RRMS merely acknowledges that most cases of CIS qualify as RRMS under the revised McDonald criteria. Retroactive CIS indications are unlikely to increase insurance approvals for DMTs for substantially more people. However, in the case of siponimod, adding a CIS approval for a medication studied only in SPMS undermines the intrinsic idea that therapeutic approvals are based directly on clinical trial data.

Drs. Eckert and Weinstock-Guttman also argue that the aSPMS designation may allow for DMT approvals in more advanced stages of MS. I am concerned that the opposite is true: that the new aSPMS category still does preclude DMT access for those with late-stage disease. This is particularly concerning because although the FDA did not provide a timeframe for assessments of disease activity, they implied this occurs only within the first few years of the transition to SPMS. The categories thus maintain distinctions that we have little basis to support. For instance, the category of aSPMS assumes that there is an identifiable moment in time when RRMS transitions to SPMS,¹⁴ or when disease activity truly ceases and a patient develops nonactive SPMS. Most patients with SPMS are still left without treatment options, and by approving siponimod only for aSPMS, even after a legitimately successful SPMS trial,⁶ the FDA’s decision may be a disincentive for other companies to pursue approvals for progressive disease.

Finally, the future of MS care is one of more robust and personalized data capture—clinically, radiographically, and through laboratory biomarkers. Even if one is unconvinced that all clinical phenotypes of MS are manifestations of a single, unified disease as proposed by the topographical model,⁸ the FDA’s approval language is built on categoric distinctions and crude assessments of disease activity and progression. The care we provide for our patients is increasingly based on more granular sensitive measures of the burden of disease that are not predicated on the FDA’s phenotype categories. Fatigue, cognitive dysfunction, pain, and diminished quality of life all occur irrespective of the classification as having CIS, RRMS, aSPMS, nonactive SPMS, or PPMS. Our goal is to provide treatment to optimize individual outcomes. The more we keep putting patients in artificial boxes from the perspective of drug approvals, the less we attend to those nuances, which is clearly the direction our field is moving.