A towering edifice of MS therapies has been built in recent years, but the structure has an exceedingly narrow base of evidence when it comes to how to sequence these agents. How does one construct something that won’t come crashing down – either because of too little efficacy or efficacy deferred, or too great a risk for the benefits that can be achieved? And in devising a blueprint, how does one manage the risk conundrum: younger, less disabled patients generally wish to avoid risk when a more aggressive approach may well change the slope of progression; whereas older, more disabled patients are ready to assume greater risks, when it is likely too late to change the outcome?
In the absence of evidence, some general principles have emerged in recent years that provide a rough guide to sequencing therapies in patients with relapsing-remitting MS.
1. Starting a treatment is more important than selecting a specific treatment. All of the front-line options have modest efficacy, so there is little advantage in choosing one over another on that basis. The choice of therapy will be influenced by patient preference (route of administration, dosing frequency, side effect profile), and reimbursement. The main hurdle at this stage is treatment refusal. Providing a menu of options is likely to reinforce the notion that treatment is optional; deferring the decision to the next visit communicates that there is no particular urgency to starting treatment. A more tolerable starting therapy will cause less confusion: switching decisions will be more appropriately based on efficacy; and the efficacy will be less likely to be compromised by poor adherence. Treatment refusal will be a lesser consideration in patients who present with more aggressive disease (see below); a more aggressive treatment plan will also be easier to implement in this subgroup since patients with severe symptoms or early disability will perceive the need.
2. Plan to switch early, as needed. The challenge with a start-anything approach is inertia (Saposnik et al. BMC Neurol 2016;16:58): there is a tendency on the part of both physician and patient not to deviate from the Newtonian course once a treatment has been rolled out. However, after a treatment has had sufficient time to demonstrate effectiveness (6-9 months), close monitoring is needed to identify ongoing disease activity, with the view to switching as soon as possible to an agent with greater potency. An analysis by MSBase reported that switching to a higher-efficacy agent (e.g. fingolimod, dimethyl fumarate, MAbs) resulted in better long-term outcomes compared to a lateral switch (Lizak et al. ECTRIMS 2015; abstract P636). Moreover, the risk of further accumulation of disability was reduced with more prolonged exposure to a higher-efficacy agent, underscoring the need to escalate earlier in the treatment course.
3. Don’t limit your options. In the escalation model (rather than induction), the next treatment is generally one with an incremental gain in efficacy. There is little advantage to a lateral switch to another front-line therapy: the therapeutic window is narrow and too much time will be spent before an effective therapy is found. In practice, the choice of an intermediate-efficacy agent, such as fingolimod, has shown good long-term safety and tolerability (Cohen et al. ECTRIMS 2015; abstract P591). Improvements in patient-reported outcomes have also been reported after switching from an injectable to fingolimod (Fox et al. Mult Scler Relat Disord 2014;3:607-619).
A key consideration is whether the choice of any given therapy will limit subsequent treatment options. At issue are carry-over effects from one agent to another, as well as cumulative toxicities resulting from repeated interventions that may adversely affect liver/kidney function or immune cell repopulation. As noted in the previous installment of this series, a CBC with lymphocytes, and liver function testing, may be advised prior to initiating another agent (see MS Sequencing: Part 6: Safety considerations when sequencing, NeuroSens, July 26, 2017). Accordingly, an initial therapy with a low potential for toxicities (e.g. glatiramer acetate) or which can be actively eliminated (e.g. teriflunomide) may be preferred to one that may be associated with elevated hepatic enzymes or persistent lymphopenia, which will delay initiation of the next agent in sequence.
4. Have an exit strategy: With respect to limiting treatment options, an important consideration is whether a given therapy has a next step, or whether there is a need to tread exceedingly carefully. This is the problem posed by natalizumab: it may be viewed more as a rescue therapy rather than as part of a sequencing plan. There are two key concerns. Since natalizumab sequesters pathologic immune cells rather than modifies them, there is a risk of rebound disease activity once natalizumab has been withdrawn. Rebound has been reported even after another agent, such as fingolimod, dimethyl fumarate or daclizumab, has been initiated, with sequelae that Gonzalez-Suarez and colleagues have called catastrophic (Gonzalez-Suarez et al. Brain Behav 2017;7:e00671. Patti et al. BMC Neurol 2015;15:252. Uphaus et al. J Neurol 2017; epublished October 3, 2017); Indeed, rebound has been reported even after natalizumab itself is restarted (Beume et al. J Clin Neurosci 2015;22:400-401).
The second concern is the risk of subclinical PML, which may manifest only after the next therapy has been initiated. This has occurred with fingolimod, and is likely the reason for the PML case recently reported with ocrelizumab. This suggests that the methods used to rule out PML were not (or could not be) rigorous enough. Outcomes would be expected to be even worse if PML were to develop after a switch to an agent with prolonged effects on immune function (e.g. alemtuzumab, cladribine). A natalizumab-alemtuzumab switch has been reported to be effective (Bertolotto et al. AAN 2017; abstract P2.407), but may be considered by some to be a high-risk strategy. Accordingly, some authors have recommended a bridging strategy when switching from natalizumab to alemtuzumab (Giovannoni et al. Pract Neurol 2016;16:389-393).
Rebound has also been seen after fingolimod withdrawal, which reportedly occurs during the washout period (Hatcher et al. JAMA Neurol 2016;73:790-794). A shorter washout would generally not be recommended, since it would be preferable for safety reasons to allow lymphocytes to recover before initiating another agent. A further consideration is that a cell-depleting therapy may be less effective after fingolimod if initiated when pathogenic lymphocytes remain sequestered in secondary lymphoid organs. For example, a recent observational study reported ongoing relapses in patients after switching from fingolimod to alemtuzumab (Willis et al. Neurol Neuroimmunol Neuroinflamm 2017;4:e320). The suggestion was prolonged lymphocyte sequestration enabled T cells to escape the depleting effects of alemtuzumab; in effect, the duration of sequestration was longer than the half-life of alemtuzumab. A short-term bridging strategy, similar to the one proposed for a natalizumab-alemtuzumab switch, might be appropriate in patients who experience breakthrough disease after discontinuing fingolimod.
Patients with more aggressive MS may well receive a more potent agent, such as alemtuzumab or ocrelizumab, earlier in the treatment course. After a two-year course of treatment, the question will then arise: what next? In the case of alemtuzumab, an estimated 50% of patients at six-year follow-up may require additional therapy (LaGanke et al. ECTRIMS 2016; abstract P681). However, it is unclear if additional courses of alemtuzumab would be required, whether continued dosing would be associated with an unfavorable safety profile, or if patients would be more appropriately de-escalated and maintained on a less potent therapy. Similarly, additional six-monthly courses of ocrelizumab may be administered, if required, but there are no long-term data on this approach. The concern is that safety issues may arise with chronic suppression of B cell function. More data from observational studies or database analyses are needed to identify what are the safest exit strategies for both of these high-potency agents.
Dr. Paul S. Giacomini: Treatment sequencing remains among the foremost challenges facing MS physicians and patients today. With an ever-expanding repertoire of therapeutic options, choice has become increasingly overwhelming. This complex decision has been further complicated by an emerging paradigm shift of induction therapy versus treatment escalation. Induction, using higher-efficacy agents albeit with greater potential safety considerations, has the advantage of striking early, and potentially resetting the disease trajectory. However, this approach seems to carry greater potential long-term risks compared to escalation, the older approach of incrementally switching to agents of slightly greater efficacy and risk. The lack of definitive comparative trials between therapies, or long-term sequencing safety data, has further complicated this conundrum, and even expert physicians can only thoughtfully speculate or rely on personal experience to help govern these decisions. MSBase and other large-scale “real-world” database studies have been somewhat instructive but, despite statistical tools to aid comparisons, the data generated by these studies are not definitive evidence that one drug is superior to the other, or what treatment sequencing order is optimal.
Emerging higher-efficacy therapies (cladribine, ocrelizumab, alemtuzumab) could potentially rewrite the long-term outcomes in MS – but when, in whom, and for how long should these potent therapies be used, remains a subject of intense discussion with insufficient data to guide our hand. The treatment sequencing debate seems destined to carry on, at least until we have head-to-head comparison studies, long-term safety and efficacy data with newer agents, or reliable biomarkers to help personalize therapeutic decision-making.
NeuroSens survey on sequencing – Part 7