In the presidential lecture series at CMSC 2021, Ellen M. Mowry, MD, called into question whether or not the induction treatment approach in multiple sclerosis does in fact present a higher chance of preventing long-term disability.
In the treatment of multiple sclerosis (MS), the discussion surrounding the ideal treatment approach for patients who are newly diagnosed landed most in 1 of 2 camps: induction—the process of beginning with high-efficacy disease-modifying therapy (DMT) as soon as possible—or escalation—the process of beginning on a lower-efficacy DMT and only adjusting therapy as needed.
In recent years, the data that have been published suggesting that using higher-efficacy therapy early on, despite higher risk of adverse events (AEs), better prevents long-term disability have led many toward the induction approach. But some have called this into question, including Ellen M. Mowry, MD, director, Multiple Sclerosis Experimental Therapeutics Program, and professor of neurology, Johns Hopkins Medicine. In her presidential lecture at the 2021 Consortium of Multiple Sclerosis Centers (CMSC) Annual Meeting, October 25-28, Mowry noted that many of the studies done that support this approach may have been impacted by a number of factors, namely residual confounding and selection bias. Additionally, she noted, many of these cohorts include individuals who were not diagnosed with MS under the updated 2017 McDonald criteria, which may have led to delays or misdiagnoses prior to beginning on MS DMT.1
“When I think about this question [of induction vs escalation], I really try to make it clear to myself about what I'm actually trying to do in treating a person with multiple sclerosis. I, personally, find that most people can have control of relapses and the development of new lesions—those efficacy outcomes that were targeted by clinical trials—within a relatively short period of time, even if they escalate,” Mowry said. “I don't find that there are too many individuals these days who don't achieve control of their inflammatory activity, at least in my clinic. But I'm really interested in understanding whether the treatment choices a person with MS and myself make at the time of their diagnosis matters with respect to how they're doing several years down the road. Can I change the way I practice or advise people to use these disease-modifying therapies to prevent sustained disability over time?”
Mowry raised a number of questions about the use of DMT to prevent sustained disability in patients with MS, one being that if a patient was started on a moderate-efficacy DMT, monitored, and switched to a high-efficacy DMT upon breakthrough disease, would that work as well as initiating high-efficacy treatment from the beginning? Additionally, she noted the diversity and individuality among MS patients, ultimately questioning whether a one-size-fits-all approach is best for MS care and pointing to the possible safety implications of such an approach.
She also noted that the conversation about deciding on a DMT for a given patient is affected by a number of factors, including patient-, disease-, treatment-, and system-related aspects of the disease. Mowry explained that because of this, taking care in approaching the conversation with patients is crucial (FIGURE). Additionally, she noted, the importance of patients’ inclusion in the shared decision-making process is not only apparent but impactful: In 2 ongoing trials, TREAT-MS and DELIVER-MS, between 40%-50% of patients who declined randomization ultimately chose to take an escalation approach to their disease. Much of this, Mowry noted, appears driven by the safety concerns of those who are newly diagnosed.
One of the studies Mowry brought up was work from Sorensen et al in 2020, which suggested that the overall disease course of MS in the last decade may be milder due to the complex interplay of a number of factors, such as the overall effectiveness of DMT use and new treatment and diagnostic guidance.2 Additional data from the University of California, San Francisco MS-EPIC cohort (n = 547) suggest that the risk of patients achieving an Expanded Disability Status Scale (EDSS) score of 6 or higher was only 4.7% (95% CI, 2.6-6.8) over 10 years, and that risk increased to just 16.2% (95% CI, 11.5-20.7) after 20 years.3
“These numbers are dramatically lower than what we learned about in medical school several years ago, when we would tell people their risk of becoming disabled was closer to 50%,” Mowry explained. “Now, there are a lot of reasons that more modern MS studies show a lower risk of disability over time than in years past. One is that the way we define multiple sclerosis has changed dramatically over time. Of course, several years ago, the diagnosis of a relapsing-remitting MS course required having 2 attacks of multiple sclerosis over time, so you had to have a pretty substantial amount of clinical disease activity to be diagnosed, whereas these days, with modern criteria, a person can be 24 or 48 hours into their very first MS relapse, have had an MRI and perhaps a spinal fluid analysis, and immediately be diagnosed with MS if they meet the right criteria.”
Mowry noted that many have cited the work from Brown et al in 2019, which utilized some novel approaches to attempt to minimize the risk of confounding and bias compared to older studies. Those data suggest that treatment with high-efficacy therapies—fingolimod (Gilynea; Novartis), natalizumab (Tysabri; Biogen, or alemtuzumab (Lemtrada; Sanofi Genzyme)—resulted in a roughly one-third lowered risk of developing secondary progressive MS compared with glatiramer acetate or interferon ß (HR, 0.66; 95% CI, 0.44-0.99; P =.046).4
“However, I do think it's important to point out a few things about the study when we cite it, or studies like it, because there are still some limitations that I think are important to consider when we decide whether or not this is going to change our practice dramatically,” Mowry said. She explained that the cohort utilized for the study were primarily from the MSBASE cohort, which includes nearly 45,000 individuals with MS, as well as the addition of those from different cohorts to include alemtuzumab—but criteria led to only 615 individuals being eligible for inclusion, or roughly 0.013% of the total cohort. “We can see already that there's probably a pretty significant selection bias in terms of who was included in this particular study,” Mowry said.
Additionally, she noted that those in the glatiramer acetate or interferon ß groups were only included in the study if they were seen in the period before they were escalated. Although Mowry noted that the reasons provided for taking that approach were reasonable, it creates a potential issue of difference in disease characteristics between the groups, as those in the glatiramer acetate or interferon ß group may have been diagnosed under older criteria and may have had more severe disease from the start.
“There's also a concern in this and other analyses that there's residual confounding,” Mowry said. “The thing that we worry most about is what's called confounding by indication—that the reason that people were assigned to these medications is not random, but rather is related to the prognosis of their illness. While some of these analyses tried to adjust for the propensity of treatment, there's a strong likelihood that the studies aren't fully able to account for this type of confounding.”
These issues of residual confounding and selection bias, Mowry noted, are inherent in many of the studies that have been used to justify the induction approach. After sharing a few other examples, she noted the difficulties that are inherent in extrapolating clinical trial data to the real-world clinical setting. These, she explained, are driven by a few factors that are not generalizable to the larger MS population. The first is that trial populations tend to mostly include those patients who are white, of Eastern European descent, who have met specific MS disease activity criteria and who have minimal comorbid conditions. As well, trial conditions are not often generalizable, as they offer minimal time to treatment parameters and offer extra personnel for safety monitoring that is not available in standard practice. Mowry added that in clinical trials, the shorter length of treatment may hide a likely greater risk of AEs in real-world practice.
“Finally, and perhaps most importantly, to me, I'd have difficulty extrapolating evidence from clinical trials because I don't think that they're really testing treatment strategies,” she said. “If I enroll somebody in a trial, whether it's comparing a DMT versus placebo or a stronger-efficacy DMT versus a moderate-efficacy DMT, the small segment of people who are doing well are not then offered escalation in those trials. That's not really mirroring the decisions we need to make in the real world. It's not really a treatment strategy study.”
She concluded by urging physicians to get involved in a trial aimed at improving the available data on induction vs escalation in the real world: the aforementioned TREAT-MS and DELIVER-MS, the latter of which is a pragmatic trial aiming to assess 900 patients with MS who were diagnosed under the 2017 McDonald criteria over the course of 60 months, with a primary outcome of disability progression measured by EDSS plus (EDSS along with 9-Hole Peg Test and Timed 25-Foot Walk Scores). The former, DELIVER-MS, is a parallel trial being conducted in both the US and the UK.
For more coverage of CMSC 2021, click here.