Multiple sclerosis (MS) is an inflammatory neurological disorder caused by demyelination of neurons and characterized by lesions in the central nervous system (CNS).1 After the onset of MS, as early treatment initiation with disease-modifying therapies (DMTs) has been shown to reduce long-term disability.2 However, with the development of numerous effective therapeutic agents, the choice of which treatment to select poses challenges for both clinicians and patients. Currently, no biomarkers exist to identify the best treatment for a particular patient, therefore treatment selection relies on disease characteristics, as well as clinician and patient preferences.1 Moreover, head-to-head trials are limited in the MS treatment spectrum.2
There are currently 2 widely accepted treatment strategies for MS: escalation and induction therapy. Escalation therapy begins with agents that are associated with less pronounced risk, but also may be less effective. It then escalates to potentially more effective therapies that are associated with increased safety markers. In contrast, induction therapy begins with more effective agents at the time of diagnosis. Given the challenges of measuring long-term impact, it determining the more effective strategy is still a topic of debate.3,4
Although DMTs are often categorized by mode of administration (eg, oral, intravenous, injection), they can also be divided by their general mode of action. The 3 categories of currently available DMTs are immunosuppressive, immunomodulating, and anti-trafficking. Immunosuppressive agents are considered more aggressive therapies, capable of removing disease-causing cells, and are used at the start of induction therapeutic strategies. Immunomodulating and anti-trafficking agents are considered less aggressive; they alter immune cell function or prevent their entrance into the brain, respectively.2 Understanding these therapeutic methods and drug classifications is important for treatment selection, but this is made more difficult by the lack of head-to-head comparative clinical trials, which require large sample sizes and financial resources. Additionally, there is a lack of sensitive, reliable outcome measures for these trials; the currently utilized measures, attack frequency, and increasing Expanded Disability Status Scale (EDSS) score both lack sensitivity.4 The EDSS score is currently the only scale allowed by the FDA for measuring disability progression; however, this method has been criticized, as it does not assess cognition or memory.5
Given the extensive treatment landscape and noted limitations regarding direct treatment comparisons in determining an optimal regimen, clinicians and payers may review the efficacy and/or value of individual agents by indirectly comparing their respective trial data and findings. Indirect treatment comparisons (ITCs) offer additional perspective on the therapeutic potential of MS therapies, but clinical heterogeneity regarding trial design and findings raise questions about the utilization of such comparisons when assessing the potential of individual agents, notably when agents are compared within individual drug classes based on mode of action.6 One particular mode of action in the MS treatment landscape that has evolved rapidly in recent years that is worthy of consideration as it relates to ITCs is the sphingosine-1-phosphate (S1P) inhibitor class.
S1P inhibitors bind to S1P receptors (S1PRs) expressed on the surface of lymphocytes to prevent T cells from exiting lymphoid organs.7,8 The precise mechanism of how S1P modulators exert a therapeutic effect in MS patients is not completely understood, despite improvements in comprehending how S1P modulators impact lymphocyte egression.8
Currently, 3 S1P inhibitors are approved by the FDA. When reviewing clinical trials for approved S1P inhibitors for the treatment of MS, it is important to recognize the differences in study design and population. Therefore, an effective comparison of these agents may not be based on efficacy comparisons, but rather on understanding the results of each study based on its individual goals. To demonstrate both the challenges and potential value of ITCs, ahead is a brief review of individual drug profiles, as well as key elements of study design and findings from each S1P inhibitor’s respective clinical trials.
Sphingosine-1-Phosphate Inhibitors: Class Review
Fingolimod. The first of the S1P-modulating agents, fingolimod, inspired the breadth of research surrounding S1P-mediated signaling. Fingolimod was the first oral drug approved by the FDA for relapsing-remitting MS (RRMS) therapy. Fingolimod is phosphorylated in vivo by sphingosine kinase to the active metabolite, fingolimod phosphate. Early studies demonstrated that fingolimod localizes to the CNS white matter with preferential dissemination along the myelin sheath. The active metabolite binds with high affinity to 4 of the 5 available S1PRs: S1PR1, S1PR3, S1PR4, and S1PR5. Once bound, it induces internalization of the S1PR1 receptor, preventing the lymphocyte from egressing the lymphoid tissue, decreasing inflammatory cell infiltration into the CNS.9-11
To assess the long-term efficacy of fingolimod as compared with placebo, the FREEDOMS study (NCT00289978) recruited patients with RRMS for a 24-month study to assess annualized relapse rate (ARR), disability progression, MRI measures of inflammation, disease burden, and tissue destruction.12 At the trial’s conclusion, all clinical and MRI-related efficacy end points favored fingolimod over placebo, regardless of fingolimod dosage.12 The ARR for fingolimod patients was 0.16 and 0.18 for the 1.25 mg and 0.5 mg doses, respectively, compared with 0.4 in the placebo group, showing a respective 60% and 54% relative reductions. Decreased risk of disability progression and decreased volume of lesions were also observed.12 The cumulative probability of disability progression after 3 months was 17.7% and 16.6% for 0.5 mg and 1.25 mg fingolimod doses, respectively, compared with 24.1% probability with placebo. Additionally, the mean number of gadolinium (Gd)-enhancing lesions at 24 months was 0.2 in both fingolimod dosage groups (SD, 1.1 and 0.8 in the 1.25 mg and 0.5 mg groups, respectively), while placebo-controlled patients had an average of 1.1 lesions.12 Although patients taking fingolimod showed less risk of relapse and disability progression, researchers were concerned by several adverse effects (AEs) of the agent, including infections, cardiovascular events, macular edema, and elevated liver-enzyme levels, demonstrating the necessity for further longer-term assessment.12
In one of the few head-to-head comparison trials available, the TRANSFORMS trial (NCT00340834) compared oral fingolimod with intramuscular (IM) interferon beta-1a (IFNβ-1a), which is a well-established treatment as one of the first drugs approved for RRMS.1,13 The primary endpoint of the study was ARR, and the number of 12-month T2-weighted MRI scans and time to confirmed disability progression were also assessed as secondary endpoints.13 Fingolimod’s efficacy was superior compared with that of IM IFNβ-1a, as demonstrated by ARRs of 0.16 to 0.20 with fingolimod compared with 0.33 with IFNβ-1a, corresponding to a 38% to 52% relative reduction. Although there was no observed difference in time to confirmed progression of disability, compared with IFNβ-1a, fingolimod more effectively reduced lesion activity and brain volume loss as shown on MRI.13 Despite higher efficacy, fingolimod may be associated with a higher risk for infection, and some AEs may be dose-related. This prompted further evaluation of dose-related alterations with fingolimod in 2-year, placebo-controlled phase 3 trials.13 The results were published the same year as the aforementioned FREEDOMS trial, in which the effectiveness of fingolimod was evaluated over 2 years. Results suggested that although infections were prevalent (69%) in the fingolimod groups, they were similar compared with placebo (72%).12 Consistent with previous clinical experiences, transient, dose-related decreases in heart rate occurred after the first administration of fingolimod.12
Siponimod. Siponimod is a second-generation S1P-modulating agent that was approved in early 2019.3 It is indicated for the treatment of relapsing forms of MS to include clinically isolated syndrome, RRMS, and active secondary progressive MS (SPMS). Siponimod induces lymphopenia by halting lymphocyte egression from lymph nodes. It is selective toward S1PR1 and S1PR5 only, and it has a much shorter half-life than fingolimod, as lymphocyte levels return to baseline within a week of treatment discontinuation.11,14 The majority of the effects of siponimod are attributed to its effect on S1PR1 and subsequent prevention of lymphocyte egression; however, S1PR1 and S1PR5 are commonly expressed in cells present in the brain, including neurons, microglia, and oligodendrocytes.14 In the event that AEs occur, its short half-life allows for flexibility in MS treatment with a shorter washout.15
As mentioned in the first article of this publication (see page 9), more than half of patients with RRMS eventually transition to SPMS. However, most therapies used for the treatment of SPMS have been investigated specifically in patients with RRMS, measuring relapse rates as primary end points.16 As the first of a new generation of S1P inhibitors, siponimod was evaluated in the EXPAND trial (NCT01665144) for its safety and efficacy in patients with active SPMS using the primary endpoint of confirmed disability progression (CDP) at 3 months. In the EXPAND trial, CDP was defined as a 1-point or 0.5-point increase in the EDSS score, depending on the patient’s starting EDSS score.16 Patients in the siponimod group had a CDP of 26%, compared with 32% of patients in the placebo group (HR, 0.79; 95% CI, 0.65 to 0.95), for a relative risk reduction of 21% (P = .013). Findings also indicated that the relative risk reduction of 6-month CDP was 26%. Additionally, the percentage of patients in the siponimod group who were free from decreased Gd-enhancing lesions was higher than in the placebo group (89% versus 67%, respectively), and a higher percentage of patients in the siponimod group were free from new or enlarging T2 lesions compared with placebo (57% versus 37%, respectively).16 Siponimod had a similar safety profile to other S1P modulators.16
Ozanimod. The most recent S1P-modulating agent available to MS patients is ozanimod, which was approved in early 2020.It selectively binds S1PR1 and S1PR5 with high affinity, preventing lymphocyte egression and reducing peripheral lymphocytes.17 Ozanimod differs from fingolimod in that it does not require phosphorylation for activation. It also has a half-life of 19 hours, allowing for once-daily dosing and rapid lymphocyte recovery after treatment discontinuation.11 Additionally, no first-dose observation is required before treatment initiation with ozanimod.18-20
Ozanimod was compared with IFNβ-1a for RMS in a 24-month head-to-head trial called RADIANCE (NCT02047734). Another study, SUNBEAM (NCT02294058), was being completed concurrently comparing the same therapies; it concluded at 12 months.21,22 The primary end point used in both studies was ARR. Other factors that were assessed included the number of new or enlarging T2 lesions, number of Gd-enhancing lesions, and time to onset of disability progression using EDSS.21,22 The RADIANCE trial results indicated that ozanimod was associated with a lower ARR thanIFNβ-1a at 0.17 and 0.22 for 1.0 mg and 0.5 mg doses of ozanimod, respectively, compared with 0.28 in IFNβ-1a patients.21 There was no observed difference in the time to confirmed disability progression, but there were fewer T2 and Gd-enhancing lesions and decreased brain volume loss with ozanimod.21 The SUNBEAM trial had similar results, with ozanimod ARRs at 0.18 and 0.24 for 1.0 mg and 0.5 mg doses, respectively, compared with 0.35 for IFNβ-1a. Fewer lesions and smaller loss of brain volume were also observed in the ozanimod 1 mg group.22 Additionally, in both trials, fewer instances of AEs leading to discontinuation occurred in patients receiving ozanimod compared with IFNβ-1a, leading both studies to conclude that ozanimod is an effective oral therapy for patients with RRMS.21,22
Ponesimod. Currently in development and not yet FDA approved, ponesimod is another S1P-modulating agent selective toward only S1PR1. Initial studies have shown promising results for decreasing active lesions; also, one trial found that only 2% of patients experienced bradycardia, a common AE of other S1P-modulators that necessitates first-dose monitoring.11,15 The OPTIMUM trial (NCT02425644) compared the efficacy of oral ponesimod and oral teriflunomide, with results showing that ponesimod was associated with a reduction in ARR of 30.5% up to week 108.23
Indirect Treatment Comparisons: Challenges and Opportunities
Based on the clinical trial data and the differences in each trial’s design, patient populations, and assessment of end point results, it is difficult to evaluate which S1P-modulating agent is the most efficacious (Table).12,13,16,21,22,24 To that end, the differences between the clinical trials are worth noting. For instance, in EXPAND (NCT01665144), patients with SPMS were recruited to assess the efficacy of siponimod on slowing disability progression. Additionally, it was the only study to utilize 3-month CDP as the primary end point, while all other trials utilized ARR. It is also significant that the patient population in EXPAND was significantly older, with greater disability progression and fewer relapses relative to patients in other studies.13,16, 21,22,24 Meanwhile, SUNBEAM and RADIANCE included patients with broadly relapsing disease and used ARR as the primary end point in evaluating ozanimod. Thus, while siponimod and ozanimod were found to be effective, the context for each is quite different, which has implications for the utilization of these agents.
Broadening the discussion to the utility of ITCs in the wider field of MS treatment, the feasibility of specific ITCs must be assessed, given the heterogeneity across trials even within the same drug class. It is also worth noting that comparing ITCs is particularly difficult when comparing newer trials with older ones, as treatment history, baseline characteristics, and eligibility criteria may differ as time progresses.6 For example, the McDonald diagnostic criteria for MS were first introduced in 2001 and have been subsequently updated and expanded in 2005, 2010, and 2017. This could introduce heterogeneity among trials completed years apart in which different versions of the McDonald criteria were utilized.25 Furthermore, because the treatment landscape of MS has changed considerably, caution should be exercised when comparing the results among modern-day trials with those of older trials regarding to treatment history, as there were far fewer medications patients could have received in the 1990s and 2000s compared with today.6
Evaluating the heterogeneity among trials highlights the shortcomings of current clinical trial design. For example, EDSS is commonly utilized, yet there are clear differences in how it is used across studies to outline progression. Similar to what has been discussed here with regard to some studies evaluating ARR compared with 3-month CDP, Samjoo, et al, have suggested the adoption of a consistent definition of MS end points for a regular interpretation and analysis of clinical trials in MS to better facilitate cross-trial comparisons.6 This recent publication evaluated the heterogeneity among several different studies to compare the effectiveness of siponimod with that of IFNβ and ocrelizumab. Although they were able to complete an ITC for siponimod and IFNβ, this same comparison analysis was not plausible for siponimod and ocrelizumab due to substantial variation in study design, patient characteristics, outcome definitions, and predefined treatment effect modifiers. The ITC completed for siponimod and IFNβ was in favor of siponimod in the treatment of SPMS.6
Other ITCs have been published to indirectly compare MS therapeutics, including 2 different oral therapies, fingolimod and dimethyl fumarate. At the conclusion of one analysis, no statistically significant difference was found between the therapies as they showed similar ARR and 12-week CDP, although the authors noted that the results should be translated with caution, given that there could be unobserved or unknown differences between the trials.26 An ITC was also completed to compare 2 non–S1P-modulating therapies, ocrelizumab and IFNβ-1a. This comparison found ocrelizumab to be more effective, as it resulted in a significantly lower ARR, lower rate of disability progression, and a higher rate of disability improvement.27
As the treatment spectrum for MS continues to expand, ITCs will likely continue to play an important role in assessing the utility of individual agents. Despite their notable limitations, ITCs may help clinicians and payers better understand treatment pathways and disease pathophysiology. For instance, within the S1P drug class, although it is difficult to accurately compare the efficacy of S1P modulators against each other and other MS agents in broader patient populations, the variety in S1P modulators (reflected in their respective trial designs and results) highlights the varying specificity across different S1P receptors and demonstrates the potential of this therapy class and individual agents to address the unique needs of patients with MS. Indirect treatment comparisons therefore illustrate the importance of evaluating differences among individual agents and trials, as well as individualizing treatment selection based on specific drug and patient factors.