Сравнительная эффективность Diroximel Fumarate vs. Диметил Фумарат у лиц с множественным склерозом: анализ претензий США в отношении релапсовых результатов.

Key Summary Points

Introduction

Multiple sclerosis (MS) is a chronic, acquired, inflammatory disease of the central nervous system in which demyelination and axonal loss lead to progressive neurological impairment [ 1 ]. Relapsing multiple sclerosis (RMS) is a common subtype of MS characterized by recurrent clinical relapses separated by periods of remission, with gradual accumulation of disability over time. For patients with RMS, the primary goal of treatment is to reduce the number and severity of relapses as well as delay disability progression [ 1 , 2 ]. In controlled clinical trial settings, relapse efficacy is typically quantified through annualized relapse rate (ARR) and percentage of patients with MS relapse. The availability of over 20 disease-modifying therapies (DMTs) for MS in the USA, including infusions, injectables, and oral medications, underscores the importance of optimizing treatment [ 3 ]. Among these, fumarates, including dimethyl fumarate (DMF) and diroximel fumarate (DRF), have demonstrated robust efficacy in reducing relapse burden in pivotal clinical trials and long-term extension studies [ 4 – 7 ]. As of 31 July 2025, at least 645,172 persons with MS (pwMS) globally have been treated with DMF, and at least 54,755 pwMS have been treated with DRF [ 8 – 11 ].

When administered orally, DRF and DMF undergo rapid esterase hydrolysis to convert to monomethyl fumarate (MMF), which is believed to ameliorate MS pathophysiology by anti-inflammatory modulation and Nrf-2 activation while preserving the humoral immune response [ 12 – 15 ]. DRF’s distinct larger chemical structure is hypothesized to minimize gastrointestinal (GI) irritation via lower production of methanol and reduced reactivity with off-target proteins and receptors [ 16 – 18 ]. Consequently, DRF improves upon the incidence of GI adverse events (e.g., nausea, diarrhea, abdominal pain, and flushing) and thus the tolerability profile of DMF, which may impact treatment adherence in real-world settings, particularly during therapy initiation [ 7 , 19 – 23 ]. Given the critical importance of treatment adherence in chronic conditions such as MS, treatment interruptions or suboptimal adherence are strongly correlated with an increased risk of relapse and progression and undermine the therapeutic potential of the drug [ 24 , 25 ].

While clinical trials have confirmed the bioequivalence and improved tolerability profile of DRF [ 4 , 16 , 17 ], real-world evidence may further help to clarify if these advantages translate into tangible clinical benefits. In an analysis of pwMS from a specialty pharmacy dataset, DRF resulted in a high rate of persistence in adherence, in a population which included a large proportion of patients who had switched from prior DMF treatment [ 8 ]. In addition, pwMS who switched from DMF to DRF because of persistent GI adverse effects showed higher levels of adherence after switching [ 8 ]. However, it remains unclear whether DRF’s tolerability-associated improvements deliver advantageous relapse outcomes vs. DMF, as prior analyses have focused primarily on pharmacy claim-driven adherence metrics without a further view on relapse-linked medical claim outcomes.

This study aims to evaluate the real-world comparative effectiveness of DRF versus DMF in a large US RMS population sourced from the Forian claims database. Given DRF’s improved GI tolerability profile and potential for related improved treatment adherence, we hypothesize that there may be differences in real-world effectiveness of DRF vs. DMF [ 8 , 16 , 17 ].

Methods

Data Source

This retrospective cohort study utilized the Forian administrative claims database to compare the effectiveness of DRF and DMF in pwMS. The Forian database comprises a diverse, commercially available dataset of open and closed medical and pharmacy claims, capturing a broad representation of payer types and geographical regions. Forian’s medical claims comprise 59% commercial, 40% Medicare/Medicaid, and 2% VA or other payers, while its pharmacy claims comprise 50% commercial, 49% Medicare/Medicaid, and 1% VA or other payers. The payer distribution of the Forian claims database closely mirrors that of the overall US population (i.e., 63% commercial, 35% Medicare/Medicaid, and 1% VA or other payers), indicating strong representativeness of the larger US population [ 26 ]. The study period spanned from January 1, 2018 to June 30, 2025.

PwMS Selection

PwMS were eligible for inclusion if they were at least 18 years of age and had at least one pharmacy claim for DRF or DMF between January 1, 2019 and June 30, 2024. This selection window ensured that all included pwMS had a minimum of 12 months of baseline period and 12 or 24 months of follow-up (i.e., for inclusion in the 1-year or 2-year follow-up cohorts, respectively) prior to the study end date cutoff in June 2025.

To ensure a clean comparator design, pwMS were required to have at least two claims for their respective index drug (DRF or DMF) and no claims for any other MS DMT (Table S1 ) at any point during the entire follow-up period. PwMS were required to show at least one medical and one pharmacy claim at any point prior to the baseline period, during the 12-month baseline period, and during each year of the follow-up periods.

The “index date” was defined as the date of the first pharmacy claim for either DRF or DMF. To ensure adequate data availability for characterization and outcome assessment, the study timeline was divided into a baseline period (12 months prior to the index date) and a follow-up period (12 or 24 months post-index).

Persistence criteria were applied to isolate the effect of the study drugs. PwMS were required to remain on their index treatment for the entire duration of the analysis (1 or 2 years) without discontinuation, defined as a gap in therapy of ≥ 90 days (calculated as the next start date minus the sum of the previous start date and days’ supply). PwMS who switched to or received any other DMT during the 1- or 2-year follow-up periods were excluded. Female pwMS with any evidence of pregnancy in the year prior to the index date were also excluded to rule out pregnancy-related treatment interruptions or confounders.

PwMS in the 1-year follow-up population received exclusively DRF or DMF treatment for MS for a period of at least 12 months post-index date, while pwMS in the 2-year follow-up population received exclusively DRF or DMF treatment for MS for a period of at least 24 months post-index date.

Study Endpoints

The primary assessed outcome was ARR. As there is no specific ICD-10 diagnosis code for an MS relapse, a validated algorithm was employed to identify relapse events based on healthcare utilization patterns. A relapse was defined as the occurrence of either (1) a MS-related hospitalization or (2) an outpatient MS relapse event.

MS-related hospitalization was defined as an acute inpatient admission with a primary diagnosis of MS (i.e., ICD-10 code G35). Outpatient MS relapse event was defined as an outpatient visit with MS diagnosis (i.e., ICD-10 code G35) accompanied by a claim for relapse-associated treatment within 7 days of the visit [ 27 ]. Qualifying relapse-associated treatments included intravenous (IV) corticosteroids, adrenocorticotropic hormone (ACTH), total plasma exchange, or high-dose oral corticosteroids (≥ 500 mg/day), using pre-defined coding rules (Table S3 ).

To avoid double counting a single relapse episode, claims meeting the definition within 30 days of an initial event were considered part of the same relapse event.

Secondary outcomes included (1) time to first relapse and (2) treatment refill adherence, as measured by the proportion of days covered (PDC) based on days’ supply data from pharmacy fills.

Statistical Analysis

To mitigate selection bias inherent in observational claims analysis, propensity score matching was utilized to balance the DRF and DMF cohorts. PwMS were matched based on their probability of receiving DRF conditional on characteristics assessed at baseline.

Baseline covariates incorporated in the propensity score assessed during the 12-month pre-index period included age, sex, and clinical history. To account for disease complexity, two composite scores were calculated for each pwMS: (1) Charlson Comorbidity Index (CCI)—a weighted score derived from the presence of comorbid conditions [ 28 ], and (2) MS Severity Score—a claims-based severity proxy calculated by summing weighted indicators of disease burden (Table S3 ) [ 29 ].

Each DRF-treated patient was matched to three patients treated with DMF with similar baseline characteristics, forming matched clusters centered on patients treated with DRF. Matching quality was assessed using absolute standardized mean differences (|SMD|), with |SMD|< 0.10 indicating negligible imbalance between groups.

Treatment effectiveness was compared between the matched cohorts using two statistical methods: (1) independent z test or t test, and (2) the Average Treatment Effect on the Treated (ATT) framework.

Differences in ARR of DRF vs. DMF cohorts were assessed using independent t tests and ATT analysis across 1-year and 2-year follow-up populations. ARR and p value were reported via independent t test analysis, while DRF/DMF Poisson rate ratio and 95% CI were reported via ATT analysis.

Time to first relapse was analyzed using Kaplan–Meier methods for DMF and DRF cohorts across 1-year and 2-year follow-up populations. Kaplan–Meier curves were compared between groups using the log-rank test.

All statistical tests were two-sided, with a p value < 0.05 considered statistically significant. Data transformation and analyses were conducted using SQL and Python.

Ethical Considerations

A license was obtained from Forian to access and use the Forian claims database. All patient information was anonymized, and patient confidentiality was maintained through compliance with Health Insurance Portability and Accountability Act regulations. This analysis is based on previously collected data and does not involve any new studies of human or animal subjects.

Results

Baseline Cohort Characteristics

Within the 1-year follow-up population, a sample of 7120 pwMS was identified, consisting of 1780 DRF-treated and 5340 DMF-treated pwMS. Demographic and clinical variables were well balanced via propensity score matching, with |SMD| < 0.05 for all covariates. The mean age for DRF and DMF cohorts was 49.50 years (SD 11.80 years) and 49.90 years (SD 11.66 years), respectively. Both cohorts consisted of 19.2% male pwMS. Mean MS Severity Score was 2.51 for DRF and 2.48 for DMF. CCI Score was 0.66 for DRF and 0.70 for DMF. The baseline proportion of pwMS with a relapse was identical at 15.1% for both cohorts (Table 1 ).

The 2-year analysis was conducted on 2956 pwMS, with 739 pwMS in the DRF cohort and 2217 pwMS in the DMF cohort. Similar to the 1-year follow-up population, baseline characteristics were well balanced between groups via propensity score matching, with |SMD| < 0.05 for all covariates (Table 1 ).

Of note, the difference in patient numbers between the 1-year and 2-year analysis is due to several different reasons. It is possible for patients from the 1-year analysis to subsequently change insurance or have insufficient claims activity after the 1-year timepoint, and some remained persistent on the index DMT at the end of the data window (June 30, 2025) but had not yet reached 2 years of follow-up, precluding their inclusion in the 2-year analysis (Table S4 ).

The DMF cohort was further assessed to understand the distribution of pwMS receiving generic vs. branded (i.e., Tecfidera) medication. The majority of pwMS within the DMF cohort received generic DMF (i.e., either receiving generic DMF exclusively or had most recently switched to generic DMF through the follow-up period) across both the 1-year (55%) and 2-year (69%) follow-up populations (Table 2 ).

Annualized Relapse Rate

Within the 1-year follow-up population, ARR was significantly lower in the DRF group (0.12) compared to the DMF group (0.16), representing a difference of 0.04 in ARR ( p = 0.03; Fig. 1 ). The DRF/DMF rate ratio was 0.79 ( p = 0.03; 95% CI 0.64–0.97; Fig. 1 ), indicating that pwMS on DRF experienced a 21% reduction in ARR compared to pwMS treated with DMF only.

Within the 2-year follow-up population, the DRF cohort maintained a significant reduction in ARR compared to the DMF cohort. ARR was 0.16 for DMF compared to 0.11 for DRF, representing a difference of 0.05 in ARR ( p = 0.02; Fig. 2 ). The DRF/DMF rate ratio was 0.69 (95% CI 0.54–0.87; Fig. 2 ), indicating a 31% reduction in ARR for DRF-treated pwMS over the 2-year period.

Proportion of PwMS Remaining Relapse-Free

Kaplan–Meier analysis demonstrated that a slightly higher proportion of pwMS treated with DRF remained relapse-free compared to those treated with DMF.

In the 1-year follow-up population (Fig. 3 ), Kaplan–Meier curves for DRF and DMF treatment groups were statistically divergent as assessed by the log-rank test over the 12-month follow-up period ( p = 0.03; Fig. 3 ), potentially indicating an early and sustained benefit in delaying or preventing relapse onset.

In the 2-year follow-up population (Fig. 4 ), Kaplan–Meier curves for DRF and DMF groups achieved consistent numerical separation post-index. However, curves did not statistically differ within the 2-year population over the observation period when assessed with log-rank analysis ( p = 0.12; Fig. 4 ).

Treatment Refill Adherence

Treatment refill adherence was high across both DRF and DMF cohorts. In the 1-year analysis, mean PDC was 82.8% for DRF and 83.1% for DMF. Similarly, in the 2-year analysis, mean PDC remained stable (82.3% for DRF vs. 82.4% for DMF) with no statistically significant difference observed. These findings indicate that within the specific study population, enriched for pwMS persisting on long-term treatment by nature of the inclusion criteria, pharmacy refill data showed equivalent coverage (Table 3 ).

Discussion

This retrospective real-world claims analysis evaluates the comparative effectiveness of diroximel fumarate (DRF) versus dimethyl fumarate (DMF) in a large, representative US pwMS population. PwMS treated with DRF experienced a significant reduction in relapse burden compared to those treated with DMF in a real-world setting. This was evident across multiple endpoints, including a significant reduction in ARR at both 1-year and 2-year follow-ups, and a significant improvement in time to first relapse at 1 year of treatment. However, additional research is warranted to corroborate these findings and further elucidate factors driving these differences in real-world effectiveness.

1-Year vs. 2-Year Follow-up Populations

The divergence in outcomes between the 1-year and 2-year analyses offers important insights into the real-world utilization of these therapies. In the 1-year analysis, DRF demonstrated reduction in ARR vs. DMF (21% lower rate, corresponding to a rate ratio of 0.79; p = 0.03). The Kaplan–Meier analysis corroborates this early benefit, demonstrating significant separation of the Kaplan–Meier time to first relapse curve.

In the 2-year follow-up population, the reduction in ARR was also robust and statistically significant (31% lower rate in DRF vs. DMF, corresponding to a rate ratio of 0.69; p = 0.01). However, it is important to note that the Kaplan–Meier proportion of pwMS remaining relapse-free for this 2-year population did not achieve significant divergence between the DRF and DMF cohorts ( p = 0.12). This difference may be influenced by the survivorship effect inherent in the study design. To be included in the 2-year analysis, pwMS were required to persist on therapy for the full 24 months. DMF pwMS who successfully remain on therapy for 2 years are likely a self-selected group who may tolerate the drug well, as those with severe GI issues may have discontinued earlier and been excluded from the 2-year cohort. Consequently, the 2-year DMF cohort represents the “best-case” users. Despite this selection pressure, DRF still significantly reduced the total frequency of relapses, even among pwMS persisting on longer-term treatment.

Refill Adherence and Pharmacy Data

Adherence, as measured by proportion of days covered, showed no significant difference between the cohorts (both approximately 82%; Table 3 ). This is not unexpected given the strict inclusion criteria requiring continuous enrollment and persistence. However, the disconnect between the equivalent PDC despite a difference in relapse outcomes highlights a known limitation of claims-based adherence metrics, in that pharmacy fills do not always equate to on-time doses of medication actually taken. A patient may refill their prescription diligently (i.e., high PDC) but skip individual doses to manage daily GI symptoms (i.e., low actual adherence).

This discrepancy may be mechanistically explained by the specific composition of the DMF cohort, where the majority of pwMS within the DMF cohort received generic DMF (Table 2 ). Generic maintenance medications are primary candidates for pharmacy automatic refill programs, which may increase PDC measures by automating medication possession independent of actual patient consumption. Conversely, branded DMF (i.e., Tecfidera) and DRF (i.e., Vumerity) may not be subject to the same degree of refill automation given distribution via specialty pharmacy. In this context, the high prevalence of generic utilization in the DMF cohorts may have inflated fill-based adherence metrics via automated dispensing cycles and masked potential interruptions in dosing caused by tolerability differences. Thus, the finding that pwMS treated with DRF achieved improved relapse outcomes despite equivalent fill-based adherence suggests that true adherence (i.e., taking the medication as prescribed every day) may have been higher in the DRF group, consistent with its established tolerability profile.

Methodological Considerations and Limitations

The results of this study should be considered alongside several limitations inherent to retrospective claims analyses, consistent with those noted in comparable real‑world studies of MS therapies.

Because there is no specific ICD-10 code for MS relapse, the identification of MS relapses relied on claims-based business rules rather than clinical adjudication. We utilized a validated algorithm requiring an MS diagnosis accompanied by hospitalization or defined outpatient relapse criteria. This approach, while necessary to define MS relapse, may misclassify relapse events; for instance, a pwMS treated with oral steroids for a non-MS condition but coded with an MS diagnosis may present a false positive. However, such misclassification was likely controlled similarly between the matched DRF and DMF cohorts and minimizes the likelihood of generating a false treatment effect.

The Forian data sample used in this study comprises both open and closed claims and may introduce potential data fragmentation. In an open claims environment, data is captured from clearinghouses rather than a single payer. Therefore, if a pwMS seeks care out-of-network or through a provider not captured by the clearinghouse, medical or pharmacy activity may be missed. While this could result in an underestimation of ARR, the claims capture rates and the “open” vs. “closed” claims distribution across the DRF and DMF cohorts were consistent, which ensures that the relative comparison between the cohorts remains valid; in other words, if there is any missing data or claims activity in this study sample, under-capture would be expected to apply similarly to both the DRF and DMF cohorts.

The DMF cohort included a mix of branded Tecfidera and various generic equivalents, as the study period (2018–2025) covers the entry of generic DMF in the USA. This heterogeneity in drug manufacturers is not present in the single-source DRF arm, as only branded Vumerity is commercially available.

Additionally, while propensity score matching balanced the cohorts on observable characteristics (e.g., age, baseline relapse risk, CCI), it may not account for unmeasured or unobservable confounders such as lifestyle factors (e.g., smoking, BMI) or specific MS phenotypes, which may influence outcomes.

Conclusion

In this retrospective analysis of pwMS in the USA, treatment with DRF was associated with a reduction in relapses compared to DMF. These findings suggest that the improved tolerability profile of DRF versus DMF may contribute to improved real-world effectiveness for pwMS.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 284 KB)

Acknowledgments

Medical Writing, Editorial, and Other Assistance

Nicholas Hong wrote the first draft of the manuscript with input from all authors and styled the manuscript per journal requirements; Anne Littlewood copyedited the manuscript—both from ClearView Healthcare Partners (Newton, MA, USA). Medical writing and editing were funded by Biogen (Cambridge, MA, USA).

Author Contributions

All named authors meet the International Committee of Medical Journal Editors criteria for authorship for this article, take responsibility for the integrity of the work as a whole. The authors had full editorial control of the manuscript and provided their final approval of all content. All authors contributed to writing (reviewing and editing). Daniel Bandari, Devon Conway, and Mirla Avila contributed to the analysis and interpretation of data. Mike Roy, Liang Dong, Minxing Li, and Chencan Zou contributed to the methodology, data curation, statistical analysis, and data interpretation. Anne Littlewood and Nicholas Hong contributed to data analysis, historical literature review, and manuscript development. Sam Ulin led the initial experimental design and conceptualization, was responsible for project initiation, and contributed to the analysis and interpretation of data. James B. Lewin, Jason P. Mendoza, Nicholas Belviso, and Babak Amerian-Williams contributed to the conceptualization, study design, analysis, and interpretation of data. All authors reviewed and approved the final manuscript for submission.

Funding

This study, the Rapid Service Fee, and the Open Access Fee were funded by Biogen.

Data Availability

Forian’s claims database is a proprietary resource accessible via licensing. ClearView holds an enterprise license, and Biogen funded a one-time use for this analysis.

Declarations

Conflict of Interest

James B. Lewin, Jason P. Mendoza, Babak Amerian-Williams, and Nicholas Belviso are employees of Biogen and may hold stock in the company. Daniel Bandari: consultant for Biogen, EMD Serono; speaker bureaus for Biogen and EMD Serono; research support from EMD Serono (ended) and Genentech. Mirla Avila: consulting fees from Biogen, Genentech, Novartis, and TG Therapeutics (all ended); research support from Biogen (ongoing) and Novartis (ended). Devon Conway: consultant for Alexion, Amgen, Biogen, Bristol Myers Squibb, and Genentech; speaker bureau for Biogen; research support from Biogen and Novartis. Liang Dong, Nicholas Hong, Minxing Li, Anne Littlewood, Mike Roy, Sam Ulin, and Chencan Zou are employees of ClearView Healthcare Partners. ClearView Healthcare Partners was hired by and received compensation from Biogen for the data analysis and writing support associated with this manuscript. They were not compensated for authorship.

Ethical Approval

A license was obtained from Forian to access and use the Forian claims database. All patient information was anonymized, and patient confidentiality was maintained through compliance with Health Insurance Portability and Accountability Act regulations. This analysis is based on previously collected data and does not involve any new studies of human or animal subjects.

Footnotes

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Doshi A, Chataway J. Multiple sclerosis, a treatable disease. Clin Med. 2016;16(6):s53–9.

2. Boster A, Nicholas J, Wu N, et al. Comparative effectiveness research of disease-modifying therapies for the management of multiple sclerosis: analysis of a large health insurance claims database. Neurol Ther. 2017;6(1):91–102.

3. National Multiple Sclerosis Society. Disease-modifying therapies. https://www.nationalmssociety.org . Accessed 23 Jan 2026.

4. Wray S, Then Bergh F, Wundes A, et al. Efficacy and safety outcomes with diroximel fumarate after switching from prior therapies or continuing on DRF: results from the phase 3 EVOLVE-MS-1 study. Adv Ther. 2022;39(4):1810–31.

5. Pandey KS, Giles K, Balashov K, et al. Long-term safety and effectiveness of dimethyl fumarate in patients with multiple sclerosis treated in routine medical practice: final analysis of the ESTEEM study. Neurol Ther. 2025;14(1):243–60.

6. Gold R, Kappos L, Arnold DL, et al. Placebo-controlled phase 3 study of oral BG-12 for relapsing multiple sclerosis. N Engl J Med. 2012;367(12):1098–107.

7. Gold R, Arnold DL, Bar-Or A, et al. Long-term safety and efficacy of dimethyl fumarate for up to 13 years in patients with relapsing-remitting multiple sclerosis: final ENDORSE study results. Mult Scler J. 2022;28(5):801–16.

8. Lager B, Liseno J, Božin I, et al. Real-world analysis affirms the high persistence and adherence observed with diroximel fumarate in patients with multiple sclerosis. Neurol Ther. 2023;12(1):145–59.

9. Mao-Draayer Y, Bar-Or A, Balashov K, et al. Real-world safety and effectiveness of dimethyl fumarate in patients with MS: results from the ESTEEM phase 4 and PROCLAIM phase 3 studies with a focus on older patients. Adv Ther. 2025;42(1):395–412.

10. Roman C, Garabedian M, Schobel VR, et al. Patient and healthcare provider experience of diroximel fumarate: considerations for selecting disease-modifying therapy. Neurodegener Dis Manag. 2025;15(6):271–81.

11. Biogen Internal Data. 2025.

12. Chen H, Assmann JC, Krenz A, et al. Hydroxycarboxylic acid receptor 2 mediates dimethyl fumarate’s protective effect in EAE. J Clin Investig. 2014;124(5):2188–92.

13. Parodi B, Rossi S, Morando S, et al. Fumarates modulate microglia activation through a novel HCAR2 signaling pathway and rescue synaptic dysregulation in inflamed CNS. Acta Neuropathol. 2015;130(2):279–95.

14. Schulze-Topphoff U, Varrin-Doyer M, Pekarek K, et al. Dimethyl fumarate treatment induces adaptive and innate immune modulation independent of Nrf2. Proc Natl Acad Sci. 2016;113(17):4777–82.

15. Longbrake EE, Mao-Draayer Y, Cascione M, et al. Dimethyl fumarate treatment shifts the immune environment toward an anti-inflammatory cell profile while maintaining protective humoral immunity. Mult Scler J. 2021;27(6):883–94.

16. Naismith RT, Wolinsky JS, Wundes A, et al. Diroximel fumarate (DRF) in patients with relapsing–remitting multiple sclerosis: interim safety and efficacy results from the phase 3 EVOLVE-MS-1 study. Mult Scler J. 2020;26(13):1729–39.

17. The EVOLVE-MS-2 Study Group, Naismith RT, Wundes A, et al. Diroximel fumarate demonstrates an improved gastrointestinal tolerability profile compared with dimethyl fumarate in patients with relapsing–remitting multiple sclerosis: results from the randomized, double-blind, phase III EVOLVE-MS-2 study. CNS Drugs. 2020;34(2):185–96.

18. Palte MJ, Wehr A, Tawa M, et al. Improving the gastrointestinal tolerability of fumaric acid esters: early findings on gastrointestinal events with diroximel fumarate in patients with relapsing-remitting multiple sclerosis from the phase 3, open-label EVOLVE-MS-1 Study. Adv Ther. 2019;36(11):3154–65.

19. Chinea A, Amezcua L, Vargas W, et al. Real-world safety and effectiveness of dimethyl fumarate in Hispanic or Latino patients with multiple sclerosis: 3-year results from ESTEEM. Neurol Ther. 2020;9(2):495–504.

20. Fox RJ, Miller DH, Phillips JT, et al. Placebo-controlled phase 3 study of oral BG-12 or glatiramer in multiple sclerosis. N Engl J Med. 2012;367(12):1087–97.

21. Gold R, Arnold DL, Bar-Or A, et al. Safety and efficacy of delayed-release dimethyl fumarate in patients with relapsing-remitting multiple sclerosis: 9 years’ follow-up of DEFINE, CONFIRM, and ENDORSE. Ther Adv Neurol Disord. 2020;13:1756286420915005.

22. Williams MJ, Amezcua L, Okai A, et al. Real-world safety and effectiveness of dimethyl fumarate in Black or African American patients with multiple sclerosis: 3-year results from ESTEEM. Neurol Ther. 2020;9(2):483–93.

23. Mallucci G, Annovazzi P, Miante S, et al. Two-year real-life efficacy, tolerability and safety of dimethyl fumarate in an Italian multicentre study. J Neurol. 2018;265(8):1850–9.

24. Belotti LMB, Di Martino M, Zenesini C, et al. Impact of adherence to disease-modifying drugs in multiple sclerosis: a study on Italian real-world data. Mult Scler Relat Disord. 2023;80:105094.

25. Burks J, Marshall TS, Ye X. Adherence to disease-modifying therapies and its impact on relapse, health resource utilization, and costs among patients with multiple sclerosis. Clin Outcomes Res CEOR. 2017;28(9):251–60.

26. Bureau UC. Census.gov. Health Insurance in the United States: 2024—Tables. https://www.census.gov/data/tables/2025/demo/health-insurance/p60-288.html . Accessed 23 Jan 2026.

27. Ben-Zacharia AB, Feng JJ, Moss BP, et al. Clinical characteristics and treatment outcomes in multiple sclerosis patients treated with anti-CD20s who switched to fumarates: a retrospective analysis of a US healthcare claims database. J Comp Eff Res. 2025;14(3):e240071.

28. Quan H, Li B, Couris CM, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. 2011;173(6):676–82.

29. Roxburgh RHSR, Seaman SR, Masterman T, et al. Multiple Sclerosis Severity Score: using disability and disease duration to rate disease severity. Neurology. 2005;64(7):1144–51.

Associated Data

Supplementary Materials

Supplementary file1 (PDF 284 KB)

Data Availability Statement

Forian’s claims database is a proprietary resource accessible via licensing. ClearView holds an enterprise license, and Biogen funded a one-time use for this analysis.