Реальный опыт использования озанимода у взрослых с множественной скlerозом.

Introduction

Ozanimod is a second-generation sphingosine 1-phosphate receptor (S1PR) modulator that was approved by the U.S. Food and Drug Administration for treatment of relapsing multiple sclerosis (MS) in 2020. S1PR modulators are thought to exert their therapeutic effects through sequestration of lymphocytes in the lymph nodes [ 1 ], although they can cross the blood brain barrier and may exert direct CNS effects as well [ 2 ]. There are five subtypes of S1PRs, which are widely expressed throughout the body, including in neurons, smooth muscle, and in the heart’s conduction system. Modulation of S1PR in other tissues can have significant adverse effects, such as bradycardia and hypertension. Ozanimod and other second-generation S1PR modulators selectively bind S1PR1 and S1PR5, thereby improving their safety profiles [ 3 , 4 ].

Ozanimod approval was based on two randomized phase three clinical trials, SUNBEAM and RADIANCE [ 5 , 6 ], which showed that ozanimod decreased annualized relapse rates compared to interferon beta-1a treatment, while also showing a favorable safety profile. However, confirmed disability progression at 3 and 6 months was not significantly different between treatment groups in a pooled analysis of the SUNBEAM and RADIANCE clinical trials [ 5 ]. In a retrospective analysis using a matching-adjusted indirect comparison, ozanimod demonstrated comparable effictiveness to fingolimod but had fewer adverse events [ 7 ]. However, the safety and effectiveness of ozanimod in real-world practice, particularly in a broader population and outside of the closely regulated clinical trial environment, remains unclear. Of special interest is ozanimod use in older patients and in those with comorbidities such as hypertension and diabetes mellitus.

We conducted a retrospective study to investigate ozanimod prescribing patterns and to describe its performance in a real-world cohort with respect to persistence, tolerability, safety, and effectiveness.

Methods

Patient selection and data collection

This retrospective study was approved by the Cleveland Clinic Institutional Review Board (IRB 23–660). People with MS (pwMS) were identified through a search of the electronic medical record (EMR) for any individual seen at a United States Cleveland Clinic facility and prescribed ozanimod for MS. Data from March 2020 to September 2023 were collected. All data were stored within an encrypted REDCap database on Cleveland Clinic servers.

Inclusion criteria included a diagnosis of MS or clinically isolated syndrome per the 2017 revisions of the McDonald Criteria [ 8 ], at least one dose of ozanimod outside of a clinical trial, and age ≥18 years at the time of ozanimod initiation. Exclusion criteria included another neurodegenerative condition that, in the opinion of the investigators, could confound study results, ozanimod use only through a clinical trial, or if ozanimod was taken for an indication other than MS.

Data collected included demographics, comorbidities (hypertension, hyperlipidemia, type 2 diabetes mellitus, coronary artery disease, peripheral vascular disease, depression, anxiety, asthma, COPD), MS disease history, disease-modifying therapy (DMT) history, ozanimod initiation and discontinuation dates, characterization of relapses, tolerability and adverse events, Patient-Determined Disease Steps (PDDS) [ 9 ], MRI and optical coherence tomography (OCT) data, and serum studies, including complete blood count and metabolic panels. Individuals are asked to complete Multiple Sclerosis Performance Test (MSPT) at each visit, which is an electronic adaptation of the Multiple Sclerosis Functional Composite [ 10 , 11 ], and includes the Walking Speed Test (WST), the Manual Dexterity Test (MDT), and the Processing Speed Test (PST) [ 10 ]. Clinical relapses were defined as new or worsening symptoms lasting at least 24 hours outside of a fever or infection that were attributed to MS relapse by the treating neurological team. Radiological disease activity was classified as new or enlarging T2 lesions or gadolinium enhancing (GdE) lesions and was taken directly from structured radiology interpretations that specifically report on these findings. If MRIs were obtained without gadolinium, only data on new or enlarging T2 lesions were collected. Ozanimod tolerability and side effects were determined based on data documented in the EMR, defined as any symptom reported by the patient that correlated temporally to the initiation of ozanimod. Adverse effects (AE) were defined as undesirable clinical events occurring after initiation of ozanimod, including infection or worsening of hypertension.

Data were reviewed through 12 months of ozanimod treatment or until ozanimod discontinuation. Data were grouped into four epochs: baseline, 3 months, 6 months, and 12 months post-initiation. Data were assigned to the temporally closest epoch, within a window of ±3 months, which allowed flexibility given the inherently variable nature of retrospective observational data. If there were multiple available data, those closest to a particular epoch were chosen.

Statistical analysis

Continuous variables were summarized as mean (standard deviation (SD)) or median (interquartile range [IQR]), and categorical variables as counts (percentages). Baseline was defined as the date of ozanimod initiation. Comparisons were conducted for pwMS with available data at one or more time points: baseline, 3 months, 6 months, and 12 months post-treatment. For each continuous variable, linear mixed-effects regression models (assuming normal distribution) with a random intercept were used for comparison to account for multiple observations per participant over time. Model residuals were visually assessed for normality. Generalized linear mixed-effects models (GLMM) with a logit link were used for binary variables. Time was treated as a categorical variable. Relapses were identified during the year before and the 12 months after treatment initiation. The annualized relapse rate (ARR) and 95% confidence interval (CI) pre- and post-treatment were estimated and compared using a GLMM with a Poisson link and the log-transformed observation duration as the offset variable. All tests were two-sided, and statistical significance was set at p  < 0.05. Findings were considered statistically significant after false discovery rate (FDR) correction across all measures used for follow-up and baseline comparisons (laboratory values, clinical and radiological disease activity, and disability progression). FDR adjusted p-values were reported. R version 4.5.1 was used for all analyses.

Results

Baseline demographics

A total of 206 pwMS were included in the study. Baseline characteristics and demographics are summarized in Table 1 . The mean age of pwMS starting ozanimod was 47.3 ± 11.5 years, with the majority being female ( n  = 140, 68%) and White ( n  = 175, 85%). Most had a relapsing-remitting disease course ( n  = 182, 88%). The mean time from diagnosis at ozanimod initiation was 9.9 ± 10.1 years. Baseline PDDS data were available for 68 individuals, with the majority reporting no disability ( n  = 31, 46%).

Table 1. Baseline characteristics of the study sample prior to starting ozanimod ( N  = 206). SD: standard deviation. RRMS: relapsing-remitting MS. Characteristic Mean (SD)/N (%) Age, years: mean (SD) 47.3 (11.5) Sex: female (%) 140 (68) Race: (%) White Black Native Hawaiian/Pacific Islander Asian Unknown 175 (85.0) 18 (8.7) 5 (2.4) 2 (1.0) 6 (2.9) Ethnicity: (%) Not Hispanic Hispanic Unknown 193 (93.7) 9 (4.4) 4 (1.9) Disease duration since MS symptom onset, years: mean (SD) Missing 12/206 pwMS 12.5 (11.0) Disease duration since MS diagnosis, years: mean (SD) Missing 9/206 pwMS 9.9 (10.1) Most recent disease course: (%) RRMS/CIS SPMS PPMS PRMS Unknown 182 (88.3) 13 (6.3) 1 (0.5) 6 (2.9) 4 (1.9) PDDS: (%) 3 0 Normal 1 Mild disability 2 Moderate disability 3 Gait disability 4 Early cane 5 Late cane 6 Bilateral support 7 Wheelchair bound Missing 31 (45.6) 17 (25.0) 7 (10.3) 9 (13.2) 2 (2.9) 0 (0) 0 (0) 2 (2.9) 138 Comorbidities: (%) Hypertension Hyperlipidemia Type 2 diabetes mellitus Depression/anxiety Prescribed a serotonergic medication 51 (24.8) 60 (29.1) 17 (8.3) 97 (47.1) 76 (37.8) DMT experience Previously taken DMT: (%) Most recent prior DMT: (%)  High efficacy   ocrelizumab   natalizumab   rituximab   ofatumumab  Medium/low efficacy   fingolimod   dimethyl fumarate   glatiramer acetate   teriflunomide   interferon beta   diroximel fumarate   siponimod   cladribine   ponesimod   IVIG 146 (70.9) 25 (17.1) 14 (9.6) 5 (3.4) 4 (2.7) 2 (1.4) 121 (82.9) 56 (38.4) 22 (15.1) 20 (13.7) 7 (4.8) 7 (4.8) 4 (2.7) 2 (1.4) 1 (0.7) 1 (0.7) 1 (0.7) Abbreviations: CIS: clinically isolated syndrome. SPMS: secondary progressive MS. PPMS: primary progressive MS. PRMS: progressive relapsing MS. PDDS: patient-determined disease steps.

Baseline characteristics of the study sample prior to starting ozanimod ( N  = 206). SD: standard deviation. RRMS: relapsing-remitting MS.

Abbreviations: CIS: clinically isolated syndrome. SPMS: secondary progressive MS. PPMS: primary progressive MS. PRMS: progressive relapsing MS. PDDS: patient-determined disease steps.

Most of the cohort had previously been treated with DMT ( n  = 146, 71%), with most switching from a medium/low efficacy DMT ( n  = 121, 83%) to ozanimod. Overall, the most common DMT taken prior to ozanimod was fingolimod ( n  = 56, 38%), followed by dimethyl fumarate ( n  = 22, 15%) ( Table 1 ). For those with prior DMT experience ( n  = 146), 129 individuals had a recorded stop date for their prior DMT. The average time between stopping the prior DMT and starting ozanimod was 256 days (SD = 723). 68 individuals had the same date for stopping the prior DMT and starting ozanimod. The maximum gap was 5215 days.

Individuals switched to ozanimod most often for insurance reasons ( n  = 48, 33%) followed by poor tolerance of prior DMT ( n  = 29, 20%) ( Figure 1 ). Of the 48 pwMS who switched to ozanimod due to cost, 47 were most recently treated with fingolimod, perhaps due to the loss of copay support programs when generic fingolimod entered the market. Among the 25 pwMS who switched from a high efficacy DMT to ozanimod, the most common switch was from ocrelizumab ( n  = 14, 9.6%). Individuals most often switched from a high efficacy DMT due to side effects/poor tolerance ( n  = 7, 28%), followed by clinical worsening ( n  = 4, 16%). A proportion of pwMS discontinued high efficacy DMT more than one year prior to starting ozanimod ( n  = 8, 32%). The remainder of the patients discontinued high efficacy DMT for reasons including colitis ( n  = 2), JC virus positivity on natalizumab ( n  = 1), infection ( n  = 1), concern for development of Crohn’s disease from ocrelizumab and hypogammaglobulinemia ( n  = 1), or MRI activity ( n  = 1). Supplementary Figure S1 details the time course of the discontinuation of high efficacy DMT prior to switch to ozanimod.

Reason for switch from previous DMT to ozanimod.

Ozanimod persistence, tolerability, and safety

Most pwMS ( n  = 149, 72%) remained on ozanimod at last data collection. In those who stopped ozanimod, it was discontinued at a median of 204 [IQR 57; 383] days. Reasons for discontinuation are summarized in Figure 2 . The most common reason for discontinuation was side effects or AEs, which included worsening hypertension ( n  = 5), gastrointestinal issues such as nausea or heartburn ( n  = 4), heart palpitations ( n  = 3), headache ( n  = 2), fatigue ( n  = 1), dizziness ( n  = 1), mood effects ( n  = 1), heart failure exacerbation ( n  = 1), and tachycardia ( n  = 1). Bilateral macular edema was identified in one individual at 12 months of treatment with ozanimod; the treating clinician reported that co-morbid type 1 diabetes mellitus was likely contributory. One pwMS developed necrotizing fasciitis, which led to discontinuation of ozanimod.

Discontinuation pattern of ozanimod. *9 pwMS with new T2 hyperintense or enhancing brain lesions. **7 pwMS with elevated liver function tests. 1 with lymphopenia. Other reasons for discontinuation include cost ( n  = 5), clinical relapse ( n  = 5), infections ( n  = 5), starting cancer treatment ( n  = 5), convenience/patient choice ( n  = 4), disability progression ( n  = 3), family planning ( n  = 1).

Following discontinuation of ozanimod, individuals most often switched to ofatumumab ( n  = 14, 25%), followed by ocrelizumab ( n  = 12, 21%). Ten individuals (17.5%) remained off DMT at last follow-up.

Side effects were reported in 28%, 11%, and 7.5% of the population at 3, 6, and 12 months, respectively, after starting ozanimod ( Table 2 ). The most commonly reported side effect was headache, although the proportion of pwMS reporting headache decreased over the course of 12 months of treatment (15% at 3 months vs 3.3% at 12 months). No AEs were observed in 76%, 83%, and 77% of the cohort at 3, 6, and 12 months. The most common AE was worsening hypertension, which was reported in 5 (5.4%), 3 (2.7%), and 6 (5%) individuals at 3, 6, and 12 months, respectively ( Table 2 ). A subgroup analysis of the combined group of pwMS with hypertension, hyperlipidemia, or diabetes demonstrated similar rates of side effects, AEs, and discontinuation of ozanimod (26/90 pwMS, 28.9%) as the overall cohort (Supplementary Table S1). Serotonergic medications were prescribed for 38–44% of the cohort during the first 12 months of ozanimod treatment. There were no cases of serotonin syndrome.

Table 2. Side effect and adverse event profile by time point.   3 Months, N 1  = 93 6 Months, N 1  = 112 12 Months, N 1  = 120 Did the patient report any tolerability concerns?       No 67 (72%) 100 (89%) 111 (93%) Yes 26 (28%) 12 (11%) 9 (7.5%) Heart palpitations 1 (1.1%) 2 (1.8%) 1 (0.8%) Dizziness 1 (1.1%) 0 (0%) 1 (0.8%) Headache 14 (15%) 4 (3.6%) 4 (3.3%) Other 16 (17%) 8 (7.1%) 3 (2.5%) Did any adverse events occur?       No 71 (76%) 93 (83%) 92 (77%) Yes 22 (24%) 19 (17%) 28 (23%) UTI 3 (3.2%) 3 (2.7%) 5 (4.2%) URI 8 (8.6%) 4 (3.6%) 14 (12%) Other Infection 2 (2.2%) 6 (5.4%) 3 (2.5%) Hypertension Worsening 5 (5.4%) 3 (2.7%) 6 (5%) a Number of patients within each epoch. Abbreviations: UTI: urinary tract infection; URI: upper respiratory infection.

Side effect and adverse event profile by time point.

a Number of patients within each epoch.

Abbreviations: UTI: urinary tract infection; URI: upper respiratory infection.

Overall, there was a significant decrease in mean white blood cell count at 12 months compared to baseline (7.19 k/µL (SD = 3.42) at baseline vs 5.56 k/µL (2.32) at 12 months, mean change: −1.52; 95% CI: −1.97 to −1.06; p  < 0.001, Figure 3 , Supplementary Table S2). Furthermore, the mean absolute lymphocyte count (ALC) decreased over the same period (1.60 k/µL (SD = 0.94) at baseline vs 0.61 k/µL (SD = 0.53) at 12 months, mean change: −1.02, 95% CI: −1.15 to −0.88; p  < 0.001, Figure 3 , Supplementary Table S2). However, when considering only individuals who switched from fingolimod to ozanimod ( n  = 56), there was no significant change in mean white blood cell count (5.47 k/µL (SD = 2.89) at baseline vs 5.24 k/µL (SD = 1.58) at 12 months, p  = 0.21). Among this group of pwMS who switched to fingolimod, there was a significant decrease in ALC over the same period (0.76 k/µL (SD = 0.52) at baseline vs 0.55 k/µL (SD = 0.33) at 12 months, mean change: −0.19, 95% CI: −0.33 to −0.04; p  = 0.034, Figure 3 , Supplementary Table S2).

Change over time of WBC, ALC, AST, and ALT in all pwMS (solid line) and in the subgroup that switched from fingolimod (dashed line). The error bars show the standard deviation. * represents p  < 0.01 for the comparison of each individual epoch to the baseline for all pwMS. # represents p  < 0.05 for the comparison of each individual epoch with baseline for the subgroup that switched from fingolimod.

Abbreviations: WBC: white blood cell count. ALC: absolute lymphocyte count. AST: aspartate aminotransferase. ALT: alanine transaminase.

Clinical and radiological disease activity

The number of clinical relapses decreased from 35 in the 12 months before starting ozanimod to 13 in the 12 months after DMT initiation. The annualized relapse rate (ARR) decreased from 0.11 (95% CI: 0.06–0.20) in the 12 months prior to treatment to 0.05 (95% CI: 0.02–0.10) after 12 months of treatment, representing approximately a 56% reduction (incidence rate ratio [IRR]: 0.44, 95% CI: 0.34–0.83; p  = 0.034). There was also a reduction in the odds of pwMS with new or enlarging T2 intracranial lesions (40%, n  = 39 vs 6.6%, n  = 5; odds ratio [OR]: 0.10, 95% CI: 0.04–0.28; p  < 0.001) and in the percentage of pwMS with GdE lesions (27%, n  = 36 vs 3.9%, n  = 3, p  = 0.027) when comparing baseline to 12 months ( Table 3 ). There were few new or enlarging T2 or enhancing cervical spinal cord lesions at any epoch ( Table 3 ).

Table 3. Brain and cervical spinal cord radiographic disease by epoch. Generalized linear mixed models with a logit link were used to compare each epoch to baseline for new brain T2 lesions. McNemar’s test was used for new brain GdE lesions; p-values were not available for cervical spinal cord lesions due to too few discordant pairs. FDR (False Discovery Rate) adjusted p-values were reported. The number of events in each column reflects events within that epoch only.   Baseline 3 Months 6 Months 12 Months INTRACRANIAL LESIONS           N  = 97 N  = 30 N  = 77 N  = 76 New or enlarging brain MRI T2 lesions 39 (40%) 10 (33%) 12 (16%) 5 (6.6%) Odds ratio (95% CI)   0.74 (0.31 – 1.76) 0.27 (0.13 – 0.57) 0.10 (0.04 – 0.28) FDR adjusted p-value 0.72 0.003  <0.001 Number of new or enlarging brain MRI T2 lesions         0 58 (60%) 20 (67%) 65 (84%) 71 (93%) 1 16 (16%) 10 (33%) 9 (12%) 2 (2.6%) 2 10 (10%) 0 (0%) 1 (1.3%) 2 (2.6%) 3 4 (4.1%) 0 (0%) 1 (1.3%) 1 (1.3%) >3 9 (9.3%) 0 (0%) 1 (1.3%) 0 (0%)   N  = 132 N  = 31 N  = 77 N  = 76 New GdE brain lesions 36 (27%) 1 (3.2%) 3 (3.9%) 3 (3.9%) Absolute change in n (%) to baseline   3 (19%), n  = 16, 14 (26%), n  = 53 10 (19%), n  = 52 FDR adjusted p-value   0.18 0.001 0.027 Number of new GdE brain MRI lesions         0 96 (73%) 30 (97%) 74 (96%) 73 (96%) 1 20 (15%) 1 (3.2%) 3 (3.9%) 2 (2.6%) 2 4 (3%) 0 (0%) 0 (0%) 1 (1.3%) 3 1 (0.8%) 0 (0%) 0 (0%) 0 (0%) >3 11 (8.3%) 0 (0%) 0 (0%) 0 (0%) CERVICAL SPINAL CORD LESIONS           N  = 34 N  = 14 N  = 32 N  = 28 New or enlarging cervical cord T2 lesions 5 (15%) 4 (29%) 1 (3.1%) 0 (0%) Absolute change in n (%) to baseline   1 (50%), n  = 2 0 (0%), n  = 4 2 (29%), n  = 7 Number of new or enlarging cervical cord T2 lesions         0 29 (85%) 10 (71%) 31 (97%) 28 (100%) 1 2 (5.9%) 4 (29%) 1 (3.1%) 0 (0%) 2 0 (0%) 0 (0%) 0 (0%) 0 (0%) 3 0 (0%) 0 (0%) 0 (0%) 0 (0%) >3 3 (8.8%) 0 (0%) 0 (0%) 0 (0%)   N  = 77 N  = 14 N  = 33 N  = 27 New GdE cord lesions 8 (10%) 2 (14%) 0 (0%) 0 (0%) Absolute change in n (%) to baseline   1 (14%), n  = 7 0 (0%), n  = 13 2 (15%), n  = 13 Number of new GdE cord lesions         0 69 (90%) 12 (86%) 33 (100%) 27 (100%) 1 8 (10%) 2 (14%) 0 (0%) 0 (0%) 2 0 (0%) 0 (0%) 0 (0%) 0 (0%) 3 0 (0%) 0 (0%) 0 (0%) 0 (0%) >3 0 (0%) 0 (0%) 0 (0%) 0 (0%) N  = total observations at each epoch; n  = number of complete pairs used for comparison. Abbreviation: GdE: gadolinium enhancing.

Brain and cervical spinal cord radiographic disease by epoch. Generalized linear mixed models with a logit link were used to compare each epoch to baseline for new brain T2 lesions. McNemar’s test was used for new brain GdE lesions; p-values were not available for cervical spinal cord lesions due to too few discordant pairs. FDR (False Discovery Rate) adjusted p-values were reported. The number of events in each column reflects events within that epoch only.

N  = total observations at each epoch; n  = number of complete pairs used for comparison.

Abbreviation: GdE: gadolinium enhancing.

Disability progression

Evaluation of measures of disability across the first 12 months of ozanimod use were impacted by low sample size ( N  = 27–70 for each epoch). There were no significant differences in PDDS when comparing each epoch to baseline. Similarly, there was no difference in walking speed or manual dexterity in either hand after starting ozanimod compared to baseline. There was a small increase in the mean number of correct responses on the processing speed test from baseline to 12 months after ozanimod initiation (estimated mean change: 2.7; 95% CI:0.51 to 4.89, p  = 0.043) ( Table 4 ).

Table 4. Disability metrics across 12 months of ozanimod. Generalized linear mixed model with a logit link was used to compare each epoch to baseline for PDDS. The change (Δ) from baseline and 95% confidence interval (CI) were estimated from linear mixed models. FDR (False Discovery Rate) adjusted p-values were reported. The number of events in each column is representative of the number of events within each epoch.   Baseline 3 Months 6 Months 12 Months PDDS N  = 70 N  = 27 N  = 30 N  = 34 Moderate or less 58 (83%) 25 (93%) 21 (70%) 27 (79%) Gait disability or severe 12 (17%) 2 (7.4%) 9 (30%) 7 (21%) Odds ratio (95% CI)   0.39 (0.06 – 1.56) 2.07 (0.75 – 5.62) 1.25 (0.43 – 3.48) FDR adjusted p-value 0.40 0.29 0.81 Walking Speed Test (seconds) N  = 66 N  = 24 N  = 29 N  = 30 Mean (SD) 6.99 (6.80) 6.82 (1.92) 8.47 (6.42) 6.63 (2.33) Δ from baseline (95% CI)   0.15 (−0.77 – 1.07) 0.26 (−0.72 – 1.23) 0.41 (−0.51 – 1.33) FDR adjusted p-value 0.81 0.81 0.60 Manual Dexterity Test, Right Hand (seconds) N  = 76 N  = 25 N  = 30 N  = 29 Mean (SD) 27.88 (7.62) 28.97 (7.32) 27.41 (7.14) 24.88 (7.42) Δ from baseline (95% CI)   0.92 (−1.92 – 3.75) −0.50 (−3.26 – 2.25) −2.53 (−5.27 – 0.21) FDR adjusted p-value 0.73 0.81 0.16 Manual Dexterity Test, Left Hand (seconds) N  = 76 N  = 25 N  = 30 N  = 28 Mean (SD) 28.59 (9.42) 29.47 (8.21) 29.36 (7.89) 26.75 (13.03) Δ from baseline (95% CI)   0.77 (−3.17 – 4.72) 0.67 (−3.10 – 4.45) −1.66 (−5.50 – 2.17) FDR adjusted p-value 0.81 0.81 0.60 Processing Speed Test: Number Correct N  = 85 N  = 27 N  = 30 N  = 32 Mean (SD) 52.46 (11.42) 49.52 (12.52) 50.60 (14.70) 54.88 (13.25) Δ from baseline (95% CI)   −0.25 (−2.44 – 1.94) 0.46 (−1.86 – 2.78) 2.70 (0.51 – 4.89) FDR adjusted p-value 0.84 0.81 0.043 N  = total observations at each epoch. Abbreviations: PDDS: patient-determined disease steps. IQR: interquartile range.

Disability metrics across 12 months of ozanimod. Generalized linear mixed model with a logit link was used to compare each epoch to baseline for PDDS. The change (Δ) from baseline and 95% confidence interval (CI) were estimated from linear mixed models. FDR (False Discovery Rate) adjusted p-values were reported. The number of events in each column is representative of the number of events within each epoch.

N  = total observations at each epoch.

Abbreviations: PDDS: patient-determined disease steps. IQR: interquartile range.

Discussion

This study provides comprehensive evidence evaluating the persistence, tolerability, safety, and effectiveness of ozanimod in a real-world cohort. The results suggest good clinical and radiographic effectiveness, with fewer pwMS experiencing clinical relapses on ozanimod than in the year prior to ozanimod initiation (IRR = 0.44, p  = 0.034). Furthermore, radiographic and disability outcomes also indicate a good response to ozanimod. The medication was well-tolerated, with 72% of the cohort remaining on treatment at last follow-up. No serious adverse events requiring hospitalization were observed.

While observational studies have inherent limitations, well-designed examples remain essential in guiding treatment strategies for multiple sclerosis. When combined with the rigor of randomized controlled trial data, they provide a more comprehensive understanding of treatment effectiveness in real-world settings. This integration allows for personalized patient care, by providing guidance for more nuanced clinical decisions than can be answered from clinical trial data alone. Observational data can more closely reflect the complexity of MS in everyday practice [ 12 , 13 ].

A number of differences between our cohort and that of the pivotal phase 3 trials are worth noting. In particular, pwMS in our study were older (mean age 47.3 years) than in SUNBEAM and RADIANCE (mean age 34.8 to 38 years) [ 5 , 6 ] and had longer disease duration (9.9 vs 3.5–4 years in SUNBEAM and RADIANCE) [ 5 , 6 ]. Our cohort was also more diverse with 15% non-White individuals versus 0.2–1% in the phase 3 trials [ 5 , 6 ]. A significant proportion of our cohort had type 2 diabetes mellitus (8%) or hypertension (25%) and worsening hypertension while on ozanimod was reported in only 2–6% of the cohort across the 12 months of this study. People with uncontrolled diabetes mellitus type 2 (hemoglobin A1c > 7 [ 5 ] or 9% [ 6 ]) or significant comorbidities such as retinopathy or nephropathy were excluded from SUNBEAM and RADIANCE, but were included in this study. Further, individuals with prior exposure to lymphocyte-depleting or lymphocyte-trafficking agents were excluded from SUNBEAM and RADIANCE, and only 28–29% of participants were DMT experienced, as compared to 71% in our study [ 5 , 6 ]. Similar to the pivotal phase 3 trials, no pwMS in our cohort developed bradycardia or heart block temporally associated with initiation of ozanimod. Overall, our older, sicker, and more racially diverse cohort included populations that are often underrepresented in clinical trials. Importantly, ozanimod remained safe and well-tolerated, highlighting its potential suitability for a wide range of real-world patients.

With regards to inflammatory activity, RADIANCE and SUNBEAM required a clinical relapse within the past 12 months or a clinical relapse within 24 months and a GdE lesion within the last 12 months, which resulted in an average of 1.3 relapses in the year prior for the 1 mg ozanimod group in both phase 3 trials vs an average of 0.17 relapses in the year prior for our study. Furthermore, the ARR at 12 months of treatment was 0.06 for the current study vs 0.18 at 12 months (SUNBEAM) and 0.17 at 24 months (RADIANCE). These findings suggest that our older population was less clinically active than those in the phase 3 trials but still responded well to ozanimod for relapse reduction.

Our findings represent the first published real-world data on the use of ozanimod for the treatment of multiple sclerosis (MS). A 5-year, prospective, observational study known as the OzEAN trial [ 14 ] is currently ongoing to further assess the real-world effectiveness, safety, and treatment patterns of ozanimod in pwMS. However, data from this trial have not yet been published. In contrast, multiple observational studies have been conducted on the use of ozanimod in ulcerative colitis [ 15 , 16 ]. These studies consistently demonstrate that ozanimod is generally well-tolerated and associated with a favorable safety profile in that patient population.

Hypertension was reported as an adverse event in 2.7–5.4% of our cohort across the 12 months of this study. In contrast, a study of individuals with secondary progressive MS taking siponimod reported that 16.2% experienced hypertension as an adverse event over an 18-month follow-up period, although this may be partially attributable to the study’s older population (mean age 53.4 [8.5]) [ 17 ]. Among pwMS treated with fingolimod, real-world studies that reported on hypertension showed rates ranging from approximately 4%-6% over follow-up periods of 1–5 years [ 14 , 18 , 19 ]. Furthermore, recent evidence suggests that fingolimod may be associated with an increased risk of cardiovascular disease, specifically hypertension, compared to natalizumab. Given the relationship between fingolimod and hypertension, the relationship between hypertension and ozanimod also deserves further future study [ 20 ].

The most common reason for switching DMT to ozanimod was insurance related. For those pwMS transitioning from fingolimod to ozanimod, nearly all (47 of 48) cited cost as the reason for the change in DMT. Generic fingolimod became commercially available in 2022, and subsequently, patient assistance programs linked to the brand name version ended or changed, leading to increased costs for patients. We suspect this change in out-of-pocket costs for pwMS receiving fingolimod drove most of the insurance-related switches.

Ozanimod previously carried an FDA drug interaction warning for serotonergic medications based on in vitro evidence showing that one of its active metabolites inhibits monoamine oxidase B (MAO-B), potentially increasing the risk of serotonin syndrome when used concurrently with serotonergic agents. However, recent data have cast doubt on the significance of this potential interaction [ 21 , 22 ], with no clinical evidence supporting an increased risk of serotonin syndrome with ozanimod. As a result, the FDA recently removed the interaction warning for ozanimod and serotonergic agents. Our findings further confirm the lack of a link between serotonin syndrome and ozanimod, as no pwMS developed serotonin syndrome over 12 months of follow-up.

S1PR modulators, through the mechanism of lymphocyte sequestration, inherently cause lymphopenia, which may require treatment discontinuation in more severe cases. Previous studies suggested that lymphopenia associated with ozanimod may be less severe compared to fingolimod and that ALC can rebound when switching from fingolimod to ozanimod [ 23 , 24 ]. However, in a subgroup analysis, we did not find a significant rebound in lymphocyte count among those switching from fingolimod to ozanimod. In fact, there was a small but significant decrease in ALC relative to baseline at months 3, 6, and 12 after switching from fingolimod to ozanimod. The reason for this discrepancy is unclear but could be related to small subgroup sample size( n  = 17–25) or differences in clinical features. For example, in an observational analysis performed by Caliendo et al. [ 23 ], pwMS who were switched from fingolimod to ozanimod due to lymphopenia demonstrated a moderate improvement in lymphocyte count after six months of ozanimod (0.46 n/μL at fingolimod discontinuation to 0.75 n/μL at six months of ozanimod). Of the 18 pwMS in this study, 13 had discontinued fingolimod due to grade 3 or 4 lymphopenia. By contrast, in our study, no patients discontinued fingolimod prior to initiating ozanimod because of lymphopenia. Taken together, these studies suggest that switching from fingolimod to ozanimod may be helpful in the setting of severe lymphopenia, but ALC increases are less likely in those transitioning to ozanimod with more typical lymphocyte counts.

There are several limitations of the current study. Although we were able to collect data on all pwMS or CIS at Cleveland Clinic U.S. sites who took at least one dose of ozanimod outside of a clinical trial, the sample size was still limited. Due to the nature of retrospective observational studies, there were substantial missing data, but we found no evidence to suggest the missingness was systematic in a way that might introduce bias. Standard practice at our centers is to see a patient in follow-up three months after changing a disease-modifying therapy, but this can vary between clinicians, leading to missing data. The Multiple Sclerosis Performance Test (MSPT) is also sometimes bypassed for reasons including online visits or patients arriving late to their appointments. Accordingly, data were particularly limited for disability and functional outcomes measures, weakening the robustness of these findings. We also chose to bin data into a single epoch for each time point to maximize the sample size and allow for an understanding of ozanimod’s effects over time. As a result, some data were excluded, but this occurred quite infrequently due to the typical pattern of follow-up visits at our centers. Another concern is that regression to the mean may have occurred between the pre-ozanimod and post-ozanimod ARR assessments, and we are actively engaged in more longitudinal data collection data to help mitigate this possibility. In addition, while many pwMS switched directly from a prior DMT to ozanimod, many had an extended interval between treatments, with such individuals being better characterized as new starts than as switchers. Although these 12-month treatment data represent an important initial step in evaluating real-world ozanimod tolerance and effectiveness, longitudinal data are essential to more fully characterize treatment response and safety.

A strength of our study is the inclusion of pwMS from multiple geographically diverse sites, which increases the generalizability of our results. The study also involved a manual chart review, which allows for collection of details that might be missed in other study types, such as analysis of claims-based data. As discussed above, this study also provides guidance for the use of ozanimod in pwMS who would not have qualified for the pivotal trials, including older individuals, those with comorbidities, and individuals taking SSRIs. Finally, the inclusion of multiple outcome types, including clinical, imaging, and safety metrics, allows for a comprehensive evaluation of ozanimod performance in a real-world cohort.

Conclusions

This retrospective study provides real-world evidence on the prescribing patterns, persistence, tolerability, safety, and effectiveness of ozanimod in pwMS. The findings suggest that ozanimod is a well-tolerated and effective treatment option with most individuals remaining on treatment at 12-month follow-up and experiencing reductions in clinical relapses and MRI activity compared to baseline. While some individuals experienced side effects, AEs were generally uncommon, and treatment discontinuation was infrequent. Compared to the pivotal phase 3 trials, this cohort represented an older and more diverse population with more DMT experience, longer disease duration, and lower relapse rates at baseline, yet demonstrated robust safety outcomes.

Supplementary Material

Supplementary Table 1.docx

Supplementary Figure 1.docx

Supplementary Table 2 .docx

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Funding Statement

This work was supported by the Bristol Myers Squibb Foundation [IM047-1055].

Article highlights

Real-world data is important for understanding the use of newer multiple sclerosis treatments, such as the second-generation sphingosine-1 receptor modulator, ozanimod, in larger populations than are observed in clinical trials.

In this analysis, 206 persons with multiple sclerosis who were treated with ozanimod were retrospectively followed for one year to determine safety, effectiveness, and tolerability of ozanimod.

The annualized relapse rate on ozanimod was reduced by 56% compared to the year prior to starting ozanimod.

Radiographic disease activity was significantly reduced on ozanimod and no significant differences in disability metrics were observed after one year of treatment versus baseline.

Ozanimod was well tolerated, with 72% of individuals remaining on treatment at last data collection.

Worsening hypertension and upper respiratory infections were the most common adverse events.

CRediT authorship contributions

Jessica Cooperrider, MD: Methodology, Investigation, Data curation, Validation, Writing – original draft.

Jeffrey Lambe, MD: Investigation, Data curation, Writing – review and editing.

Michelle Chu, MD: Investigation, Data curation, Writing – review and editing.

Mengke Du, MSc: Methodology, Formal Analysis, Visualization, Writing – review and editing.

Jennifer Reardon, MSN: Conceptualization, Writing – review and editing.

Burhan Chaudry, MD: Conceptualization, Writing – review and editing.

Andrew Thorpe, PhD: Conceptualization, Writing – review and editing.

Carrie Hersh, DO, MS: Conceptualization, Supervision, Funding Acquisition, Writing – review and editing.

Devon Conway, MD, MS: Conceptualization, Supervision, Investigation, Data Curation, Funding Acquisition, Writing – review and editing.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/17582024.2026.2622078

References

Associated Data

Supplementary Materials

Supplementary Table 1.docx

Supplementary Figure 1.docx

Supplementary Table 2 .docx