1102584 research-article20222022 MSJ0010.1177/13524585221102584Multiple Sclerosis JournalBAC Cree, KW Selmaj MULTIPLE SCLEROSIS MSJ JOURNAL Original Research Paper Long-term safety and efficacy of ozanimod Multiple Sclerosis Journal 1­–19 in relapsing multiple sclerosis: Up to 5 years DOI: 10.1177/ https://doi.org/10.1177/13524585221102584 13524585221102584 https://doi.org/10.1177/13524585221102584 of follow-up in the DAYBREAK open-label © The Author(s), 2022. extension trial Article reuse guidelines: sagepub.com/journals- permissions Bruce AC Cree , Krzysztof W Selmaj, Lawrence Steinman, Giancarlo Comi, Amit Bar-Or, Douglas L Arnold , Hans-Peter Hartung, Xavier Montalbán, Eva K Havrdová, James K Sheffield, Neil Minton, Chun-Yen Cheng, Diego Silva, Ludwig Kappos and Jeffrey A Cohen Abstract Correspondence to: BAC Cree Background: Ozanimod, an oral sphingosine 1-phosphate receptor 1 and 5 modulator, is approved in Weill Institute for multiple countries for treatment of relapsing forms of MS. Neurosciences, Department of Neurology, University of Objective: To characterize long-term safety and efficacy of ozanimod. California San Francisco, 675 Nelson Rising Lane, San Methods: Patients with relapsing MS who completed a phase 1‒3 ozanimod trial were eligible for an Francisco, CA 94158, USA. open-label extension study (DAYBREAK) of ozanimod 0.92 mg/d. DAYBREAK began 16 October bruce.cree@ucsf.edu Bruce AC Cree 2015; cutoff for this interim analysis was 2 February 2021. Weill Institute for Results: This analysis included 2494 participants with mean 46.8 (SD 11.9; range 0.033‒62.7) months Neurosciences, Department of Neurology, University of of ozanimod exposure in DAYBREAK. During DAYBREAK, 2143 patients (85.9%) had treatment- California San Francisco, San emergent adverse events (TEAEs; similar in nature to those in the parent trials), 298 (11.9%) had a seri- Francisco, CA, USA Krzysztof W Selmaj ous TEAE, and 75 (3.0%) discontinued treatment due to TEAEs. Serious infections (2.8%), herpes zoster Center for Neurology, Łódź, infections (1.7%), confirmed macular edema cases (0.2%), and cardiac TEAEs (2.8%) were infrequent. Poland and Collegium Medicum, Department of Adjusted annualized relapse rate was 0.103 (95% confidence interval, 0.086‒0.123). Over 48 months, Neurology, University of Warmia and Mazury in 71% of patients remained relapse free. Adjusted mean numbers of new/enlarging T2 lesions/scan and Olsztyn, Olsztyn, Poland gadolinium-enhancing lesions were low and similar across parent trial treatment subgroups. Lawrence Steinman Conclusions: This long-term extension of ozanimod trials confirmed a favorable safety/tolerability pro- Department of Neurology and Neurological Sciences, file and sustained benefit on clinical and magnetic resonance imaging measures of disease activity. Beckman Center for Molecular Medicine, Stanford University Medical Center, Stanford, CA, USA Keywords: Multiple sclerosis, sphingosine 1-phosphate receptor modulators, adverse events, clinical Giancarlo Comi efficacy, ozanimod, extension study Vita-Salute San Raffaele University and Casa di Cura del Policlinico, Milan, Italy Date received: 23 December 2021; revised: 22 April 2022; accepted: 5 May 2022 Amit Bar-Or Center for Neuroinflammation and Introduction magnetic resonance imaging (MRI) and slowed brain Experimental Therapeutics, and Department of Ozanimod, a sphingosine 1-phosphate (S1P) receptor volume loss relative to intramuscular interferon β-1a Neurology, Perelman School 1 and 5 modulator, is approved in multiple countries 30 μg/wk.3,4 Ozanimod was well tolerated, with fewer of Medicine, University of for treatment of adults with either relapsing forms of treatment-emergent adverse events (TEAEs) leading Pennsylvania, Philadelphia, PA, USA multiple sclerosis (RMS) or moderately to severely to discontinuation than interferon β-1a.3,4 Patients Douglas L Arnold active ulcerative colitis. Four clinical trials of NeuroRx Research and ozanimod in RMS were completed, including a phase who completed any RMS trial were eligible to enroll Montréal Neurological 1 study (unpublished), a phase 2 study with an exten- in DAYBREAK, an open-label extension trial of Institute, McGill University, Montréal, QC, Canada sion period (RADIANCE),1,2 and two phase 3 trials ozanimod 0.92 mg/d (equivalent to ozanimod HCl (RADIANCE and SUNBEAM).3,4 In phase 3 trials, 1 mg). This is an interim analysis of safety (primary ozanimod 0.92 mg/d for up to 24 months significantly objective) and efficacy (secondary objective) in reduced clinical relapses and lesion counts on brain DAYBREAK. journals.sagepub.com/home/msj 1 Multiple Sclerosis Journal 00(0) Hans-Peter Hartung Department of Neurology, Methods TEAEs, including those of special interest (infections, Medical Faculty, Heinrich- malignancies, macular edema, cardiac events, pul- Heine University, Düsseldorf, Germany/Brain and Mind Parent trials monary and liver function test abnormalities), were Centre, The University of DAYBREAK participants originated from four monitored throughout the trial. Depression and sui- Sydney, Sydney, Australia/ Department of Neurology, “parent” ozanimod trials (Figure 1). The phase 1 study cide, lymphopenia, and post-treatment discontinua- Medical University of was a 12-week, open-label pharmacokinetic/pharmaco- tion relapse also are reported. Safety follow-up after Vienna, Vienna, Austria/ Department of Neurology, dynamic study of ozanimod 0.92 or 0.46 mg/d. Phase discontinuation was expanded from 28 to 75 days in Palacky University Olomouc, 2 RADIANCE was a randomized, double-blind, 2018 and to 90 days in 2019. External experts Olomouc, Czech Republic Xavier Montalbán 24-week study of ozanimod 0.92 or 0.46 mg/d versus reviewed hepatic dysfunction cases. An independent Department of Neurology- placebo,1 followed by a 24-month dose-blinded Macular Edema Review Panel (MERP) reviewed Neuroimmunology, Centre d’Esclerosi Múltiple de extension period in which patients initially assigned to TEAEs related to macular edema, cases of central Catalunya (Cemcat), Hospital ozanimod remained on their assigned dose while those foveal thickness greater than 20% of baseline, and Universitari Vall d’Hebron, Barcelona, Spain in the placebo group were randomized to ozanimod OCT abnormalities. Eva K Havrdová 0.92 or 0.46 mg/d.2 Two phase 3, randomized, double- Department of Neurology blind, double-dummy trials compared oral ozanimod COVID-19 cases were captured as TEAEs in the pri- and Center for Clinical Neuroscience, First Medical 0.92 or 0.46 mg/d with intramuscular interferon β-1a mary safety analysis using the 2 February 2021 data Faculty, Charles University, 30 µg/wk; phase 3 RADIANCE lasted 24 months,3 cutoff. In addition, we performed a separate post hoc Prague, Czech Republic whereas SUNBEAM continued until the last enrolled analysis to better analyze any potential effects of oza- James K Sheffield Neil Minton participant was treated for 12 months.4 Parent trial par- nimod on COVID-19 outcomes. Confirmed and sus- Chun-Yen Cheng ticipants were aged 18‒55 years with RMS, brain MRI pected COVID-19 cases occurring between 1 Diego Silva Bristol Myers Squibb, lesions consistent with multiple sclerosis (MS), and November 2019 and 10 May 2021 were identified Princeton, NJ, USA an Expanded Disability Status Scale (EDSS) score based on adverse event preferred terms related to cor- Ludwig Kappos of 0‒5.0 (phases 2 and 3) or 0‒6.0 (phase 1).1,3,4 onavirus infection or coronavirus test positive. Cases Research Center for Clinical Neuroimmunology were assessed for seriousness, action taken with oza- and Neuroscience Basel nimod, and infection outcome. (RC2NB), Departments of Head, Spine and DAYBREAK extension Neuromedicine, Clinical Patients who completed a parent trial were eligible to Efficacy outcomes included annualized relapse rate Research, Biomedicine, and Biomedical Engineering, enter the single-arm, open-label, phase 3 DAYBREAK (ARR), time to first confirmed relapse, relapse-free University Hospital and trial (ClinicalTrials.gov ID: NCT02576717; EudraCT: rate, number of new/enlarging T2 lesions per scan University of Basel, Basel, Switzerland 2015-002500-91). All DAYBREAK participants re- relative to DAYBREAK baseline and number of gad- Jeffrey A Cohen consented. DAYBREAK began 16 October 2015, and olinium-enhancing (GdE) lesions on brain MRI, and Mellen Center for MS Treatment and Research, is being conducted in 25 countries in Europe and in time to onset of 3- and 6-month confirmed disability Cleveland Clinic, Cleveland, North America, plus South Africa and New Zealand. progression (CDP-3, CDP-6) (⩾1-point increase in OH, USA Data cutoff for the primary safety and efficacy analy- EDSS from DAYBREAK baseline). ses was 2 February 2021. Patients from the phase 3 trials underwent a 1-week Statistical analyses dose escalation upon entry into DAYBREAK. Dose Patients who received ⩾1 dose of ozanimod in escalation was not performed for patients from the DAYBREAK composed the safety and intent-to-treat other trials unless there was a ⩾14-day gap in populations. Where provided, parent trial data are treatment. reported only for patients who entered DAYBREAK. Results are reported for the overall DAYBREAK pop- Physical and skin examinations, electrocardiogram, ulation, and parent trial treatment assignment sub- pulmonary function tests (PFT), optical coherence groups (Supplemental Figure S1). MRI lesion counts tomography (OCT), and brain MRI were performed were analyzed in patients from the phase 3 parent tri- at baseline, every 12 months, and at termination. als. Unless otherwise specified, TEAEs and ARR are Laboratory testing was performed and vital signs reported from DAYBREAK baseline through data were measured at baseline, every 3 months for 3 years cutoff; other efficacy outcomes and laboratory abnor- and every 6 months thereafter, at termination, and malities are reported from DAYBREAK baseline during post-treatment follow-up. EDSS and neuro- through month 48 because data thereafter are limited logic examination were performed at baseline, every to a small number of patients. Because malignancies 3 months for 3 years and then every 6 months, upon are infrequent and have a variable latency period, all suspected relapse, and at study termination. Unless malignancies from date of first ozanimod exposure in otherwise specified, “baseline” refers to day 1 of any trial to DAYBREAK data cutoff are reported in DAYBREAK. patients exposed to either ozanimod dose; the subset 2 journals.sagepub.com/home/msj Parent Study Parent Study Treatment Perioda DAYBREAK Baseline/ Randomization 12 weeks, Study Baseline open-label Ozanimod n=11 n=10 journals.sagepub.com/home/msj Phase 1 0.92 mg/d RPC01-1001 Ozanimod n=13 n=10 (NCT02797015) 0.46 mg/d 24 weeks, 24 months, dose-blind extension double-blind Placebo n=88 n=109 Ozanimod 0.92 mg/d n=123 Phase 2 RADIANCE Ozanimod n=83 n=109 0.92 mg/d Ozanimod 0.46 mg/d n=126 (NCT01628393) Ozanimod n=87 0.46 mg/d Ongoing DAYBREAK OLE Study (NCT02576717) 24 months, double-blind Ozanimod 0.92 mg/da,b n=2494 n=344 IM IFN β-1a 30 µg/wk n=440 Phase 3 n=364 RADIANCE Ozanimod 0.92 mg/d n=434 (NCT02047734) n=345 Ozanimod 0.46 mg/d n=439 ≥12 months, double-blind n=392 IM IFN β-1a 30 µg/wk n=445 Phase 3 n=398 SUNBEAM Ozanimod 0.92 mg/d n=448 (NCT02294058) n=413 Ozanimod 0.46 mg/d n=453 Figure 1. Parent trials feeding into the DAYBREAK extension. Source: Adapted (colors revised, placebo arms pooled, and size of boxes adjusted to reflect study duration) from Selmaj et al.5 under Creative Commons Attribution 4.0 International License. HCl: hydrochloride; IFN: interferon; IM: intramuscular; OLE: open-label extension; PO: per os (oral). aIn all trials, upon initiation of ozanimod, patients received 0.23 mg PO (equivalent to ozanimod HCl 0.25 mg) on days 1–4, 0.46 mg PO (equivalent to ozanimod HCl 0.5 mg) on days 5–7, and then their assigned dose of 0.46 mg or 0.92 mg PO (equivalent to ozanimod HCl 1 mg) on day 8 and thereafter. All patients entering the phase 2 dose-blinded extension period underwent dose escalation, even if treated with ozanimod in the parent trial, to maintain the blind. bIn DAYBREAK, dose escalation was performed for all patients entering from one of the active-controlled phase 3 trials, irrespective of prior treatment assignment (to maintain the blinding in the parent trials); dose escalation was not performed for those entering from the phase 1 or 2 trials, unless the last dose of ozanimod was greater than 14 days before entering DAYBREAK. 3 BAC Cree, KW Selmaj et al. Multiple Sclerosis Journal 00(0) of malignancies occurring in DAYBREAK are also DAYBREAK population is about 2 years older on reported. average, as expected given the duration of the parent trials (Supplemental Table S2). In addition, mean TEAE incidence was summarized by Medical number of GdE lesions was lower and mean number Dictionary for Regulatory Activities version 22.1 of T2 lesions was higher at DAYBREAK baseline preferred term. Incidence rates (IRs) per 1000 person- than at parent trial baseline, which are expected find- years (PY) (100,000 PY for malignancies) were cal- ings given the treatment history and nature of the culated to adjust for time on study. lesions. ARR was calculated using a negative binomial regres- sion model adjusted for region (Eastern Europe ver- Safety sus rest of world), age at parent trial baseline, and During DAYBREAK, 85.9% of patients experi- parent trial baseline number of GdE lesions. The natu- enced ⩾1 TEAE. Most were mild or moderate; rates ral log transformation of time on treatment was used of severe TEAEs, serious TEAEs, and TEAEs lead- as an offset term to adjust for patients having different ing to permanent treatment discontinuation were exposure times. Time to first confirmed relapse and 7.6%, 11.9%, and 3.0%, respectively, and similar relapse-free rates were analyzed using Kaplan–Meier across parent trial treatment subgroups (Table 1). The analysis, with censoring of patients for whom follow- most common TEAEs were nasopharyngitis, head- up ended before a relapse occurred. ache, and upper respiratory tract infection (Table 1). Prior to the 2 February 2021 data cutoff, three patients The adjusted mean (95% confidence interval (CI)) died from malignancies, two from accidents, and one number of new/enlarging T2 brain MRI lesions per each from community-acquired pneumonia, cerebral scan relative to baseline (parent baseline for parent hemorrhage probably due to rupture of a preexist- trials, DAYBREAK baseline for DAYBREAK) was ing occult aneurysm, sudden death due to unknown derived based on a negative binomial regression cause, and pulmonary embolism following leg frac- model, and adjusted for study, region, age at baseline, ture surgery. There were three deaths associated with and baseline number of GdE lesions, with the log of COVID-19 (Supplemental Figure S2). the number of postbaseline scans as an offset term. The adjusted mean (95% CI) number of GdE lesions Infections―most commonly affecting the respiratory was derived using a negative binomial regression or urinary tract―occurred in 56.7% of DAYBREAK model, adjusted for study, region, age at baseline, and participants. The incidence of serious infection was baseline number of GdE lesions. Time to onset of 2.8% (Table 2). Opportunistic infections occurred in CDP-3 and CDP-6 was analyzed using Kaplan–Meier 5.6%; the most common were oral herpes (2.0%; IR analysis. 5.4/1000 PY) and herpes zoster (1.7%; IR 4.6/1000 PY) (Table 2). There were no serious opportunistic infections before data cutoff; however, a case of pro- Results gressive multifocal leukoencephalopathy (PML) was reported in March 2021 in a 46-year-old woman who Disposition and demographics had received ozanimod 0.92 mg for about 4 years.6 Of 2639 eligible patients, 2494 (94.5%) enrolled and She discontinued ozanimod and had a nonfatal out- were treated in DAYBREAK; 2055 (82.4%) were come with neurologic sequelae. Details regarding this continuing at data cutoff (Figure 2). Mean duration of case are provided in Supplemental Materials. ozanimod exposure in DAYBREAK was 46.8 months (standard deviation (SD) 11.9; range 0.033‒62.7; Between 1 November 2019 and 10 May 2021, 190 9725.6 PY). Mean ozanimod exposure during parent (8.7%) of 2181 DAYBREAK participants experi- trials plus DAYBREAK was 60.7 months (SD 16.5; enced confirmed (n = 160; 7.3%) or suspected 12,617.1 PY); maximum was 98.8 months. (n = 30; 1.4%) COVID-19 while receiving ozani- mod. Their demographic characteristics are shown Baseline demographics and disease characteristics in Supplemental Table S3. Most cases (n = 176/190 were generally consistent across parent trial treatment (92.6%)) were nonserious, and most patients (179/190 subgroups (Supplemental Table S1). Baseline charac- (94.2%)) recovered without sequelae while either teristics of participants in DAYBREAK and those of continuing or temporarily interrupting ozanimod participants in the phase 2 and phase 3 parent trials (Supplemental Figure S2). One patient died from pul- (who compose over 99% of the DAYBREAK popu- monary embolism associated with COVID-19; one lation) are similar, with the exception that the died from suspected COVID-19 pneumonia; and one 4 journals.sagepub.com/home/msj BAC Cree, KW Selmaj et al. Patients with RMS who completed a parent study (see Fig. 1) and were eligible to enroll in DAYBREAK N=2639a Did not enroll N=144 (5.5%) Enrolled from Phase 1 Enrolled from Phase 2 Enrolled from Phase 3 Enrolled from Phase 3 RPC01-1001 RADIANCE RADIANCE SUNBEAM N=20 N=218 N=1054b N=1203 Enrolled in DAYBREAK and received open-label ozanimod 0.92 mg (Safety and ITT population) N=2494b Discontinued N=439 (17.6%) • Voluntarily withdrew: 219 (8.8%) • Adverse event: 67 (2.7%) • Lack of efficacy: 60 (2.4%) • Lost to follow-up: 24 (1.0%) • Physician decision: 15 (0.6%) • Died: 6 (0.2%) • Sponsor decision: 3 (0.1%) • Protocol violation: 3 (0.1%) • Other: 42 (1.7%) Participants ongoing in DAYBREAK at data cutoff N=2055 (82.4%) Figure 2. Flow diagram for DAYBREAK extension trial. ITT: intent-to-treat; RMS: relapsing multiple sclerosis. aSee Figure 1 for n values by treatment group in each of the parent trials. bOne participant entering DAYBREAK from phase 3 RADIANCE did not receive ozanimod 0.92 mg. died from a lung abscess following COVID-19 from interferon β-1a to ozanimod. Based on review of (Supplemental Figure S2). TEAEs and OCTs, the MERP confirmed four macular edema cases (0.2%; IR 0.4/1000 PY (95% CI 0.2‒1.1)) During the parent trials and DAYBREAK, 38 (1.4%) during DAYBREAK. Three confirmed cases occurred of 2787 patients exposed to either ozanimod dose– within a year of switching from interferon β-1a to developed malignancies (Table 3). The most frequent ozanimod in patients with preexisting risk factors was basal cell carcinoma (n = 11 (0.4%)). Nine (0.5%) or potentially predisposing conditions, as previously of 1868 women developed breast cancer. Three reported.5 The fourth occurred in a patient who malignancy-related deaths (from metastatic pancre- received ozanimod 0.92 mg in phase 3 RADIANCE atic carcinoma, disseminated cancer with unknown and DAYBREAK for approximately 6 years total with primary, and glioblastoma) occurred in DAYBREAK; no apparent risk factors for macular edema. Three there were no malignancy-related deaths in the parent cases resolved after ozanimod discontinuation; one trials. The subset of malignancies that occurred dur- was ongoing at data cutoff. ing DAYBREAK is also shown in Table 3. Cardiac TEAEs occurred in 69 (2.8%) patients There were nine (0.4%) reports of macular edema (Supplemental Table S4), seven of whom had serious during DAYBREAK―five in patients who switched events (three myocardial infarctions (IR 0.3/1000 PY, journals.sagepub.com/home/msj 5 Multiple Sclerosis Journal 00(0) Table 1. Overall safety during DAYBREAK by parent trial treatment subgroup and overall DAYBREAK population. IFN β-1a 30 μg Ozanimod Continuous All DAYBREAK (Ozanimod 0.92 mg) → Ozanimod 0.46 mg → Ozanimod Ozanimod (N = 2494) 0.92 mg 0.92 mg 0.92 mg (N = 736) (N = 877) (N = 881) n (%) IR/1000 PYa n (%) n (%) n (%) (95% CI) Any TEAE 647 (87.9) 750 (85.5) 746 (84.7) 2143 (85.9) 693.4 (664.6‒723.4) Severe TEAEs 68 (9.2) 69 (7.9) 53 (6.0) 190 (7.6) 20.7 (18.0‒23.9) Serious TEAEs 84 (11.4) 111 (12.7) 103 (11.7) 298 (11.9) 33.3 (29.8‒37.4) TEAEs leading to permanent 27 (3.7) 24 (2.7) 24 (2.7) 75 (3.0) 7.9 (6.3‒9.9) treatment discontinuation Individual TEAEs in ⩾5% of all DAYBREAK population Nasopharyngitis 150 (20.4) 169 (19.3) 170 (19.3) 489 (19.6) 59.3 (54.2‒64.8) Headache 124 (16.8) 132 (15.1) 137 (15.6) 393 (15.8) 46.1 (41.8‒50.9)  Upper respiratory tract 88 (12.0) 94 (10.7) 96 (10.9) 278 (11.1) 31.5 (28.0‒35.5) infection Lymphopeniab 84 (11.4) 96 (10.9) 77 (8.7) 257 (10.3) 29.4 (26.0‒33.3) ALC decreasedb 62 (8.4) 78 (8.9) 78 (8.9) 218 (8.7) 24.5 (21.5‒28.0) Back pain 60 (8.2) 74 (8.4) 71 (8.1) 205 (8.2) 22.7 (19.8‒26.0) Hypertensionc 65 (8.8) 74 (8.4) 48 (5.4) 187 (7.5) 20.7 (17.9‒23.9) GGT increased 65 (8.8) 54 (6.2) 52 (5.9) 171 (6.9) 18.9 (16.2‒21.9) Bronchitis 35 (4.8) 59 (6.7) 51 (5.8) 145 (5.8) 15.8 (13.4‒18.6) Respiratory tract infection 44 (6.0) 53 (6.0) 48 (5.4) 145 (5.8) 15.9 (13.5‒18.7) Urinary tract infection 38 (5.2) 54 (6.2) 51 (5.8) 143 (5.7) 15.6 (13.3‒18.4)  Respiratory tract infection 39 (5.3) 49 (5.6) 44 (5.0) 132 (5.3) 14.4 (12.1‒17.0) viral Depression-related TEAEsd 38 (5.2) 48 (5.5) 43 (4.9) 129 (5.2) 14.0 (11.8‒16.6) ALC: absolute lymphocyte count; CI: confidence interval; GGT: gamma-glutamyl transferase; IFN: interferon; IR: incidence rate; PY: person-years; TEAE: treatment-emergent adverse event. aIR per 1000 person-years is calculated as number of persons/person-years × 1000 for specific system organ class category or preferred term subcategory. Person-years for each category/subcategory: for a person in a particular category/subcategory, the time on study is calculated based on the date the person first has a TEAE within the category/subcategory (date of first TEAE − first dose date of study drug + 1)/365.25; for persons who do not have a TEAE in the category/subcategory, the time on study is the study duration (last date on study − first dose date of study drug + 1)/365.25. bALC reductions are an expected pharmacodynamic effect related to the mechanism of ozanimod; although investigators were not required to report ALC reductions as TEAE, preferred terms of lymphopenia and ALC decreases may have been reported as TEAEs according to investigator determination. cIncludes preferred terms of hypertension, essential hypertension, labile hypertension, systolic hypertension, and hypertensive crisis. dIncludes preferred terms of depression, depressed mood, and depressive symptoms. 95% CI 0.1‒1.0), and one each of myocardial ischemia, 5.1 (11.8) and 1.7 (9.8) mm Hg, respectively, in the unstable angina, bradycardia, and coronary artery ste- overall population. nosis). Six of seven patients who developed bradycar- dia had switched from interferon β-1a or ozanimod Forced expiratory volume in one second (FEV1) and 0.46 mg to ozanimod 0.92 mg in DAYBREAK. forced vital capacity (FVC) were less than 80% of Vascular TEAEs occurred in 236 (9.5%) participants baseline at any postbaseline visit in 9.2% and 7.6% of (Supplemental Table S4). Hypertension was the most patients, respectively. PFT abnormalities did not common (n = 187 (7.5%)) and included two serious increase over time, and there were no differences by cases of hypertension (0.08%) and four (0.2%) non- parent trial treatment subgroup. One patient each dis- serious cases reported as “hypertensive crisis” (with- continued treatment due to pulmonary sarcoidosis, out sequelae). During DAYBREAK, 66/2490 (2.7%) dyspnea, decreased FEV1, and decreased FVC. patients had a recorded systolic blood pressure greater than 180 mm Hg or an increase of greater Seventy-nine (3.2%) and 34 (1.4%) patients had a than 40 mm Hg from DAYBREAK baseline. From maximal alanine aminotransferase (ALT) or aspartate DAYBREAK baseline to month 48, mean (SD) aminotransferase, respectively, ⩾3 times the upper change in systolic and diastolic blood pressure was limit of normal (ULN); 70 (2.8%) had bilirubin 6 journals.sagepub.com/home/msj BAC Cree, KW Selmaj et al. Table 2. Serious and opportunistic infections during DAYBREAK by parent trial treatment subgroup and overall DAYBREAK population. IFN β-1a 30 μg Ozanimod Continuous All DAYBREAK → Ozanimod 0.46 mg → Ozanimod Ozanimod 0.92 mg (Ozanimod 0.92 mg) 0.92 mg 0.92 mg (N = 881) (N = 2494) (N = 736) (N = 877) n (%) n (%) n (%) n (%) IR/1000 PYa (95% CI) Any serious infections 19 (2.6) 25 (2.9) 25 (2.8) 69 (2.8) 7.4 (5.8‒9.3) Serious infections in ⩾2 patients in the all DAYBREAK populationb   Pneumoniac 2 (0.3) 1 (0.1) 5 (0.6) 8 (0.3) 0.8 (0.4‒1.7)   Appendicitis 3 (0.4) 3 (0.3) 1 (0.1) 7 (0.3) 0.7 (0.4‒1.6)   Coronavirus infectiond 0 (0.0) 4 (0.5) 3 (0.3) 7 (0.3) 0.7 (0.4‒1.5)   Pyelonephritis acute 4 (0.5) 1 (0.1) 1 (0.1) 6 (0.2) 0.6 (0.3‒1.4)   Pneumonia viral d 1 (0.1) 2 (0.2) 2 (0.2) 5 (0.2) 0.5 (0.2‒1.3)   Bronchitis 1 (0.1) 1 (0.1) 0 (0.0) 2 (0.1) 0.2 (0.1‒0.8)   Dacryocystitis 0 (0.0) 2 (0.2) 0 (0.0) 2 (0.1) 0.2 (0.1‒0.8)   Endometritis 0 (0.0) 0 (0.0) 2 (0.2) 2 (0.1) 0.2 (0.1‒0.8)   Escherichia UTI 1 (0.1) 1 (0.1) 0 (0.0) 2 (0.1) 0.2 (0.1‒0.8)   Lyme disease 0 (0.0) 0 (0.0) 2 (0.2) 2 (0.1) 0.2 (0.1‒0.8)   Measles 0 (0.0) 1 (0.1) 1 (0.1) 2 (0.1) 0.2 (0.1‒0.8)   Pyelonephritis 1 (0.1) 0 (0.0) 1 (0.1) 2 (0.1) 0.2 (0.1‒0.8) Any OI e 41 (5.6) 59 (6.7) 40 (4.5) 140 (5.6) 15.2 (12.9‒18.0) OIs in ⩾0.5% of patients in the all DAYBREAK population   Oral herpes 15 (2.0) 23 (2.6) 13 (1.5) 51 (2.0) 5.4 (4.1‒7.2)   Herpes zosterf 11 (1.5) 20 (2.3) 12 (1.4) 43 (1.7) 4.6 (3.4‒6.2)   Herpes simplex 6 (0.8) 4 (0.5) 3 (0.3) 13 (0.5) 1.4 (0.8–2.4) CI: confidence interval; IFN: interferon; IR: incidence rate; MedDRA: Medical Dictionary for Regulatory Activities; ITT, intent to treat; TEAE, treatment- emergent adverse event; SDEI, sponsor-designated event of interest; ; OI: opportunistic infection; PY: person-years; UTI: urinary tract infection. aIR per 1000 person-years is calculated as number of persons/person-years × 1000 for specific system organ class category or preferred term subcategory. Person-years for each category/subcategory: for a person in a particular category/subcategory, the time on study is calculated based on the date the person first has a TEAE within the category/subcategory (date of first TEAE − first dose date of study drug + 1)/365.25; for persons who do not have a TEAE in the category/subcategory, the time on study is the study duration (last date on study − first dose date of study drug + 1)/365.25. bAdditional serious infections that occurred in a single participant each (0.04%) in the all DAYBREAK population included abdominal wall infection, limb abscess, acute sinusitis, arthritis infective, chronic hepatitis B, chronic sinusitis, diverticulitis, echinococciasis, epididymitis, gastroenteritis, gastroenteritis Staphylococcal, gastroenteritis viral, HIV infection, hepatitis A, injection site abscess, acute otitis media, peritonitis, peritonsillar abscess, chronic pyelonephritis, salpingitis, salpingo-oophoritis, sepsis, septic shock, tonsillitis, toxic shock syndrome, upper respiratory tract infection, urinary tract infection, and vestibular neuronitis. cIn addition to the deaths related to COVID-19 discussed in “Results” section, there was one other infection-related death in a bed-ridden 46-year-old woman with greater than 5 years of ozanimod 0.92 mg exposure. On DAYBREAK day 977, the participant was hospitalized for community-acquired pneumonia; ozanimod was discontinued; death from respiratory failure occurred 11 days later. dThe seven TEAEs reported using the MedDRA preferred term “coronavirus” and 4 of the 5 reported as “pneumonia viral” are COVID-19 cases reported prior to 2 February 2021 ITT data cutoff and represent a subset of the cases reported in the extended COVID-19 analysis reported in “Results” section and Supplemental Materials. The reported IRs for these preferred terms likely underestimate the true risk because they are based on all DAYBREAK participants in the ITT population from study inception through data cutoff, some of whom discontinued the study prior to the start of the COVID-19 pandemic. eTEAEs were coded using MedDRA version 22.1; the SDEI search strategy for opportunistic infections was based on MedDRA version 18.1. fVaricella zoster virus infection. > 2 × ULN (Table 4). Eighteen patients (0.7% of the to liver enzyme elevations. No cases of severe drug- overall population) had ALT ⩾ 3 × ULN on consecu- induced liver injury occurred. tive postbaseline assessments (Table 4), suggesting that most elevations were transient. First occurrence One hundred twenty-nine patients (5.2%) had a of ALT elevations > 1 × ULN while patients were depression-related TEAE (Table 1). Five had suicidal enrolled in DAYBREAK primarily occurred during ideation (IR 0.5/1000 PY (95% CI 0.2‒1.3)). Two year 1 (Figure 3). Hepatobiliary TEAEs were observed attempted suicide (IR 0.2/1000 PY (95% CI 0.1‒0.8)). in 85 (3.4%) patients (Supplemental Table S5). There was one serious hepatic TEAE (chronic hepatitis). Patients who received ozanimod 0.92 mg during the One patient discontinued ozanimod due to acute hep- parent trials exhibited stable absolute lymphocyte atitis (nonviral, nonautoimmune) and nine (0.4%) due counts (ALCs) (mean of 0.7‒0.8×109/L at each time journals.sagepub.com/home/msj 7 Multiple Sclerosis Journal 00(0) Table 3. Treatment-emergent malignancies during the parent trials or DAYBREAK (all patients who were exposed to ozanimod 0.46 or 0.92 mg) and treatment-emergent malignancies occurring in DAYBREAK only, by preferred term, in patients with RMS. All ozanimod-exposed patients DAYBREAK (ozanimod 0.92 mg) (ozanimod 0.46 or 0.92 mg) (N = 2494) (N = 2787) n (%) IR/100,000 PYa n (%) IR/100,000 PYa (95% CI) (95% CI) Any treatment-emergent 38 (1.4) 358.6 (253.7‒492.2) 29 (1.2) 306.6 (205.3‒440.4) malignancy Cutaneous 14 (0.5) 132.0 (72.1‒221.4) 9 (0.4) 95.0 (43.5‒180.4)    Nonmelanoma skin cancer 13 (0.5) 122.5 (65.2‒209.5) 8 (0.3) 84.5 (36.5‒166.4)    Basal cell carcinoma 11 (0.4) 103.6 (51.7‒185.5) 7 (0.3) 73.9 (29.7‒152.2)    Bowen’s disease 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7)    Keratoacanthoma 1 (0.04) 9.4 (0.2‒52.4) 0 (0.0) –    Squamous cell carcinoma 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7)   Malignant melanoma 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7) Noncutaneous 24 (0.9) 225.9 (144.7‒336.1) 20 (0.8) 211.0 (128.9‒325.9)    Breast cancer (women only)b,c 9/1868 (0.5) 126.8 (58.0‒240.8) 6/1668 (0.4) 94.9 (34.8‒206.6)    Papillary thyroid cancer 3 (0.1) 28.2 (5.8‒82.4) 3 (0.1) 31.6 (6.5‒92.4)    Seminomas (men only)c,d 2/919 (0.2) 56.6 (6.8‒204.3) 1/826 (0.1) 31.6 (0.8‒175.9)   Cervical cancer (women 2/1868 (0.1) 28.2 (3.4‒101.8) 2/1668 (0.1) 31.6 (3.8‒114.3) only)c,e    Chronic myeloid leukemia 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7)   Clear cell renal cell 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7) carcinoma   Endometrial adenocarcinomac 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7)    Gastrointestinal stromal 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7) tumor   Glioblastoma 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7)   Myxoid liposarcoma 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7)   Neoplasm malignant 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7)    Pancreatic carcinoma 1 (0.04) 9.4 (0.2‒52.4) 1 (0.04) 10.5 (0.3‒58.7) metastatic RMS: relapsing multiple sclerosis; IR: incidence rate; PY: person-years; CI: confidence interval. aIR/100,000 PY, study duration‒adjusted incidence rate per 100,000 person-years, calculated as number of persons having the malignancy of interest/person-years × 100,000, where person-years = (date first malignancy of interest was documented − date of first dose of study drug + 1)/365.25; for patients not having the malignancy of interest, the time on study is the study duration (last date on study − first dose date of study drug + 1)/365.25. bIncludes preferred terms: breast cancer, invasive breast carcinoma, and invasive ductal breast carcinoma. cGender-specific IRs were calculated only for gender-specific malignancies with >1 event. dIncludes preferred terms: seminoma and testicular seminoma (pure) stage I. eIncludes preferred terms: cervix carcinoma and squamous cell carcinoma of the cervix. point) throughout 48 months of DAYBREAK (Figure or 1.2% of the overall safety population (30/2488)) 4). After patients switched from interferon β-1a, mean had a consecutive ALC < 0.2×109/L on retest, sug- ALC showed an expected decline over the first gesting most instances were transient. Seven (0.3%) 3 months of ozanimod in DAYBREAK, to 43.9% of patients permanently discontinued treatment due to baseline; this reduction was largely sustained through TEAEs reported as lymphopenia (n = 3) or lympho- month 48 (Figure 4). During DAYBREAK, ALC was cyte count decreased (n = 4). less than 0.2 × 109/L in 244/2488 (9.8%) patients (Table 5). Persistent reductions were uncommon: 187 Of 439 (17.6%) patients who discontinued ozanimod in of the patients with ALC < 0.2×109/L had a repeat DAYBREAK for any reason, 10 (2.3%) subsequently ALC assessment within 30 days, of whom 30 (16.0%, experienced confirmed relapses, which all occurred 8 journals.sagepub.com/home/msj BAC Cree, KW Selmaj et al. Table 4. Hepatic laboratory abnormalities during DAYBREAK by parent trial treatment subgroup and overall DAYBREAK population. Based on laboratory testing IFN β-1a 30 μg Ozanimod Continuous All DAYBREAK → Ozanimod 0.46 mg → Ozanimod Ozanimod 0.92 mg (Ozanimod 0.92 mg) 0.92 mg 0.92 mg (N = 881) (N = 2494) (N = 736) (N = 877) n = 734 n = 875 n = 880 n = 2489 Maximum ALT, n (%) >1 × ULN 285 (38.8) 321 (36.7) 322 (36.6) 928 (37.3) ⩾3 × ULN 31 (4.2) 28 (3.2) 20 (2.3) 79 (3.2) ⩾5 × ULN 7 (1.0) 11 (1.3) 0 (0.0) 18 (0.7) ⩾10 × ULN 3 (0.4) 5 (0.6) 0 (0.0) 8 (0.3) Mean (SD) maximal change from 30.6 (72.27) 26.6 (73.72) 17.9 (23.03) 24.7 (60.53) baseline in ALT, U/L ALT ⩾ 3 × ULN on consecutive 7 (1.0) 7 (0.8) 4 (0.5) 18 (0.7) postbaseline assessments, n (%) ALT ⩾ 5 × ULN on consecutive 2 (0.3) 4 (0.5) 0 (0.0) 6 (0.2) postbaseline assessments, n (%) Maximum AST, n (%) >1 × ULN 163 (22.2) 162 (18.5) 166 (18.9) 491 (19.7) ⩾3 × ULN 9 (1.2) 20 (2.3) 5 (0.6) 34 (1.4) ⩾5 × ULN 6 (0.8) 7 (0.8) 1 (0.1) 14 (0.6) ⩾10 × ULN 2 (0.3) 2 (0.2) 0 (0.0) 4 (0.2) Mean (SD) maximal change from 15.4 (56.35) 13.4 (44.47) 9.1 (13.88) 12.5 (41.29) baseline in AST, U/L Maximum GGT, n (%) >1 × ULN 373 (50.8) 452 (51.7) 460 (52.3) 1285 (51.6) >2.5 × ULN 136 (18.5) 151 (17.3) 135 (15.3) 422 (17.0) >5 × ULN 39 (5.3) 38 (4.3) 45 (5.1) 122 (4.9) >20 × ULN 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Mean (SD) maximal change from 51.0 (68.68) 38.7 (52.03) 32.0 (50.41) 40.0 (57.42) baseline in GGT, U/L Maximum bilirubin, n (%) >1 × ULN 166 (22.6) 190 (21.7) 189 (21.5) 545 (21.9) >2 × ULN 23 (3.1) 22 (2.5) 25 (2.8) 70 (2.8) >3 × ULN 2 (0.3) 3 (0.3) 4 (0.5) 9 (0.4) >10 × ULN 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Mean (SD) maximal change from 6.4 (5.69) 5.2 (4.76) 5.4 (5.38) 5.6 (5.29) baseline in bilirubin, µmol/L Severe drug-induced liver injury (Hy’s 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Law cases),a n (%) IFN: interferon; ALT: alanine aminotransferase; ULN: upper limit of normal; SD: standard deviation; AST: aspartate aminotransferase; GGT: gamma-glutamyl transferase. aHy’s Law, defined as ALT or AST ⩾ 3 × ULN plus total bilirubin > 2 × ULN without cholestasis and without alternative explanation, is used by the US Food and Drug Administration to identify drugs likely to cause severe drug-induced liver injury.7 >1 month (34‒141 days) after ozanimod discontinu- consider any of the relapses to be severe, and no relapses ation in the absence of any disease-modifying ther- resulted in a persistent severe increase in disability. apy (DMT). These relapses were associated with EDSS increases of 0.0‒2.5 points compared with the previ- ous visit. Seven patients, including the patient with Efficacy the 2.5-point EDSS increase, had complete recovery The adjusted ARR during DAYBREAK among all and three had partial recovery. Investigators did not participants was 0.103 (95% CI 0.086‒0.123). ARR journals.sagepub.com/home/msj 9 10 10 >1 to <3 x ULN Multiple Sclerosis Journal 00(0) ≥3 to <5 x ULN 9 ≥5 to <10 x ULN ≥10 x ULN 8 7.2 7 7 6 5 4.1 4 3.1 3 specified ALT elevation 2.3 2.1 1.9 Percentage of patients with 2 1.5 1.5 1.6 1.5 0.9 1 1 0.9 1 0.8 1 0.4 <0.1 0.2 <0.1 <0.1 0.2 <0.1 0.2 0.3 <0.1 0.2 <0.1 0.2 0.2 <0.1 0 0 0 0 0 0 0 0 <0.1 0.1 0.2 <0.1 0.2 <0.1 <0.1 0 <0.1 0 <0.10 0 0 0 0.1 0 0 0 0 0 0 0.1 0 0 0 <0.1 0 0 n= 2494 2487 2465 2432 2405 1896 2003 2199 2293 2254 2236 2217 2178 2089 1876 1904 1759 Baseline 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 Months in DAYBREAK trial Figure 3. First occurrence of ALT elevations during DAYBREAK as a percentage of patients (ITT population).a ALT: alanine aminotransferase; ITT: intent to treat; ULN: upper limit of normal. aNote that >70% of subjects had ozanimod exposure prior to DAYBREAK entry. journals.sagepub.com/home/msj journals.sagepub.com/home/msj 2 IFN -1a 30 g→ Ozanimod 0.92 mg Ozanimod 0.46 mg→ Ozanimod 0.92 mg 1.8 Continuous Ozanimod 0.92 mg 1.6 All DAYBREAK (Ozanimod 0.92 mg) 1.4 1.2 1 0.8 0.6 Mean (SE) ALC x109/L 0.4 0.2 0 BL 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 Months in DAYBREAK trial Number of patients IFN -1a 30 g → 736 710 722 714 696 599 634 660 669 655 655 644 622 597 540 540 508 ozanimod 0.92 mg Ozanimod 0.46 mg → 877 859 860 851 840 644 681 757 795 782 778 776 763 734 664 673 619 ozanimod 0.92 mg Continuous ozanimod 881 856 872 853 840 623 682 768 816 801 799 794 780 743 671 686 622 0.92 mg Total DAYBREAK 2494 2425 2454 2418 2376 1866 1997 2185 2280 2238 2232 2214 2165 2074 1875 1899 1749 Figure 4. Mean ALC values during DAYBREAK in the overall population and by parent trial treatment subgroup. BL: baseline; ALC: absolute lymphocyte count; IFN: interferon; SE: standard error. 11 BAC Cree, KW Selmaj et al. Multiple Sclerosis Journal 00(0) Table 5. Absolute lymphocyte count based on minimal value during DAYBREAK by parent trial treatment subgroup and overall DAYBREAK population. IFN β-1a 30 μg Ozanimod 0.46 mg Continuous All DAYBREAK → Ozanimod 0.92 mg → Ozanimod 0.92 mg Ozanimod 0.92 mg (Ozanimod 0.92 mg) (N = 736) (N = 877) (N = 881) (N = 2494) Number of patients n = 734 n = 874 n = 880 n = 2488 with an assessment Patients with minimum ALC, a n (%) 50% had assessments; n = 493). either interferon β-1a or ozanimod 0.46 mg experi- enced reductions in ARR and brain MRI lesion counts. Adjusted mean number of new/enlarging T2 lesions per scan (Figure 6(a)) and GdE lesions (Figure 6(b)) Some DMTs with immunomodulatory effects, includ- at month 48 of DAYBREAK—when all patients had ing S1P modulators, are associated with an increased received ⩾4 years of ozanimod 0.92 mg—was low incidence of herpes infection and reactivation.8–10 and similar across parent trial treatment subgroups. Herpes infections were the most common opportunis- Patients initially treated with interferon β-1a experi- tic infection; none were serious or disseminated. The enced a decline in lesion counts upon switching to IRs for oral herpes (5.4/1000 PY) and herpes zoster ozanimod in DAYBREAK. Among parent trial treat- (4.6/1000 PY) in DAYBREAK were similar to values ment subgroups, 35.0% of the interferon β-1a group reported in an analysis of the overall ozanimod MS and 37.7% of both the ozanimod 0.46 and 0.92 mg clinical development program that used an earlier groups had no new/enlarging T2 lesions from DAYBREAK data cutoff (oral herpes: 5.7/1000 PY; DAYBREAK baseline through month 48; 88.9%, herpes zoster: 5.3/1000 PY), suggesting that risk of 88.5%, and 87.0%, respectively, had no GdE lesions these infections does not increase with longer ozani- on the month 48 scan (Supplemental Table S6). mod use.5 Serious herpetic infections have occurred in patients taking other S1P receptor modulators.11,12 During DAYBREAK, 346/2494 (13.9%) patients had The IRs for oral herpes and herpes zoster were lower CDP-3 and 285/2494 (11.4%) had CDP-6 by data cut- in DAYBREAK than previously reported with fin- off; rates were similar across parent trial treatment golimod (oral herpes: 10.3/1000 PY; varicella zoster subgroups. Kaplan–Meier analysis of time to CDP-3 virus: 7‒11/1000 PY).13,14 is shown in Figure 7. Among 2494 DAYBREAK participants, one patient developed PML shortly after data cutoff.6 This is the Discussion and conclusions first known case in an ozanimod-treated patient This interim analysis of the DAYBREAK extension across 16,512 PY of exposure in the ozanimod devel- study characterizes safety and efficacy of ozanimod opment program, including MS and inflammatory 12 journals.sagepub.com/home/msj BAC Cree, KW Selmaj et al. 0.16 0.14 Adjusted ARR (95% CI)a 0.12 0.1 0.08 0.06 0.04 0.02 0.103 0.110 0.095 0.103 0 n= 740 875 879 2494 IFN -1a 30 g→ Ozanimod 0.46 mg→ Continuous All DAYBREAK Ozanimod 0.92 mg Ozanimod 0.92 mg Ozanimod 0.92 mg (Ozanimod 0.92 mg) Analysis group (a) IFN β-1a 30 µg→Ozanimod 0.92 mg Ozanimod 0.46 mg→Ozanimod 0.92 mg Continuous Ozanimod 0.92 mg 0.4 Estimated proportion of patients with first confirmed relapse 0.3 0.2 0.1 0.0 Day 1 M6 M12 M18 M24 M30 M36 M42 M48 Time Parent Trial Treatment Group Number of patients at risk IFN -1a 30 g 740 653 610 567 536 508 481 450 339 Ozanimod 0.46 mg 875 792 728 683 641 614 587 559 441 Ozanimod 0.92 mg 879 805 741 703 663 632 609 571 446 (b) Figure 5. (a) Annualized relapse rate in DAYBREAK, ITT population. (b) Kaplan–Meier curve for time to first confirmed relapse during DAYBREAK, ITT population. ARR: annualized relapse rate; CI: confidence interval; GdE: gadolinium-enhancing; IFN: interferon; ITT: intent to treat. aBased on the negative binomial regression model, adjusted for region (Eastern Europe vs rest of world), age at parent baseline, and the parent baseline number of GdE lesions. The natural log transformation of time on treatment is used as an offset term to adjust for persons having different exposure times. The negative binomial regression model was used to account for possible overdispersion, and results in a more robust CI in a single-arm open-label study compared with the Poisson model that were used to calculate ARR in the primary analysis of the phase 3 trials. bowel disease, and 6455 PY of postmarketing expo- COVID-19 cases during DAYBREAK were largely sure to ozanimod, as of 19 November 2021. PML nonserious and a vast majority of patients recovered has been reported with MS DMTs, including S1P from COVID-19 without sequelae. The 18-month cum­ modulators.15–17 ulative incidence of total and confirmed COVID-19 journals.sagepub.com/home/msj 13 Multiple Sclerosis Journal 00(0) Parent Trial Month 12 DAYBREAK Month 12 DAYBREAK Month 24 DAYBREAK Month 36 DAYBREAK Month 48 4.5 New/Enlarging T2 Lesions (Phase 3 ITT Populationa) new/enlarging T2 lesions per MRI scanb 4 Adjusted mean (95% CI) number of 3.5 3 2.5 2 1.5 1 0.5 3.46 1.54 1.26 1.14 0.85 2.34 1.17 1.01 1.02 0.88 1.86 1.32 1.20 1.15 1.03 0 n= 683 696 651 605 488 706 722 683 643 514 705 726 698 655 509 IFN β-1a 30 µg Ozanimod 0.92 mg Ozanimod 0.46 mg Ozanimod 0.92 mg Continuous Ozanimod 0.92 mg Αnalysis group (a) Parent Trial Month 12 DAYBREAK Month 12 DAYBREAK Month 24 DAYBREAK Month 36 DAYBREAK Month 48 0.6 GdE Lesions (Phase 3 ITT Populationa) 0.5 Adjusted mean (95% CI) number 0.4 of GdE lesionsc 0.3 0.2 0.1 0 0.44 0.11 0.14 0.20 0.06 0.25 0.13 0.15 0.16 0.08 0.17 0.21 0.22 0.25 0.12 n= 683 695 651 604 487 706 723 683 642 513 705 725 698 653 509 IFN β-1a 30 µg Ozanimod 0.92 mg Ozanimod 0.46 mg Ozanimod 0.92 mg Continuous Ozanimod 0.92 mg Αnalysis group (b) Figure 6. (a) Number of new/enlarging T2 lesions per scan, and (b) number of GdE lesions at month 12 of the active- controlled phase 3 parent trials and months 12, 24, 36, and 48 of the DAYBREAK extension by phase 3 parent trial treatment subgroup. CI: confidence interval; GdE: gadolinium-enhancing; IFN: interferon; ITT: intent to treat. aNew/enlarging T2 lesions and GdE lesions were analyzed only in the subset of patients who entered DAYBREAK from an active- controlled phase 3 trial. bBased on negative binomial regression model, adjusted for study, region (Eastern Europe vs rest of world), age at baseline, and baseline number of GdE lesions, with the log of the number of postbaseline scans as an offset term in the model. cBased on negative binomial regression model, adjusted for study, region (Eastern Europe vs rest of world), age at baseline, and the baseline number of GdE lesions. among DAYBREAK participants was 8.7% and cumulative incidence of COVID-19 in the general 7.3%, respectively. The cumulative incidence of population was 0.05%‒10.0% in countries participat- COVID-19 varies worldwide,18 and data show that ing in DAYBREAK.18,21 Over 90% of DAYBREAK MS patients do not have a higher risk of infection participants are from Eastern Europe; published with SARS-CoV-2.19,20 As of 9 May 2021, the COVID-19 incidence data among MS patients in 14 journals.sagepub.com/home/msj BAC Cree, KW Selmaj et al. IFN -1a 30 g→Ozanimod 0.92 mg Ozanimod 0.46 mg→Ozanimod 0.92 mg Continuous Ozanimod 0.92 mg 1.0 progression confirmed at 3 months Estimated probability of disability 0.8 0.6 0.4 0.2 0.0 Day 1 M6 M12 M18 M24 M30 M36 M42 M48 Time Parent Trial Treatment Group Number of patients at risk IFN -1a 30 g 740 711 678 639 619 593 564 532 407 Ozanimod 0.46 mg 875 844 810 773 738 712 683 654 521 Ozanimod 0.92 mg 879 852 806 780 746 720 698 674 534 Figure 7. Kaplan–Meier analysis of time to onset of 3-month confirmed disability progression during DAYBREAKa (ITT population). IFN: interferon; ITT: intent-to-treat. aThe study period includes DAYBREAK day 1 through the data cutoff date. Eastern Europe are sparse or lack robust sample size. breast cancer among ozanimod-treated women with COVID-19-related mortality in DAYBREAK (3/190 RMS in the parent trials or DAYBREAK; the expected (1.6%)) was similar to case-fatality rates in the gen- incidence of breast cancer was seven events over the eral populations of participating countries treatment period, derived by applying the breast can- (0.7%‒4.4%)18 and within range of recent studies of cer IR (92.4/100,000 PY) in an age-matched female COVID-19 mortality among MS patients population from the Surveillance, Epidemiology, and (0.3%‒4.2%).22–26 In a review of MS patients, End Results to the female RMS participants with COVID-19 mortality based on data through 18 exposure to ozanimod (7096 PY). Malignancies con- December 2020 was 3.0% overall and 0.5%‒2.5% tinue to be monitored in the ozanimod clinical devel- among DMT users.27 opment program. The IR for malignancy with ozanimod during the par- Macular edema is a TEAE with S1P modulators.12,13,30 ent trials plus DAYBREAK is 358.6/100,000 PY The incidence of macular edema in DAYBREAK was (95% CI 253.7‒492.2), and the IR in an earlier analy- low. Most confirmed cases occurred in patients with sis of the same trials was 289.3/100,000 PY (95% CI predisposing conditions or risk factors within 1 year 187.2‒427.1). The IR in the phase 3 trial population of ozanimod initiation. One case without risk factors was 298.2/100,000 PY (95% CI 128.7‒587.6).5 No occurred after 6 years of ozanimod exposure; delayed particular pattern in malignancy type occurred. onset of ME is unusual but has been reported after Malignancies are a potential concern with any immune- 7 years of fingolimod use.36,37 modulating therapy, including S1P modulators.28–30 A study of S1P modulators suggests there may be an Bradycardia was infrequent, occurring mostly in those association with an increase in risk of skin cancers, who switched from interferon β-1a or ozanimod particularly basal cell carcinomas.31 Similar to the 0.46 mg to ozanimod 0.92 mg. Similarly, bradycardia general population, basal cell carcinomas were the is uncommon after the first year of fingolimod use.13 most frequent cancer type reported in ozanimod clini- The IR for myocardial infarction in DAYBREAK cal trials (IR 103.6/100,000 PY). The estimated inci- (0.3/1000 PY (95% CI 0.1‒1.0)) was similar to that dence of basal cell carcinoma in the general population reported in the LONGTERMS study of fingolimod (IR ranges from 25 to 321 per 100,000 PY in European 0.2/1000 PY).13 Cardiovascular abnormalities associ- and US populations.32–35 There were nine cases of journals.sagepub.com/home/msj 15 Multiple Sclerosis Journal 00(0) ated with S1P modulators12,13,30 are thought to be medi- 0.92 mg for up to 5 years are consistent with the ated by S1P receptors expressed by cardiomyocytes.38 established safety profile and demonstrate sustained control of disease activity and disability progression PFT abnormalities, observed in a minority of patients, with continued ozanimod use. Adverse events, includ- were rarely clinically relevant, as discontinuation for ing COVID-19, were generally manageable, infre- dyspnea or PFT abnormalities was uncommon. Dose- quently leading to permanent treatment discontinuation dependent reductions from baseline in FEV1 have or adverse outcomes. occurred with other S1P modulators.39–41 Acknowledgements The proportion of patients with hepatic laboratory Support for third-party writing assistance for this abnormalities during DAYBREAK was similar to that manuscript was provided by Lauren A Cerruto and in SUNBEAM4 and RADIANCE.3 These abnormali- Jessica D Herr, PharmD, of Peloton Advantage, ties were generally not of clinical consequence. Liver LLC, an OPEN Health company, and funded by enzyme elevations are a known effect of S1P modula- Bristol Myers Squibb. The authors thank Sonia Ariana tors, and cases of acute liver failure requiring trans- Afsari, MD, and Jenna Hoogerheyde for reviewing plantation have been reported in patients treated with the manuscript and also thank the members of the fingolimod.40–42 Data Monitoring Committee, the Macular Edema Review Panel, the expert hepatic safety advisors, the ALC reduction to below normal (based on labora- patients, their families, and trial investigators for their tory values), which occurred in 95.6% of patients, is contributions. an expected pharmacodynamic effect of ozanimod, and therefore was not required to be reported as a Declaration of Conflicting Interests TEAE. For those who continued ozanimod 0.92 mg The author(s) declared the following potential con- from parent trial into DAYBREAK, mean ALC flicts of interest with respect to the research, author- remained stable and consistent with ALC levels at ship, and/or publication of this article: B.A.C.C. that dose in SUNBEAM4 and RADIANCE.3 In a reports personal compensation for consulting for prior pooled safety analysis of DAYBREAK and the Alexion, Atara, Autobahn, Avotres, Biogen, EMD parent trials, 182 patients had a minimal postbase- Serono, Gossamer Bio, Horizon, Neuron23, Novartis, line ALC < 0.2 × 109/L, but ALC was less than Sanofi, TG Therapeutics, and Therini, and received 0.2 × 109/L around the time of onset of only 1/44 research support from Genentech. K.W.S. reports serious infections and 1/113 nonserious opportunis- consulting for Biogen, Celgene, Genzyme, Merck, tic infections; thus, no relationship was found Novartis, Ono Pharma, Roche, Synthon, and Teva. between ALC < 0.2 × 109/L and serious or oppor- L.S. reports consulting for AbbVie, Atreca, Celgene, tunistic infections.5 Novartis, Teva, Tolerion, and EMD Serono, and research support from Atara, Biogen, and Celgene. Patients discontinuing any MS DMT are vulnerable G.C. reports compensation for consulting and/or to posttreatment relapse. Incidence of relapse after speaking activities from Almirall, Biogen, Celgene, ozanimod discontinuation was low in DAYBREAK, EXCEMED, Forward Pharma, Genzyme, Merck, and none of the relapses resulted in persistent, severe Novartis, Roche, Sanofi, and Teva. A.B.-O. partici- disability. Cases of severe disease recurrence have pated as a speaker in meetings sponsored by and been reported after discontinuation of other S1P received consulting fees and/or grant support from modulators.43,44 Atara Biotherapeutics, Biogen, BMS-Celgene, EMD Serono, Sanofi-Genzyme, Novartis, and Roche- Limitations typical of open-label extension studies Genentech. D.L.A. reports personal fees for consult- include lack of a control group or blinding; therefore, ing and/or grants from Albert Charitable Trust, comparisons to blinded study data may be limited. Biogen, Novartis, Celgene, F. Hoffmann-La Roche, Patients with less favorable outcomes may have Frequency Therapeutics, MedDay, Merck Serono, dropped out of the parent trials, creating potential Novartis, Population Council, and Sanofi-Aventis; selection bias, but the number of subjects who did not grants from Biogen, Immunotec, and Novartis; and an enter the OLE was low (5.5% of eligible patients) and equity interest in NeuroRx Research. H.-P.H. reports only 17.6% of enrolled patients discontinued from personal fees for consulting, serving on steering com- DAYBREAK. mittees, and speaking from Bayer Healthcare, Biogen, Celgene, GeNeuro, Genzyme, Merck, MedImmune, Data from this ongoing, long-term observational Novartis, Octapharma, Roche, Sanofi, and Teva. study of patients with RMS treated with ozanimod X.M. has received speaking honoraria and travel 16 journals.sagepub.com/home/msj BAC Cree, KW Selmaj et al. expenses for participation in scientific meetings, Ludwig Kappos https://orcid.org/0000-0003-4175 has been a steering committee member of clinical -5509 trials or participated in advisory boards of clinical Jeffrey A Cohen https://orcid.org/0000-0001-9245 trials in the past 3 years with Actelion, Alexion, -9772 Bayer, Biogen, Bristol Myers Squibb/Celgene, EMD Serono, EXCEMED, Genzyme, Hoffmann-La Roche, Data Sharing Immunic, Janssen Pharmaceuticals, MedDay, Merck, Bristol Myers Squibb policy on data sharing may be Mylan, MSIF, Nervgen, NMSS, Novartis, Roche, found at https://www.bms.com/researchers-and-part- Sanofi-Genzyme, Teva Pharmaceuticals, and TG ners/independent-research/data-sharing-request- Therapeutics. E.K.H. reports personal compensation process.html for consulting and speaking for Actelion, Biogen, Celgene, Merck, Novartis, Roche, Sanofi, and Teva, Supplemental Material and is supported by Czech Ministry of Education, Supplemental material for this article is available project PROGRES Q27/LF1. J.K.S., N.M. C.-Y.C online. and D.S. are employees and shareholders of Bristol Myers Squibb. L.K. has received institutional research support: steering committee, advisory board, consul- tancy fees: Actelion, Bayer HealthCare, Biogen, References Bristol Myers Squibb, Genzyme, Janssen, Japan 1. Cohen JA, Arnold DL, Comi G, et al. Safety and Tobacco, Merck, Novartis, Roche, Sanofi, Santhera, efficacy of the selective sphingosine 1-phosphate Shionogi, and TG therapeutics; speaker fees: Bayer receptor modulator ozanimod in relapsing multiple HealthCare, Biogen, Merck, Novartis, Roche, and sclerosis (RADIANCE): A randomised, placebo- Sanofi; support of educational activities: Allergan, controlled, phase 2 trial. Lancet Neurol 2016; 15(4): Bayer HealthCare, Biogen, CSL Behring, Desitin, 373–381. Genzyme, Merck, Novartis, Roche, Pfizer, Sanofi, 2. Cohen JA, Comi G, Arnold DL, et al. Efficacy and Shire, and Teva; license fees for Neurostatus prod- safety of ozanimod in multiple sclerosis: Dose- ucts; and grants: Bayer HealthCare, Biogen, European blinded extension of a randomized phase II study. Union, Innosuisse, Merck, Novartis, Roche, Swiss Mult Scler 2019; 25(9): 1255–1262. MS Society, and Swiss National Research Foundation. 3. Cohen JA, Comi G, Selmaj KW, et al. Safety and J.A.C. reports personal compensation for consulting efficacy of ozanimod versus interferon beta-1a for Adamas, Atara, Bristol Myers Squibb, Convelo, in relapsing multiple sclerosis (RADIANCE): A MedDay, and Mylan; and serving as an Editor of multicentre, randomised, 24-month, phase 3 trial. Multiple Sclerosis Journal. Lancet Neurol 2019; 18(11): 1021–1033. Some of the data in this manuscript were previously 4. Comi G, Kappos L, Selmaj KW, et al. Safety and presented at the 37th Congress of the European efficacy of ozanimod versus interferon beta-1a Committee for Treatment and Research in Multiple in relapsing multiple sclerosis (SUNBEAM): A Sclerosis (ECTRIMS); 13–15 October 2021; The multicentre, randomised, minimum 12-month, phase Digital Experience. 3 trial. Lancet Neurol 2019; 18(11): 1009–1020. Funding 5. Selmaj KW, Cohen JA, Comi G, et al. Ozanimod in relapsing multiple sclerosis: Pooled safety results The author(s) disclosed receipt of the following finan- from the clinical development program. Mult Scler cial support for the research, authorship, and/or publi- Relat Disord 2021; 51: 102844. cation of this article: The ozanimod RMS trials were supported by Celgene International II. The sponsor 6. Sriwastava S, Chaudhary D, Srivastava S, et al. was involved in data analysis and interpretation, and Progressive multifocal leukoencephalopathy and manuscript preparation, review, and approval. All sphingosine 1-phosphate receptor modulators used in multiple sclerosis: An updated review of literature. authors vouch for data accuracy, reviewed all drafts, J Neurol 2022; 269(3): 1678–1687. and approved the final manuscript. 7. Guidance for industry: Drug-induced liver injury: ORCID iDs Premarketing clinical evaluation. Silver Spring, MD: Bruce AC Cree https://orcid.org/0000-0001-7689 U.S. Department of Health and Human Services, -2533 Food and Drug Administration, 2009. Douglas L Arnold https://orcid.org/0000-0003 8. Moiola L, Barcella V, Benatti S, et al. The risk -4266-0106 of infection in patients with multiple sclerosis journals.sagepub.com/home/msj 17 Multiple Sclerosis Journal 00(0) treated with disease-modifying therapies: A Delphi 20. Zabalza A, Cárdenas-Robledo S, Tagliani P, consensus statement. Mult Scler 2021; 27(3): et al. COVID-19 in multiple sclerosis patients: 331–346. Susceptibility, severity risk factors and serological response. Eur J Neurol 2021; 28(10): 3384–3395. 9. Sharma K, Chaudhary D, Beard K, et al. A comprehensive review of varicella-zoster virus, 21. World population projections. Dover, DE: herpes simplex virus and cryptococcal infections Worldometer. https://www.worldometers.info/world- associated with sphingosine-1-phosphate receptor population/world-population-projections/. Accessed modulators in multiple sclerosis patients. Mult Scler June 13, 2022. Relat Disord 2022; 59: 103675. 22. Sormani MP, De Rossi N, Schiavetti I, et al. Disease- 10. Zingaropoli MA, Pasculli P, Iannetta M, et al. modifying therapies and coronavirus disease 2019 Infectious risk in multiple sclerosis patients treated severity in multiple sclerosis. Ann Neurol 2021; with disease-modifying therapies: A three-year 89(4): 780–789. observational cohort study. Mult Scler J Exp Transl Clin 2022; 8(1): 20552173211065731. 23. Louapre C, Collongues N, Stankoff B, et al. Clinical characteristics and outcomes in patients with 11. Cohen JA, Barkhof F, Comi G, et al. Oral fingolimod coronavirus disease 2019 and multiple sclerosis. or intramuscular interferon for relapsing multiple JAMA Neurol 2020; 77: 1079–1088. sclerosis. N Engl J Med 2010; 362: 402–415. 24. Reder AT, Centonze D, Naylor ML, et al. COVID-19 12. Kappos L, Bar-Or A, Cree BAC, et al. Siponimod in patients with multiple sclerosis: Associations with versus placebo in secondary progressive multiple disease-modifying therapies. CNS Drugs 2021; 35(3): sclerosis (EXPAND): A double-blind, randomised, 317–330. phase 3 study. Lancet 2018; 391: 1263–1273. 25. Czarnowska A, Brola W, Zajkowska O, et al. Clinical 13. Cohen JA, Tenenbaum N, Bhatt A, et al. Extended course and outcome of SARS-CoV-2 infection in treatment with fingolimod for relapsing multiple multiple sclerosis patients treated with disease- sclerosis: The 14-year LONGTERMS study modifying therapies—the Polish experience. Neurol results. Ther Adv Neurol Disord 2019; 12: Neurochir Pol 2021; 55(2): 212–222. 1756286419878324. 26. Salter A, Fox RJ, Newsome SD, et al. Outcomes 14. Arvin AM, Wolinsky JS, Kappos L, et al. Varicella- and risk factors associated with SARS-CoV-2 zoster virus infections in patients treated with infection in a North American registry of patients fingolimod: Risk assessment and consensus with multiple sclerosis. JAMA Neurol 2021; 78: recommendations for management. JAMA Neurol 699–708. 2015; 72(1): 31–39. 27. Barzegar M, Mirmosayyeb O, Gajarzadeh M, et al. 15. Oshima Y, Tanimoto T, Yuji K, et al. Drug-associated COVID-19 among patients with multiple sclerosis: progressive multifocal leukoencephalopathy in A systematic review. Neurol Neuroimmunol multiple sclerosis patients. Mult Scler 2019; 25(8): 1141–1149. Neuroinflamm 2021; 8: e1001. 16. Patel A, Sul J, Gordon ML, et al. Progressive 28. Alping P, Askling J, Burman J, et al. Cancer risk multifocal leukoencephalopathy in a patient for fingolimod, natalizumab, and rituximab in MS with progressive multiple sclerosis treated with patients. Ann Neurol 2020; 87: 688–699. ocrelizumab monotherapy. JAMA Neurol 2021; 78: 29. Kappos L, Li DK, Stuve O, et al. Safety and 736–740. efficacy of siponimod (BAF312) in patients with 17. Gerevini S, Capra R, Bertoli D, et al. Immune relapsing-remitting multiple sclerosis: Dose-blinded, profiling of a patient with alemtuzumab-associated randomized extension of the phase 2 BOLD study. progressive multifocal leukoencephalopathy. Mult JAMA Neurol 2016; 73: 1089–1098. Scler 2019; 25(8): 1196–1201. 30. Kappos L, Fox RJ, Burcklen M, et al. Ponesimod 18. COVID-19 weekly epidemiological update. Geneva: compared with teriflunomide in patients with World Health Organization, 2021. https://www.who. relapsing multiple sclerosis in the active-comparator int/publications/m/item/weekly-epidemiological- phase 3 OPTIMUM study: A randomized clinical update-on-covid-19---11-may-2021. Accessed June trial. JAMA Neurol 2021; 78: 558–567. 13, 2022. 31. Stamatellos VP, Rigas A, Stamoula E, et al. 19. Moreno-Torres I, Meca Lallana V, Costa-Frossard S1P receptor modulators in multiple sclerosis: L, et al. Risk and outcomes of COVID-19 in patients Detecting a potential skin cancer safety signal. with multiple sclerosis. Eur J Neurol 2021; 28(11): Mult Scler Relat Disord 2022; 59. DOI: 10.1016/j. 3712–3721. msard.2022.103681. 18 journals.sagepub.com/home/msj BAC Cree, KW Selmaj et al. 32. Bielsa I, Soria X, Esteve M, et al. Population-based 38. Li N and Zhang F. Implication of sphingosin-1- incidence of basal cell carcinoma in a Spanish phosphate in cardiovascular regulation. Front Biosci Mediterranean area. Br J Dermatol 2009; 161(6): 2016; 21: 1296–1313. 1341–1346. 39. Calabresi PA, Radue EW, Goodin D, et al. Safety 33. Celić D, Lipozencić J, Jurakić Toncić R, et al. The and efficacy of fingolimod in patients with incidence of basal cell carcinoma in Croatia: An relapsing-remitting multiple sclerosis epidemiological study. Acta Dermatovenerol Croat (FREEDOMS II): A double-blind, randomised, 2009; 17(2): 108–112. placebo-controlled, phase 3 trial. Lancet Neurol 34. de Vries E, Micallef R, Brewster DH, et al. Population- 2014; 13(6): 545–556. based estimates of the occurrence of multiple vs first 40. Mayzent (package insert). East Hanover, NJ: Novartis primary basal cell carcinomas in 4 European regions. Pharmaceuticals, 2021. Arch Dermatol 2012; 148(3): 347–354. 41. Ponvory (package insert). Titusville, NJ: Janssen 35. Muzic JG, Schmitt AR, Wright AC, et al. Incidence Pharmaceuticals, 2021. and trends of basal cell carcinoma and cutaneous squamous cell carcinoma: A population-based study 42. Gilenya (package insert). East Hanover, NJ: Novartis in Olmsted County, Minnesota, 2000 to 2010. Mayo Pharmaceuticals Corporation, 2019. Clin Proc 2017; 92(6): 890–898. 43. Barry B, Erwin AA, Stevens J, et al. Fingolimod 36. Camós-Carreras A, Alba-Arbalat S, Dotti-Boada rebound: A review of the clinical experience and M, et al. Late onset macular oedema in a patient management considerations. Neurol Ther 2019; 8(2): with multiple sclerosis treated with fingolimod. 241–250. Neuroophthalmology 2021; 45(1): 61–64. 44. Litwin T, Smoliński Ł and Członkowka A. 37. Zarbin MA, Jampol LM, Jager RD, et al. Ophthalmic Substantial disease exacerbation in a patient with Visit SAGE journals online evaluations in clinical studies of fingolimod relapsing-remitting multiple sclerosis after withdrawal journals.sagepub.com/ home/msj (FTY720) in multiple sclerosis. Ophthalmology 2013; from siponimod. Neurol Neurochir Pol 2018; 52(1): 120(7): 1432–1439. 98–101. SAGE journals journals.sagepub.com/home/msj 19