No evidence of disease activity (NEDA) analysis by epochs in patients with relapsing multiple sclerosis treated with ocrelizumab vs interferon beta-1a

Background No evidence of disease activity (NEDA; defined as no 12-week confirmed disability progression, no protocol-defined relapses, no new/enlarging T2 lesions and no T1 gadolinium-enhancing lesions) using a fixed-study entry baseline is commonly used as a treatment outcome in multiple sclerosis (MS). Objective The objective of this paper is to assess the effect of ocrelizumab on NEDA using re-baselining analysis, and the predictive value of NEDA status. Methods NEDA was assessed in a modified intent-to-treat population (n = 1520) from the pooled OPERA I and OPERA II studies over various epochs in patients with relapsing MS receiving ocrelizumab (600 mg) or interferon beta-1a (IFN β‐1a; 44 μg). Results NEDA was increased with ocrelizumab vs IFN β-1a over 96 weeks by 75% (p < 0.001), from Week 0‒24 by 33% (p < 0.001) and from Week 24‒96 by 72% (p < 0.001). Among patients with disease activity during Weeks 0‒24, 66.4% vs 24.3% achieved NEDA during Weeks 24‒96 in the ocrelizumab and IFN β-1a groups (relative increase: 177%; p < 0.001). Conclusion Superior efficacy with ocrelizumab compared with IFN β-1a was consistently seen in maintaining NEDA status in all epochs evaluated. By contrast with IFN β-1a, the majority of patients with disease activity early in the study subsequently attained NEDA status with ocrelizumab.


Introduction
No evidence of disease activity (NEDA) is a composite measure of the absence of confirmed disability worsening (as measured by the Expanded Disability Status Scale (EDSS)) and of clinical and magnetic resonance imaging (MRI) measures of disease activity in relapsing multiple sclerosis (RMS). NEDA was originally described in a post-hoc analysis of the placebo-controlled, two-year Phase III pivotal trial for natalizumab. 1 NEDA has since been widely used in analyses of other disease-modifying therapies (DMTs), [2][3][4][5][6][7][8][9] and has been adopted as an outcome measure in RMS clinical trials. [10][11][12][13] NEDA may provide a more sensitive and comprehensive measure of capturing overall treatment benefit and has been proposed as a primary endpoint in future pivotal Phase III clinical studies. 14,15 NEDA is increasingly recognized as an important treatment goal for patients with RMS 16 and has been shown to be informative in the prediction of long-term disability progression both independent of DMT type 17 (cohort study) and in the core and extension periods of clinical trial patient populations. 18 As a binary outcome, NEDA status and its failure are often driven by MRI components of the measure and hence are particularly influenced by the frequency of MRI assessments. 8,14 The analysis of NEDA is also affected by the pharmacodynamics of the particular therapy assessed. Re-baselining, wherein a poststudy baseline time point is utilized as the new baseline reference for subsequent assessment of disability worsening and disease activity, therefore may reflect a truer representation of a DMT's steady state of efficacy 16 unconfounded by any initial disease activity carried over from baseline and recent prebaseline disease state.
B cells are a significant contributor to the pathogenesis of MS. 19,20 CD20 is a cell-surface antigen expressed on pre-B cells, mature B cells and memory B cells, but not lymphoid stem cells and plasma cells. [21][22][23] Ocrelizumab is a recombinant, humanized monoclonal antibody that selectively depletes CD20-expressing B cells 24,25 while preserving the capacity for B-cell reconstitution and maintaining pre-existing humoral immunity. 26,27 In the two identical Phase III trials, OPERA I and OPERA II, ocrelizumab significantly reduced all individual components of NEDA compared with high-dose, high-frequency interferon beta-1a (IFN b-1a) at Week 96 in patients with RMS. 28 The objective of the current analyses was to assess the effect of ocrelizumab on the proportion of patients with NEDA and determine the predictive value of NEDA over time in multiple epochs across the pooled OPERA I and OPERA II studies.

Patients and methods
Trial design and patients NEDA was determined in the pooled population of the two identical Phase III, multicenter, randomized, double-blind, double-dummy, parallel-group OPERA I and OPERA II trials (OPERA I/NCT01247324 and OPERA II/NCT01412333). The study protocol was approved by each center's independent ethics committee. The study design was written in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, and all enrolled patients provided written informed consent. The OPERA I and OPERA II trials were deemed poolable based on their identical protocols and according to formal poolability testing results. 28 Baseline demographics and disease characteristics were not different within each study arm or between studies. Study details have been reported previously. 28 Key eligibility criteria included an age of 18 to 55 years, diagnosis of RMS (2010 revised McDonald criteria) 29 and an EDSS score of 0 to 5.5 at screening. Patients were randomized (1:1) to receive either ocrelizumab 600 mg by intravenous infusion every 24 weeks or subcutaneous IFN b-1a three times per week at a dose of 44 lg throughout the 96-week treatment period (including a per-label incremental titration scheme in the first four weeks). Patients were stratified by region (United States (US) vs rest of world (ROW)) and baseline EDSS score (<4.0 and 4.0) in the randomization for OPERA I and OPERA II.
Clinical and MRI endpoints, including NEDA EDSS scores were determined at screening, baseline and every 12 weeks: Confirmed disability progression (CDP) was defined as a 1-point increase in EDSS score from a baseline EDSS score of 0-5.5, or a 0.5-point increase in EDSS score from a baseline EDSS score greater than 5.5, sustained for at least 12 weeks (12-week CDP). Protocol-defined relapses were new or worsening neurological symptoms attributable to MS: Symptoms must (i) persist for greater than 24 hours, should not be attributable to confounding clinical factors, and be immediately preceded by a stable or improving neurological state for at least 30 days; and (ii) be accompanied by objective neurological worsening consistent with an increase of at least half a step on the EDSS scale, or two points on at least one of the appropriate Functional Systems Scores (FSSs), or one point on two or more of the appropriate FSSs. Brain MRI was performed at baseline and Weeks 24, 48 and 96; new or enlarging T2 lesions and/or T1 gadoliniumenhancing lesions on any scan during the epoch investigated were considered evidence of MRI disease activity. NEDA status was defined as the combined absence of: protocol-defined relapses; 12-week CDP; new or enlarging T2 lesions; and T1 gadolinium-enhancing lesions.

Statistical analyses
NEDA outcome was primarily assessed in a modified intent-to-treat (mITT) population over the controlled treatment phase (baseline to 96 weeks), 28 which included all patients in the intent-to-treat population, but patients who discontinued treatment early for reasons other than lack of efficacy or death and had NEDA before early discontinuation were excluded. Further analyses evaluated the proportion of patients with NEDA for several epochs, including: Weeks 0-48; Weeks 48-96 (where all components of NEDA including 12-week CDP were re-baselined to Week 48); Weeks 0-24; Weeks 24-48 (where all components of NEDA including 12-week CDP were re-baselined to Week 24); and Weeks 24-96 (where all components of NEDA including 12-week CDP were re-baselined to Week 24). NEDA and its components were compared in patients treated with ocrelizumab with those receiving IFN b-1a in a post-hoc analysis using the Cochran-Mantel-Haenszel test stratified by study, geographic region (US vs ROW) and baseline EDSS score (<4.0 vs 4.0). Patients who discontinued treatment early with at least one event (i.e. protocol-defined relapse, 12-week CDP, new or enlarging T2 lesion or T1 gadolinium-enhancing lesion) before early discontinuation were considered as having evidence of disease activity (EDA). Even if a patient did not report an event before early discontinuation, the patient was considered as having EDA if the reason for early discontinuation was lack of efficacy or death.
Probability of attaining or maintaining NEDA relative to earlier EDA or NEDA status, and the predictive value of NEDA status Subgroups of patients by NEDA status in Weeks 0-24 were assessed for subsequent NEDA status in Weeks 24-96 and in Weeks 24-48. Similarly, subgroups of patients by NEDA status in Weeks 24-48 were assessed for subsequent NEDA status in Weeks 48-96. Using the pooled treatment arms, NEDA status in Weeks 0-48 was used to predict time to (i) first protocol-defined relapse, and (ii) first 12-week and 24-week CDP in Weeks 48-96, and was evaluated based on the Cox proportional hazard regression model which included the NEDA during Weeks 0-48 as a factor, stratified by study, geographical region (US vs ROW) and baseline EDSS score (<4.0 vs 4.0).

Patient demographics and disease characteristics
In the OPERA I and OPERA II trials, the pooled intention-to-treat population comprised 1656 patients (IFN b-1a, n ¼ 829; ocrelizumab, n ¼ 827). The mITT reference population used for the NEDA analysis comprised 759 and 761 patients randomized to high-dose, high-frequency IFN b-1a and ocrelizumab, respectively. Detailed baseline demographics and disease characteristics of (i) the mITT population and (ii) patients with EDA and NEDA over 96 weeks are presented by treatment arm in Table 1. All major baseline covariates were balanced between treatment groups in the mITT population. Compared with patients who maintained NEDA over 96 weeks, patients with EDA across treatment groups had slightly more relapses recorded in the 12 months prior to baseline, a greater number of T1 gadolinium-enhanced lesions and a higher T2 lesion burden at baseline. The age and proportion of female patients were slightly lower in the EDA group. Baseline EDSS score, disease duration and normalized brain volume were similar between patients who maintained NEDA or experienced EDA events.
Overall NEDA (Week 0-96) and re-baselined NEDA (Weeks 24-96) status in the pooled OPERA I and OPERA II studies In the pooled analyses of OPERA I and OPERA II, the relative proportion of patients with NEDA was increased by 75% (47.7% vs 27.1%; p < 0.001) with ocrelizumab compared with IFN b-1a over 96 weeks (Figure 1(a)). 28 Following re-baselining at Week 24, the relative proportion of patients with NEDA was 72% higher (72.2% vs 41.9%; p < 0.001) with ocrelizumab compared with IFN b-1a during Weeks 24-96 ( Figure 1(b)).
During the 96-week epoch, a significant difference in the proportion of patients without disease activity was seen for each individual component of NEDA, including 12-week CDP, with ocrelizumab compared with IFN b-1a (p < 0.001; Table 2). This was reflected in the proportion of patients with no disability worsening and clinical disease activity (no 12-week CDP and no relapses: ocrelizumab 73.8% vs IFN b-1a 59.4%; p < 0.001) and no brain MRI measures of disease activity (no new or enlarging T2 lesions and no T1 gadolinium-enhancing lesions: ocrelizumab 62.2% vs IFN b-1a 37.6%; p < 0.001; Table 2). Following re-baselining at Week 24, similar results were seen during Weeks 24-96 for the individual components of NEDA and the pairwise combination of clinical and MRI measures (       Table S1).

Discussion
A higher proportion of patients had NEDA with ocrelizumab at the first MRI at Week 24 and in all epochs of the OPERA studies compared with highdose, high-frequency IFN b-1a. Over two years, 48% of patients treated with ocrelizumab had NEDA, and as many as 72% had NEDA in Weeks 24-96 and 82% in Weeks 48-96, which is greater than that observed with other high-efficacy therapies in similar epochs. 1,30,31 Absolute proportions of patients maintaining NEDA status should be considered in the context of the existence of a certain level of inherent background noise in the binary assessment of NEDA status because of potential false-positive detection of clinical EDA events, particularly for relapses, which may yield a treatment   ceiling effect. Pairwise combinations of disease parameters in the clinical and MRI components of NEDA showed the consistent overall benefit of ocrelizumab compared with IFN b-1a, which was further reflected in all the individual components of NEDA, including confirmed disability progression. Compared with patients who maintained NEDA over 96 weeks, the patients with EDA across treatment groups had higher brain MRI activity at baseline as measured by T1 gadolinium-enhancing lesions and higher T2 lesion burden, while no sizable difference was seen in pre-baseline relapse rate, and baseline EDSS score, disease duration and brain atrophy.
The early impact of ocrelizumab on MS disease activity was shown in a Phase II, randomized, placebo-controlled, double-blind trial in patients with relapsing-remitting MS, in which ocrelizumab demonstrated a robust effect on MRI activity as early as Week 8 after initiating treatment. 32 However, in order to represent the full efficacy of a DMT unconfounded by disease activity carried over during the first four to eight weeks from treatment initiation, particularly MRI related, a rebaselining approach at first available MRI has been suggested. 3,14,33,34 Re-baselining MS disease parameters to Week 24 showed a 72% relative increase in the proportion of ocrelizumab-treated patients with NEDA from Weeks 24 to 96 compared with IFN b-1a-treated patients. Moving into the clinical practice setting, the optimal timing of re-baselining should reflect the anticipated timing for reaching complete DMT efficacy, to give a more reliable indication of subsequent drug failure. Conclusions from cross-trial comparisons are limited because of differences including comparators, patient populations, MRI techniques, frequency of assessments, analysis methods and definitions of NEDA. Nevertheless, irrespective of the epoch chosen (including baseline-Week 24, baseline-Week 48 and Weeks 24-96), greater absolute proportions of patients with NEDA were observed in the present analysis compared with those reported with other high-efficacy therapies for RMS. 1,8,10,33 As noted above and reported in other studies, 14 When pooling treatment arms, NEDA status in the first year (Weeks 0-48) predicted a lower risk of relapse (Kaplan-Meier analysis of time to first relapse) and a lower risk of disability progression (as measured by 12-and 24-week CDP) in the second year (Weeks 48-96). The lower risk of subsequent relapse and disability worsening in patients with NEDA during the first year, combined with the high proportion of patients receiving ocrelizumab treatment maintaining NEDA in Year 2 in those patients with NEDA in Year 1, suggest NEDA status over the short term may predict longer-term benefits. Contradictory data have been reported on the prognostic value of NEDA status over two years in predicting future disability progression up to 10 years, at least in patient cohorts from real-world settings where the majority of patients were treated with self-injectable DMTs (interferons or glatiramer acetate). 17,36,37 Few patients maintained NEDA over the long term in studies in which self-injectable DMTs were used (7.9% over seven years; 17 0% over 10 years 9 ). Conversely, NEDA was enhanced in patients over the long term in studies in which more effective DMTs were used (34% over seven years with natalizumab; 38 40% over five years with alemtuzumab 39 ). These data support the notion that, balanced with any risk associated with a particular DMT, NEDA may be a realizable long-term treatment goal for patients with RMS in the era of higherefficacy therapies. Furthermore, re-baselining of NEDA status especially during the first year of DMT initiation might have value in clinical practice to assess early treatment response, and inform longer-term therapeutic decisions to optimize the control of MS disease activity.
Further evolution of the NEDA concept may include the future integration of brain volume loss, and research is ongoing to determine the optimal threshold(s) of annualized rates that may discriminate pathological atrophy at the individual patient level, while accounting for the effect of aging. 9,40 Similarly, the incorporation of cognitive, ambulation and upper extremity function measures may enable a more comprehensive ascertainment of the absence of disability progression when assessing NEDA.
Overall, ocrelizumab consistently resulted in a profound reduction of clinical and subclinical disease activity compared with IFN b-1a in patients with RMS, as measured by NEDA across various epochs. Understanding the associations between NEDA and patient-reported outcomes is warranted to better inform the day-to-day relevance of maintaining NEDA status. Data from open-label extension studies will help determine whether NEDA maintained in the two-year OPERA studies will translate into sustained NEDA and enhanced protection against accrual of disability over the long term.