Practice Points
- Unsolicited reports of the wearing-off effect related to ocrelizumab were infrequent (< 2%).
- Symptoms reported with the wearing-off effect were nonserious.
- The majority of patients reporting the wearing-off effect continued with ocrelizumab.
Multiple sclerosis (MS) is a chronic, neurologic, inflammatory disease affecting approximately 90,000 Canadians.1,2 MS can affect various parts of the central nervous system, leading to a range of possible symptoms (eg, fatigue, lack of coordination, sensory symptoms, vision problems, bladder problems, cognitive impairment, mood changes). In Canada, approximately 90% of MS cases are relapsing-remitting (RRMS),1 characterized by reversible episodes of neurological dysfunction and disability. However, damage from these events can accumulate over time, with resulting lesions in the brain and spinal cord.2 In 10% of Canadian people with MS, the disease is progressive from onset (primary progressive MS [PPMS]).1
Although there is no cure for MS, current therapies support people with MS in reducing disease activity, delaying progression, and gaining better symptom control.2,3 Among a broad class of disease-modifying therapies (DMTs) are the higher-efficacy monoclonal antibodies (mAbs) alemtuzumab, natalizumab, and 3 anti-CD20s: ocrelizumab, ofatumumab, and ublituximab.3-5 Each has a unique structure, mechanism of action, dosing regimen, and safety profile.5-9 Ocrelizumab, a humanized mAb that selectively targets CD20-expressing B cells, is approved in Canada for the treatment of RRMS and PPMS.8,10,11 After initial dosing of a split, twice-administered 300-mg infusion, ocrelizumab is administered at a single fixed dose (600 mg) every 6 months.8
Natalizumab, ocrelizumab, and ofatumumab have been linked with reports of a therapeutic phenomenon called wearing-off effect.12-18 A wearing-off effect is described as the experience of MS symptoms, such as fatigue, in the days or weeks prior to the next scheduled mAb infusion and a disappearance of such symptoms following receipt of the treatment.7 It has been reported at varying frequencies (16%-67%), which may be a result of different case definitions and/or assessment methods.12-15,19 The wearing-off effect is often referred to as a phenomenon for which research has failed to elucidate either the true cause or consequence of the effect.12,14,15,20,21 This is notably different from disease reactivation or rebound, which has been reported and studied extensively following natalizumab and fingolimod discontinuation.22,23 For natalizumab, findings from a study of the wearing-off effect associated the event with lower α4-integrin receptor occupancy on circulating leukocytes, suggesting a decreased treatment effect.24 For ocrelizumab, research suggests that the reported wearing-off phenomenon is not related to dose interval, B-cell count (remains depleted), symptom burden, or clinical efficacy but may be correlated with higher body mass index (BMI).15
Considering the differences in reported frequency as well as within patient variability of wearing-off effect reports among people with MS treated with ocrelizumab, further investigation was deemed worthwhile. In Canada, the provision of ocrelizumab for MS through a national patient support program presented an opportunity to further study wearing-off effect events in a large population of people with MS treated with ocrelizumab in a real-world setting. The primary objective of this study was to determine the frequency with which the wearing-off effect occurs during real-world treatment with ocrelizumab and, secondarily, to characterize how the event affected the course of treatment, if at all.
Methods
Study Design and Eligibility Criteria
This is a retrospective, observational study of Canadian people with MS treated with ocrelizumab enrolled in the Canadian Roche Patient Support Program (COMPASS) and reporting wearing-off effect (within the Roche Global Safety Database).
COMPASS is a patient support program in Canada, sponsored by the manufacturer of ocrelizumab, that supports the provision of the drug to people with MS. All people with MS in Canada receiving ocrelizumab would at some point in their care be enrolled in COMPASS. The Roche Global Safety Database is a centralized repository for reports of adverse events (AEs) received for products manufactured by F. Hoffmann-La Roche Ltd.
AEs were voluntarily reported, meaning event information was volunteered in an unprompted manner. Therapeutic response shortened (TRS) was the Medical Dictionary for Regulatory Activities v24.1 (MedDRA) highest-level preferred term used in the analysis to capture wearing-off events. Aligned with routine pharmacovigilance processes, medical coders selected the most appropriate lowest-level term from the MedDRA that accurately reflected the reported verbatim information. Medical judgment was applied to assign a MedDRA term that represented the reported concept when the verbatim term was not an exact match to TRS.25 Examples of verbatim terms that were coded to the preferred term (TRS) using the MedDRA included wearing-off, crap gap, wore off, and end-of-cycle phenomenon. Individual case safety reports coded to the MedDRA preferred term TRS were extracted from the Roche Global Safety Database (collected from November 5, 2008, to March 15, 2023) and linked with data from the COMPASS program using unique patient identification numbers or initials and date of birth (Figure S1).
There was no imputation of data; cases with missing identifying information were excluded from the analysis. Reports not linked to the COMPASS program (eg, those from clinical trials) were excluded from this study. Only adults (aged ≥ 18 years) with MS who had received at least 3 doses of ocrelizumab were included in the analysis of TRS. This ensured those included in the analysis had received 2 initial doses (300 mg each) and, approximately 6 months later, a third dose (600 mg; full dose), allowing for at least one 6-month dosing interval in which TRS might occur. TRS events are herein referred to as wearing-off events in keeping with the common literature, as terms synonymous with reports of wearing-off events were coded to TRS as per MedDRA.
Data Collection
The primary variables of interest for this study included reports of wearing-off effect, indication for treatment, therapy status (collected by COMPASS; eg, on therapy [within the 6-month dosing interval and intends to receive the next ocrelizumab infusion] or discontinued therapy [exceeded the 6-month dosing interval or does not intend to receive the next ocrelizumab infusion]), and reason for treatment discontinuation (eg, unknown reason, physician recommendation, AE with program drug, financial reason, switched to another therapy, or other). The Roche Global Safety Database served as the primary source for the following data: AE information (seriousness and outcome), sex at birth, and age. The COMPASS program provided baseline demographics, including data related to treatment history, current therapy status, ocrelizumab dosage, dosing interval treatment indication, and supplementary information on sex and age if unavailable in the AE report. Individuals were stratified by indication for ocrelizumab therapy (ie, RRMS or PPMS). Standard follow-up conducted by case managers in the COMPASS program includes a call after the individual is medically cleared by their neurologist, a call 5 business days prior to the scheduled infusion, a call 2 business days after the infusion, and a 4-month check-in call after the last infusion (ie, 2 months before the next).
This study involved the secondary use of personal data; hence, informed consent was not feasible nor required. A waiver of consent was granted by Advarra (May 3, 2022). All people with MS participating in the COMPASS program consented to sharing their personal information (including personal health information) with Roche and the possibility that this information would be used in an aggregated manner for publication. Still, precautions were taken to protect participant anonymity, including aggregating and anonymizing/pseudonymizing final data and ensuring linkage back to the individual participant was impossible or tightly controlled. The study protocol was reviewed and approved by the Advarra Institutional Review Board (approval number: MOD01320274).
Data Analysis and Reporting
Analyses were conducted using Python 3.8.8. For continuous variables, measures of central tendency (medians and means) and dispersion (SD and range) are reported. The COMPASS database provided a total number of individuals exposed to ocrelizumab (a denominator), and the Roche Global Safety Database, after exclusions, provided the number of individuals affected (the numerator). This enabled the reporting of frequencies and percentages for each categorical variable. All data are reported using descriptive statistics.
Results
As of March 15, 2023, a total of 10,567 people with MS had received at least 1 administration of ocrelizumab through the COMPASS program, with 10,055 having received at least 1 full dose. Of those, 138 unique cases and 140 reports of wearing-off effect (2 individuals reported wearing-off effect twice) were identified in the safety database and linked to COMPASS, representing 1.3% of the estimated total exposed population (Figure S1). Reports were primarily received by patients or other non–health care professionals (see Table S1).
Characteristics of the study and source cohorts are outlined in Table 1.
Patients reporting a wearing-off effect received an average of 7.4 infusions, with a mean time between full doses of 6.2 months. Generally, people with MS treated with ocrelizumab reporting wearing-off effect had similar disease and treatment characteristics as those of the total exposed population (Table 1). The wearing-off effect occurred in people with PPMS and RRMS in proportions similar to the overall study makeup (Table 1). The majority of individuals with PPMS were treatment naive prior to receiving ocrelizumab. Mean ocrelizumab dosing interval was similar between the groups, with the majority receiving ocrelizumab every 6 months with no variation (IQR, 0; Table 1). Duration of therapy and number of infusions received were higher in the group reporting wearing-off effect (Table 1).
All reports of wearing-off effect were nonserious (Table 2).
Overall, 11.4% of the total exposed population and 15.9% of people with MS reporting wearing-off effect discontinued treatment (Table 3).
The discontinuation rate was higher in those with wearing-off effect and PPMS vs RRMS (30.8% vs 12.5%, respectively; Table 3). Reasons for discontinuation in people with MS reporting wearing-off effect were similar to those of the overall COMPASS population (Table 3).
A comparison of people who reported wearing-off effect and continued vs discontinued ocrelizumab is outlined in Table S2.
The groups had a similar ocrelizumab dosing interval (every 6 months). People who reported wearing-off effect and discontinued ocrelizumab were more likely to be older, were more likely to be men, and had a lower overall duration of treatment than those who continued ocrelizumab.
AEs reported in conjunction with the wearing-off effect are detailed in Table 4. Generally, these were infrequent.
Discussion
We investigated reports of a wearing-off effect in people with MS in Canada by combining events from the global safety database with patient-level data from the national COMPASS database. Given that all people with MS receiving ocrelizumab outside of clinical trials in Canada receive it via COMPASS, the data set allowed for a reliable estimate of total Canadian exposure. This database, coupled with the Roche Global Safety Database, provided a useful means of analyzing the wearing-off effect phenomenon in Canada. We observed that wearing-off effects coded to TRS were reported by 1.3% of the studied population and were generally not treatment limiting, with 15.9% of people with MS who reported wearing-off effect discontinuing ocrelizumab.
Studies of wearing-off effect published to date have ranged from case series17 to prospective, phase 4, nonrandomized, multicenter trials (N = 376).26 When surveyed, people with MS treated with natalizumab have reported a wearing-off effect at a frequency of 37% to 67%.12-14,18,24,27-30 Similarly, a wearing-off effect has been reported in 51% to 61% of people with MS receiving ocrelizumab.15,31-33 These prospective studies related to ocrelizumab (88-100 participants) used a wearing-off questionnaire and noted the potential for overreporting due to enhanced recall and reporting bias.15,31,32
In contrast, our study presents the largest cohort of people with MS yet studied for the wearing-off effect (N = 10,055) using spontaneous AE reporting rather than topic-specific questionnaires. Spontaneous AE reporting has been shown to underrepresent the prevalence of an AE compared with active monitoring (or true incidence).34 Although in some cases, rare AEs may be subject to less underreporting,34 in the case of wearing off, underreporting may be amplified given its subjective nature and relative nonseriousness. Interestingly, a retrospective chart review from British Columbia, Canada, found a wearing-off phenomenon recorded in 16% (11 of 69) of patient charts.19 Our data show that the spontaneous reporting of wearing off in a large real-world population is lower than that found in prospective studies. We suspect the true incidence of wearing off lies somewhere between the rate of spontaneous reports seen in our study and the frequency of reports solicited in prospective studies.
In the prospective study by Toorop and colleagues, treatment satisfaction did not differ between those reporting wearing off and those not reporting it15; however, in the multi-mAB study noted previously, wearing off was associated with reduced treatment satisfaction scores, and 20% of participants noted they had considered switching to another DMT.33 Uniquely, our data report on treatment status and reasons for ocrelizumab discontinuation. Importantly, wearing-off events were not reported as the reason for ocrelizumab discontinuation, and most people with MS affected remained on treatment (84.1%). Furthermore, wearing-off events were confirmed to be nonserious in nature for all in our study. The majority received ocrelizumab at 6-month intervals, and overall duration of therapy was similar in the total COMPASS population as it was among those who reported wearing off and discontinued treatment.
The potential cause of wearing off remains unknown. Findings from studies related to ocrelizumab used at extended dosing intervals have shown a maintenance of clinical efficacy, even with altered B-cell repopulation.35,36 In our study, a median dosing interval of 6 months was observed in both people with MS reporting wearing-off effect as well as the total exposed population, which is in line with the recommended dosing interval of every 6 months.8,10,11 Given this consistency between groups, we were unable to observe any impact of dosing interval on wearing off.
Another theory to explain wearing off that has been explored is a decreased pharmacological effect. In the SymBOLS study (NCT04855617), no correlation was found between serum neurofilament light chain (sNfL) or ocrelizumab serum levels and symptom burden, suggesting wearing off is not a result of subclinical disease activity.31 Monteiro et al32 found wearing-off reports were associated with higher levels of CD8 and CD3/CD27 lymphocytes (ie, reduced immunomodulation) and increased plasma NfL levels (ie, neuroaxonal damage); however, these were not correlated with clinical or radiological disease activity nor ocrelizumab serum levels, so a causal effect was not established. Results from a cross-sectional study by Meca-Lallana et al found no correlation between sNfL, CD19+, baseline Expanded Disability Status Scale score, age, progression, or relapse in people with MS treated with ocrelizumab experiencing a wearing-off effect.37 Finally, Toorop and colleagues observed a correlation between higher BMI and a wearing-off effect that did not appear correlated with B-cell count.15 They speculated that higher rates of clinical disability experienced by people with MS and obesity may be responsible for greater MS symptom fluctuation.15
The prevailing hypothesis, based on current evidence, is that the wearing-off phenomenon may be a subjectively perceived lack of effect or disease-related symptom fluctuations misattributed to treatment.12-14,31 The low rates of unprompted reports observed in our study support this. In CASTING, a phase 3b study (NCT02861014), the patient-reported outcome tool SymptoMScreen, which includes elements such as fatigue, cognitive function, and sensory symptoms—all in common with wearing-off effect symptoms—was applied.38 People with MS who previously had a suboptimal response to 1 or 2 DMTs had significant improvement in SymptoMScreen scores, which were observed 24 weeks into ocrelizumab treatment and maintained over 96 weeks.38 Furthermore, results from the SymBOLS study of 103 people with MS treated with ocrelizumab failed to show correlation between wearing off and meaningful symptom changes throughout the infusion cycle as measured by Quality of Life in Neurological Disorders short forms, SymptoMScreen, and the Work Productivity and Activity Impairment Questionnaire at the beginning, middle, and end of the ocrelizumab treatment cycle.31 Importantly, the significant intra- and interindividual symptom variability associated with the MS disease course (regardless of treatment)39 is likely to affect the ability of prospective studies to objectively and consistently capture wearing-off phenomenon within a group over time. Results from a recent study of wearing-off effect with 4 mAbs found wearing off to be associated with symptoms of depression, which may support the theory that it is a psychological rather than biological phenomenon.33
The present work is characterized by some limitations. Inherent to the nature of safety databases, the extracted data require careful interpretation due to the varying quality of AE reports, which could in some instances have missing, nonverifiable, and/or incomplete information. This also precluded the analysis of comorbidities (such as BMI and laboratory parameters) that may overlap with wearing-off effect. However, the COMPASS data supplement many other data points potentially missing from AE reports.
The calculation of population-based AE rates was not feasible because voluntary AE reporting is unprompted and, particularly for nonserious events, prone to be underreported.34 Given the unsolicited and voluntary nature of the information retrieval, there is a possibility that some wearing-off symptoms were not reported, potentially leading to an underestimation of the true frequency as discussed previously. Furthermore, although the preferred term TRS captured a broad range of synonymous terms, other common MS symptoms (eg, fatigue) reported independently of the analyzed preferred term may not have been coded in the Roche Global Safety Database as TRS, further contributing to the potential underestimation of the frequency of wearing off.
Despite the above considerations, our data provide insights into a phenomenon that is difficult to objectively measure. Wearing-off event analysis can vary by the definition of the reported event, use of a biological/symptomatic surrogate, timing/interval of data collection, study design and resulting biases, and individual awareness of the effect. These are all factors at play when considering the results of this study in the context of the limited literature on the topic.
Although these findings support the provision of ocrelizumab according to the recommended dosing schedule, with no observed impact due to wearing-off effect, an opportunity to better understand the phenomenon remains. In addition to routine pharmacovigilance assessment for lack of efficacy (including wearing-off effect), longitudinal studies designed to minimize reporting bias and monitor symptom recrudescence and biological markers over the course of treatment may be insightful. Two ongoing phase 3 studies investigating ocrelizumab administered at higher doses may provide additional insights into symptom control under other clinical circumstances.40
Conclusions
In this large and representative population of people with MS in Canada treated with ocrelizumab, reports of wearing-off effect were infrequent (< 2%). Observed symptoms and treatment patterns in people with MS reporting wearing-off effect were nonserious and support the continuation of therapy according to label recommendations.