Safety and efficacy of rituximab as first- and second line treatment in multiple sclerosis – A cohort study

Background Rituximab is increasingly used as off-label therapy in multiple sclerosis (MS). More data are needed on safety and efficacy of rituximab, particularly in cohorts of de novo patients and patients in early therapy escalation. Objective To investigate the safety and efficacy of off-label treatment with rituximab in an MS-cohort of predominantly de novo patients or as therapy escalation. Methods We retrieved safety and efficacy data from the Norwegian MS-registry and biobank for all MS-patients treated with rituximab at Haukeland University Hospital, Bergen, Norway, during a four year period. Results In the 365 MS-patients (320 relapsing-remitting MS (RRMS), 23 secondary progressive MS (SPMS), and 22 primary progressive MS (PPMS)), the overall annualized relapse rate (ARR) was 0.03 and annualized drug discontinuation rate (ADDR) was 0.05. NEDA-3 was achived in 79% of patients with available data (n=351). Sixty-one patients experienced infusion-related adverse events of which two were serious (CTCAE grade 3–4). Eighteen patients experienced serious non-infusion related adverse events, of which 16 were infections. Infections (n = 34; 9.3%, CTCAE grade 2-5), hypogammaglobulinemia (n = 19, 5.2%) and neutropenia (n = 16; 4.4%) were the most common non-infusion-related adverse events. Conclusion Rituximab was a safe and highly efficient disease modifying therapy in this cohort of MS-patients; however, infections and neutropenia need to be monitored.


Introduction
Increasing evidence suggests that B-cell depletion is a safe and highly effective therapy option in multiple sclerosis (MS). 1 More than a decade ago, the first randomized, controlled phase II trial of an anti-CD20-antibody (rituximab) demonstrated high efficacy in MS. 2 Rituximab was not included in further development programs in relapsing-remitting MS (RRMS), but the humanized monoclonal anti-CD20-antibody ocrelizumab has recently proved effective in phase II/III trials, both in RRMS and primary progressive MS (PPMS). 3,4 Emerging real world data indicate a comparable and possibly superior effect and tolerability of rituximab compared to other standard MS-treatments. [5][6][7][8][9][10][11][12] Rituximab has to a large extent been used as a lastresort MS-treatment, but high efficacy treatments are now more frequently used at an early disease stage. At Haukeland University Hospital, Western Norway, rituximab has been the preferred therapeutic option for highly active MS in treatment naive patients since 2016, as well as an equivalent option for therapy switch or escalation in MS. Hence, rituximab has become the most commonly used MSimmunomodulatory therapy at our hospital. Based on the experience from this treatment approach, we conducted a retrospective cohort study to describe safety and efficacy of rituximab therapy in MS patients at Haukeland University Hospital during 2015-2019.

Material and methods
This study was a single-centre, retrospective cohort study of all MS-patients who initiated rituximab therapy from 01.01.15 until 01.01. 19 and consented to registration in the National MS registry at the Department of Neurology, Haukeland University Hospital, Bergen, Norway. Patients were identified through the Norwegian MS registry and crosschecked with the Department's list of patients treated with rituximab. From 2016 rituximab became the drug of choice for highly active disease both in therapy escalation and in treatment naïve patients, i.e. patients with two or more of the following characteristics: young age, severe function loss or extensive MRI-lesions. Observation time was set from the date of treatment initiation (at latest 01.01.19) until 01.07.19 if rituximab was not discontinued. If patients discontinued rituximab, observation time was set until six months after the latest rituximab infusion, or the date of initiation of a new disease modifying therapy (DMT).

Study variables
Data were retrieved from the Norwegian MS Registry and biobank and included age, sex, date of disease onset, number of relapses in total and during the last two years prior to rituximab therapy, previous treatment, and reason for initiation or switch to rituximab. Efficacy recording included date of any relapse or new magnetic resonance imaging (MRI) T1 gadolinium henhancing (T1Gdþ) lesion, or new T2 lesion, and change in Kurtzke Expanded Disability Status Scale (EDSS) from baseline to latest observation, at least six months from rituximab initiation. Safety recording included date and reason for rituximab discontinuation, and adverse events.

Treatment regimen and monitoring
The majority of patients received a single infusion of rituximab 1000 mg intravenously (i.v.) at initiation, followed by 500 mg i.v. every 6 months. Patients who started rituximab treatment between 01.01.2015-31.12.2016 normally received a dose of total 2000 mg i.v. at initiation (given as two single infusions of 1000 mg i.v. within an interval of two weeks) followed by 500 mg i.v. every six months (Table 1).
MRI scans were performed on 1.5 T or 3 T scanners. The most recent MRI before rituximab initiation was defined as the baseline MRI scan. A re-baseline scan was planned 6 months after rituximab treatment initiation, followed by a new MRI scan at 12 months, and further every 12 months, or if clinically indicated. Brain MRI scan was performed routinely, and a spinal MRI scan was performed if clinically indicated. An EDSS was planned as part of clinical evaluation in the outpatient clinic at latest six months after rituximab initiation, and at regular follow-up every 6-12 months. Laboratory tests including hemoglobin (Hb), white blood cells with differential count, platelet count, ESR (erythrocyte sedimentation rate), CRP (C-reactive protein), liver function parameters (GGT (Gamma-Glutamyl Transferase), ALT (alanine transaminase, AST (Aspartate transaminase), ALP (alkaline phosphatase)), kidney function (creatinine), IgG and IgM (immunoglobuline G and M) were performed every 6 months.

Study outcomes
Efficacy outcomes included (i) annual relapse rate, (ii) number of patients experiencing new MRIdisease activity during observation, defined as new T2 or T1Gdþ lesions, and (iii) disability progression, measured as change in EDSS (0.5 points or more), and (iv) proportion of patients with no evidence of disease activity (NEDA-3). NEDA-3 was defined as a composite score comprising absence of clinical relapses and disability progression, in addition to no new MRI disease activity (new T1Gdþ or new/enlarging T2-lesions) on MRI examinations for the given period. Treatment outcomes were evaluated for the period before the re-baseline MRI (i.e. the first six months of therapy) and after the re-baseline MRI (i.e. >six months of therapy). Safety outcomes included all reported adverse events, including infusion-related adverse events, except mild infections (Common Terminology Criteria for adverse events (CTCAE) grade 1), which were not included due to suspected incomplete registration in the registry and patient records.

Statistical methods
Survival (relapse-free survival, MRI event-free survival, progression free survival) was estimated with Kaplan-Meier survival curves. Analyses were performed in SPSS version 24 (IBM Corp., Armonk, NY). Descriptive data were described as means, medians and percentages.
Approval and patient consent All data were retrieved from the Norwegian MS Registry and biobank, where participants have given a prospective informed consent. The study was approved of by Regional Committee for Medical Research Ethics, Western Norway (REK no 87388).

Patient characteristics
A total of 365 MS-patients (320 RRMS, 23 SPMS, 22 PPMS) who initiated disease modifying treatment in Haukeland University Hospital, Department of neurology with rituximab during the predefined study period, and consented to registration in the Norwegian MS registry, were included. One patient (n¼1, RRMS) was identified through the Department's list of patients but had refused consent to the registry and was not included. Table 2 presents baseline characteristics. Mean (SD) age was 42.3 (AE 12.2) years and median (range) disease duration was 5.3 (AE 7) years at rituximab initiation. Mean observation time was 610 days (AE277).

Treatment response
Relapse rate, frequency of new MRI lesions, EDSS change and drug discontinuation rate were calculated and are all displayed in Table 3.
Observation time for patients in the study was <12 months in 74 patients, and !12 months in 291 patients. In 133 patients, observation time was !24 months.
Clinical relapses. The overall annualized relapse rate (ARR) in this cohort was 0.03 (RRMS 0.03, SPMS 0.03, PPMS 0). During the first six months of treatment, fourteen patients experienced a relapse of which two had a further relapse. After six months   Two patients discontinued rituximab due to treatment failure; both experienced relapses and more than one MRI event.  Disability progression. EDSS was available at baseline and after at least 6 months of treatment for 352 patients (96%). In 316 (89.8%) patients EDSS was unchanged between their first and last available score. In RRMS patients, EDSS improved in 18 (5.7%), and worsened in ten (3.2%) patients. In the progressive disease subgroup, EDSS improved in one (3.4%) and worsened in seven (16.7%) patients. In progressive disease courses, observation time was less than a year in 13 patients, 1-2 years in 32 patients, and ! 2 years in 13 patients. Drug survival. Treatment was discontinued in 33 (9%) patients, resulting in an annual drug discontinuation rate (ADDR) of 0.05 (RRMS 0.04, SPMS 0.20, PPMS 0.10). The most frequent reason for discontinuation was adverse events (n ¼ 18, 4.9%, CTCAE grade 1-2 in 14 cases, grade 3-4 in 4 cases). In two patients (0.5%), discontinuation was due to treatment failure (new MRI-lesions and clinical relapses). Six patients (1.1%) requested rituximab discontinuation in order to receive HSCT abroad. Figure 2 demonstrates MRI event-free survival, relapse-free survival and drug survival on rituximab during observation time.
Safety. Safety data results are displayed in Table 4. 118 patients (32.3%) experienced a total of 156 registered adverse events.
Subgroup analyses. In patients with neutropenia (n ¼ 16), 5 patients were registered with any infection as an adverse event. In patients without neutropenina (n¼349), 31 patients were registered with an infection.

Discussion
We present one of the largest cohorts of MS-patients treated with rituximab to date, utilizing real world data from the Norwegian MS Registry.
Efficacy data are in line with results from several studies of similar design 6,8,12 including a larger cohort study of similar design and study population. 10 Another large, recent multicentric observational study reported similar results; one difference to notice is the higher proportion of treatment naive patients in our cohort, pointing to the differences in therapeutic approach and patient selection in reported off-label MS-treatment with rituximab. Our cohort includes a relatively high proportion (28.5%) of newly diagnosed patients starting rituximab as their first choice DMT compared to many studies based on real-world data.
The overall ARR of 0.03 was low, and so was the ARR of 0.02 calculated in >6 months of treatment, when assuming a transient initial periode of suboptimal treatment effect. The calculated ARR for patients receiving a higher starting dose of rituximab was 0.05; this subgroup of patients also started treatment earlier (2015-2016) while rituximab was not concidered a first choice treatment, and could possibly represent a subgroup with more persistent disease activity in the cohort.
The total number of patients with new MRI lesions was lowafter initiation of rituximab, and after >6 months of treatment. A large proportion of patients starting rituximab had a recent history of disease activity (newly diagnosed patients, recent treatment failure on other DMTs, recent disease progression without ongoing therapy), which strengthens the efficacy results and indicate low inflammatory disease activity during rituximab treatment, further underlined by the low proportion of patients discontinuing treatment, and a very low proportion of patients discontinuing rituximab due to new disease activity.
EDSS was mainly unchanged during rituximab treatment. Notably, the baseline EDSS score might have been transiently worsened in some patients at rituximab initiation due to recent relapses, which could lead to overestimation of improvement in EDSS during treatment. EDSS was recorded during regular clinical follow up and a validation of change in EDSS score was not demanded in the study design.
We recorded a clinically significant number of patients in progressive disease courses registered with EDSS progression during rituximab treatment (7 out of 44 in SPMS and PPMS with valid EDSS scores). This is in line with previous studies reporting that disability progression may be reduced, but not completely stopped, during rituximab or ocrelizumab treatment in progressive MS disease courses. 4,14 For progressive disease courses, the observation time was relatively short, and the number of patients (n ¼ 45) was low, which limits further conclusions.
One of the main goals was to evaluate safety of rituximab treatment in MS. The rates of adverse events and the main categories of side effects in this cohort were comparable to those seen in populations of rheumatological and other autoimmune diseases. 15,16 Among serious adverse events, infections was the most common, corroborating a recent report from a Swedish MS-population. 17 Moderate and severe neutropenia (count <1.0 x 10 9 ) was seen in 1.6% of patients in this cohort and may represent a serious risk to the patient's health. As neutrophil counts were registered only at baseline and every 6 months, we could not estimate the onset time of neutropenia or the full extent. Late onset neutropenia is a well-known side effect of rituximab treatment 18 but has only been documented in few MS-patients treated with rituximab in retrospective studies. 19,20 Our results indicate that neutropenia may be underreported if not monitored.
The present study has several strengths. As a singlecentre study, the treatment-and observation protocol was standardized, and all safety data and clinical records were fully available. Thus, serious adverse events should be well monitored and recorded, including laboratory monitoring and all hospitalizations, though recording of mild adverse events offer challenges. The dosing regimen in this cohort was largely homogenous, only a few patients (12.1%) received a higher initial dosing, and almost 90% of the patients receiving the regular maintenance dose regimen with no postponed doses. The study population included a relatively high proportion of treatment naive patients, contributing novel data for this subgroup. An important limitation of the present study is the retrospective, uncontrolled design, without adjustments for possibly confounding factors. The relatively short observation time, especially in progressive disease, is another limitation, and a control group would have strengthened the study.
Our data indicate that rituximab is a safe and highly effective treatment option in MS, both as first-line and escalation therapy. The extent of treatment with rituximab is still limited, most likely because it is off-label therapy, and because of the lack of phase III treatment study documentation. Systematic safety documentation is therefore important. Infections, including serious infections, seem to represent an important safety issue during rituximab therapy in MS, and the frequency of both mild and severe neutropenia could be underestimated as there is a lack of knowledge about onset time, and patients might be asymptomatic. Further studies should focus on treatment naive patients, and also how to prevent infections through improved screening or adjusted dosing or dosing intervals.