Methods

The trial was a phase 2, randomized, double‐blind, placebo‐controlled, 48‐week trial conducted at 32 centres in the United States and Canada

Study start date: December 2004

Study completion date: December 2006

Follow‐up period: 48 weeks

Participants

Inclusion criteria: (1) patients 18 to 55 years of age with a diagnosis of RRMS; (2) at least one relapse during the preceding year; (3) an entry score of 0 to 5.0 on the EDSS

Exclusion criteria: (1) disease categorized as secondary progressive, primary progressive, or progressive relapsing disease; (2) relapse within 30 days; (3) cyclophosphamide or mitoxantrone treatment within 12 months; (4) systemic corticosteroid therapy within 30 days; (5) treatment with interferon‐beta, glatiramer acetate, natalizumab, plasmapheresis, or intravenous immune globulin within 60 days; (6) or non‐lymphocyte‐depleting immunosuppressive therapies within 90 days

Baseline characteristics were generally balanced between the two groups except for the proportion of patients without gadolinium‐enhancing lesions (higher in the placebo group than in the rituximab group, 85.7% versus 63.8%, P = 0.02)

Summary of patient characteristics at baseline (placebo: G1 (n = 35), rituximab: G2 (n = 69)):

Age: G1 = 41.5±8.5 years, G2 = 39.6±8.7 years

Sex: female G1 = 29 (82.9%), G2 = 52 (75.4%); male G1 = 6 (17.1%), G2 = 17 (24.6%)

Disease duration: since onset G1 = 9.6±7.1 years, G2 = 9.6±6.4 years; since diagnosis G1 = 6.9±6.2 years, G2 = 6.2±5.2 years

Relapses in past year: 0 relapse G1 = 2 (5.7%), G2 = 4 (5.8%); 1 relapse G1 = 27 (77.1%), G2 = 52 (75.4%); 2 relapse G1 = 5 (14.3%), G2 = 8 (11.6%); 3 relapse G1 = 0, G2 = 4 (5.8%); ≥4 relapse G1 = 1 (2.9%), G2 = 1 (1.4%); median (range) G1 = 1.0 (0‐5), G2 = 1.0 (0‐4)

EDSS score: 0 G1 = 1 (2.9%), G2 = 2 (2.9%); 1.0‐1.5 G1 = 8 (22.9%), G2 = 9 (13.0%); 2.0‐2.5 G1 = 10 (28.6%), G2 = 24 (34.8%); 3.0‐3.5 G1 = 7 (20%), G2 = 20 (29.0%); 4.0‐4.5 G1 = 7 (20%), G2 = 11 (15.9%); 5.0 G1 = 2 (5.7%), G2 = 3 (4.3%); median (range) G1 = 2.5 (0‐5), G2 = 2.5 (0‐5)

Previous treatment with interferon‐beta or glatiramer acetate: none or discontinued >6 months before study entry G1 = 21 (60%), G2 = 44 (63.8%); treatment within 6 months before study entry G1 = 14 (40%), G2 = 25 (36.2%)

Any key therapy for MS in previous 2 years: glatiramer acetate G1 = 8 (22.9%), G2 = 18 (26.1%); interferon beta‐1a G1 = 16 (45.7%), G2 = 24 (34.8%); interferon beta‐1b G1 = 5 (14.3%), G2 = 13 (18.8%); methylprednisolone or methylprednisolone sodium succinate G1 = 8 (22.9%), G2 = 19 (27.5%); natalizumab G1 = 2 (5.7%), G2 = 5 (7.2%)

Any key therapy for MS: G1 = 27 (77.1%), G2 = 54 (78.3%)

Gadolinium‐enhancing lesions: 0 lesions G1 = 30 (85.7%), G2 = 44 (63.8%); 1 lesion G1 = 2 (5.7%), G2 = 7 (10.1%); 2 lesions G1 = 2 (5.7%), G2 = 4 (5.8%); 3 lesions G1 = 0, G2 = 2 (2.9%); ≥4 lesions G1 = 1 (2.9%), G2 = 11 (15.9%); mean lesions G1 = 0.3±0.8, G2 = 2.1±5.6; median lesions (range) G1 = 0 (0‐4), G2 = 0 (0‐36)

Volume of lesions detected on MRI: on gadolinium‐enhanced MRI, mean G1 = 29.2±127.5 mm³, G2 = 211.6±702.2 mm³; median G1 = 0, G2 = 0; on T2‐weighted MRI, mean G1 = 5723.1±5514.8 mm³, G2 = 6452.2±8022.2 mm³; median G1 = 4032.0, G2 = 2878.5; on T1‐weighted MRI, mean G1 = 717.6±1025.0 mm³, G2 = 784.2±1206.4 mm³; median G1 = 369.0, G2 = 211.0

Interventions

Experimental group: 1000 mg intravenous infusions of rituximab on study days 1 and 15, before each infusion, acetaminophen (at a dose of 1 g) and diphenhydramine hydrochloride (at a dose of 50 mg) were administered orally 30 to 60 minutes beforehand as premedication

Control group: 1000 mg intravenous infusions of placebo on study days 1 and 15, before each infusion, acetaminophen (at a dose of 1 g) and diphenhydramine hydrochloride (at a dose of 50 mg) were administered orally 30 to 60 minutes beforehand as premedication

Outcomes

The primary outcome measures: the sum of the number of gadolinium‐enhancing lesions on serial T1‐weighted MRI brain scans at weeks 12, 16, 20, and 24. Lesions that persisted for more than 4 weeks were counted more than once.

Secondary outcome measures: (1) the total number of new gadolinium‐enhancing lesions observed on serial T1‐weighted MRI brain scans at weeks 12, 16, 20, and 24 (lesions persisting for more than 4 weeks were counted only once); (2) the change from the baseline lesion volume on T2‐weighted MRI scans; (3) the proportion of patients with relapses at week 24 and week 48; (4) the annualised rate of relapse from week 0 to week 24 and week 0 to week 48; (5) safety. Relapse was defined as new or recurrent neurologic symptoms that were consistent with MS, lasted for at least 48 hours, and were preceded by a relatively stable or improving neurologic state for at least 30 days.

Notes

The study was originally designed to enroll 280 patients but before the primary endpoint at week 24, the sample‐size target was reduced to 99 patients. With such a sample size the study would have 70% power at a two‐sided significance level of 0.05.

The study was designed jointly by Genentech and the investigators and supported by Biogen Idec and Genentech. Data were collected by the investigators and held and analysed by Genentech. ClinicalTrials.gov Identifier: NCT00097188.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: "Patients were randomly assigned in a 2:1 ratio to receive rituximab or placebo, and they were hierarchically stratified according to study site, status with respect to previous treatment with interferon beta or glatiramer acetate (either no treatment or discontinuation of medication ≥6 months previously versus treatment within the previous 6 months), and baseline disease severity according to the Expanded Disability Status Scale (EDSS) score (≤2.5 versus >2.5)".

Comment: the authors did not mention the specific method of random sequence generation. We wrote to the principal author ([email protected]) for the methodological information, but we got no reply. The risk of selection bias was unclear.

Allocation concealment (selection bias)

Unclear risk

The authors did not mention the specific method of allocation concealment. We wrote to the principal author but we got no reply. The risk of selection bias was unclear.

Blinding (performance bias and detection bias)
All outcomes

Low risk

Quote: "This trial was double‐blinded. Each site had both a treating investigator and an examining investigator. The treating investigator was the safety assessor, and the examining investigator was the efficacy assessor. Staff members from a central MRI reading center who were unaware of the data evaluated all scans. Data were collected by the investigators and held and analyzed by Genentech."

Comment: the participants, personnel and the outcomes assessors in this study were blinded to the allocated interventions. The risks of performance bias and detection bias were low.

Incomplete outcome data (attrition bias)
All outcomes

High risk

According to the table showing study sample and reasons for study discontinuation, of the 104 patients 96 (92.3%) completed 24 weeks, and 79 patients (76.0%) completed 48 weeks; 66 (95.7%) in the rituximab group and 30 (85.7%) in the placebo group completed 24 weeks; 58 (84.1%) in the rituximab group and 21 (60.0%) in the placebo group completed 48 weeks; 11 patients (15.9%) in the rituximab group and 14 patients (40.0%) in the placebo group discontinued before week 48. Reasons for discontinuation (placebo: G1, rituximab: G2): death G1 0, G2 1 (1.4%); adverse events G1 = 0, G2 = 1 (1.4%); pregnancy G1 = 1 (2.9%), G2 0; lost to follow‐up G1 = 2 (5.7%), 2 (2.9%); patient's decision G1 = 4 (11.4%), G2 = 0; physician's decision G1 = 3 (8.6%), G2 = 3 (4.3%); relapse G1 = 2 (5.7%), G2 = 2 (2.9%); initiation of excluded therapy G1 = 2 (5.7%), G2 = 2 (2.9%). A last observation carried forward imputation strategy was used for the missing data. The intention‐to‐treat principle was used for analyses.

Comment: a risk of attrition bias could have resulted from the high rate of dropouts (24%) over the period of 48 weeks and the imbalance in the reasons for discontinuation between the two groups. Although a last observation carried forward imputation strategy was used for the missing data and the intention‐to‐treat principle was used for analyses, the high rate of dropouts had a potential impact on the results. Generally, the higher the ratio of participants with missing data to participants with events, the greater potential there is for bias, especially for the high frequency of events. The potential impact of missing continuous outcomes increases with the proportion of participants with missing data. Both MS relapse and the occurrence of gadolinium‐enhancing T1‐weighted lesions are common events. The risk of attrition bias was high.

Selective reporting (reporting bias)

Low risk

All listed outcomes were reported adequately. The risk of reporting bias was low

Other bias

Unclear risk

Quote: "The study was designed jointly by Genentech and the investigators"

Comment: conflict of interests may exist

Quote: "At baseline the proportion of patients without gadolinium‐enhancing lesions was greater in the placebo group than in the rituximab group (85.7% versus 63.8%, P = 0.02)"

Comment: this may have led to biased results