ORIGINAL ARTICLES

Rituximab in Patients with Primary

Progressive Multiple Sclerosis
Results of a Randomized Double-Blind Placebo-Controlled
Multicenter Trial
Kathleen Hawker, MD,1 Paul O’Connor, MD,2 Mark S. Freedman, MD,3 Peter A. Calabresi, MD,4
Jack Antel, MD,5 Jack Simon, MD,6 Stephen Hauser, MD,7 Emmanuelle Waubant, MD,7
Timothy Vollmer, MD,8 Hillel Panitch, MD,9 Jiameng Zhang, PhD,10 Peter Chin, MD,10 and
Craig H. Smith, MD,10 for the OLYMPUS trial group

Objective: Rituximab, a monoclonal antibody selectively depleting CD20⫹ B cells, has demonstrated efficacy in reducing
disease activity in relapsing-remitting multiple sclerosis (MS). We evaluated rituximab in adults with primary progressive MS
(PPMS) through 96 weeks and safety through 122 weeks.
Methods: Using 2:1 randomization, 439 PPMS patients received two 1,000mg intravenous rituximab or placebo infusions
every 24 weeks, through 96 weeks (4 courses). The primary endpoint was time to confirmed disease progression (CDP), a
prespecified increase in Expanded Disability Status Scale sustained for 12 weeks. Secondary endpoints were change from baseline
to week 96 in T2 lesion volume and total brain volume on magnetic resonance imaging scans.
Results: Differences in time to CDP between rituximab and placebo did not reach significance (96-week rates: 38.5% placebo,
30.2% rituximab; p ⫽ 0.14). From baseline to week 96, rituximab patients had less ( p ⬍ 0.001) increase in T2 lesion volume;
brain volume change was similar ( p ⫽ 0.62) to placebo. Subgroup analysis showed time to CDP was delayed in rituximab-
treated patients aged ⬍51 years (hazard ratio [HR] ⫽ 0.52; p ⫽ 0.010), those with gadolinium-enhancing lesions (HR ⫽ 0.41;
p ⫽ 0.007), and those aged ⬍51 years with gadolinium-enhancing lesions (HR ⫽ 0.33; p ⫽ 0.009) compared with placebo.
Adverse events were comparable between groups; 16.1% of rituximab and 13.6% of placebo patients reported serious events.
Serious infections occurred in 4.5% of rituximab and ⬍1.0% of placebo patients. Infusion-related events, predominantly mild
to moderate, were more common with rituximab during the first course, and decreased to rates comparable to placebo on
successive courses.
Interpretation: Although time to CDP between groups was not significant, overall subgroup analyses suggest selective B-cell
depletion may affect disease progression in younger patients, particularly those with inflammatory lesions.
Ann Neurol 2009;66:460 – 471

Multiple sclerosis (MS), a chronic inflammatory demy- Relapsing forms of MS are characterized by periods of
elinating disease of the central nervous system, is a neurological deterioration (relapses) and partial or
leading cause of neurologic disability in young adults.1 complete recovery (remission), with (secondary pro-

1
Department of Neurology, The Ohio State University Medical
Center, Columbus, OH; 2Department of Medicine, University of
Toronto, Toronto, ON, Canada; 3Department of Medicine, Ottawa
Hospital-General Campus, Ottawa, ON, Canada; 4Department of
Neurology, John Hopkins University, Baltimore, MD; 5Department
of Neurology and Neurosurgery, McGill University, Montreal, QC,
Canada; 6Department of Neurology, Portland VA Medical Center,
Portland, OR; 7Department of Neurology, University of California,
San Francisco, San Francisco, CA; 8Department of Neurology, Uni-
versity of Colorado, Denver, CO; 9Department of Neurology, Uni-
versity of Vermont, Burlington, VT; and 10Genentech Inc., South
San Francisco, San Francisco.
Address correspondence to Dr Hawker, Eli Lilly, 450 South Merid-
ian Street, Indianapolis, IN 46225, E-mail: hawkerks@lilly.com
Potential conflict of interest: Dr Hawker has received honoraria
from Biogen IDEC, Teva, Bayer, Genentech and GSK; Dr
O’Connor has received honoraria from Bio MS, BiogenIDEC,
Teva, Serono, Roche, Genentech, Bayer; Dr Freedman has received
honoraria from Sanofi-Aventis, BiogenIdec, EMD Serono, Teva,
Bayer Healthcare, Eli Lilly, Eisai, Novartis, Genentech; Dr Calabrasi
has received honoraria for consulting from: Teva, Biogen-IDEC,
Genentech, Serono, Novartis, Amplimmune, Vertex, and Centacor,
PC has received research support from Bayer, Serono, Biogen-
IDEC, Teva, and Genentech; Dr Waubant’s research has been sup-
ported by Biogen-Idec, Pfizer, Sanofi-Aventis, the National MS So-
ciety, the Immune Tolerance Network and the Nancy Davis
Foundation. Dr Waubant has been a consultant for Artielle and
Zymogenetics. She has received honorarium for one educational
presentation from Biogen Idec; Dr Hauser, Dr Panitch, Dr Vollmer
have nothing to disclose.
Additional Supporting Information may be found in the online ver-
sion of this article.
Other investigators who participated in the OLYMPUS Trial Group
are listed in the Appendix on page xxx.
Received Aug 5, 2009, and in revised form Aug 19. Accepted for
publication Sep 1, 2009. Published online in Wiley InterScience
(www.interscience.wiley.com). DOI: 10.1002/ana.21867

460 © 2009 American Neurological Association

 MRI:

Study
Drug
Dose:

1000 1000 1000 1000

mg x2 mg x2 mg x2 mg x2

Week: -4 -2 0 2 24 26 48 50 72 74 96 122

Treatment Follow-up
Screen Pre-Treatment 96 wks 26 wks
2 wks 2 wks

Fig 1. Study design. The study consisted of 4 periods: a screening period (week ⫺4 to week ⫺2 [days ⫺28 to ⫺15]), a pretreat-
ment period (week ⫺2 to week 0 [days ⫺14 to ⫺1]), a treatment period (week 0 [day 1] to week 96), and a follow-up period
(weeks 97 to 122). The first course of study drug treatment was administered on days 1 and 15. Patients were dosed every 24
weeks, with the last course of treatment administered at weeks 72 and 74. When patients completed the 96-week treatment period,
they entered the follow-up period. Patients who withdrew or discontinued from the study early were followed for collection of safety
data for 48 weeks from the time of their last dose of study drug. Primary endpoint: time to confirmed disease progression during
96-week treatment period (requires 12-week confirmation); safety and tolerability; subgroup analysis for age and gadolinium lesion.
Secondary endpoint: change in T2 lesion volume from baseline to week 96; change in brain volume from baseline to week 96.
MRI ⫽ magnetic resonance imaging.

gressive MS) or without (relapsing remitting MS
[RRMS]) progression between exacerbations. RRMS is
the most common type, seen in ⬎80% of patients.2
The remaining MS patients present with gradually in-
creasing neurological disability, primarily affecting am-
bulation, without discernable relapses; this type, pri-
mary progressive MS (PPMS), has a mean age of onset
a decade later than RRMS.1,3 To date, there are no
proven disease-modifying treatments that slow the
course of PPMS. A double-blind placebo-controlled
trial with glatiramer acetate4 and small exploratory tri-
als with interferon agents5,6 failed to demonstrate effi-
cacy in PPMS patients. An effective treatment for
PPMS remains an unmet medical need.
In contrast to earlier disease concepts suggesting that
pathogenic T cells were sufficient for full expression of
MS, autoimmune B cells and humoral immune mech-
anisms are now believed to play key roles in disease
pathogenesis.7 A humoral component in MS has been
implicitly recognized for decades, evidenced by inclu-
sion of cerebrospinal fluid (CSF) oligoclonal bands and
increased intrathecal immunoglobulin (Ig) G synthesis
in MS diagnostic criteria.8 –10 Antibody deposition and
immune complement activation associated with vesicu-
lar disintegration of the myelin membrane is present in
most MS lesions,11–13 and autoantibody responses
against many antigens can also be detected in the CSF
of many MS patients.14 The evidence for diffuse pa-
thology with demyelination and axonal loss but less
overt inflammation may indicate an antibody-mediated
pathogenesis in PPMS.15
To date, therapies partially effective in relapsing
forms of MS have failed to alter the course of PPMS.
Based on what we know about the immunopathology
of MS, and our hypothesis that PPMS is at least in
part a B-cell–mediated disease, we initiated a 96-week
double-blind placebo-controlled trial with rituximab,
an anti-CD20 monoclonal antibody. Rituximab (Rit-
uxan) is a genetically engineered chimeric monoclonal
antibody that targets the CD20 antigen expressed on B
lymphocytes16 from the pre–B-cell stage through dif-
ferentiation into B cells, but excluding plasma cells.
Our objectives in this trial (OLYMPUS) were to as-
sess the efficacy of rituximab relative to placebo, as
measured by the time to confirmed disease progression
over a 96-week treatment period, and to evaluate the
safety and tolerability of rituximab in patients with
PPMS.
Subjects and Methods
Study Design and Randomization
This was a phase II/III randomized, double-blind, placebo-
controlled trial from 60 centers in the US and Canada de-
signed jointly by Genentech, Biogen Idec, and the investiga-
tors. The protocol was approved by the institutional review
board and the ethics committee of each institution. Written
informed consent was obtained from each patient or the pa-
tient’s legal guardian. Patients were randomly assigned 2:1 to
receive rituximab or placebo (Fig 1), and hierarchically strati-
fied according to study site, previous MS therapies with
interferon-beta or glatiramer acetate (no prior MS therapy,
MS therapies that ended ⬎90 days prior to randomization,
MS therapies that ended within 90 days prior to randomiza-
tion), and baseline disease severity according to the Expanded
Disability Status Scale (EDSS) score (ⱕ4.0 vs ⬎4.0), an or-
dinal scale ranging from 0 (normal neurologic examination) to
10.0 (death) in 0.5-step intervals.17 Patients received 1,000mg
intravenous infusions of rituximab or placebo on study weeks
0, 2, 24, 26, 48, 50, 72, and 74 (see Fig 1). Data collected by
the investigators was held and analyzed by Genentech. Mem-
bers of the publication committee had full access to data, and

Hawker et al: Rituximab in PPMS 461

all authors vouch for the veracity and completeness of the data
and data analysis. The study was powered to detect a 50%
relative reduction in the 2-year disease progression rate, with
about 90% power and 0.01 type I error level. The assumed
confirmed disease progression rate for the placebo arm was
32% at 96 weeks, estimated from the 2-year placebo disease
progression rate observed in the PROMiSe trial.4
Patients
Enrollment was limited to patients 18 to 65 years old with a
PPMS diagnosis18; disease duration of ⱖ1 year; EDSS base-
line between 2.0 and 6.5 points, inclusive; Functional Sys-
tems scale score of ⱖ2.0 for the pyramidal system or gait
impairment due to lower extremity dysfunction; and pres-
ence of IgG oligoclonal bands or elevated CSF IgG index, or
both, from CSF specimen obtained at screening or docu-
mented during the previous 24 months. Key exclusion crite-
ria included history of MS exacerbation or neuromyelitis op-
tica; history of myelopathy or neurodegenerative central
nervous system conditions; systemic autoimmune disorders;
recurrent or chronic infections; recent treatment with immu-
nomodulating or immunosuppressant therapies; and meta-
bolic, hematologic, or immunologic laboratory abnormalities.
Study Procedures and Endpoints
The primary efficacy outcome measure was time to confirmed
disease progression (CDP), defined as a sustained EDSS in-
crease of ⱖ1.0 point from baseline EDSS if the baseline EDSS
was between 2.0 and 5.5 points (inclusive), or an EDSS in-
crease of ⱖ0.5 point if the baseline EDSS was ⬎5.5 points,
for which change was not attributable to another etiology (eg,
fever, concurrent illness, injury, adverse reactions to concur-
rent medications, or relapse) sustained for ⱖ12 weeks. Sec-
ondary efficacy outcome measures were change from baseline
to week 96 in volume of T2 lesions and change in brain vol-
ume, as measured by brain parenchymal fraction, on magnetic
resonance imaging (MRI) brain scans. Key exploratory out-
comes included time to CDP sustained for ⱖ24 weeks;
change in Multiple Sclerosis Functional Composite (MSFC)
summary and component scale scores; pharmacodynamic as-
sessments of B-cell counts and immunoglobulin levels; and
prespecified subgroup analysis by age, gender, race/ethnicity,
use of prior therapy, EDSS at baseline, duration since MS
symptom onset, and presence of gadolinium-enhancing lesions
on brain MRI at baseline. EDSS and MSFC scores were as-
sessed at least quarterly by a neurologist blinded to treatment
and other patient data including imaging and laboratory re-
sults. MRI scans were blindly evaluated at the University of
Colorado Brain Imaging Research Laboratory, an independent
MRI reading facility. An independent Data Monitoring Com-
mittee met formally every 3 months to review unblinded
safety data, including MRI outcomes, clinical outcomes, and
adverse events. Patients had an extra EDSS determination and
MRI assessment 4 weeks after completing the first course of
study drug treatment for safety to assess for activation or wors-
ening of MS. An interim futility analysis was carried out after
about 50% patients finished 48 weeks of treatment.
Statistical Analysis
Efficacy outcomes were analyzed according to the intention-
to-treat principle and included all patients randomized.
462 Annals of Neurology Vol 66 No 4 October 2009
Safety assessments were made on patients who received any
amount of study drug (rituximab or placebo).
For the primary efficacy analysis, the stratified log-rank
test, using pretreatment severity measured by EDSS score
and status of prior interferon-beta or glatiramer acetate ther-
apies for MS as stratifying factors, was applied to compare
differences in time to CDP between the 2 treatment groups.
The alpha level at the time of the final analysis was adjusted
to 0.0499 for the interim futility analysis at 48 weeks to
maintain an overall type I error level of 0.05. A Cox propor-
tional hazard model using the same stratifying factors as the
primary analysis was used to estimate the reduction in hazard
associated with treatment compared with placebo. Ranked
analysis of variance was used to compare the difference be-
tween the rituximab and placebo group in endpoints related
to change from baseline.
There was no imputation of missing data for assessment of
time to CDP. If a patient missed a scheduled visit that was
84 days after the initial detection of disease progression,
CDP was based on the assessment at the next scheduled visit.
Patients with an initial disease progression who subsequently
discontinued the study treatment before a subsequent confir-
matory assessment could be obtained were considered to
have CDP. Data from patients who discontinued the study
early or were lost to follow-up and had not shown an initial
disease progression were censored at time of their last visit.
Patients who demonstrated an initial disease progression at
the end of the treatment period (week 96) were censored at
that time and were not considered to have CDP. The EDSS
assessments within 30 days of protocol-defined relapse were
excluded for the confirmation of CDP.
Hochberg-Bonferroni procedure was used to control the
type I error level in testing the 2 secondary endpoints. No
statistical adjustment for multiple comparisons was made for
exploratory analyses. For endpoints related to change from
baseline, the last-observed value before the treatment period
discontinuation was used to impute the missing values. Sub-
group analysis for continuous variables were conducted using
median values as cut-points, as well as additional prespecified
thresholds. Adverse events were graded according to the Na-
tional Cancer Institute Common Terminology Criteria for
Adverse Events, version 3.0.19 Summaries of adverse events,
serious adverse events, deaths, laboratory test results, vital
signs, and antibodies to rituximab were presented.
Results
Study Population
A CONSORT diagram showing the flow of partici-
pants through each stage of this randomized, placebo-
controlled trial is illustrated in Figure 2. Baseline char-
acteristics (Table 1) of 439 patients were balanced
between treatment groups. Patients had longstanding
disease, with ⬃65% having no prior MS therapies;
mean EDSS score was 4.8, and over half had an EDSS
score of ⱖ4.0. Approximately 50% of study patients
were males, and the median age of patients was 51
years. At baseline, 25% of patients had gadolinium le-
sions. Of 439 patients, 84.4% in the placebo group
and 82.5% in the rituximab group completed the 96
 608 patients assessed for
eligibility 138 not meeting
inclusion criteria
11 withdrew consent
1 lost to follow up
19 other reasons
439 patients enrolled and
randomized to study groups

292 assigned to receive rituximab 147 assigned to receive placebo

(ITT population) (ITT population)

2 lost to follow up
2 lost to follow up
49 withdraw treatments
8 had adverse event 21 withdraw treatments
27 patient’s decision 1 death
7 physician’s decision 11 patient’s decision
1 init of excluded therapy 4 physician’s decision
1 subject non-compliance 5 disease progression
1 pregnancy
4 disease progression

292 patients included in the 147 patients included in the

analysis analysis

Among them, 241 completed 96 Among them, 124 completed 96

weeks treatment period weeks treatment period

Fig 2. CONSORT diagram showing flow of participants through each stage of the OLYMPUS trial. ITT ⫽ intention-to-treat
population.

weeks (Table 2). Of 365 patients who completed the
96-week treatment period, 340 completed the 122-
week safety follow-up (93.5% placebo, 92.9% ritux-
imab). Six (2.05%) patients in the rituximab group
and 5 (3.40%) patients in the placebo group had
protocol-defined relapses during the trial.
Efficacy
PRIMARY ENDPOINT
During the 96-week treatment period, there was no ev-
idence of significant difference in time to CDP between
the rituximab and placebo groups ( p ⫽ 0.1442; Table
3, Fig 3A), although rituximab tended to delay the time
to CDP compared with placebo as measured by a strat-
ified hazard ratio of 0.77. Week 96 CDP rates were
38.5% for the placebo and 30.2% for the rituximab
group.
SECONDARY ENDPOINTS
From baseline to week 96, compared with patients re-
ceiving placebo, those receiving rituximab had signifi-
cantly less ( p ⬍ 0.001) increase in T2 lesion volume
on brain MRI scans, whereas brain volume change (de-
crease) was similar ( p ⫽ 0.62; see Table 3). The me-
dian increase in T2 lesion volume was 301.95 mm3 for
rituximab and 809.50 mm3 for placebo.
KEY EXPLORATORY ENDPOINTS
Compared with placebo, rituximab patients had less
worsening in the MSFC timed 25-foot walk test ( p ⫽
0.04, p ⫽ 0.076, and p ⫽ 0.015 at weeks 48, 96, and
122, respectively, see Table 3). Median increase from
baseline to week 96 in the timed 25-foot walk test was
1.48 seconds for placebo patients and 0.9 seconds for
rituximab patients. Results for other key exploratory
endpoints were similar between treatment arms (time
to CDP sustained for ⱖ24 weeks; change of MSFC
score; 9-Hole Peg Test; Paced Auditory Serial test from
baseline to weeks 48, 96, and 122).
SUBGROUP ANALYSES
Planned subgroup analyses of the primary endpoint (see
Fig 3B) indicated that age and presence of gadolinium
lesions were predictors for treatment effect. Baseline char-
acteristics of these patients are listed in the Supplemental
Table. Compared with placebo, time to CDP was delayed

Hawker et al: Rituximab in PPMS 463

 Table 1. Patient Baseline Demographic Characteristics
Demographic Placebo, nⴝ147 Rituximab, nⴝ292 Total, Nⴝ439

Age, mean yr (SD) 49.6 (8.7) 50.1 (9.0) 49.9 (8.9)

Gender, No. (%)
Male 66 (44.9) 152 (52.1) 218 (49.7)
Female 81 (55.1) 140 (47.9) 221 (50.3)
Ethnicity, No. (%)
White 134 (91.2) 268 (91.8) 402 (91.6)
Other 13 (8.8) 24 (8.2) 37 (8.4)
Disease duration, mean yr (SD)
Since onset 9.0 (6.8) 9.2 (6.4) 9.1 (6.6)
Since diagnosis 3.8 (4.2) 4.1 (4.2) 4.0 (4.2)
Prior MS treatment with IFN-beta or glatiramer acetate, No. (%)
No prior MS therapies 96 (65.3) 189 (64.7) 285 (64.9)
Ended ⬎90 days before trial entry 45 (30.6) 87 (29.8) 132 (30.1)
Ended ⱕ90 days before trial entry 6 (4.1) 16 (5.5) 22 (5.0)
EDSS score
Median (min, max) 4.5 (2.0, 6.5) 5.0 (2.0, 6.5) 5.0 (2.0, 6.5)
Mean (SD) 4.7 (1.4) 4.8 (1.4) 4.8 (1.4)
EDSS stratification, No. (%)
ⱕ4.0 68 (46.3) 124 (42.5) 192 (43.7)
⬎4.0 79 (53.7) 168 (57.5) 247 (56.3)
Total gadolinium-enhancing lesion count, No. (%)
⫽0 110 (74.8) 220 (75.9) 330 (75.5)
⬎0 37 (25.2) 70 (24.1) 107 (24.5)
Brain volume, cm3
Median 1,209.5 1,204.0 1,207.0
Minimum–maximum 642.0–1,522.0 712.60–1,508.0 642.0–1,522.0
Mean (SD) 1,210.91 (128.89) 1,202.92 (120.23) 1,205.75 (123.25)
T2 lesion volume, mm3
Median 5,199.50 5,240.50 5,220.70
Minimum–maximum 73.83–74,534.0 174.00–155,303.0 73.83–155,303.0
Mean (SD) 8,850.86 (11,808.95) 9,336.66 (13,744.94) 9,173.25 (13,113.98)
MSFC timed 25-foot walk test, 7.38 8.35 8.08
median seconds
SD ⫽ standard deviation; IFN ⫽ interferon; MS ⫽ multiple sclerosis; EDSS ⫽ Expanded Disability Status Scale; MSFC ⫽ Multiple
Sclerosis Functional Composite.

in rituximab-treated patients aged ⬍51 years (hazard ratio
[HR] ⫽ 0.52; p ⫽ 0.010) or those with gadolinium brain
lesions at baseline MRI (HR ⫽ 0.41; p ⫽ 0.007). More-
over, an additive predictive effect of age and gadolinium
lesions at baseline was found; for patients aged ⬍51 years
with baseline gadolinium lesions treated with placebo, the
risk of having CDP was 3 times higher than for patients
treated with rituximab (HR ⫽ 0.33; p ⫽ 0.009). In ad-
dition, there appeared to be a greater treatment effect of
rituximab on T2 lesions for patients aged ⬍51 years with
baseline gadolinium lesions, with a 61.6% ( p ⫽ 0.021)
relative reduction in total T2 lesion volume accumulation
from baseline to week 96, compared with 50.7% ( p ⫽
0.163) for patients aged ⱖ51 years with baseline gadolin-
ium lesions, ⫺13.1% ( p ⫽ 0.261) for patients aged ⬍51
years with no baseline gadolinium lesions, and 34.8%

464 Annals of Neurology Vol 66 No 4 October 2009

 Table 2. Patient Disposition, Reasons for Discontinuation, and Safety Follow-up in Intent-to-Treat Patients
Variable Placebo, Rituximab, Total,
nⴝ147, nⴝ292, Nⴝ439,
No. (%) No. (%) No. (%)

Patients randomized 147 (100) 292 (100) 439 (100)

Completed 24 weeks 136 (92.5) 269 (92.1) 405 (92.3)
Completed 48 weeks 130 (88.4) 256 (87.7) 386 (87.9)
Completed 72 weeks 125 (85.0) 246 (84.2) 371 (84.5)
Completed 96 weeks 124 (84.4) 241 (82.5) 365 (83.1)
(completion of treatment
period)
Discontinued treatment period 23 (15.6) 51 (17.5) 74 (16.9)
Reasons for discontinuation of treatment period
Adverse events 0 8 (15.7) 8 (10.8)
a
Death 1 (4.3) 0 1 (1.4)a
Lost to follow-up 2 (8.7) 2 (3.9) 4 (5.4)
Patient’s decision 11 (47.8) 27 (52.9) 38 (51.4)
Physician’s decision 4 (17.4) 7 (13.7) 11 (14.9)
Initiation of excluded 0 1 (2.0) 1 (1.4)
therapy
Subject noncompliance 0 1(2.0) 1(1,4)
Pregnancy 0 1 (2.0) 1 (1.4)
Disease progression 5 (21.7) 4 (7.8) 9 (12.2)
Safety follow-up period
Did not enter 5 (4.0) 5 (2.1) 10 (2.7)
Entered but did not 3 (2.4) 12 (5.0) 15 (4.1)
complete
Completed (week 122) 116 (93.5) 224 (92.9) 340 (93.2)
*This death was due to cardiopulmonary failure.

( p ⫽ 0.022) for patients aged ⱖ51 with no baseline gad-
olinium lesions.
Subgroup analyses indicated that patients with a short
disease duration (ⱕ3 years) receiving rituximab tended to
have a longer time to CDP compared with the patients
receiving placebo. However, a treatment effect on time to
CDP was seen with younger patients (⬍51 years) despite
a relatively longer disease duration (ⱖ7.7 years), and no
effect on CDP was seen in older patients (⬎51 years)
with relative shorter disease duration (⬍7.7 years).
Pharmacodynamics and Immunogenicity
Treatment with rituximab was associated with a rapid
and near-complete depletion (⬎95%) of CD19⫹ (sur-
rogate marker for CD20) peripheral B cells starting at
week 2 through week 96 of treatment. B-cell counts
above the lower limit of normal (LLN, 80 cells/␮l)
were used to measure recovery. At week 122, 35% of
rituximab-treated patients had recovered peripheral
B-cell counts. Of patients who discontinued the treat-
ment period early, 40% recovered peripheral B cells af-
ter 48 weeks from their last dose. Median CD3 T-cell
counts were not appreciably altered by rituximab.
At all times during the study, IgG and IgA levels
were below LLN range in ⬍5% of patients in either
treatment arm. IgM levels were below LLN at all times
during the trial in 31.7% of patients receiving ritux-
imab and 5.9% of patients receiving placebo (Table 4).
There was no evidence of increased incidence of infec-
tion or other adverse events in patients with immuno-
globulin levels below LLN.
During the treatment or safety follow-up, 20 of 286
(7.0%) patients receiving rituximab and 9 of 143
(6.3%) patients receiving placebo were positive for hu-
man antichimeric antibodies (HACA). Among them,
18 (7.5%) patients receiving rituximab and 4 (3.4%)
patients receiving placebo were HACA-positive at ap-
proximately 1 year after last infusion (see Table 4).
There was no apparent association between HACA
positivity and the type of adverse events or efficacy re-

Hawker et al: Rituximab in PPMS 465

 Table 3. Efficacy Analysis for Primary, Secondary, and Exploratory Endpoints
Endpoint Placebo, Rituximab, p
nⴝ147 nⴝ292

Primary endpoint: time to CDP

Proportion of patients with CDP (K-M estimates), %a 0.1442b
At week 48 19.3 20.2
At week 96 38.5 30.2
Secondary endpoints: T2 lesion and brain volume
T2 volume change from baseline to week 96
(LOCF)
Mean (SD) 2,205 (4306) 1,507 (3739) ⬍0.001c
Median 809.5 302.0
Brain volume change from baseline to week 96 0.62c
(LOCF)
Mean (SD) ⫺9.9 (37.0) ⫺10.8 (40.3)
Median ⫺14.0 ⫺13.1
Key exploratory endpoints
Proportion of patients with CDP (24-mo 30.4 27.3 0.59b
confirmation) at week 96, %
EDSS change from baseline to week 96 (LOCF) 0.34c
Mean (SD) 0.45 (1.0) 0.33 (1.0)
Median 0 0
Median MSFC change from baseline (LOCF)
To week 48 ⫺0.05 ⫺0.00 0.057c
To week 96 ⫺0.04 ⫺0.05 0.846c
To week 122 ⫺0.10 ⫺0.06 0.089c
Median MSFCS (timed 25-foot walk) change from
baseline (LOCF)
To week 48 ⫺0.05 ⫺0.02 0.037c
To week 96 ⫺0.11 ⫺0.07 0.076c
To week 122 ⫺0.14 ⫺0.08 0.015c
a
Stratified hazard ratio (relative to placebo) with 95% confidence interval: 0.77 (0.55–1.09). bLog-rank test stratifying baseline EDSS
and prior MS treatment. cFriedman ranked analysis of variance model adjusting baseline EDSS and prior MS treatment. CDP ⫽
confirmed disease progression; K-M ⫽ Kaplan-Meier; LOCF ⫽ last observation carried forward; SD ⫽ standard deviation; EDSS ⫽
Expanded Disability Status Scale; MSFCS ⫽ Multiple Sclerosis Functional Composite score; MS ⫽ multiple sclerosis.

sponse. The biologic relevance of HACA has yet to be
determined.
Safety
The incidence of adverse events (AEs) was comparable
between groups (see Table 4). Serious AEs were pro-
portionally higher in rituximab patients, attributable to
patients who experienced multiple serious AEs. Using
the Common Terminology Criteria for Adverse Events
grading system, 58.6% of rituximab-treated patients
experienced grade 1-2 (mild-moderate) AEs. Of 292
rituximab-treated patients, 8.6% reported severe (grade
3) or disabling (grade 4) drug-related AEs.
Infusion-related AEs were more common with ritux-
imab, which were primarily mild to moderate in sever-
ity, decreasing with successive infusions. At week 74,
infusion-related reactions were lower in rituximab
(4.9%) versus placebo (7.2%) patients. Grade 3
infusion-associated events were reported in 17 patients.
No grade 4 infusion-associated events were reported.
Infections reported in ⬎10% of either group in-
cluded upper respiratory infections, urinary tract infec-
tions, and nasopharyngitis. Of 13 rituximab patients
who reported infection-associated serious AEs, 9 (69%)
were ⱖ55 years of age.
Three deaths occurred. One patient with a history of

466 Annals of Neurology Vol 66 No 4 October 2009

A
70 All Intent-to-Treat Patients (N=439)
60 Placebo
Proportion of Patients

Rituximab
50
P=0.1442
40

30

20

10

0
0 12 24 36 48 60 72 84 96 108
Time to Confirmed Disease Progression (weeks)

B
70 Age ≥51 Placebo 70 Age ≥51
Gd Lesion=0 Rituximab Gd Lesion ≥1
60 N=187 60 N=37
HR: 1.27 HR: 0.52
Proportion of Patients

(95% CI: 0.71–2.27) (95% CI: 0.18–1.52)

50 P=0.4256 50 P=0.2243

40 40

30 30

20 20

10 10

0 0
0 12 24 36 48 60 72 84 96 108 0 12 24 36 48 60 72 84 96 108
Time to Confirmed Disease Progression (weeks) Time to Confirmed Disease Progression (weeks)

70 70
Age <51 Age <51
60
Gd Lesion=0 Gd Lesion ≥1
N=143 60 N=72
Proportion of Patients

HR: 0.63 HR: 0.33

50 (95% CI: 0.34–1.18) 50 (95% CI: 0.14–0.79)
P=0.1427 P=0.0088
40 40

30 30

20 20

10 10

0 0
0 12 24 36 48 60 72 84 96 108 0 12 24 36 48 60 72 84 96 108
Time to Confirmed Disease Progression (weeks) Time to Confirmed Disease Progression (weeks)

Fig 3. Efficacy endpoint. (A) Time to confirmed disease progression, all patients; (B) time to confirmed disease progression by age
and baseline gadolinium-enhancing lesions. HR ⫽ hazard ratio; CI ⫽ confidence interval.

brainstem lesions and aspiration received 2 infusions of contracted pneumonia and died after withdrawing
rituximab and withdrew early from the study after con- from the trial.
tracting pneumonia; he was subsequently hospitalized
multiple times for recurrent aspiration and died ap- Discussion
proximately 10 weeks after the initial event. Two pa- PPMS is thought to incorporate elements of inflamma-
tients receiving placebo died; 1 suffered from cardio- tion and neurodegeneration; however, currently ap-
pulmonary failure during the study, and the other proved first-line MS therapies have not been shown to

Hawker et al: Rituximab in PPMS 467

 Table 4. AEs and Immunogenicity in the Safety Population
Events Placebo, nⴝ147, Rituximab, nⴝ292,
No. (%) No. (%)

Total AEs 147 (100) 289 (99.0)

AEs leading to discontinuation of study drug 1 (⬍1.0) 9 (3.1)
a
AE grades
Grade 1 20 (13.6) 44 (15.1)
Grade 2 71 (48.3) 127 (43.5)
Grade 3 51 (34.7) 105 (36.0)
Grade 4 3 (2.0) 12 (4.1)
b
Grade 5 2 (1.4) 1 (⬍1.0)c
Serious AEs 20 (13.6) 48 (16.4)
Infusion-associated events, % patientsd
1st Infusion, week 0 23.1 67.1
2nd Infusion, week 2 15.1 22.6
3rd Infusion, week 24 9.6 20.3
4th Infusion, week 26 8.9 10.9
5th Infusion, week 48 6.2 13.7
6th Infusion, week 50 6.3 7.1
7th Infusion, week 72 5.6 7.8
8th Infusion, week 74 7.2 4.9
Infusion-associated AEs ⱖ5% in either group
Nausea 8 (5.4) 31 (10.6)
Fatigue 6 (4.1) 33 (11.3)
Chills 4 (2.7) 32 (11.0)
Pyrexia 2 (1.4) 28 (9.6)
Headache 4 (2.7) 37 (12.7)
Dizziness 4 (2.7) 22 (7.5)
Throat irritation 0 21 (7.2)
Pharyngolaryngeal pain 0 16 (5.5)
Pruritus 1 (⬍1.0) 54 (18.5)
Rash 3 (2.0) 22 (7.5)
Flushing 6 (4.1) 37 (12.7)
Hypotension 12 (8.2) 27 (9.2)
Infection-associated events
All infection-associated AEs and SAEs 96 (65.3) 199 (68.2)
Infection-associated SAEs 1(⬍1.0) 13 (4.5)e
Specific infection-associated SAEs
Pneumonia 1 (⬍1) 4 (1.4)
Bronchitis 0 1 (⬍1)

substantially alter disease course. Although the primary neurodegeneration in MS, insights that may also assist
endpoint was not met in this trial, our findings suggest in our understanding of other autoimmune disorders.
key insights into the interplay of inflammation and Despite the lack of statistical significance between

468 Annals of Neurology Vol 66 No 4 October 2009

 Table 4. Continued
Events
Urinary tract infection
Pyelonephritis
Cellulitis
Diverticulitis
Meningitis
Escherichia urinary tract infection
Eschericia bacteraemia
Influenza
Patients positive for HACA during treatment or follow-up period
Ig below the lower limit of laboratory normal
IgG
IgM
IgA
a
Common Terminology Criteria for Adverse Events classification system. bCardiopulmonary failure at week 74 and pneumonia at week
2. cPneumonia at week 2. dAny AE occurring during or within 24 hours of rituximab infusion. eMultiple occurrences of the same event
for a subject were counted only once. fHACA data collected on 286 patients during treatment period or safety follow-up. AE ⫽ adverse
event; SAE ⫽ adverse event; HACA ⫽ human antichimeric antibodies; Ig ⫽ immunoglobulin.
treatment groups in the overall PPMS population,
there was a trend in treatment-related delay, as evi-
denced by confirmed EDSS progression and in the
timed 25-foot walk test, a key exploratory endpoint. In
the planned subgroup analyses of both the placebo and
rituximab groups, there was a statistically significant
treatment effect by both these measures in younger pa-
tients (aged ⬍51 years) and particularly in those with
inflammation based on gadolinium-enhancing lesions
on MRI. This treatment effect was independent of
disease duration in younger patients; however, treat-
ment effects were primarily driven by faster progres-
sion rates in the younger cohort (and particularly in
those with gadolinium-enhancing lesions) than the
progression rates of older patients. gadolinium scans
were done at baseline and week 6 (for safety) only, so
the effect on gadolinium lesions of rituximab was not
assessed.
In contrast to the PROMiSe trial, in which a trend
for a treatment effect was noted in males, gender did
not appear to play a role in our trial. The trend toward
a treatment effect in PROMiSe was postulated to be
driven by the more rapid progression rates in male pa-
tients.4 Although the results from the 2 trials are at
odds in terms of gender, the treatment effect in our
trial was similarly driven by a more rapidly progressing
cohort, albeit a cohort characterized by age and gado-
linium positivity. Although the characteristics of the
patients we recruited may have played a role (positive
CSF and higher prevalence of gadolinium lesions on
Placebo, nⴝ147, Rituximab, nⴝ292,
No. (%) No. (%)
0 2 (⬍1)
0 1 (⬍1)
0 2 (⬍1)
0 1 (⬍1)
0 1 (⬍1)
0 1 (⬍1)
0 1 (⬍1)
0 1 (⬍1)
f
9/143 (6.3) 20/286 (7.0%)
7/142 (4.9) 13/274 (4.7)
8/135 (5.9) 83/262 (31.7)
1/145 (⬍1.0) 3/207 (1.0)
MRI in OLYMPUS) in the differing gender outcome,
it seems that the rate of progression may be a crucial
factor in predicting outcomes of therapy over a short
time span.
These findings provide insight that may change our
concepts of the pathogenesis and treatment of PPMS
but also of progression in all subtypes of MS. Our cur-
rent thinking is that a significant proportion of neuro-
logical disease progression in PPMS is driven by slow
degeneration of axons after their immune-mediated de-
myelination.20 However, the finding that even a small
amount of inflammation, as evidenced by gadolinium-
enhancing lesions, influenced the rate of progression
and the response to treatment in this trial argues for a
reversible component of progression as well. This ob-
servation is supported in a study in which evidence of
even mild inflammation evident on MRI in PPMS was
associated with greater disability.21 PPMS patients
demonstrate greater MRI enhancement than previously
believed when scanned earlier in the course of their dis-
ease.21–23 More sensitive, quantitative MRI imaging
techniques and analyses may be helpful to define pa-
tient populations at risk for rapid disease progression.
These findings, coupled with our data, suggest sev-
eral conclusions. First, early treatment of PPMS, aimed
at decreasing disease progression, may be beneficial, be-
fore the advent of irreversible neuronal loss. This has
been well documented in RRMS trials, in which treat-
ment at the time of diagnosis was shown to reduce the
risk of progressive disability.24,25
Hawker et al: Rituximab in PPMS 469
 Second, in our study, the slowing of disability was
driven by an effect on ambulation, as evidenced by the
effect on the timed 25-foot walk. Gadolinium-
enhancing lesions, reflecting classic focal, macroscopic
inflammation, are relatively infrequent in the spinal
cord.21 Thus, one may postulate that other compo-
nents of the immune system, such as dendritic cells as
well as B cells, likely play a role in perpetuating the
disease.21,26,27
Third, the treatment effect in younger patients, par-
ticularly those with evidence of acute blood-brain bar-
rier breakdown on neuroimaging, irrespective of disease
duration, argues for a change in disease neurobiology
over time. This influence of age as a risk factor for an
increased relapse rate has also been recently reported
for RRMS.28 Several factors may be implicated, includ-
ing alteration in the immune system with aging (im-
munosenescence29 –33). Thus, identifying patients ame-
nable to treatment, with a greater risk of progression
and thus an acceptable risk-to-benefit ratio, has impor-
tant implications for treating MS as well as other au-
toimmune diseases in which a similar age-related phe-
nomenon has been reported.34
In conclusion, although our study was negative, cou-
pled with our recent reports of selective B-cell deple-
tion in RRMS,35,36 our results provide us with new
insights into PPMS, and demonstrate the need for ex-
ploring the potential of pharmacological depletion of B
cells in MS patients with or without relapses. Our aim
is to reduce the accumulation of disability early in the
progressive phase of the disease, when a targeted anti-
inflammatory intervention may still have the potential
to retard downstream neurodegenerative processes.
This study was supported by Clinicaltrials.gov under “A Study to
Evaluate the Safety and Efficacy of Rituximab in Adults With Pri-
mary Progressive Multiple Sclerosis” (grant NCT00087529), by Ge-
nentech, Inc., and by Biogen Idec, Inc.
We thank Genentech, Inc. for assistance in data anal-
ysis and preparation of the manuscript.
Appendix
We thank each of the OLYMPUS investigators, including
Drs. Virender Bhan, Michael P. Biber, Staley A. Brod, Peter
Calabresi, Denise Campagnolo, Mark Cascione, Nelson
(Kanter) Cooke, Joanna A. Cooper, John Corboy, Bruce
Cree, Anne Cross, Pierre Duquette, Stanton Elias, Corey
Ford, Mark S. Freedman, Mitchell S. Freedman, Suzanne K.
Gazda, Francois Grand’Maison, Michael (Lava) Gruenthal,
Stuart Hoffman, William Honeycutt, John Huddlestone,
Bruce Hughes, George J. Hutton, Daniel Jacobs, Francois
H. Jacques, Lloyd H. Kasper, Lorne Kastrukoff, Michael D.
Kaufman, Omar Khan, Bhupendra O. Khatri, Mariko Kita,
Yves La Pierre, Sharon Lynch, Clyde Markowitz, Michele
Mass, David Mattson, Aaron Miller, Harold Moses, Paul
470 Annals of Neurology Vol 66 No 4 October 2009
O’Connor, Hillel Panitch, Mary Ann Picone, Kottil Ram-
mohan, Anthony T. Reder, Peter Riskind, Syed W. Rizvi,
Howard Rossman, Steven R. Schwid, Stuart Shafer, Virginia
Simnad, Michael Stein, William H. Stuart, Olaf Stuve, Ben
Thrower, Stephen E. Thurston, Michael (Mihai) Vertino,
Leslie P. Weiner, Dean Wingerchuk, and Michael M. C.
Yeung.
References
1. Hauser SL, Oksenberg JR. The neurobiology of multiple
sclerosis: genes, inflammation, and neurodegeneration. Neuron
2006;52:61–76.
2. Noseworthy JH, Lucchinetti C, Rodriguez M, Weinshenker
BG. Multiple sclerosis. N Engl J Med 2000;343:938 –952.
3. Miller DH, Leary SM. Primary-progressive multiple sclerosis.
Lancet Neurol 2007;6:903–912.
4. Wolinsky JS, Narayana PA, O’Connor P, et al. Glatiramer ac-
etate in primary progressive multiple sclerosis: results of a mul-
tinational, multicenter, double-blind, placebo-controlled trial.
Ann Neurol 2007;61:14 –24.
5. Leary SM, Miller DH, Stevenson VL, et al. Interferon beta-1a
in primary progressive MS: an exploratory, randomized, con-
trolled trial. Neurology 2003;60:44 –51.
6. Montalban X. Overview of European pilot study of interferon
beta-Ib in primary progressive multiple sclerosis. Mult Scler
2004;10(suppl 1):S62; discussion 62– 64.
7. Owens GP, Bennett JL, Gilden DH, Burgoon MP. The B cell
response in multiple sclerosis. Neurol Res 2006;28:236 –244.
8. Andersson L, Bolling M, Wirestam R, et al. Combined diffu-
sion weighting and CSF suppression in functional MRI. NMR
Biomed 2002;15:235–240.
9. Link H, Huang Y-M. Oligoclonal bands in multiple sclerosis
cerebrospinal fluid: an update on methodology and clinical use-
fulness. J Neuroimmunol 2007;180:7–28.
10. Siden A. Isoelectric focusing and crossed immunoelectrofocusing
of CSF immunoglobulins in MS. J Neurol 1979;221:39 –51.
11. Genain CP, Cannella B, Hauser SL, Raine CS. Identification of
autoantibodies associated with myelin damage in multiple scle-
rosis. Nat Med 1999;5:170 –175.
12. O’Connor KC, Appel H, Bregoli L, et al. Antibodies from in-
flamed central nervous system tissue recognize myelin oligoden-
drocyte glycoprotein. J Immunol 2005;175:1974 –1982.
13. O’Connor KC, Chitnis T, Griffin DE, et al. Myelin basic
protein-reactive autoantibodies in the serum and cerebrospinal
fluid of multiple sclerosis patients are characterized by low-
affinity interactions. J Neuroimmunol 2003;136:140 –148.
14. Kanter JL, Narayana S, Ho PP, et al. Lipid microarrays identify
key mediators of autoimmune brain inflammation. Nat Med
2006;12:138 –143.
15. Pender MP. The pathogenesis of primary progressive multiple
sclerosis: antibody-mediated attack and no repair? J Clin Neu-
rosci 2004;11:689 – 692.
16. Reff ME, Carner K, Chambers KS, et al. Depletion of B cells in
vivo by a chimeric mouse human monoclonal antibody to
CD20. Blood 1994;83:435– 445.
17. Kurtzke JF. Rating neurologic impairment in multiple sclerosis:
an expanded disability status scale (EDSS). Neurology 1983;33:
1444 –1452.
18. McDonald WI, Compston A, Edan G, et al. Recommended
diagnostic criteria for multiple sclerosis: guidelines from the In-
ternational Panel on the diagnosis of multiple sclerosis. Ann
Neurol 2001;50:121–127.
19. National Institutes of Health. Common Terminology Criteria for
Adverse Events v3.0 online instructions and guidelines. In: Can-
cer Therapy Evaluation Program DoCTaDD. Bethesda, MD:
National Cancer Institute, National Institutes of Health; 2006.
https://webapps.ctep.nci.nih.gov/webobjs/ctc/webhelp/welcome_
to_ctcae.htm
20. Dutta R, Trapp BD. Pathogenesis of axonal and neuronal dam-
age in multiple sclerosis. Neurology 2007;68:S22–S31; discus-
sion S43–S54.
21. Ingle GT, Sastre-Garriga J, Miller DH, Thompson AJ. Is in-
flammation important in early PPMS? A longitudinal MRI
study. J Neurol Neurosurg Psychiatry 2005;76:1255–1258.
22. Khaleeli Z, Ciccarelli O, Manfredonia F, et al. Predicting pro-
gression in primary progressive multiple sclerosis: a 10-year
multicenter study. Ann Neurol 2008;63:790 –793.
23. Rovaris M, Judica E, Sastre-Garriga J, et al. Large-scale, multicen-
tre, quantitative MRI study of brain and cord damage in primary
progressive multiple sclerosis. Mult Scler 2008;14:455– 464.
24. Kappos L, Freedman MS, Polman CH, et al. Effect of early
versus delayed interferon beta-1b treatment on disability after a
first clinical event suggestive of multiple sclerosis: a 3-year
follow-up analysis of the BENEFIT study. Lancet 2007;370:
389 –397.
25. Kinkel RP, Kollman C, O’Connor P, et al. IM interferon
beta-1a delays definite multiple sclerosis 5 years after a first de-
myelinating event. Neurology 2006;66:678 – 684.
26. Lyons JA, San M, Happ MP, Cross AH. B cells are critical to
induction of experimental allergic encephalomyelitis by protein
but not by a short encephalitogenic peptide. Eur J Immunol
1999;29:3432–3439.
27. Moore W. Evidence for Secondary Neurogeneration in MS.
Montreal, Canada: ECTRIMS/ACTRIMS; 2008.
28. Mowry EM, Pesic M, Grimes B, et al. Clinical predictors of
early second event in patients with clinically isolated syndrome.
J Neurol 2009;256:1061–1066.
29. Aw D, Silva AB, Maddick M, et al. Architectural changes in the
thymus of aging mice. Aging Cell 2008;7:158 –167.
30. Chen J, Buchanan JB, Sparkman NL, et al. Neuroinflammation
and disruption in working memory in aged mice after acute
stimulation of the peripheral innate immune system. Brain Be-
hav Immun 2008;22:301–311.
31. Godbout JP, Chen J, Abraham J, et al. Exaggerated neuroin-
flammation and sickness behavior in aged mice following acti-
vation of the peripheral innate immune system. FASEB J 2005;
19:1329 –1331.
32. Hasler P, Zouali M. Immune receptor signaling, aging, and au-
toimmunity. Cell Immunol 2005;233:102–108.
33. Sawada M, Sawada H, Nagatsu T. Effects of aging on neuro-
protective and neurotoxic properties of microglia in neurode-
generative diseases. Neurodegener Dis 2008;5:254 –256.
34. Prelog M. Aging of the immune system: a risk factor for auto-
immunity? Autoimmun Rev 2006;5:136 –139.
35. Bar-Or A, Calabresi P, Arnold D, et al. Safety and efficacy of
rituximab in adults with relapsing remitting multiple sclerosis: a
phase I open-label multicenter trial. Ann Neurol 2008;63:
395– 400.
36. Hauser SL, Waubant E, Arnold DL, et al. B-cell depletion with
rituximab in relapsing-remitting multiple sclerosis. N Engl
J Med 2008;358:676 – 688.
Hawker et al: Rituximab in PPMS 471