Treatment of rheumatoid arthritis with a recombinant human tumor necrosis factor receptor (p75)-Fc fusion protein.

Abstract

Tumor necrosis factor (TNF) is a proinflammatory cytokine involved in the pathogenesis of rheumatoid arthritis, and antagonism of TNF may reduce the activity of the disease. This study evaluated the safety and efficacy of a novel TNF antagonist - a recombinant fusion protein that consists of the soluble TNF receptor (p75) linked to the Fc portion of human IgG1 (TNFR:Fc).
In this multicenter, double-blind trial, we randomly assigned 180 patients with refractory rheumatoid arthritis to receive subcutaneous injections of placebo or one of three doses of TNFR:Fc (0.25, 2, or 16 mg per square meter of body-surface area) twice weekly for three months. The clinical response was measured by changes in composite symptoms of arthritis defined according to American College of Rheumatology criteria.
Treatment with TNFR:Fc led to significant reductions in disease activity, and the therapeutic effects of TNFR:Fc were dose-related. At three months, 75 percent of the patients in the group assigned to 16 mg of TNFR:Fc per square meter had improvement of 20 percent or more in symptoms, as compared with 14 percent in the placebo group (P<0.001). In the group assigned to 16 mg per square meter, the mean percent reduction in the number of tender or swollen joints at three months was 61 percent, as compared with 25 percent in the placebo group (P<0.001). The most common adverse events were mild injection-site reactions and mild upper respiratory tract symptoms. There were no dose-limiting toxic effects, and no antibodies to TNFR:Fc were detected in serum samples.
In this three-month trial TNFR:Fc was safe, well tolerated, and associated with improvement in the inflammatory symptoms of rheumatoid arthritis.

Full text

Page 1

Volume 337Number 3

?

141

The New England
Journal
of



Medicine

© Copyright, 1997, by the Massachusetts Medical Society

VOLUME 337

J

ULY

17, 1997

NUMBER 3

TREATMENT OF RHEUMATOID ARTHRITIS WITH A RECOMBINANT HUMAN
TUMOR NECROSIS FACTOR RECEPTOR (p75)–Fc FUSION PROTEIN

L

ARRY

W. M
OY
M. F
ILLIAM

ORELAND
LEISCHMANN
J. K
OOPMAN

, M.D., S
, M.D., A
, M.D., K

COTT

W. B
RTHUR
M

AUMGARTNER
L. W
EAVER
OHLER
, P

, M.D., M
, M.D., R
.D., M
ICHAEL

ICHAEL
OBERT
B. W
R

H. S
E. E
IDMER

CHIFF

, M.D., E
, M.D., S
.D.,
AND

LIZABETH
TANLEY
ONSUELO

A. T
C
OHEN
M. B

INDALL
, M.D.,
LOSCH

, M.D.,
R











TTLINGER
, P





W





ENDALL




H






H



C




, M.D.

A

BSTRACT
Background
inflammatory cytokine involved in the pathogenesis
of rheumatoid arthritis, and antagonism of TNF may
reduce the activity of the disease. This study evalu-
ated the safety and efficacy of a novel TNF antago-
nist — a recombinant fusion protein that consists of
the soluble TNF receptor (p75) linked to the Fc por-
tion of human IgG1 (TNFR:Fc).
Methods
In this multicenter, double-blind trial, we
randomly assigned 180 patients with refractory rheu-
matoid arthritis to receive subcutaneous injections of
placebo or one of three doses of TNFR:Fc (0.25, 2, or
16 mg per square meter of body-surface area) twice
weekly for three months. The clinical response was
measured by changes in composite symptoms of ar-
thritis defined according to American College of Rheu-
matology criteria.
Results
Treatment with TNFR:Fc led to significant
reductions in disease activity, and the therapeutic ef-
fects of TNFR:Fc were dose-related. At three months,
75 percent of the patients in the group assigned to
16 mg of TNFR:Fc per square meter had improve-
ment of 20 percent or more in symptoms, as com-
pared with 14 percent in the placebo group (P
In the group assigned to 16 mg per square meter,
the mean percent reduction in the number of ten-
der or swollen joints at three months was 61 percent,
as compared with 25 percent in the placebo group
(P
?
0.001). The most common adverse events were
mild injection-site reactions and mild upper respira-
tory tract symptoms. There were no dose-limiting tox-
ic effects, and no antibodies to TNFR:Fc were detect-
ed in serum samples.
Conclusions
In this three-month trial TNFR:Fc was
safe, well tolerated, and associated with improve-
ment in the inflammatory symptoms of rheumatoid
arthritis. (N Engl J Med 1997;337:141-7.)
©1997, Massachusetts Medical Society.


Tumor necrosis factor (TNF) is a pro-





?

0.001).






From the University of Alabama at Birmingham, Birmingham (L.W.M.,
W.J.K.); the Physician’s Clinic of Spokane, Spokane, Wash. (S.W.B.); the
Denver Arthritis Clinic, Denver (M.H.S.); Portland Medical Associates,
Portland, Oreg. (E.A.T.); the Metroplex Clinical Research Center, Dallas
(R.M.F., S.C.); the Arthritis Center of Nebraska, Lincoln (A.L.W.); Taco-
ma, Wash. (R.E.E.); and Immunex Corporation, Seattle (K.M., M.B.W.,
C.M.B.). Address reprint requests to Dr. Moreland at the Arthritis Clinical
Intervention Program, Division of Clinical Immunology and Rheumatol-
ogy, University of Alabama at Birmingham, 1717 6th Ave. S., Rm. 068,
Birmingham, AL 35294-7201.

HEUMATOID arthritis is a common dis-
ease, and it produces substantial morbidity
as well as an increase in mortality.
though the causes of rheumatoid arthritis
are not fully understood, laboratory and clinical evi-
dence suggests that proinflammatory cytokines, partic-
ularly tumor necrosis factor (TNF), have an important
role in its pathogenesis.
TNF induces the release of
matrix metalloproteases from neutrophils, fibroblasts,
and chondrocytes
; induces the expression of en-
dothelial adhesion molecules involved in the migration
of leukocytes to extravascular sites of inflammation
and stimulates the release of other proinflammatory
cytokines.
TNF concentrations are increased in the
synovial fluid of persons with active rheumatoid arthri-
tis,
and increased plasma levels of TNF are as-
sociated with joint pain. Administration of TNF an-
tagonists to patients with rheumatoid arthritis has
been shown to reduce symptoms.
There are two distinct cell-surface TNF receptors
(TNFRs), designated p55 and p75.
cated versions of membrane TNFRs, consisting of
only the extracellular, ligand-binding domain, are
present in body fluids and are thought to be in-
volved in regulating TNF activity.
have been detected in synovial tissue and at the junc-
tion between cartilage and pannus.
are increased in serum and synovial fluid in rheuma-
toid arthritis
and in many other autoimmune
and inflammatory conditions.

1-4

Al-

5,6


7-9


10

;

11,12


13,14


15


16-19



20,21

Soluble, trun-

22,23

Soluble TNFRs

24,25

Their levels

26-29


30-44


Copyright © 1997 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at FAC DE FARMACIA on June 18, 2005 .

Page 2

142

?

July 17, 1997

The New England Journal of Medicine

A recombinant human TNFR p75–Fc fusion pro-
tein (TNFR:Fc) (Enbrel, Immunex, Seattle) has been
developed for therapeutic neutralization of TNF.
DNA encoding the soluble portion of human TNFR
p75 was linked to DNA encoding the Fc portion of
a human IgG1 molecule, and the combined DNA
was then expressed in a mammalian cell line. The re-
sulting immunoglobulin-like dimer, composed exclu-
sively of human amino acid sequences, acts as a com-
petitive inhibitor of TNF and prevents binding of
TNF to the cell-surface TNFR, thereby reducing the
biologic activity of TNF. Safety studies in normal
human volunteers revealed no adverse effects after
the intravenous administration of TNFR:Fc.
were trends toward a reduction in disease activity
in a safety and dose-finding study of TNFR:Fc ad-
ministered for four weeks to a small number of pa-
tients with refractory rheumatoid arthritis.
basis of these findings, we undertook a multicenter,
randomized, double-blind, placebo-controlled trial of
TNFR:Fc in patients with active, refractory rheuma-
toid arthritis.

45


46

There

47

On the

METHODS

Patients
Men and women 18 years of age or older were eligible if they
met the criteria of the American Rheumatism Association for
rheumatoid arthritis and were in functional class I, II, or III
according to the criteria of the American College of Rheumatol-
ogy. All candidates had been unsuccessfully treated (lack of effi-
cacy) with between one and four of the following disease-modify-
ing antirheumatic drugs: hydroxychloroquine, oral or injectable
gold, methotrexate, azathioprine, penicillamine, and sulfasalazine.
No such therapy was allowed during a four-week washout period
before the first dose of study drug. Patients receiving nonsteroi-
dal antiinflammatory drugs (NSAIDs), corticosteroids (
per day), or both were eligible if the dosage had been stable for
at least four weeks before day 1 of the washout period and re-
mained so throughout the study and follow-up period. Requisite
base-line laboratory values included a hemoglobin level of at
least 8.5 per liter, platelet count of at least 125,000 per cubic
millimeter, white-cell count of at least 3500 per cubic millime-
ter, a serum creatinine level of not more than 2 mg per deciliter
(177
m
mol per liter), and liver aminotransferase levels not more
than twice the laboratory’s upper limit of normal. The necessary
degree of disease activity at enrollment (before washout) was
confirmed by a finding of 10 or more swollen joints, 12 or more
tender joints, and one of the following two criteria: a Westergren
erythrocyte sedimentation rate of at least 28 mm per hour or a
serum C-reactive protein level of more than 2.0 mg per deciliter;
or morning stiffness for at least 45 minutes. Sexually active men
and sexually active premenopausal women, except for men or
women who had undergone surgical sterilization, were required
to use a medically accepted form of contraception by the time of
enrollment and to continue its use through follow-up. Women
in this category also had to have a negative serum pregnancy test
within five days before the first dose of study drug.


48


49


?

10 mg



Study Protocol
The study protocol was approved by the human-research com-
mittee at each participating center. Before the start of the washout
period, patients gave written informed consent, had a complete
medical history taken, and underwent a complete physical examina-
tion. A hematology profile (complete blood count, differential

count, and platelet count), serum chemical profile (blood urea ni-
trogen, creatinine, alanine aminotransferase, aspartate aminotrans-
ferase, total protein, and albumin), and analysis of a clean-catch
urine specimen with microscopical analysis were also completed.
Base-line clinical assessments included the following: complete
count of swollen and tender joints (71 joints evaluated; cervical
spine and hips evaluated only for tenderness); duration of morning
stiffness; health-assessment questionnaire
global assessment, on a scale from 0 (asymptomatic) to 10 (severe
symptoms); patient’s assessment of pain, on a visual-analogue scale
from 0 (no pain) to 10 (severe pain)
imentation rate; and C-reactive protein level.
sessments were repeated on day 1 of treatment and every two
weeks or monthly throughout the study. Assessments were per-
formed by rheumatologists or trained nurse coordinators. To min-
imize variation between observers, each patient’s disease activity was
assessed primarily by the same person throughout the study.
A patient could be withdrawn from the trial at any time after
enrollment for the following reasons: the patient’s request, preg-
nancy, serious infection, severe or life-threatening adverse event,
or inadequate control of arthritis symptoms (
crease in the total number of swollen or tender joints) necessitat-
ing an increase in the systemic corticosteroid dosage or reinstitu-
tion of therapy with disease-modifying antirheumatic drugs.
Follow-up evaluations after the discontinuation of therapy at
three months were completed every two weeks for one month,
then once a month until the patient required new or previous an-
tirheumatic therapy or until the total count of swollen and tender
joints returned to the base-line value. Hematologic testing, serum
chemistry, urinalysis, and anti–TNFR:Fc antibody testing were re-
peated periodically during the trial, including follow-up, and on
the day the patient required initiation or reinstitution of therapy
with disease-modifying antirheumatic drugs or the joint count re-
turned to the base-line value. All patients were evaluated for side
effects and laboratory abnormalities.

50

; physician’s and patient’s

50

; Westergren erythrocyte sed-
51
Disease-activity as-



?

50 percent in-

Treatment
Patients were randomly assigned to one of four treatment
groups: placebo; 0.25 mg of TNFR:Fc per square meter of body-
surface area (the 0.25-mg group); 2 mg of TNFR:Fc per square
meter (the 2-mg group); or 16 mg of TNFR:Fc per square meter
(the 16-mg group). The study drug or placebo was injected
subcutaneously twice weekly for three months. TNFR:Fc was sup-
plied as a sterile lyophilized powder containing 10 mg of TNFR:Fc,
40 mg of mannitol, 10 mg of sucrose, and 1.2 mg of TRIS
(tromethamine) per vial. The placebo was a lyophilized powder con-
taining 40 mg of mannitol, 10 mg of sucrose, and 1.2 mg of TRIS.
The injection volume was standardized for all patients by dilution
with bacteriostatic water for injection (two 1.5-ml injections per
dose). Injections were given at approximately the same time in the
morning on a Monday–Thursday or Tuesday–Friday schedule.


Concomitant Medications
In addition to stable doses of NSAIDs and corticosteroids, the
following analgesic medications were allowed throughout the tri-
al, except on the day before a joint evaluation: acetaminophen with
codeine phosphate, acetaminophen with propoxyphene napsylate,
and acetaminophen with oxycodone hydrochloride.


Antibody Testing
Serum samples were collected for antibody testing on day 1 (be-
fore treatment), at three months (end of treatment), and two to
four weeks after treatment ended.
Wells of polystyrene microtiter plates (Maxisorp, Nunc, Roskilde,
Denmark) were coated with 63 ng of TNFR:Fc per milliliter in
0.01 M phosphate-buffered saline, pH 7.2, and incubated over-
night at 2 to 8°C. The plates were washed with phosphate-buffered
saline, and controls and samples diluted in phosphate-buffered sa-
line with 5 percent normal goat serum were added and incubated

Copyright © 1997 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at FAC DE FARMACIA on June 18, 2005 .

Page 3

TREATMENT OF RHEUMATOID ARTHRITIS WITH RECOMBINANT HUMAN TNFR:Fc

Volume 337 Number 3

?

143

in the wells for one hour at room temperature. After a second wash-
ing with phosphate-buffered saline, a peroxidase-labeled F(ab
specific goat antihuman immunoglobulin conjugate (Jackson Im-
munoresearch Laboratories, West Grove, Pa.) was added, and the
plates were incubated for another hour at room temperature. Af-
ter a final washing with phosphate-buffered saline,
diamine dihydrochloride peroxidase substrate–chromogen solution
was added to the wells for a 10-minute incubation at room tem-
perature for color development. The reaction was stopped with the
addition of 1 M phosphoric acid, and optical densities were deter-
mined at 490 nm. Test samples were run in duplicate at dilutions
of 1:50, 1:100, 1:200, and 1:400. A rabbit anti-TNFR polyclonal
serum, along with a heterologous antirabbit immunoglobulin de-
tection reagent, was used as a positive control for detection of
TNFR:Fc coated onto the plates. Wells coated with human IgG
served as controls for the conjugate. Samples were scored as posi-
tive if two or more of the dilutions of the post-treatment sample
or samples had optical-density values at least four times as high as
those of the corresponding pretreatment-sample dilutions. The
ability to detect antibodies to TNFR:Fc was confirmed by the pres-
ence of antibodies in serum samples obtained from monkeys treat-
ed with TNFR:Fc and tested with the same reagents as those used
for the detection of human antibodies.
Development of autoantibodies has been reported with the ad-
ministration of other inhibitors of TNF.
were not performed.

?

)

2

-

o

-phenylene-

52

Tests for autoantibodies

Statistical Analysis
The percent change from base line to three months (day 85)
in the swollen-joint count, tender-joint count, and total count of
swollen or tender joints was the primary measure of efficacy. Sec-
ondary end points included pain, quality of life, duration of
morning stiffness, erythrocyte sedimentation rate, C-reactive pro-
tein level, and physician’s and patient’s global assessments. If a
subject withdrew from the study, the last available value was used
as the three-month value. The data were also analyzed to deter-
mine the number of patients meeting American College of Rheu-
matology criteria for 20 and 50 percent improvement, which
specify 20 and 50 percent reductions in the number of swollen
or tender joints and the same degree of improvement in at least
three of five other variables: pain, degree of disability according
to the health-assessment questionnaire, patient’s global assess-
ment, physician’s global assessment, and erythrocyte sedimenta-
tion rate or C-reactive protein level.
jects who dropped out were considered to have had no response.
For changes from base line, the four treatment groups were
compared by analysis of variance, with adjustment for treatment
and study site. Treatment differences were consistent across study
sites. Response rates for American College of Rheumatology cri-
teria were compared by the chi-square test. Six pairwise compar-
isons of efficacy were made for each end point. With the Bonfer-
roni correction for multiple comparisons, the P value had to be
less than 0.008 for significance at the 0.05 level to be retained.
The sample size of 180 subjects (45 in the placebo group and
135 in the three TNFR:Fc groups) was chosen on the basis of the
variability in the percent change from base line in the total count
of swollen or tender joints determined in a previous trial.
suming a standard deviation of 40 for the percent change in the
total count and assuming a difference in this change between the
placebo and treatment groups of 20 percentage points (that is, a
mean reduction of 20 percent in the placebo group and 40 per-
cent in the treated patients), the sample size was sufficient to de-
tect such differences between treatment (n
(n
?
45) groups with 82 percent power.


51

For these end points, sub-

47

As-

?

135) and placebo



RESULTS

Characteristics of the Patients
The characteristics of the patients before treat-
ment are summarized in Table 1. Forty-eight men

and 132 women were enrolled in the trial. Their
mean age was 53 years, and 77 percent had disease
of more than 5 years’ duration. No significant dif-
ferences between groups were detected in pretreat-
ment characteristics or base-line disease activity. Sev-
enty-six percent of the patients completed TNFR:Fc
treatment (61 percent in the 0.25-mg group, 78
percent in the 2-mg group, and 93 percent in the
16-mg group), as compared with 52 percent of the
patients assigned to placebo. The primary reason for
withdrawal was inadequate control of arthritis symp-
toms. Among the patients receiving TNFR:Fc, the
proportions of patients who withdrew because of in-
adequate symptom control were 35 percent in the
0.25-mg group, 17 percent in the 2-mg group, and
5 percent in the 16-mg group; among the patients
receiving placebo, it was 43 percent.

Efficacy
TNFR:Fc produced significant improvement in all
measures of disease activity (Table 2). A clear dose–
response relation was observed in the numbers of
swollen or tender joints, and patients who received
the highest dose of TNFR:Fc had the greatest im-
provement. In the 16-mg group, the mean percent
reduction in the total count at three months was 61
percent, as compared with 25 percent in the placebo
group (P
?
0.001). Figures 1 and 2 depict the num-
bers of swollen and tender joints as a function of
time in each treatment group. In the placebo and
0.25-mg groups, there was an initial response, but
no improvement was noted thereafter. The 16-mg
dose of TNFR:Fc was associated with the greatest re-
duction in the number of swollen or tender joints.
This difference was apparent by the end of week 2 and
was most pronounced at the end of treatment (at
three months).
TNFR:Fc treatment was also associated with sig-
nificant reductions in pain and duration of morning
stiffness, significant improvement in the quality of
life and physician’s and patient’s global assessments,
and significant reductions in disease activity as as-
sessed by objective laboratory measures (erythrocyte
sedimentation rate and C-reactive protein level) (Ta-
ble 2). According to American College of Rheuma-
tology criteria, at three months 57 percent of the
16-mg group had at least 50 percent improvement,
as compared with 7 percent of the placebo group
(P
?
0.001); 75 percent of the 16-mg group had at
least 20 percent improvement, as compared with 14
percent of the placebo group (P
Measures of disease activity moved toward base-line
levels after the cessation of TNFR:Fc therapy.






?

0.001) (Table 3).

Safety and Tolerability
TNFR:Fc was well tolerated; no dose-limiting tox-
ic effects were observed. Only one patient withdrew
because of an adverse event related to TNFR:Fc

Copyright © 1997 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at FAC DE FARMACIA on June 18, 2005 .

Page 4

144

?

July 17, 1997

The New England Journal of Medicine

(a mild injection-site reaction). Adverse events relat-
ed or potentially related to TNFR:Fc were limited to
mild injection-site reactions (erythema or erythema
plus discomfort) and mild upper respiratory tract
symptoms (cough, rhinitis, sinusitis, upper respirato-
ry tract infection, and pharyngitis). Injection-site re-
actions generally occurred with only one of the sever-
al injections given in the three-month treatment
period and resolved in two to three days. Upper res-
piratory tract symptoms were transient, resolved with-
out interruption of TNFR:Fc dosing, and occurred
most frequently in the 2-mg and 16-mg groups; oth-
er, nonrespiratory infections were rare and never seri-
ous. One patient died during the study, a 72-year-old
patient receiving placebo.
No major abnormalities in hematologic findings or
serum chemical profiles were noted during or after
the study; in fact, dose-related improvements in ane-
mia and dose-related decreases in platelet counts were
seen, which reflect a reduction in disease activity.
No antibodies to TNFR:Fc were detected in se-
rum samples from any patient tested.

DISCUSSION

The results of this randomized, double-blind trial
show the clinical efficacy of a soluble recombinant
human TNFR p75–Fc fusion protein in patients with
active rheumatoid arthritis. Treatment with TNFR:Fc
for three months was associated, in a dose-related
fashion, with a reduction in disease activity as assessed
by a number of clinical end points, biochemical mark-
ers of disease, and quality of life. Taken together with
results obtained in preclinical models
with monoclonal antibodies to TNF,
show that TNF antagonism is a valid approach to the
treatment of rheumatoid arthritis. In addition, these
data demonstrate the usefulness of a soluble cytokine
receptor in human disease.
The mechanism of action of TNFR:Fc in rheuma-
toid arthritis probably involves its ability to inhibit
competitively TNF binding to cell-surface TNFR.
On the basis of results of preclinical and clinical
studies in which TNF concentrations were deter-
mined in biologic fluids after the administration of
TNFR:Fc, it is clear that TNFR:Fc does not pro-

53

and trials
16-19
these data



*HAQ denotes health-assessment questionnaire, DMARDs disease-modifying antirheumatic drugs,
and NSAIDs nonsteroidal antiinflammatory drugs. Plus–minus values are means
†On this scale, 0 is best and 10 worst.
‡On this scale, 45 is best and 245 worst.

?

SD.

T

ABLE

1.

D

EMOGRAPHIC



AND
THE

B
S

ASE
TUDY

-L
P

INE
ATIENTS

C

LINICAL
.*

C

HARACTERISTICS



OF









C

HARACTERISTIC

P

LACEBO

TNFR:F
2 mg/m

C

0.25 mg/m

2


2

16 mg/m

2

No. of patients
Mean age (yr)
Female sex (%)
White race (%)
Duration of disease (%)
?
2 yr
2–5 yr
?
5 yr
Measures of arthritis activity
Swollen-joint count (no.)
Tender-joint count (no.)
Total count (no.)
Morning stiffness (hr)
Physician’s assessment†
Patient’s assessment†
Pain (visual analogue)†
Quality of life (HAQ)‡
Erythrocyte sedimentation rate (mm/hr)
C-reactive protein (mg/dl)
Medications in year before study (%)
Corticosteroids
DMARDs
Methotrexate
NSAIDs
Concurrent medications (%)
Corticosteroids
NSAIDs
Acetaminophen with codeine phosphate
Acetaminophen with propoxyphene napsylate
Acetaminophen with oxycodone hydrochloride
44
55
82
91
46
54
70
96
46
52
61
100
44
52
82
93





11
18
71
7
17
76
4
15
80
4
16
80
22
29
51

?
?
?
4.9
7.0
6.7
6.4

9
13
19




146
40
3.9
24
32
56

?
?
?
4.3
7.4
7.1
6.9

11
14
23




153
44
4.1
24
32
55?21
5.2
7.2
6.9
6.7
138
36
3.6

?
?13

9

24?11
30?15
54?25
4.9
6.5
6.5
6.3
135
35
3.6
75
100
34
75
76
100
41
80
80
100
30
70
91
100
27
75
66
73
25
34
11
59
70
33
39
24
65
80
33
30
26
77
75
18
30
7
Copyright © 1997 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at FAC DE FARMACIA on June 18, 2005 .

Page 5

TREATMENT OF RHEUMATOID ARTHRITIS WITH RECOMBINANT HUMAN TNFR:Fc
Volume 337Number 3
?
145
*VAS denotes visual-analogue scale, and HAQ health-assessment questionnaire.
†Data reflect an intention-to-treat analysis. The P values were obtained from an analysis of variance
comparing all four treatment groups in terms of the percent change from base line to three months
with a model that contained main effects of treatment and center. For each patient, missing values
were replaced by the last available value. In the analyses of erythrocyte sedimentation rate and
C-reactive protein level, only patients with observed values were included.
‡On this scale, 0 is best and 10 worst.
§On this scale, 45 is best and 245 worst.
¶Percent improvement from base line was calculated as the average of the changes in the individual
patients, not as the difference between the group means at base line and month 3.
TABLE 2. EFFECT OF TREATMENT ON MEASURES OF DISEASE ACTIVITY AT THREE MONTHS.
MEASURE OF ACTIVITY*PLACEBO
TNFR:FC
P
VALUE†
0.25 mg/m22 mg/m2
16 mg/m2
No. of patients
Swollen-joint count (no.)
Tender-joint count (no.)
Total count (no.)
Morning stiffness (hr)
Physician’s assessment‡
Patient’s assessment‡
Pain (VAS)‡
Quality of life (HAQ)§
Erythrocyte sedimentation rate (mm/hr)
C-reactive protein (mg/dl)
Improvement from base line (%)¶
Swollen-joint count
Tender-joint count
Total count
44
17
22
39
46
19
24
43
46
17
17
34
44
11
13
24
?0.001
?0.001
?0.001
0.004
?0.001
?0.001
0.001
?0.001
?0.001
?0.001
4.1
5.9
6.2
6.1
5.3
5.6
5.8
5.6
2.6
4.3
4.6
4.6
1.1
2.7
3.2
3.1
141
40
137
39
123
27
104
21
2.62.42.00.9
24
28
25
16
25
22
32
46
40
58
64
61
?0.001
?0.001
?0.001
Figure 1. Mean Swollen-Joint Count.
The shaded bar represents the treatment period. For each pa-
tient, missing values were replaced by the last available value.
25
22
19
16
13
10
012345
Month
Mean Swollen-Joint Count
Placebo
0.25 mg of TNFR:Fc/m2
2 mg of TNFR:Fc/m2
16 mg of TNFR:Fc/m2
Figure 2. Mean Tender-Joint Count.
The shaded bar represents the treatment period. For each pa-
tient, missing values were replaced by the last available value.
10
35
30
25
20
15
012345
Month
Mean Tender-Joint Count
Placebo
0.25 mg of TNFR:Fc/m2
2 mg of TNFR:Fc/m2
16 mg of TNFR:Fc/m2
mote rapid removal of TNF.45,54 In animals and in
humans, the administration of TNFR:Fc has been
shown to prolong the half-life of TNF; however, it
also renders the TNF biologically unavailable.45,54
Thus, TNFR:Fc acts as both a cytokine “carrier” and
a TNF antagonist. Another characteristic of TNFR:Fc
that might influence its activity is the ability to bind
to Fc receptors. However, the potential for Fc-recep-
tor binding of TNFR:Fc in vivo is probably minimal
because of the high concentration of immunoglob-
ulin in human plasma. Another potentially impor-
tant consideration is the capacity of TNFR:Fc to
bind to another inflammatory cytokine in the TNF
family, lymphotoxin-a (TNF-b).55 It is conceivable
Copyright © 1997 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at FAC DE FARMACIA on June 18, 2005 .