ARTHRITIS & RHEUMATISM

Vol. 58, No. 5, May 2008, pp 13241331

DOI 10.1002/art.23471
2008, American College of Rheumatology

Radiographic Progression of Ankylosing Spondylitis After

Up to Two Years of Treatment With Etanercept
D. van der Heijde,1 R. Landewe,2 S. Einstein,3 P. Ory,4 D. Vosse,2 L. Ni,5 S.-L. Lin,5
W. Tsuji,5 and J. C. Davis, Jr.6
Objective. To investigate the effect of etanercept
therapy on radiographic progression in patients with
ankylosing spondylitis (AS).
Methods. Patients with AS who had previously
participated in a 24-week randomized, double-blind,
placebo-controlled trial of etanercept therapy were enrolled in a 72-week open-label extension. Radiographs of
the cervical and lumbar spine from patients who received etanercept (25 mg twice weekly) for up to 96
weeks were compared with radiographs from patients in
a large prevalence cohort (Outcome Assessments in
Ankylosing Spondylitis International Study [OASIS])
who had not been treated with antitumor necrosis
factor (anti-TNF) agents. Radiographs obtained at 2
time points up to 96 weeks apart from patients in both
study populations were digitized and read by 2 indepen-

dent readers who were blinded with regard to patient

group and sequence. The primary end point was the
96-week change in the modified Stoke AS Spine Score
(mSASSS).
Results. A total of 257 patients treated with
etanercept were compared with 175 unselected patients
from the OASIS study. There was no significant difference in the change in the mSASSS from baseline among
patients who received etanercept (mean SD 0.91
2.45) versus those from the OASIS group (0.95 3.18).
Conclusion. Unlike other inflammatory rheumatic diseases such as rheumatoid arthritis and psoriatic arthritis, structural progression in AS seems to be
independent of TNF, despite the fact that TNF is
responsible for the signs and symptoms due to inflammation in this disease.

ClinicalTrials.gov identifier: NCT00356356.

Supported by Immunex Corporation, a wholly owned subsidiary of Amgen Inc., and by Wyeth Pharmaceuticals. The Rheumatology Department of the Academic Hospital Maastricht received compensation from Amgen for the use of the OASIS data set.
1
D. van der Heijde, MD, PhD: Leiden University Medical
Center, Leiden, The Netherlands; 2R. Landewe, MD, D. Vosse, MD:
University Hospital Maastricht, Maastricht, The Netherlands; 3S.
Einstein, MSE: Bio-Imaging Technologies Inc., Newtown, Pennsylvania; 4P. Ory, MD: University of Washington, Seattle; 5L. Ni, MS, S.-L.
Lin, MD, PhD, W. Tsuji, MD: Amgen Inc., Thousand Oaks, California; 6J. C. Davis, Jr., MD, MPH: University of California, San
Francisco.
Dr. van der Heijde has received consulting fees, speaking fees,
and/or honoraria (less than $10,000 each) from Abbott, Amgen,
Centocor, Wyeth, Chugai, UCB, Roche, and Schering-Plough. Dr.
Landewe has received consulting fees, speaking fees, and/or honoraria
(less than $10,000 each) from Abbott, Amgen, Bristol-Myers Squibb,
Centocor, Wyeth, UCB, and Schering-Plough. Dr. Ory has received
consulting fees, speaking fees, and/or honoraria (more than $10,000 each)
from Amgen, Abbott, and Targeted Genetics. Drs. Ni, Lin, and Tsuji own
stock or stock options in Amgen. Dr. Davis has received consulting fees,
speaking fees, and/or honoraria from Abbott and Wyeth.
Address correspondence and reprint requests to D. van der
Heijde, MD, PhD, Korte Raarberg 46, 6231 KR Meerssen, The
Netherlands. E-mail: d.vanderheijde@kpnplanet.nl.
Submitted for publication August 14, 2007; accepted in
revised form February 11, 2008.

Ankylosing spondylitis (AS) belongs to a family

of rheumatic diseases known as spondylarthritides that
characteristically cause spinal joint inflammation and
bony fusion of the spine. AS is the prototype of the
spondylarthritides and is typified by ankylosis of the
axial skeleton. Radiographic damage known to result
from AS primarily includes fusion of entheses of the
sacroiliac joints and of the posterior articulations and
ligaments of the spine. These fusions can lead to impaired spinal mobility and in turn decreased ability to
perform daily activities and severely reduced healthrelated quality of life (1).
Tumor necrosis factor (TNF) has been shown
to play an important role in the inflammatory response
observed in AS. It has been found at increased levels in
the serum and synovium of patients with AS (2,3), and
treatment with TNF-blocking agents (etanercept,
adalimumab, and infliximab) has been shown to safely
and effectively reduce the signs and symptoms of AS
(46) and significantly improve health-related quality of
life (1). In addition, these agents have been shown to
1324

RADIOGRAPHIC PROGRESSION OF AS AFTER ETANERCEPT TREATMENT FOR UP TO 96 WEEKS

suppress bony inflammation as detected on magnetic

resonance imaging (79).
TNF also plays a significant proinflammatory
role in rheumatoid arthritis (RA) and psoriatic arthritis
(PsA), 2 inflammatory rheumatic diseases that are dominated by bone destruction rather than bone formation.
TNF-blocking agents have been effective in reducing
disease activity as well as halting the destructive process
of these diseases (1013). This relationship between
disease activity (inflammation) and bone damage in RA
is well established, and prevention of radiographic damage through the suppression of the inflammatory process
is a widely recognized treatment goal.
A similar relationship between inflammation and
bone damage has not been demonstrated in AS (14). In
fact, studies in animal models of AS have suggested an
uncoupling of inflammation and bone formation in the
spine (15,16), and there is increasing evidence that bone
formation in the spine is under the influence of bone
morphogenetic protein (16,17) and the Wnt signaling
pathway (18). The role of TNF in this process is still not
fully elucidated. Findings of a number of uncontrolled
clinical studies have suggested that, analogous to the
situation in RA and PsA, anti-TNF agents may inhibit
progression of structural damage in AS (19,20). To date,
however, these anecdotal observations are not corroborated by solid evidence from controlled studies. In the
present study, using a 3-way blinded radiograph reading
design with 2 independent readers and adjudication
methodology, we investigated whether 96-week radiographic progression in a cohort of patients treated with
the TNF-blocking agent etanercept differed from the
96-week radiographic progression in an unrelated observational cohort of patients who had never been treated
with anti-TNF agents.
PATIENTS AND METHODS
Patients and study design. Patients in the etanercept
arm of this controlled study were enrolled in a 24-week
multicenter, double-blind, placebo-controlled randomized
controlled trial (RCT) followed by a 72-week open-label
extension during which all patients received etanercept
(ClinicalTrials.gov NCT00356356) (5). In the RCT, 277 AS
patients (139 in the placebo arm and 138 in the etanercept
arm) were followed up for 24 weeks. Eligible patients were
then given the option to enroll in an open-label extension to
evaluate the safety and efficacy of etanercept treatment for up
to 96 weeks. Patients were treated with etanercept 25 mg
subcutaneously twice weekly, and were allowed to receive
concomitant nonsteroidal antiinflammatory drugs (NSAIDs),
analgesics, and disease-modifying antirheumatic drugs
(DMARDs) including corticosteroids. All patients from the

1325

RCT and the open-label extension who received at least 1 dose

of etanercept and had baseline radiographs of the cervical and
lumbar spine were analyzed. Therefore, patients who received
etanercept in the RCT could have received up to 96 weeks of
continuous etanercept treatment, and patients who received
placebo in the RCT could have received up to 72 weeks of
continuous etanercept treatment.
Patients in the control arm of the present study were
enrolled in the Outcome Assessment in Ankylosing Spondylitis
International Study (OASIS) (21). The OASIS study was an
international observational study on outcome in AS patients
from 3 different countries (Belgium, France, and The Netherlands). It included consecutive patients who were followed up
for 10 years according to a predefined protocol, including
assessment of radiographs of the cervical and lumbar spine at
baseline, at 1-year and 2-year followup, and every 2 years
thereafter. Patients were treated according to common practice guidelines including the use of NSAIDs, analgesics, and
regular exercise therapy. All patients from the OASIS study
with radiographs at baseline and at 2 years, except patients
with complete spinal fusion at baseline (n 5), were evaluated
as controls for the present study (n 175).
Procedures and end points. Radiographs of the lateral
cervical and lateral lumbar spine, obtained at baseline and at
96 weeks in patients in the etanercept arm and the control arm,
were digitized and patient identifiers and temporal indicators
removed, to ensure blinding. Radiographs were scored independently by 2 trained readers who were blinded with regard to
treatment, temporal sequence, and patient group, using a
computer-assisted masked reading system. Radiographs were
scored using the modified Stoke AS Spine Score (mSASSS)
(22), identified as the preferred radiographic scoring method
in AS by the ASsessment in Ankylosing Spondylitis International Working Group (ASAS) and Outcome Measures in
Rheumatology Clinical Trials group (22,23). In the mSASSS,
all anterior corners (from the lower corner of T12 to the upper
corner of S1 and from the lower corner of C2 to the upper
corner of T1) are scored for the presence of erosions, sclerosis,
and/or squaring (1 point per site), nonbridging syndesmophytes (2 points per site), and bridging syndesmophytes (3
points per site). The mSASSS, which is the sum of the scores at
all individual sites, ranges from 0 to 72. In order to assess
intrareader variability, 12.5% of the patients radiographs were
reread by each reader. These patients were selected to equally
represent 4 quartiles of change from baseline, determined in
the initial reading. Intrareader reliability for status scores
(assessed by intraclass correlation coefficient) was 0.9 for
each of the readers, and interreader reliability was 0.81.
The primary analysis was a direct comparison of the
96-week change from baseline in the mSASSS between all
patients who had baseline radiographs in the etanercept group
and those in the control group, adjusted for baseline mSASSS.
Missing information on 96-week change from baseline in the
mSASSS due to missing postbaseline radiographs was imputed
using the median change from baseline score observed among
control patients with the same baseline score.
A secondary sensitivity analysis included a comparison
between all patients who had baseline and 96-week radiographs in the etanercept group and those in the control group.
Two sensitivity analyses were performed on a subpopulation: a
comparison between all patients who had baseline radiographs

1326

VAN DER HEIJDE ET AL

in the etanercept group and those from the OASIS study who
would have fulfilled the entry criteria for the RCT, and a
comparison between all patients who had baseline and 96-week
radiographs in the etanercept group and those from the
OASIS study who would have fulfilled the entry criteria for
the RCT.
Other sensitivity analyses involved simultaneous adjustment for multiple characteristics of disease activity at
baseline (i.e., mSASSS, Bath Ankylosing Spondylitis Functional Index [BASFI] [24], Bath Ankylosing Spondylitis Disease Activity Index [BASDAI] [25], and C-reactive protein
level) in the primary and secondary patient populations described above, or comparison of patients in the etanercept
group and in the control group after stratification according to
whether they regularly took NSAIDs during the study. Regular
NSAID use was defined as taking usual antiinflammatory
doses of medications throughout the duration of the study.
Patients who discontinued regular NSAIDs, took NSAIDs only
as needed for control of symptoms, or took NSAIDs at less
than the usual antiinflammatory doses were not considered to
be regular NSAID users. Further comparisons included analyses in which patients were stratified by duration of treatment
(etanercept patients who received etanercept for 48 weeks,
72 weeks, or 96 weeks), by treatment during the initial 24
week RCT (patients who received placebo or etanercept
during the RCT), and by response status (according to the
ASAS 40% response criteria [ASAS40]) (26).
Statistical analysis. Baseline characteristics of the
patients in the etanercept trial and patients in the OASIS study
were compared by chi-square test for categorical variables and
t-test for continuous variables. Radiographic progression was
compared between the patients in the etanercept trial and
those in the OASIS study using Quade rank analysis of
covariance, adjusted for baseline mSASSS on the change in
mSASSS from baseline to 96 weeks.

RESULTS
Patient characteristics at enrollment. A total of
93% of the patients (257 of 277) who had previously
received either placebo (n 139) or etanercept (n
138) in the RCT enrolled in the open-label extension
and had data available for radiographic analysis. Of the
257 patients, 76% received etanercept for at least 48
weeks and 50% for at least 72 weeks. Figure 1 shows the
disposition of patients in the RCT and open-label extension. A total of 175 of the 219 original OASIS patients
had available radiographs for analysis. Of these 175
patients, none had received therapy with etanercept or
another anti-TNF agent. Forty-three percent of them
(n 76) would have met the inclusion criteria for the
RCT at baseline.
Demographic and baseline clinical data on the
etanercept-treated patients and the control patients are
shown in Table 1. Mean age, disease duration, male/
female distribution, and frequency of HLAB27 positiv-

Figure 1. Patient disposition in the initial randomized controlled trial

(RCT) and the open-label extension (OLE). AE adverse event;
ISR injection site reaction.

ity were similar between the 2 groups. A significantly

smaller proportion of patients from the OASIS cohort
were taking DMARDs at baseline, compared with the
proportion of patients receiving etanercept (10% versus
32%; P 0.0001). As would be expected, OASIS
patients also had significantly less active disease, as
indicated by lower values for patient global assessment,
BASFI, and BASDAI (all P 0.0001). These values
were more comparable when the analysis was limited to
OASIS patients meeting the entry criteria for the RCT.
Notwithstanding these differences in disease activity
variables, measures assessing severity such as the
mSASSS, and spinal mobility scores were similar between patients in the 2 groups.
Efficacy. The changes in radiographic scores
among patients in the etanercept arm and those in the
control arm (OASIS) were similar between the 2 groups
(mean SD 0.91 2.45 and 0.95 3.18, respectively;
P 1.00) (Table 2). The probability plot of mean change

RADIOGRAPHIC PROGRESSION OF AS AFTER ETANERCEPT TREATMENT FOR UP TO 96 WEEKS

Table 1.

1327

Baseline characteristics of the patients*

OASIS meeting
RCT entry criteria
(n 76)

Etanercept
(n 257)

44 12.5
121 (69.1)
11 8.5
1.5 1.98
10 (5.7)

48 12.3
54 (71.1)
12 9.8
1.50 1.81
3 (3.9)

41 10.2
194 (75.5)
10 8.5
2.0 2.20
2 (0.8)

18 (10.3)
1 (0.6)
17 (9.7)
0 (0.0)
142 (81.1)
38 27.8
34 25.5
35 20.9
2.9 1.4
14 17.6

4 (5.3)
1 (1.3)
3 (3.9)
0 (0.0)
64 (84.2)
27 27.2
55 16.6
47 19.8
2.3 1.4
19 20.8

83 (32.3)
30 (11.7)
56 (21.8)
4 (1.6)
201 (78.2)
63 18.0
54 20.7
63 20.9
3.0 1.7
16 18.3

OASIS
(n 175)

Baseline characteristic
Age, mean SD years
Male, no. (%)
Duration of AS, mean SD years
CRP, mean SD mg/dl
Patients with CRP level outside normal range
of 01.00 mg/dl, no. (%)
DMARD use at baseline, no. (%)
Methotrexate
Sulfasalazine
Hydroxychloroquine
HLAB27 positive, no. (%)
Patient global assessment (0100), mean SD
BASFI, mean SD
BASDAI, mean SD
Modified Schober score, mean SD
mSASSS, mean SD

* Baseline refers to the baseline of the Outcome Assessment in Ankylosing Spondylitis International
Study (OASIS) or the randomized controlled trial (RCT) of etanercept; n values are the number of
patients with baseline radiographs. AS ankylosing spondylitis; CRP C-reactive protein; DMARD
disease-modifying antirheumatic drug; BASFI Bath Ankylosing Spondylitis Functional Index;
BASDAI Bath Ankylosing Spondylitis Disease Activity Index; mSASSS modified Stoke Ankylosing
Spondylitis Spine Score.
P 0.0001 versus etanercept-treated patients.

in the mSASSS illustrated that the change was close to 0 in

most patients (Figure 2). When mean change in the
mSASSS was compared between patients who received
etanercept and OASIS patients who met RCT entry requirements at baseline, they were again found to be similar,
as indicated by the closely overlapping probability plots
(Figure 3). In the latter group, the mean SD change was
1.27 3.64 (P not significant versus patients who received
etanercept) (Table 2). Twenty-nine patients in the etanercept group (11%) had missing postbaseline radiographs,
while no patients in the OASIS cohort had missing postbaseline radiographs. When mean change in the mSASSS
was compared between the etanercept-treated patients
without missing postbaseline radiographs (n 228) and
the OASIS patients, no significant difference was detected
(P 0.0.83).

Table 2.

Changes in mean baseline cervical and lumbar

radiography scores were also compared between patients receiving etanercept and, as 2 separate groups,
patients from the OASIS cohort overall and patients
from the OASIS cohort who met entry criteria for the
RCT. Again, there was no significant difference in the
change in cervical or lumbar radiography scores between
the patients receiving etanercept and the OASIS patients (P 0.28 and P 0.29, respectively). Similarly,
there was no significant difference between the mean
changes in cervical or lumbar radiography scores between patients receiving etanercept and the OASIS
patients who met entry criteria for the RCT (P 0.66
and P 0.17, respectively). The mean changes in
cervical or lumbar radiography scores were also found to
be similar when the OASIS patients were compared with

Changes in radiography scores at 96 weeks*

mSASSS
Cervical radiography score
Lumbar radiography score

OASIS
(n 175)

OASIS meeting
RCT entry criteria
(n 76)

Etanercept
(n 257)

0.95 3.18
0.42 2.11
0.53 1.88

1.27 3.64
0.53 2.29
0.73 2.00

0.91 2.45
0.49 1.40
0.42 1.84

* Values are the mean SD change from baseline. There were no statistically significant differences
between groups in the mean change in score. See Table 1 for definitions.

1328

Figure 2. Probability plot of 2-year progression in the modified Stoke

Ankylosing Spondylitis Spine Score (mSASSS) in patients receiving
etanercept and patients in the Outcome Assessment in Ankylosing
Spondylitis International Study (OASIS), for whom baseline and
2-year radiographs were available.

only the etanercept-treated patients who had postbaseline radiographs (as opposed to imputation of data for
etanercept-treated patients from whom postbaseline radiographs were not available) (P 0.23 and P 0.36,
respectively).
In multiple subgroup analyses, there were also no
significant differences in the mean change in the mSASSS.
When etanercept-treated patients were subgrouped according to duration of treatment, the mean change in the

Figure 3. Probability plot of 2-year progression in the modified Stoke

Ankylosing Spondylitis Spine Score (mSASSS) in patients receiving
etanercept and patients in the Outcome Assessment in Ankylosing
Spondylitis International Study (OASIS) who met the entry criteria for
the randomized controlled trial at baseline, for whom baseline and
96-week radiographs were available.

VAN DER HEIJDE ET AL

Figure 4. Distribution of the type of radiographic changes. Values are

the percentage of sites at which the given type of change (as determined by change in radiography score from baseline) was noted.
OASIS Outcome Assessment in Ankylosing Spondylitis International Study.

mSASSS in each subgroup (48 weeks, 72 weeks, or

96 weeks) was similar to the mean change in OASIS
patients (P 0.98, P 0.82, and P 0.92, respectively).
There was no significant difference in the mean change in
mSASSS when OASIS patients were compared with either
open-label etanercepttreated patients who received etanercept during the RCT (P 0.61) or those who received
placebo during the RCT (P 0.55), or when OASIS
patients were compared with either etanercept-treated
patients who were ASAS40 responders (P 0.60) or those
who were ASAS40 nonresponders (P 0.38). A total of
145 of 257 etanercept-treated patients (56%) and 92 of 175
OASIS patients (53%) were characterized as regular
NSAID users. Stratifying patients by regular use of
NSAIDs during the study did not affect the results. Additionally, adjustment for multiple covariates had no effect
on the outcome.
An exploratory analysis was performed to investigate the character of the radiographic changes observed in each cohort (Figure 4). While radiographic
changes occurred at only a small percentage of the sites
analyzed, the prevailing type of change that did occur
was compatible with syndesmophyte formation or syndesmophyte growth (bridging). There were no differences between the 2 groups in this analysis. These results
support the hypothesis that progression in AS is due to
syndesmophyte formation and growth (inappropriate
bone formation), rather than to erosions and squaring
(bone resorption).
DISCUSSION
Results of this study, the first blinded, controlled
trial of radiographic changes in AS patients treated with

RADIOGRAPHIC PROGRESSION OF AS AFTER ETANERCEPT TREATMENT FOR UP TO 96 WEEKS

the TNF antagonist etanercept, provide evidence that

etanercept does not inhibit syndesmophyte formation in
AS. Findings of previous uncontrolled studies have
suggested that TNF blockers (both etanercept and infliximab) may slow the progression of structural damage
in AS compared with that which occurs in patients not
taking TNF blockers (19,20). However, results of these
uncontrolled studies should be interpreted with great
caution for several reasons. Multiple readers who were
aware of the origin of the radiographs (patient group)
were used. In addition, these studies used scores that
were assigned by readers who were blinded with regard
to time order and compared them with results from the
literature that were obtained based on known time
order. This may have a major effect on the observed
progression rate, which is substantially lower if the
reader is blinded with regard to the sequence of the films
(27). Finally, readers may have been biased against
progression, knowing that all radiographs they read
originated from trials with TNF-blocking drugs.
A randomized placebo-controlled trial would
have the highest validity in assessing the effect of
TNF-blocking drugs on radiographic progression. However, feasibility of such a trial in patients with AS is very
limited (28). The slow rate of radiographic progression
in AS requires a minimum followup of 2 years, and an
appropriate comparator with clinical efficacy for signs
and symptoms similar to that obtained with TNF blockade is not available. In addition, appropriate prognostic
factors for radiographic progression, allowing selection
of patients prone to progression, as in RA, have not
been identified until recently (29). Also, TNF inhibitors
rapidly and effectively control clinical symptoms in a
large number of patients with AS and are now available
for their treatment. For these reasons, a 2-year placebocontrolled study is not considered feasible or ethical by
most rheumatologists. If performed, such a study would
be complicated by a high dropout rate early in the trial
with many patients crossing over to TNF blockers,
limiting the ability for detection of a treatment effect.
The best alternative to a concurrently controlled study is
comparison with existing radiographs from patients who
were not treated with TNF-blocking drugs.
Use of a historic control group imposes special
requirements. Historic control patients should be unselected and representative of the entire population of
patients with AS, should have never been treated with
TNF blockers, should be well characterized, and should
have radiographs available with a 2-year interval by
protocol. The OASIS cohort, consisting of patients who

1329

were enrolled in an observational study and were followed up for many years, fulfills these requirements.
Radiographs from the patients in the historic
control group and from the patients treated with a TNF
blocker should be scored in one session. Also, readers
should be blinded with regard to the origin of the
radiograph, the clinical data, and the sequence of the
radiographs. Once more, all of these requirements were
fulfilled in the present study.
There was a difference in disease activity between
patients in the OASIS study, who were unselected and had
a wide range of disease activity, and patients in the
etanercept group, all of whom had active disease. However,
disease activity in AS may be unrelated to structural
progression (30). In the OASIS cohort and in a separate
cohort of AS patients, signs of clinical disease activity have
proven to be unrelated to structural progression as assessed
radiographically (31). The validity of our observations is
supported by the fact that the results were not altered when
the comparison was limited to patients in the OASIS
cohort who would have fulfilled the disease activity entry
criteria for the RCT or to etanercept-treated patients who
had postbaseline radiographs (as opposed to imputing the
missing data for those without postbaseline radiographs),
or after adjustment for duration of exposure to etanercept,
disease activity at baseline, or NSAID use.
There are methodologic issues that may have confounded the results and deserve attention. DMARD use
was more frequent among patients in the etanercept group
than among those in the OASIS group. There are, however, no data showing that DMARDs can inhibit structural
progression, and if so, this would result in a lower rather
than a higher progression rate in the etanercept group. In
addition, NSAID use was more frequent in the etanercept
group at baseline, and similar proportions of patients in the
etanercept cohort and the OASIS cohort were classified as
regular NSAID users. There has been some indication that
NSAIDs, especially if used continuously, are able to inhibit
syndesmophyte formation (32), yet as noted above, no
difference was detected when our analyses were adjusted
for NSAID use.
Based on experience regarding radiographic assessments of patients with RA (and PsA), many rheumatologists expected that TNF-blocking drugs would
inhibit structural progression in AS. These expectations
may have been unrealistic. There are major differences
in pathophysiologic mechanisms between RA and AS:
structural damage in AS is dominated by inappropriate
bone formation (syndesmophytes), while in RA the
destructive process is mainly due to bone resorption.
Syndesmophytes may reflect inappropriate repair that is

1330

induced, but not necessarily maintained, by inflammatory stress in AS (33). In contrast to AS, the destructive
process that dominates in RA is well characterized and
is regulated by TNF and RANKL, leading to activation
of osteoclasts, while inhibitors of Wnt proteins, such as
Dkk-1 (18), cause a decrease in osteoblast formation.
The net result is the rapid formation of erosions without
sufficient repair. In AS, bone formation dominates the
picture. Bone formation is regulated by the transforming
growth factor/bone morphogenetic protein family as well
as the group of Wnt proteins. Wnt signaling activates
osteoprotegerin, which counteracts RANKL-induced osteoclast activation, a TNF-dependent process. It has
recently been shown that Dkk-1 levels are decreased in
AS and increased in RA (18), suggesting that Wnt
signaling cascades are switched on in AS while being
suppressed in RA.
These pathophysiologic considerations are supported by clinical observations. In a recent investigation,
as well as in this study, it was confirmed that syndesmophyte formation is the dominant feature of structural
progression in AS, while erosions at the corners of the
vertebrae play only a minor role (34). Moreover, we and
others have identified clear causal relationships between
inflammation and formation of erosions in RA (35,36),
while a relationship between inflammation and syndesmophyte formation in AS could not be established. The
current controlled study confirms what might be expected on pathophysiologic grounds, i.e., that the formation of syndesmophytes may not be influenced by
inhibition of TNF-regulated inflammation. Additional
evidence that TNF inhibition may not play a role in
syndesmophyte formation comes from a recent preliminary report that infliximab, a monoclonal antibody that
blocks TNF activity, also failed to inhibit radiographic
progression in AS patients after 2 years of use (37).
Despite all of the evidence that TNF blockade
does not inhibit syndesmophyte formation, many investigators still believe there could be a relationship between inflammation as a trigger and syndesmophyte
formation as a result. Such an assumption implies that
an alternative explanation for the lack of a treatment
effect should be sought. The average disease duration of
the etanercept-treated AS patients in this study was 10
years. Assuming that inflammation triggers the initiation
of syndesmophyte formation, it could be hypothesized
that earlier intervention in the disease process, before
the reparative processes have started, could prevent
further bone formation. Another hypothesis is that
inflammation would have to be suppressed for a longer
period of time before the inhibitory effects can be seen

VAN DER HEIJDE ET AL

on radiographs. Such a hypothesis would gain support if,

with longer followup (e.g., 4 years), a reduction in
radiographic progression over time could be detected in
comparison with the control group.
In conclusion, this first large, controlled study did
not demonstrate inhibition of structural progression of
spine disease in AS patients treated with etanercept for
nearly 2 years, even though etanercept has been shown to
be highly effective in treating clinical signs and symptoms.
This differential effect may be due to dissociation of the
TNF-dependent inflammatory processes and the TNFindependent bone formation processes in AS.
AUTHOR CONTRIBUTIONS
Dr. van der Heijde had full access to all of the data in the
study and takes responsibility for the integrity of the data and the
accuracy of the data analysis.
Study design. Van der Heijde, Landewe, Einstein, Vosse, Tsuji, Davis.
Acquisition of data. Landewe, Einstein, Ory, Vosse, Lin, Tsuji, Davis.
Analysis and interpretation of data. Van der Heijde, Landewe,
Einstein, Vosse, Ni, Lin, Tsuji, Davis.
Manuscript preparation. Van der Heijde, Landewe, Vosse, Lin, Tsuji,
Davis, and Dr. Marc. D. Kubasak (nonauthor; Amgen Inc.).
Statistical analysis. Landewe, Ni, Lin, Tsuji.
Reading of radiographs. Ory, Vosse.

ROLE OF THE STUDY SPONSORS

Marc D. Kubasak, PhD (Amgen Inc.), assisted with the
writing and preparation of the manuscript. Immunex Corporation, a
wholly owned subsidiary of Amgen Inc., and Wyeth Pharmaceuticals
facilitated the study design and the writing of the manuscript, and
reviewed and approved the manuscript prior to submission. The
authors independently collected the data, interpreted the results, and
had the final decision to submit the manuscript for publication.

REFERENCES
1. Davis JC, van der Heijde D, Dougados M, Woolley JM. Reduction
in health-related quality of life in patients with ankylosing spondylitis and improvements with etanercept therapy. Arthritis
Rheum 2005;53:494501.
2. Braun J, Bollow M, Neure L, Seipelt E, Seyrekbasan F, Herbst H, et
al. Use of immunohistologic and in situ hybridization techniques in
the examination of sacroiliac joint biopsy specimens from patients
with ankylosing spondylitis. Arthritis Rheum 1995;38:499505.
3. Francois RJ, Neure L, Sieper J, Braun J. Immunohistological
examination of open sacroiliac biopsies of patients with ankylosing
spondylitis: detection of tumor necrosis factor in two patients
with early disease and transforming growth factor in three more
advanced cases. Ann Rheum Dis 2006;65:71320.
4. Braun J, Brandt J, Listing J, Zink A, Alten R, Burmester G, et al.
Long-term efficacy and safety of infliximab in the treatment of
ankylosing spondylitis: an open, observational, extension study of a
three-month, randomized, placebo-controlled trial. Arthritis
Rheum 2003;48:222433.
5. Davis JC, van der Heijde D, Braun J, Dougados M, Cush J, Clegg
DO, et al. Sustained durability and tolerability of etanercept in
ankylosing spondylitis from 96 weeks. Ann Rheum Dis 2005;64:
155762.
6. Van der Heijde D, Kivitz A, Schiff MH, Sieper J, Dijkmans BA,

RADIOGRAPHIC PROGRESSION OF AS AFTER ETANERCEPT TREATMENT FOR UP TO 96 WEEKS

7.

8.

9.

10.

11.
12.

13.

14.
15.
16.

17.
18.
19.

20.

21.

Braun J, et al, for the ATLAS Study Group. Efficacy and safety of
adalimumab in patients with ankylosing spondylitis: results of a
multicenter, randomized, double-blind, placebo-controlled trial.
Arthritis Rheum 2006;54:213646.
Baraliakos X, Davis J, Tsuji W, Braun J. Magnetic resonance imaging
examinations of the spine in patients with ankylosing spondylitis
before and after therapy with the tumor necrosis factor receptor
fusion protein etanercept. Arthritis Rheum 2005;52:121623.
Braun J, Landewe R, Hermann KG, Han J, Yan S, Williamson P,
et al, for the ASSERT Study Group. Major reduction in spinal
inflammation in patients with ankylosing spondylitis after treatment with infliximab: results of a multicenter, randomized, doubleblind, placebo-controlled magnetic resonance imaging study. Arthritis Rheum 2006;54:164652.
Haibel H, Rudwaleit M, Brandt HC, Grozdanovic Z, Listing J,
Kupper H, et al. Adalimumab reduces spinal symptoms in active
ankylosing spondylitis: clinical and magnetic resonance imaging
results of a fifty-twoweek open-label trial. Arthritis Rheum
2006;54:67881.
Keystone EC, Kavanaugh AF, Sharp JT, Tannenbaum H, Hua Y,
Teoh LS, et al. Radiographic, clinical, and functional outcomes of
treatment with adalimumab (a human antitumor necrosis factor
monoclonal antibody) in patients with active rheumatoid arthritis
receiving concomitant methotrexate therapy: a randomized,
placebo-controlled, 52-week trial. Arthritis Rheum 2004;50:
140011.
Mease PJ, Kivitz AJ, Burch FX, Siegel EL, Cohen SB, Ory P, et al.
Etanercept treatment of psoriatic arthritis: safety, efficacy, and
effect on disease progression. Arthritis Rheum 2004;50:226472.
Kavanaugh A, Antoni CE, Gladman D, Wassenberg S, Zhou B,
Beutler A, et al. The Infliximab Multinational Psoriatic Arthritis
Controlled Trial (IMPACT): results of radiographic analyses after
1 year. Ann Rheum Dis 2006;65:103843.
Van der Heijde D, Klareskog L, Rodriguez-Valverde V, Condreanu C, Bolosiu H, Melo-Gomes J, et al, for the TEMPO Study
Investigators. Comparison of etanercept and methotrexate, alone
and combined, in the treatment of rheumatoid arthritis: two-year
clinical and radiographic results from the TEMPO study, a doubleblind, randomized trial. Arthritis Rheum 2006;54:106374.
Marzo-Ortega H, Emery P, McGonagle D. The concept of disease
modification in spondyloarthropathy. J Rheumatol 2002;29:15835.
Bardos T, Szabo Z, Czipri M, Vermes C, Tunyogi-Csapo M, Urban
RM, et al. A longitudinal study on an autoimmune murine model
of ankylosing spondylitis. Ann Rheum Dis 2005;64:9817.
Lories RJ, Daans M, Derese I, Matthys P, Kasran A, Tylzanowski
P, et al. Noggin haploinsufficiency differentially affects tissue
responses in destructive and remodeling arthritis. Arthritis Rheum
2006;54:173646.
Lories RJ, Derese I, Luyten FP. Modulation of bone morphogenetic protein signaling inhibits the onset and progression of
ankylosing enthesitis. J Clin Invest 2005;115:15719.
Diarra D, Stolina M, Polzer K, Zwerina J, Ominsky MS, Dwyer D,
et al. Dickkopf-1 is a master regulator of joint remodeling. Nat
Med 2007;13:15663.
Dijkmans BA, Wanders A, van der Heijde D, Fatenejad S.
Radiographic results from a long-term multicenter trial of etanercept (Enbrel) in patients with ankylosing spondylitis [abstract].
Arthritis Rheum 2005:52 Suppl 9:S634.
Baraliakos X, Listing J, Rudwaleit M, Brandt J, Sieper J, Braun J.
Radiographic progression in patients with ankylosing spondylitis
after 2 years of treatment with the tumour necrosis factor
antibody infliximab. Ann Rheum Dis 2005;64:14626.
Spoorenberg A, van der Heijde D, de Klerk E, Dougados M, der
Vlam K, Mielants H, et al. A comparative study of the usefulness
of the Bath Ankylosing Spondylitis Functional Index and the
Dougados Functional Index in the assessment of ankylosing
spondylitis. J Rheumatol 1999;26:9615.

1331

22. Creemers MC, Franssen MJ, van t Hof MA, Gribnau FW, van de
Putte LB, van Riel PL. Assessment of outcome in ankylosing
spondylitis: an extended radiographic scoring system. Ann Rheum
Dis 2005;64:1279.
23. Van der Heijde D, Landewe R, the ASAS Working Group. Selection
of a method for scoring radiographs for ankylosing spondylitis clinical
trials, by the Assessment in Ankylosing Spondylitis Working Group
and OMERACT. J Rheumatol 2005;32:20489.
24. Calin A, Garrett S, Whitelock H, Kennedy LG, OHea J, Mallorie
P, et al. A new approach to defining functional ability in ankylosing spondylitis: the development of the Bath Ankylosing Spondylitis Functional Index. J Rheumatol 1994;21:22815.
25. Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P,
Calin A. A new approach to defining disease status in ankylosing
spondylitis: the Bath Ankylosing Spondylitis Disease Activity
Index. J Rheumatol 1994;21:228691.
26. Brandt J, Listing J, Sieper J, Rudwaleit M, van der Heijde D,
Braun J. Development and preselection of criteria for short-term
improvement after anti-TNF therapy in ankylosing spondylitis.
Ann Rheum Dis 2004;63:143844.
27. Wanders A, Landewe R, Spoorenberg A, de Vlam K, Mielants H,
Dougados M, et al. Scoring of radiographic progression in randomized clinical trials in ankylosing spondylitis: a preference for
paired reading order. Ann Rheum Dis 2004;63:16014.
28. Van der Heijde D, Landewe R, van der Linden S. How should
treatment effect on spinal radiographic progression in patients
with ankylosing spondylitis be measured? Arthritis Rheum 2005;
52:197985.
29. Maksymowych WP, Landewe R, Conner-Spady B, Dougados M,
Mielants H, van der Tempel H, et al. Serum matrix metalloproteinase 3 is an independent predictor of structural damage progression in patients with ankylosing spondylitis. Arthritis Rheum
2007;56:184653.
30. Landewe R, Wanders AJ, Mielants H, Dougados M, van der Heijde
D. Function in patients with ankylosing spondylitis is the resultant of
patient reported disease activity and radiographic damage of the
spine [abstract]. Ann Rheum Dis 2004;63 Suppl 1:889.
31. Wanders A, Landewe R, Dougados M, Mielants H, van der Linden
S, van der Heijde D. Association between radiographic damage of the
spine and spinal mobility for individual patients with ankylosing
spondylitis: can assessment of spinal mobility be a proxy for radiographic evaluation? Ann Rheum Dis 2005;64:98894.
32. Wanders A, van der Heijde D, Landewe R, Behier JM, Calin A,
Olivieri I, et al. Nonsteroidal antiinflammatory drugs reduce
radiographic progression in patients with ankylosing spondylitis: a
randomized clinical trial. Arthritis Rheum 2005;52:175665.
33. Schett G. Landewe R, van der Heijde D. Tumor necrosis factor
blockers and structural remodeling in ankylosing spondylitis: what
is reality and what is fiction? Ann Rheum Dis 2007;66:70911.
34. Baraliakos X, Listing J, Rudwaleit M, Haibel H, Brandt J, Sieper
J, et al. Progression of radiographic damage in patients with
ankylosing spondylitis: defining the central role of syndesmophytes. Ann Rheum Dis 2007;66:9105.
35. Boers M, Kostense PJ, Verhoeven AC, van der Linden S, for the
COBRA Trial Group. Inflammation and damage in an individual
joint predict further damage in that joint in patients with early
rheumatoid arthritis. Arthritis Rheum 2001;44:22426.
36. Welsing PM, Landewe RB, van Riel PL, Boers M, van Gestel AM,
van der Linden S, et al. The relationship between disease activity
and radiologic progression in patients with rheumatoid arthritis: a
longitudinal analysis. Arthritis Rheum 2004;50:208293.
37. Van der Heijde D, Landewe R, Deodar A, Baker D, Han J, Xu W,
et al. Radiographic progression in patients with ankylosing spondylitis after 2 years of treatment not inhibited with infliximab
[abstract]. Ann Rheum Dis 2007;66 Suppl II:856.