Efficacy of adalimumab in the treatment of axial spondylarthritis without radiographically defined sacroiliitis: Results of a twelve-week randomized, double-blind, placebo-controlled trial followed by an open-label extension up to week fifty-two

Authors

  • Hildrun Haibel,

    1. Charité Medical University Hospital, Campus Benjamin Franklin, Berlin, Germany
    Search for more papers by this author
    • Dr. Haibel has received speaking fees from Abbott and Schering-Plough (less than $10,000 each).

  • Martin Rudwaleit,

    1. Charité Medical University Hospital, Campus Benjamin Franklin, Berlin, Germany
    Search for more papers by this author
    • Dr. Rudwaleit has received speaking fees from Abbott, Schering-Plough, Centocor, Wyeth, and Novartis (less than $10,000 each).

  • Joachim Listing,

    1. German Rheumatism Research Centre, Berlin, Germany
    Search for more papers by this author
  • Frank Heldmann,

    1. Centre of Rheumatology Herne, Herne, Germany
    Search for more papers by this author
  • Robert L. Wong,

    1. Abbott Laboratories, Parsippany, New Jersey
    Search for more papers by this author
    • Drs. Wong and Kupper own stock or stock options in Abbott Laboratories.

  • Hartmut Kupper,

    1. Abbott GmbH & Company KG, Ludwigshafen, Germany
    Search for more papers by this author
    • Drs. Wong and Kupper own stock or stock options in Abbott Laboratories.

  • Jürgen Braun,

    1. Centre of Rheumatology Herne, Herne, Germany
    Search for more papers by this author
    • Dr. Braun has received speaking fees, consulting fees, and/or honoraria from Schering-Plough, Wyeth, Abbott, and Centocor (less than $10,000 each).

  • Joachim Sieper

    Corresponding author
    1. Charité Medical University Hospital, Campus Benjamin Franklin, Berlin, and German Rheumatism Research Centre, Berlin, Germany
    • Medical Department I, Rheumatology, Charité Medical University Hospital, Campus Benjamin Franklin, Hindenburgdamm 30, 12200 Berlin, Germany
    Search for more papers by this author
    • Dr. Sieper has received speaking fees, consulting fees, and/or honoraria from Schering-Plough, Wyeth, Abbott, Roche, and Pfizer (less than $10,000 each).


  • ClinicalTrials.gov identifier: NCT00235105.

Abstract

Objective

To evaluate the efficacy and safety of the tumor necrosis factor (TNF) antagonist adalimumab in patients with axial spondylarthritis (SpA) without radiographically defined sacroiliitis refractory to conventional treatment.

Methods

Patients with active axial SpA (n = 46) were randomized to receive placebo or adalimumab at a dosage of 40 mg subcutaneously every other week for 12 weeks, followed by an open-label extension that continued up to week 52. The diagnosis of axial SpA required the presence of 3 of 6 diagnostic criteria, including 2 of the following 3 criteria: inflammatory back pain, HLA–B27 positivity, or acute inflammation of the spine or sacroiliac joints on magnetic resonance imaging, in the absence of radiographic evidence of sacroiliitis. The primary end point was a 40% response according to the improvement criteria of the Assessment of SpondyloArthritis international Society (ASAS40).

Results

All 46 patients (22 receiving adalimumab and 24 receiving placebo) completed the 12-week trial; 38 patients completed the extension period to week 52. At week 12, an ASAS40 response was achieved by 54.5% of the adalimumab-treated patients, as compared with 12.5% of the placebo-treated patients (P = 0.004). After switching to adalimumab, a similar degree of efficacy was also achieved by the patients who were initially treated with placebo. Efficacy was maintained in all patients until week 52. Young age at study entry and an elevated C-reactive protein concentration were the best predictors of achieving an ASAS40 response. Serious adverse events occurred in 5 patients, none of which was related to the study drug.

Conclusion

Adalimumab is the first TNF antagonist to demonstrate good clinical efficacy and safety in patients with axial SpA without radiographically defined sacroiliitis.

The tumor necrosis factor (TNF) antagonists adalimumab, etanercept, and infliximab have all been proven to be highly effective in the treatment of patients with ankylosing spondylitis (AS) that continued to be active despite treatment with nonsteroidal antiinflammatory drugs (NSAIDs) (1–4). In all of these studies, an ∼50% improvement in the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI 50) or similar outcome measures was reported. All patients included in these trials fulfilled the modified New York criteria for AS (5), which in addition to clinical criteria, require the presence of radiographic changes in the sacroiliac joints of at least grade II bilaterally or grade III unilaterally.

However, these radiographic definitions reflect only chronic changes in the axial skeleton, such as joint erosions, sclerosis, and ankylosis, which are not manifestations of active inflammation, but rather, are the consequences of previous inflammation. In fact, despite ongoing inflammation, it can take several years for these radiographic changes to become evident in some patients (6). Recent magnetic resonance imaging (MRI) studies have demonstrated the presence of active inflammation in the sacroiliac joints, spine, and other skeletal sites in patients with axial spondylarthritis (SpA) whose radiographic findings are normal (7–9). Some years ago, it was proposed that so-called undifferentiated SpA should be considered part of the entire spectrum of SpA (10, 11).

We have previously used the term pre-radiographic axial SpA instead of undifferentiated SpA to describe SpA in patients who do not have definite radiographic evidence of sacroiliac joint changes (6). Although the majority of patients with pre-radiographic axial SpA develop radiographic features typical of sacroiliitis later in the disease course, the disease does not progress in this manner in all SpA patients (6, 12, 13). Therefore, in the present study, we used the term axial SpA without radiographically defined sacroiliitis to describe these patients. We recently proposed that the disease phases with and without radiographically defined sacroiliitis should be viewed as different stages of the same disease and that despite negative radiographic findings, patients in the pre-radiographic stage of SpA could be reliably diagnosed with the use of a combination of clinical, laboratory, and imaging assessments (6, 14).

Early in the course of AS, symptoms of inflammation are more likely to prevail, whereas at later stages, symptoms are often caused by both inflammatory and secondary changes. TNF antagonists are highly antiinflammatory and would therefore be expected to provide effective treatment. Indeed, in a previous study, we found that patients with AS of shorter duration, with better physical functioning, and with elevated C-reactive protein (CRP) concentrations were more likely to experience a clinical response when treated with TNF antagonists (15). Moreover, we have also found that patients in the pre-radiographic stage of AS have at least the same level of disease activity and pain as patients with established AS (16).

The first placebo-controlled study of patients with undifferentiated axial SpA failed to demonstrate a major effect of sulfasalazine therapy (17). Until now, the efficacy of TNF antagonists in the treatment of patients with predominantly axial SpA without radiographically defined sacroiliitis has been reported in only a few open-label studies in which small numbers of patients were evaluated (18–22). The present study is the first randomized placebo-controlled trial of TNF antagonists in this unique group of patients. In this study, we sought to determine how the level of response in these patients compared with the previously reported level of response in patients with established AS.

PATIENTS AND METHODS

Study design.

In this double-blind, placebo-controlled trial, 46 patients with active axial SpA without radiographically defined sacroiliitis were randomized to receive placebo or adalimumab at a dosage of 40 mg subcutaneously every other week for 12 weeks, followed by an open-label extension that continued up to week 52. All patients were refractory to treatment with NSAIDs, which was defined as a lack of response to at least 1 NSAID at the maximum tolerated dosage, as judged by the investigator. Study drug was provided as a subcutaneous solution for injection in 1-ml prefilled syringes containing either adalimumab 40 mg or matching placebo (Abbott Laboratories, Abbott Park, IL).

This study was conducted at 2 centers in Germany (Charité Medical University Hospital, Berlin, and Centre of Rheumatology Herne) in accordance with the ethics principles of the Declaration of Helsinki. The study was approved by an independent ethics committee. Signed informed consent was obtained from each patient before any study-related procedures were performed. The ClinicalTrials.gov identifier is NCT00235105.

Diagnostic criteria for axial SpA without radiographically defined sacroiliitis.

The diagnosis of axial SpA without radiographically defined sacroiliitis was based on the presence of chronic low back pain (>3 months' duration), an age at symptom onset of <50 years, and the fulfillment of at least 3 of the following 6 criteria, including at least 2 of the first 3 criteria: 1) inflammatory back pain (23); 2) HLA–B27 positivity; 3) an MRI showing active inflammation of the spine or sacroiliac joints; 4) a history of a good response to NSAID treatment; 5) the presence (current or past) of 1 or more extraspinal manifestations (anterior uveitis, peripheral arthritis, or enthesitis); and 6) a family history of SpA (Table 1). Patients with radiographically defined sacroiliitis as required by the modified New York criteria (5) were excluded.

Table 1. Characteristics of the study patients at baseline, by treatment group*
 Placebo (n = 24)Adalimumab (n = 22)
  • *

    Patients with active inflammatory lesions in either the sacroiliac joints or the spine on magnetic resonance imaging (MRI) were considered MRI positive. Responses to nonsteroidal antiinflammatory drugs (NSAIDs) were determined by patient-based questionnaire and were scored on a scale of 1–4 (1 = very good, 2 = good, 3 = moderate, and 4 = no response); scores of 1 and 2 were counted as a good response to NSAIDs. Enthesitis was measured according to the Maastricht Ankylosing Spondylitis Enthesitis Score (range 0–13). BASDAI = Bath Ankylosing Spondylitis Disease Activity Index (range 0–10).

Age, mean (range) years37 (26–54)38 (25–64)
% female5059
Disease duration, mean (range) years8 (1–24)7 (2–16)
% HLA–B27 positive7559
% MRI positive7555
% HLA–B27 positive and MRI positive5414
% with inflammatory back pain100100
% with a history of good response to NSAIDs7582
% with extraspinal manifestations7568
% with a history of uveitis418
% with peripheral arthritis2932
% with enthesitis5482
BASDAI, mean (range)6.3 (4.2–8.9)6.7 (4.2–7.9)

The region of the axial skeleton that was evaluated by MRI (using T1, T2, and short tau inversion recovery sequences) was chosen according to the patient's symptoms, but always included the sacroiliac joints. Active inflammation was diagnosed by consensus opinion of a rheumatologist (JS) and a radiologist; no formal criteria for the diagnosis of active inflammation were defined. MRI of the sacroiliac joints was performed on 45 patients. MRI of the sacroiliac joints plus the spine was performed on 12 patients.

Inclusion and exclusion criteria.

Eligible patients were at least 18 years of age and had active disease, defined as a BASDAI ≥4, despite treatment with NSAIDs (24). Treatment with disease-modifying antirheumatic drugs or >7.5 mg/day of prednisone (or equivalent) was not permitted during the study and had to be discontinued at least 1 month before initiation of adalimumab therapy. Treatment with biologic agents had to be discontinued at least 12 weeks before study entry. Patients who were pregnant, had a history of uncontrolled concomitant diseases, or had abnormal results or clinically relevant changes on clinical and laboratory examinations were excluded from study participation.

Patients were screened for latent tuberculosis according to local guidelines, which included a tuberculin skin test (Mendel-Mantoux test with 10.0 IU of purified protein derivative) and a chest radiograph. Patients who tested positive for latent tuberculosis were required to initiate isoniazid treatment (300 mg/day for a total of 9 months) at least 4 weeks before the first injection of study drug.

Outcome assessments.

Clinical and laboratory outcome assessments were performed at screening, at baseline, and at weeks 2, 4, 8, 12, 16, 20, 28, 36, 44, and 52. These included the BASDAI, the Bath Ankylosing Spondylitis Functional Index (BASFI) (25), the Bath Ankylosing Spondylitis Metrology Index (BASMI) (26), and the patient's and physician's global assessments of disease activity. General pain and nocturnal pain levels were measured with a 0–10 numerical rating scale. Assessments of health-related quality of life included the Short Form 36 health survey (SF-36) (27), the EuroQol (EQ-5D) Index (28), and the Ankylosing Spondylitis Quality of Life (ASQoL) questionnaire (29) (measured only at baseline and at week 52). The Maastricht Ankylosing Spondylitis Enthesitis Score (MASES) (30) was used to assess 13 entheses. A swollen joint count (assessing 64 joints, excluding the maxillary joints) was performed, and CRP concentrations were measured.

The primary end point was 40% improvement in disease activity according to the Assessment of SpondyloArthritis international Society criteria (ASAS40) (31, 32). Secondary outcomes were 50% improvement in disease activity according to the BASDAI (BASDAI 50), 20% improvement according to the ASAS criteria (ASAS20), partial remission according to the ASAS criteria, and the mean improvement in disease activity measures, physical function, and health-related quality of life.

Nonresponders (those who failed to achieve an ASAS40 response) at the end of the double-blind trial (week 12) and after open-label therapy for at least 12 weeks were eligible for a dosage escalation of adalimumab to 40 mg/week.

Statistical analysis.

According to previous studies (4), we expected that at least 50% of patients treated with adalimumab and <10% of patients treated with placebo would achieve an ASAS40 response. Under this assumption, we calculated that a sample size of 23 patients per group would be needed in order to attain 80% power to detect a significant difference in the response rates by Fisher's exact test. The statistical analysis was based on the intent-to-treat population.

Fisher's exact test was applied to compare response rates and categorical variables. The Clopper and Pearson method was used to calculate 95% confidence intervals (95% CIs) of the response rates. Means were compared between groups by analysis of covariance, with the baseline value as the covariate. Because extreme values could bias the mean number of swollen joints, we calculated the 5% trimmed means (the highest and lowest 5% of the values, which were excluded from analysis), the Hodges-Lehmann estimate of the median improvement scores, and nonparametric tests based on all data (Mann-Whitney U test, Wilcoxon's test for paired comparisons). Because all patients were switched to adalimumab after 12 weeks, combined data from both groups were analyzed at 52 weeks of treatment (52 weeks of adalimumab treatment in patients randomized to receive adalimumab and 40 weeks of adalimumab treatment in those randomized to receive placebo).

For categorical response criteria (i.e., the ASAS40 and the BASDAI 50), patients who withdrew from the study were counted as nonresponders. For continuous variables (i.e., analysis of mean values for the BASDAI), the last observation carried forward method was used to impute missing values for patients who withdrew from the study as well as to include all patients in these analyses. Differences in the response rates at week 12 and week 52 were compared by McNemar's test, and mean changes between time points were compared by Student's paired t-test. Logistic regression was used to investigate the predictive value of different factors for the BASDAI 50 or the ASAS40 responses on a multivariate level. All statistical tests were 2-sided. P values less than 0.05 were considered significant.

RESULTS

Baseline characteristics.

Of the 46 patients who were enrolled in the study, 22 were randomized to receive adalimumab 40 mg every other week and 24 were randomized to receive placebo (Table 1). The mean age of the adalimumab group at baseline was 38 years and the mean disease duration was 7 years; 59% of them were women, and 59% were HLA–B27 positive. The mean age of the placebo group at baseline was 37 years and the mean disease duration was 8 years; 50% of them were women, and 75% were HLA–B27 positive. All 46 patients completed the 12-week placebo-controlled trial, and 38 patients (83%) completed the extension to week 52. Eight patients (3 in the adalimumab group and 5 in the placebo group) were positive for latent tuberculosis and were treated with isoniazid 300 mg/day for 9 months according to local guidelines. Only 1 patient (randomized to receive adalimumab) had previously been treated short term with infliximab as part of another study.

Primary and secondary end points.

After 12 weeks of therapy, 12.5% (95% CI 2.7, 32.4) of the placebo group had achieved an ASAS40 response, 25% had achieved an ASAS20 response, and none of them had achieved partial remission according to the ASAS criteria (Figure 1). Of the patients who received adalimumab, 54.5% (95% CI 32.2, 75.6) achieved an ASAS40 response (P = 0.004 versus the placebo group), 68.2% (95% CI 45.1, 86.1) achieved an ASAS20 response (P = 0.007), and 22.7% (95% CI 7.8, 45.4) achieved an ASAS partial remission response (P = 0.019).

Figure 1.

Responses on the Assessment of SpondyloArthritis international Society criteria for 20% improvement in disease activity (ASAS20), the ASAS40, the ASAS criteria for partial remission (ASAS PR), and the Bath Ankylosing Spondylitis Disease Activity Index criteria for 50% improvement (BASDAI 50) at week 12 and up to week 52 of treatment with adalimumab (40 mg every other week) or placebo in patients with active axial spondylarthritis without radiographically defined sacroiliitis. At week 12, patients taking placebo were switched to open-label adalimumab treatment.

The percentages of patients initially randomized to receive placebo who achieved ASAS20, ASAS40, and ASAS partial remission responses at week 52 (i.e., after 40 weeks of adalimumab therapy) increased significantly to 66.7% (P = 0.002), 54.2% (P = 0.004), and 37.5% (P = 0.002), respectively. Clinical improvements achieved in the adalimumab-treated patients were sustained through week 52: 54.5% achieved an ASAS20 response, 45.5% achieved an ASAS40 response, and 18.2% achieved an ASAS partial remission response (Figure 1). The BASDAI 50 response rates are shown in Figure 1. Among the entire group of 46 study patients, 23 (50%) achieved an ASAS40 response at week 52.

In the placebo group, the mean BASDAI did not change significantly between baseline and week 12 (6.2 versus 5.0) (Figure 2A). After these patients were switched to adalimumab therapy, the mean BASDAI decreased to 3.3 at week 52. In the adalimumab group, the mean BASDAI decreased from 6.5 at baseline to 3.8 at week 12 and remained low up to week 52 (score of 3.7). At week 12, there was a statistically significant difference in the mean BASDAI (P = 0.036) and the mean BASFI (P = 0.012) between the placebo-treated patients and the adalimumab-treated patients (Figures 2A and B). The difference between the 2 groups became significant at week 4 for the BASDAI and at week 8 for the BASFI.

Figure 2.

Changes in the mean scores on A, the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and B, the Bath Ankylosing Spondylitis Functional Index (BASFI) during treatment with adalimumab (40 mg every other week) or placebo in patients with active axial spondylarthritis without radiographically defined sacroiliitis. At week 12, patients taking placebo were switched to open-label adalimumab treatment. Values are the mean. P values are for the differences between treatment groups at week 12.

Improvements in the patient's and physician's global assessments, general pain, and nocturnal pain were also statistically significantly greater in the adalimumab group compared with the placebo group at week 12 (Table 2). At week 52, 52% of all study patients from both treatment groups had achieved a BASDAI 50 response. The mean CRP concentration in the placebo group changed little, from 7.8 mg/liter at baseline to 6.0 mg/liter at week 12, but the adalimumab group showed a significant decrease over the same period, from 6.2 mg/liter at baseline to 3.0 mg/liter at week 12 (P = 0.049 versus placebo). After switching to adalimumab, the mean CRP concentration in patients originally randomized to receive placebo decreased to 4.7 mg/liter at week 52, whereas in the patients who received continuous treatment with adalimumab, the mean CRP concentration remained low at 2.6 mg/liter.

Table 2. Summary of secondary outcome measures in the 2 treatment groups at week 12 and in all patients at week 52*
 Baseline, mean ± SDWeek 12, mean ± SDAll patients
Adalimumab groupPlacebo groupAdalimumab groupPlacebo groupPBaseline, mean ± SDWeek 52, mean ± SDImprovement, mean (95% CI)
  • *

    P values for the comparison of adalimumab versus placebo at week 12 were determined by analysis of covariance. 95% CI = 95% confidence interval; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; NRS = numerical rating scale; BASFI = Bath Ankylosing Spondylitis Functional Index; BASMI = Bath Ankylosing Spondylitis Metrology Index; CRP = C-reactive protein; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; SF-36 = Short Form 36 health survey; MCS = mental component summary; PCS = physical component summary; ASQoL = Ankylosing Spondylitis Quality of Life questionnaire; NA = not assessed; EQ-5D = EuroQol Index.

  • The highest and lowest 5% of the values were not included in the analysis.

  • Median of the individual improvement scores.

BASDAI (0–10 NRS)6.5 ± 1.26.2 ± 1.33.8 ± 2.55.0 ± 2.40.0366.3 ± 1.23.5 ± 2.72.8 (2.1, 3.6)
Morning stiffness (BASDAI items 5 and 6; 0–10 NRS)5.8 ± 1.56.2 ± 2.03.0 ± 2.45.2 ± 2.60.0056.0 ± 1.83.1 ± 2.72.9 (2.1, 3.7)
BASFI (range 0–10)5.4 ± 2.04.9 ± 1.63.0 ± 2.44.1 ± 2.60.0125.2 ± 1.83.1 ± 2.72.0 (1.4, 2.6)
BASMI (range 0–10)1.3 ± 1.21.3 ± 1.61.3 ± 1.41.7 ± 1.50.1791.3 ± 1.41.7 ± 1.5−0.4 (−0.7, −0.04)
CRP, mg/liter6.2 ± 5.87.8 ± 7.03.0 ± 3.86.0 ± 5.30.0497.0 ± 6.43.7 ± 7.53.4 (0.6, 6.2)
Patient's global assessment (0–10 NRS)6.6 ± 1.76.7 ± 1.53.9 ± 2.45.5 ± 2.30.0256.6 ± 1.63.9 ± 2.72.8 (2.0, 3.6)
Physician's global assessment (0–10 NRS)6.0 ± 1.25.8 ± 1.13.7 ± 2.65.0 ± 1.90.0265.9 ± 1.13.0 ± 2.42.8 (2.1, 3.4)
General pain (0–10 NRS)7.2 ± 1.17.1 ± 1.64.4 ± 2.85.8 ± 2.50.0507.2 ± 1.44.1 ± 3.03.1 (2.3, 3.9)
Nocturnal pain (0–10 NRS)7.0 ± 1.66.4 ± 2.54.0 ± 3.05.3 ± 2.50.0196.7 ± 2.13.7 ± 3.33.0 (2.1, 3.9)
No. of swollen joints (range 0–64)0.6 ± 0.91.7 ± 7.50.3 ± 0.70.2 ± 0.70.201.2 ± 6.70.1 ± 0.42.5 (1.0, 19.5)
MASES (range 0–13)3.0 ± 2.62.5 ± 3.32.5 ± 3.52.8 ± 3.40.582.7 ± 2.91.8 ± 2.60.9 (−0.02, 1.9)
SF-36 MCS score (range 0–100)41.3 ± 12.543.6 ± 11.144.6 ± 12.743.9 ± 11.80.37642.5 ± 11.747.4 ± 11.64.9 (1.6, 8.1)
SF-36 PCS score (range 0–100)28.8 ± 7.630.7 ± 6.038.8 ± 11.834.9 ± 9.60.07329.8 ± 6.840.1 ± 12.510.3 (6.9, 13.8)
ASQoL score (range 0–18)10.8 ± 3.79.5 ± 3.0NANA10.1 ± 3.44.9 ± 5.25.3 (3.8, 6.7)
EQ-5D (range 0–1)0.58 ± 0.30.58 ± 0.30.78 ± 0.20.72 ± 0.20.3210.6 ± 0.30.80 ± 0.220.22 (0.13, 0.31)

At week 12, there were no statistically significant differences between the adalimumab and the placebo groups in values on the BASMI, the mental and physical component summary scores of the SF-36, and the EQ-5D (Table 2). At week 52, the SF-36 mental and physical component summary scores, EQ-5D, and ASQoL scores were significantly improved in all patients as compared with baseline (Table 2).

Effect of treatment on peripheral arthritis, enthesitis, and other manifestations.

At baseline, 30% of the 46 study patients (29% in the placebo group and 32% in the adalimumab group) also had peripheral arthritis and 67% (54% in the placebo group and 82% in the adalimumab group) had enthesitis as measured by the MASES. The 5% trimmed mean number of swollen joints in the adalimumab group decreased from 0.6 at baseline to 0.3 at week 12. However, the mean number of swollen joints also decreased in the placebo group, from 1.7 at baseline to 0.2 at week 12 (P = 0.20) (Table 2). No significant change in the number of entheses between baseline and week 12 was observed for either treatment group (Table 2). In the entire group of 46 study patients, there was a significant decrease (P = 0.01) in the number of swollen joints, from 1.2 at baseline to 0.12 at week 52 (median improvement score 2.5 [95% CI 1.0, 19.5]), and a nonsignificant decrease in the mean MASES, from 2.7 at baseline to 1.8 at week 52 (mean improvement score 0.9 [95% CI −0.02, 1.9]).

Analysis of treatment responders and nonresponders.

Patients who were major responders (ASAS40 response) to adalimumab experienced a substantial reduction in their disease activity as well as improvement in their physical function. In the responders, the mean BASDAI decreased from 6.1 at baseline to 1.2 at week 52, and the mean BASFI decreased from 4.6 at baseline to 1.1 at week 52 (Figure 3). In the nonresponders, the mean BASDAI and the mean BASFI showed minimal changes, with scores of 5.7 and 5.1, respectively, at week 52.

Figure 3.

Changes in the mean scores on the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and the Bath Ankylosing Spondylitis Functional Index (BASFI) between baseline and week 52 in the subgroups of responders (n = 23) and nonresponders (n = 23) to adalimumab therapy, according to the Assessment of SpondyloArthritis international Society criteria for 40% improvement in disease activity (ASAS40). Numbers at the top of the bars are the percentages represented by the bars.

Baseline predictors of a major clinical response.

When the entire group of study patients from both treatment groups were analyzed together at week 52, patients with a CRP concentration of >6 mg/liter, a disease duration of ≤3 years, or younger age (≤30 years) at study entry had a significantly greater probability of achieving an ASAS40 response or a BASDAI 50 response (i.e., a major clinical response) (Table 3). Of the patients with an elevated CRP concentration at baseline (>6 mg/liter), 82.4% achieved a BASDAI 50 response (P = 0.006 versus those without an elevated CRP concentration at baseline) and 76.5% achieved an ASAS40 response (P = 0.013 versus those without an elevated CRP concentration at baseline). Of the patients with a disease duration of ≤3 years at baseline, 80% and 73.3% achieved BASDAI 50 and ASAS40 responses, respectively; by comparison, of the patients with a disease duration of >10 years, 14.3% and 0.0% achieved BASDAI 50 (P = 0.014) and ASAS40 (P = 0.006) responses, respectively (Table 3). On multivariate logistic regression analysis, younger age and an elevated CRP concentration remained significant predictors of a BASDAI 50 or an ASAS40 response, whereas a short disease duration did not achieve statistical significance.

Table 3. Summary of subgroup analyses for BASDAI 50 and ASAS40 responders at week 52*
Baseline characteristicNo. of patientsBASDAI 50 responders at week 52ASAS40 responders at week 52
No. (%) of patientsPNo. (%) of patientsP
  • *

    Only 45 patients were analyzed; 1 patient was excluded from the study (before treatment with adalimumab for at least 8 weeks after switching from placebo). BASDAI 50 = Bath Ankylosing Spondylitis Disease Activity Index criteria for 50% improvement; ASAS40 = Assessment of SpondyloArthritis international Society criteria for 40% improvement in disease activity; CRP = C-reactive protein.

  • Only 44 patients were counted: those who had a magnetic resonance imaging (MRI) scan at baseline and were treated with adalimumab for at least 8 weeks after switching from placebo.

HLA–B27     
 Positive3120 (64.5)0.1117 (54.8)0.53
 Negative145 (35.7) 6 (42.9) 
MRI     
 Positive2917 (58.6)0.7616 (55.2)0.75
 Negative158 (53.3) 7 (46.7) 
HLA–B27 and MRI     
 B27 positive, MRI negative158 (53.3) 7 (46.7) 
 B27 negative, MRI positive135 (38.5)0.136 (46.2)0.59
 B27 positive and MRI positive1612 (75.0) 10 (62.5) 
Sex     
 Male2013 (65.0)0.3711 (55.0)0.77
 Female2512 (48.0) 12 (48.0) 
CRP     
 >6 mg/liter1714 (82.4)0.00613 (76.5)0.013
 ≤6 mg/liter2811 (39.3) 10 (35.7) 
Disease duration     
 ≤3 years1512 (80.0) 11 (73.3) 
 4–10 years2312 (52.2)0.01412 (52.2)0.006
 >10 years71 (14.3) 0 (0) 
Age at study entry     
 ≤30 years1513 (86.7) 12 (80.0) 
 31–40 years1810 (55.6)0.0019 (50.0)0.005
 >40 years122 (16.7) 2 (16.7) 

No statistically significant differences between responders and nonresponders were noted among patients who were HLA–B27 positive (P = 0.11 for the BASDAI 50 and P = 0.53 for the ASAS40) or who had active inflammatory changes of the spine or sacroiliac joints (P = 0.76 for the BASDAI 50 and P = 0.75 for the ASAS40) (Table 3). Patients who were HLA–B27 positive and had active inflammatory changes shown by MRI were more frequently BASDAI 50 (75.0%) or ASAS40 (62.5%) responders, although these differences did not reach statistical significance. None of the patients enrolled in the study were HLA–B27 negative and MRI negative, which suggests that the diagnostic criteria for axial SpA without radiographically defined sacroiliitis were adhered to during this study.

Weak, nonsignificant associations were also found for the presence of peripheral arthritis at baseline (n = 14 patients) and a BASDAI 50 (71.4%) or an ASAS40 (64.3%) response. In contrast, only 45.2% of patients without peripheral arthritis at baseline (n = 32) were BASDAI 50 or ASAS40 responders.

Safety.

During the 12-week placebo-controlled phase of the study, 123 adverse events occurring in 40 patients were reported; 63 of the adverse events occurred in 19 patients in the adalimumab group, and 60 occurred in 21 patients in the placebo group. Infections of the respiratory tract were the most frequently reported adverse events (15 adalimumab-treated patients and 9 placebo-treated patients). Skin infections were reported more often in the adalimumab group (12 patients) than in the placebo group (4 patients). No serious adverse events occurred during the 12-week placebo-controlled phase.

During the open-label extension, 198 adverse events occurring in 45 patients, including 8 serious adverse events in 5 patients, were reported. None of the serious adverse events were considered to be related to the study drug. The serious adverse events were as follows: 1 patient underwent surgery to repair an ulcer of the nasal septum; 1 patient underwent laryngotracheoscopy because of Reinke's edema, had a remitting cholesteatoma of the ear, and underwent surgery for a cholesteatoma; 1 patient underwent hysterectomy because of a myoma of the uterus and an ovarian cyst; 1 patient underwent surgery for a fallopian tube adhesion; and 1 patient was hospitalized because of treatment failure with worsening of symptoms.

Response to dosage escalation of adalimumab therapy.

The adalimumab dosage was switched from 40 mg every other week to 40 mg/week in 10 patients who did not achieve an ASAS40 response at the end of the double-blind trial (week 12) and after open-label therapy for at least 12 weeks. This change in the dosing regimen occurred between weeks 28 and 36. However, this increased dosage had no major impact on the safety or efficacy responses. In patients who switched to weekly adalimumab, the mean BASDAI score did not change significantly (6.6 at baseline and 6.3 at week 52), and no clinically important change was observed in the mean BASFI (5.4 at baseline and 5.6 at week 52). Therefore, these patients were also included in the analysis at week 52. Only 1 of these 10 patients achieved an ASAS20 response as well as an ASAS40 response after switching to weekly adalimumab injections.

DISCUSSION

Our study has shown that patients with axial SpA without definite radiographic changes in the sacroiliac joints respond at least as well to TNF antagonist therapy as patients with established AS evaluated in previous clinical trials (3, 33). In these NSAID-refractory patients, an ASAS40 response was achieved in 54.5% of the adalimumab-treated patients, compared with only 12.5% of the placebo-treated patients, after 12 weeks of therapy. Statistically significant and clinically important responses on the BASDAI 50, ASAS20, and ASAS partial remission criteria were also obtained. The spinal assessments of morning stiffness and back pain also improved significantly. However, no clear difference in peripheral arthritis between the 2 treatment groups was seen.

Importantly, the response to adalimumab was sustained during long-term treatment, with an ASAS40 response rate of 50% in patients treated for up to 52 weeks. Patients with axial SpA without radiographically defined sacroiliitis also experienced statistically significant improvements in health-related quality of life (measured by the SF-36, EQ-5D, and ASQoL instruments) after up to 1 year of adalimumab treatment.

Both young age (≤30 years) at study entry and an elevated CRP concentration (>6 mg/liter) independently predicted a major clinical response (i.e., ASAS40 or BASDAI 50) by multivariate logistic regression analysis. Shorter disease duration was also predictive in a univariate analysis, although it overlapped with the patient's age, a finding that was not unexpected. For example, 80% of the patients ≤30 years of age and 73% of the patients with symptom duration of ≤3 years showed a major treatment response after up to 1 year of therapy. Neither HLA–B27 positivity (54.8% in responders versus 42.9% in nonresponders) nor active inflammation as demonstrated by MRI (55.2% in responders versus 46.7% in nonresponders) at baseline differentiated between responders and nonresponders, as defined by an ASAS40 response (Table 3). Accurate prediction of treatment response was slightly better for patients who were both HLA–B27 positive and had active inflammation demonstrated by MRI (62.5%). This might also explain the numerically greater, although statistically nonsignificant, percentage of patients achieving ASAS partial remission at week 52 after switching from placebo to adalimumab as compared with those who received continuous adalimumab (Figure 1). This finding was probably a result of the larger number of patients in the placebo group who were HLA–B27 positive and had active inflammation (Table 1). Overall, however, these characteristics had a rather small predictive value, and this finding is similar to the clinical trial results in patients with established AS reported elsewhere (15, 34).

Approximately one-half of the study patients had only a minor clinical response or even no response to the treatment regimen. While we used strict criteria for the diagnosis of active axial SpA in patients who were included in this study, the criteria should be regarded as preliminary and should be confirmed or perhaps modified in the near future. All of our study patients had inflammatory back pain, although this was not obligatory for study inclusion, and either had active inflammation demonstrated on MRI or were positive for HLA–B27. We do not believe that diagnostic inaccuracy could account for the approximately one-half of patients who did not respond to adalimumab treatment. Overall, approximately one-half of patients with axial SpA, including both those with axial SpA in our study and those with established AS in previously published studies (3, 33), can be expected to have a good response (defined as a BASDAI 50 or an ASAS40 response) to adalimumab.

The mean disease duration in our study patients was relatively high (7.5 years), considering that none of these patients with axial SpA had chronic radiographic changes in their sacroiliac joints, as defined by the modified New York criteria (5). However, it has been shown that the interval between the pre-radiographic stage of axial SpA and the appearance of typical radiographic changes may be up to 10–20 years (6, 12, 13). We found that patients with younger age and shorter disease duration benefit the most from TNF antagonist treatment. Since the mean time from the first symptoms of axial SpA to the diagnosis is typically 5–10 years (1, 2), the shortening of this interval is still a major task for the rheumatology community.

A higher percentage of female patients in this study than in AS trials is consistent with the percentages reported in studies of other cohorts with undifferentiated SpA (10, 35) and might indicate that radiographic changes of chronic disease develop later or less frequently in women than in men. Also, a lower percentage of HLA–B27 positivity has been described in other early SpA cohorts (10, 35).

The results of this randomized controlled trial indicate that patients with axial SpA without radiographically defined sacroiliitis can be reliably diagnosed early in their disease course, before chronic (radiologic) changes are seen, that approximately one-half of patients with active axial SpA can be expected to improve when treated with adalimumab, and that younger patients, patients with a shorter disease duration, and patients with elevated CRP concentrations may show an even better response. At this time, the best way to identify patients who are candidates for long-term therapy with adalimumab is to identify those who have a good response during the first 12 weeks of therapy, which is consistent with the recommendations of the ASAS International Working Group concerning the use of TNF antagonists in the treatment of AS (36). Based on the data presented here, increasing the frequency of adalimumab injections from once every other week to once a week cannot be recommended for this group of patients. Considering the relatively high costs of TNF antagonist treatment, any therapy intended for use in a group of patients with early disease should target those who have a major response, with the goal of converting high-grade disease activity to low-grade disease activity or even remission.

Based on the findings of this study, regulatory agency approval of TNF antagonists for the treatment of active AS should be extended to include as a new indication patients with active axial SpA who have not yet developed radiographic changes of the sacroiliac joints. Although there is currently no evidence that TNF inhibition can alter radiographic damage in established AS, it remains to be demonstrated that structural changes might be prevented if patients are treated before radiographic damage has occurred (37).

AUTHOR CONTRIBUTIONS

Dr. Sieper 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. Haibel, Rudwaleit, Wong, Braun, Sieper.

Acquisition of data. Haibel, Heldmann, Braun, Sieper.

Analysis and interpretation of data. Haibel, Rudwaleit, Listing, Kupper, Braun, Sieper.

Manuscript preparation. Haibel, Wong, Kupper, Braun, Sieper, Dana L. Randall (nonauthor; JK Associates, Conshohocken, PA, a medical communications company), Michael A. Nissen (nonauthor; Abbott Laboratories).

Statistical analysis. Listing.

ROLE OF THE STUDY SPONSOR

Abbott GmbH & Company KG, an affiliate of Abbott Laboratories, collaborated with Drs. Sieper, Haibel, Rudwaleit, Listing, Heldmann, and Braun to design and conduct the study. All authors analyzed the data, critically reviewed the manuscript for important intellectual content, approved the final draft, and agreed to its submission. Abbott provided editorial assistance (performed by Dana L. Randall, MS, PharmD, JK Associates, Conshohocken, PA, a medical communications company, and by Michael A. Nissen, ELS, Abbott Laboratories). Abbott did not control the analysis or interpretation of the study results.

Ancillary