Drs. Han, Yan, and Williamson own stock in Johnson & Johnson, of which Centocor, Inc., is a subsidiary.
Major reduction in spinal inflammation in patients with ankylosing spondylitis after treatment with infliximab: Results of a multicenter, randomized, double-blind, placebo-controlled magnetic resonance imaging study
Article first published online: 27 APR 2006
Copyright © 2006 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 54, Issue 5, pages 1646–1652, May 2006
How to Cite
Braun, J., Landewé, R., Hermann, K.-G. A., Han, J., Yan, S., Williamson, P. and van der Heijde, D. (2006), Major reduction in spinal inflammation in patients with ankylosing spondylitis after treatment with infliximab: Results of a multicenter, randomized, double-blind, placebo-controlled magnetic resonance imaging study. Arthritis & Rheumatism, 54: 1646–1652. doi: 10.1002/art.21790
- Issue published online: 27 APR 2006
- Article first published online: 27 APR 2006
- Manuscript Accepted: 25 JAN 2006
- Manuscript Received: 27 SEP 2005
- Centocor, Inc., Malvern, PA
- Schering-Plough Research Institute, Kenilworth, NJ
To determine whether the effects of anti–tumor necrosis factor α (TNFα) in reducing the signs and symptoms of ankylosing spondylitis (AS) coincide with a reduction in spinal inflammation as detected by magnetic resonance imaging (MRI).
Pre- and postgadolinium T1 and STIR MR images of the spine were acquired at baseline and at week 24 in patients with AS who participated in a multicenter, randomized, double-blind, placebo-controlled study. Patients were randomly assigned at an 8:3 ratio to receive infusions of infliximab (5 mg/kg) or placebo at weeks 0, 2, and 6 and then every 6 weeks thereafter. MR images were obtained and evaluated independently by 2 readers who were blinded to the treatment allocation and time sequence of the images.
A total of 194 patients in the infliximab group and 72 patients in the placebo group had evaluable images at baseline and week 24. About 80% of the patients had at least 1 active spinal lesion at baseline, as assessed by MRI. The improvement in the MRI Activity Score after 6 months was significantly greater in the patients who received infliximab (mean 5.02, median 2.72) than in those who received placebo (mean 0.60, median 0.0) (P < 0.001). Almost complete resolution of spinal inflammation was seen in most patients who received infliximab, irrespective of baseline activity.
Patients with AS who received infliximab therapy showed a decrease in spinal inflammation as detected by MRI, whereas those who received placebo showed persistent inflammatory spondylitis.
Considered to be the prototype of the spondylarthritides, ankylosing spondylitis (AS) is a frequently occurring, chronic rheumatic disease that is characterized by inflammatory back pain due to sacroiliitis, spondylitis, and enthesitis (1, 2). As the disease progresses, spinal inflammation may lead to bone erosions, syndesmophytes, and complete ankylosis of the spine. As a result, patients with AS have increased rates of work disability, unemployment (3), and mortality (4). Traditional therapeutic options for patients with AS, such as nonsteroidal antiinflammatory drugs (NSAIDs), have only limited efficacy on the signs and symptoms of the disease (5). Tumor necrosis factor α (TNFα) has been shown to play an important role in the pathogenesis of AS (6), and a new class of drugs targeting TNFα has recently emerged as an alternative for patients with persistently high levels of disease activity despite traditional therapies (7, 8).
However, the disease-specific instruments that are mainly used to evaluate the efficacy of therapeutic interventions in AS are primarily based on subjective measures, such as pain, duration and extent of morning stiffness, fatigue, and the patient's ability to cope with daily activities, and are quantified as disease activity (9) and physical function (10). Furthermore, acute-phase reactants, such as the erythrocyte sedimentation rate and C-reactive protein (CRP), are often not elevated in patients with active AS (11, 12). Thus, other methods are needed to objectively assess disease activity and response to therapy.
Magnetic resonance imaging (MRI) has emerged in recent years as a potential assessment tool because of its ability to detect inflammation in the sacroiliac joints, the spine, and other joints affected by AS (13, 14). Since the axial skeleton is the anatomic area most commonly involved in AS, the spine is the most relevant target for evaluation by MRI. Only specialized MRI techniques, such as STIR and T1 images obtained after injection of a contrast agent, are capable of detecting inflammation with a high level of specificity. Increased signal on STIR images or signal enhancement after gadolinium diethylenetriaminepentaacetic acid (Gd-DTPA) on T1 images may reflect bone marrow edema and hypervascularization, respectively, that is otherwise not detectable with conventional radiography (15, 16). MRI has been used in previous studies of anti-TNF agents in patients with AS (17–19); however, those studies were either observational without a control group or included only small numbers of patients.
In a recent large, multicenter, randomized, double-blind, placebo-controlled study (the Ankylosing Spondylitis Study for the Evaluation of Recombinant Infliximab Therapy [ASSERT]), patients with moderate-to-severe AS received either placebo (n = 78) or 5 mg/kg of infliximab (n = 201) every 6 weeks. The clinical effects on the signs and symptoms of AS identified in that trial have been previously reported (20).
To provide objective evidence that therapy with the anti-TNF monoclonal antibody infliximab had an effect on spinal inflammation, we performed MRI examinations at baseline and after 6 months of treatment in patients in the ASSERT trial. This is, to the best of our knowledge, the largest MRI study of patients with AS ever performed. Using a scoring system that had previously been used successfully (19, 21–23) and a new method of presenting the data, we identified a profound and consistent decrease in spinal inflammation attributable to the anti-TNF therapy.
PATIENTS AND METHODS
Role of the funding source.
The ASSERT study was managed by a Steering Committee, which consisted of Dr. Braun (Rheumazentrum Ruhrgebiet), Dr. van der Heijde (University Hospital Maastricht), and Dr. Williamson (Centocor). The Steering Committee and members of the Centocor clinical trial team designed the study, with input from several of the study investigators. Clinical data were collected by the investigators, and MRIs were interpreted by Dr. Landewé (University Hospital Maastricht) and Dr. Hermann (Charité Medical School). Data were analyzed by Drs. Han and Yan (Centocor). The manuscript was written primarily by Dr. Braun, with the assistance of a medical writer from Centocor. All authors reviewed and contributed to the manuscript during its development, agreed to submit the manuscript, and approved the content of the submitted manuscript.
The details of the ASSERT study design have been reported elsewhere (20). Briefly, patients with active AS for at least 3 months, a Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) score of ≥4 (range 0–10), and a spinal pain assessment score of ≥4 on a visual analog scale (VAS; range 0–10 cm) were eligible for the study. Patients were randomly assigned at an 8:3 ratio to receive infusions of 5 mg/kg of infliximab or placebo at weeks 0, 2, 6, 12, and 18. CRP levels above or below 3 times the upper limit of normal were used for stratification.
Magnetic resonance imaging.
MRIs were obtained at baseline and week 24 using 1.0T or 1.5T scanners and phased-array coils. Sagittal images of the upper (C2 through T10) and lower (T8 through S1) spine were taken, and markers were placed on the patient's skin to guide the identification of vertebrae and the orientation of the images. The field of view was 34–38 cm. The following sequences were obtained: T1-weighted turbo spin-echo, with a slice thickness of 3 mm (repetition time [TR] 500–700 msec and time to echo [TE] minimum accessible, depending on the capability of the machine); the same sequence with fat saturation after infusion of the contrast agent Gd-DTPA (0.1 mmoles/kg of body weight); and STIR (a sequence with intrinsic fat saturation), with a slice thickness of 3 mm (TR 2,000–4,000 msec and TE 35–55 msec).
The T1 image for the upper part of the spine was acquired immediately (<1 minute) after injection of the contrast agent. Since acquisition of that image takes ∼5 minutes, the T1 image for the lower part of the spine was acquired 5–6 minutes later. Thus, theoretically, the sequence for the lower part of the spine may have had a lower sensitivity for the detection of spinal inflammation.
Scoring of MR images.
Two qualified, trained, and independent readers (RL and K-GAH) who were blinded to the treatment information, patients' identities, and the chronology of the images evaluated each sequence. Most images were scored once. However, each reader scored a random sample of the images (10%) a second time at least 3 months after the first evaluation to allow for calculation of intrareader scoring variability.
Images were scored according to a previously described method (19, 21–23) in which activity was assessed at the level of the vertebral unit. A vertebral unit was defined as the area between 2 virtual horizontal lines through the middle of 2 adjacent vertebrae. Using gadolinium-enhanced T1 and STIR sequences, each vertebral unit was given an MRI Activity Score based on the amount of bone marrow edema or erosions, as follows: 0 = no erosions or bone marrow edema, 1 = minor bone marrow edema involving ≤25% of the vertebral unit, 2 = moderate bone marrow edema involving >20% but ≤50% of the vertebral unit, 3 = major bone marrow edema involving >50% of the vertebral unit, 4 = bone marrow edema and minor erosion involving ≤25% of the vertebral unit, 5 = bone marrow edema and moderate erosion involving >20% but ≤50% of the vertebral unit, and 6 = bone marrow edema and major erosion involving >50% of the vertebral unit. Thus, the MRI Activity Score for each vertebral unit ranged from 0 to 6. With 23 vertebral units assessed (from C2 to S1), the total MRI Activity Score for the spine ranged from 0 to 138. The combined information provided by gadolinium-enhanced T1 and STIR sequences was used for scoring. The mean of both readers' scores was used in the analysis.
Only patients with evaluable MRIs for both the baseline and week 24 assessments were included in the analysis. Intraclass correlation coefficients (ICCs) were calculated to evaluate the level of agreement between the scores assigned by the 2 readers. The intrareader reliability coefficient was used to evaluate agreement between the 2 readings for the 10% of patients whose images were scored twice by the same reader. Guyatt's effect size for the change in MRI Activity Scores from baseline to week 24 was calculated by dividing the mean change in the treatment group by the standard deviation of the change in the placebo group. Treatment group differences in MRI Activity Scores were evaluated using an analysis of variance on the van der Waerden normal scores. To show the data of all patients in a coherent manner, probability plots of MRI change scores were produced, as previously described (24). Double probability plots were created to illustrate the relationship between change in inflammation and baseline inflammation in every patient individually.
Baseline characteristics of the AS patients.
The baseline characteristics of the patients with evaluable MR images are provided in Table 1. The study population was typical of patients with moderate-to-severe AS. Most patients were men and were HLA–B27 positive. Despite randomization, patients in the placebo group had a greater median disease duration than patients in the infliximab group. About one-half of the patients in each treatment group had high CRP levels at baseline (at least 3 times the upper limit of normal).
|Assessment||Placebo (n = 72)||Infliximab (n = 194)|
|Men, no. (%)||63 (87.5)||153 (78.9)|
|Age, median (IQR) years||42 (35, 47)||40 (32, 47)|
|Disease duration, median (IQR) years||13.4 (4.9, 18.4)||7.7 (3.3, 14.9)|
|HLA–B27 positive, no. (%)||65 (90.3)||168 (87.0)†|
|History of uveitis, no. (%)||25 (34.7)||69 (35.6)|
|History of psoriasis, no. (%)||5 (6.9)||15 (7.7)|
|History of IBD, no. (%)||5 (6.9)||13 (6.7)|
|BASDAI score, median (IQR)||6.5 (5.2, 7.0)||6.6 (5.3, 7.6)|
|BASFI score, median (IQR)||6.3 (4.0, 7.5)||5.7 (4.5, 7.1)|
|BASMI score, median (IQR)||4.0 (2.5, 6.0)||4.0 (2.0, 5.0)|
|CRP, median (IQR) mg/dl [normal range 0–0.5 mg/dl]||1.8 (0.7, 3.2)||1.5 (0.7, 3.2)|
|CRP ≤3 times the upper limit of normal, no. (%)||34 (47.2)||90 (46.4)|
|CRP >3 times the upper limit of normal, no. (%)||38 (52.8)||104 (53.6)|
Reliability of the MRI scoring system.
The ICC for the change in MRI Activity Score from baseline to week 24, which is an estimate of the interreader variation in the scoring of a particular patient's images, was 0.74. The intrareader reliability coefficient for the change in MRI Activity Score from baseline to week 24 was 0.74.
At baseline, similar proportions of patients in the infliximab and the placebo groups had evidence of spinal inflammation, with 149 (77%) and 57 (79%) patients in the infliximab and placebo groups, respectively, having an MRI Activity Score >0. The median total MRI Activity Score at baseline was also similar in the 2 study groups (Table 2).
|Placebo (n = 72)||Infliximab (n = 194)|
|Baseline MRI Activity Scores|
|Mean ± SD||7.1 ± 8.2||6.6 ± 6.9|
|Median (IQR)||5.0 (1.0, 11.0)||4.1 (1.0, 11.0)|
|No. (%) of patients with MRI Activity Scores >0||57 (79)||149 (77)|
|Change in MRI Activity Scores from baseline to week 24|
|Mean ± SD||−0.6 ± 3.4||−5.0 ± 6.2|
|Median (IQR)||0.0 (−3.0, 0.6)||−2.7 (−9.0, 0.0)|
|Range||−8.0, 9.0||−25.5, 5.5|
|P versus placebo||<0.001|
The probability plot displayed in Figure 1A shows that the curve representing infliximab treatment lies to the right of and below the curve representing placebo, indicating that infliximab was associated with greater improvement in MRI Activity Scores. Compared with the placebo group, more patients in the infliximab group had negative changes in MRI Activity Scores at week 24 versus baseline, and fewer patients had positive changes versus baseline. Moreover, patients in the infliximab group had a significantly greater improvement in total MRI Activity Scores from baseline to week 24 (mean ± SD 5.0 ± 6.2, median 2.7) than did patients in the placebo group (0.6 ± 3.4, median 0.0) (P < 0.001) (Table 2). Guyatt's effect size was 1.5.
Representative examples of a gadolinium-enhanced T1 image obtained at baseline and at week 24 are shown in Figure 2. Some spinal inflammation (MRI Activity Score >1) was still detected at week 24 in 72 patients in the infliximab group (37.1%) as compared with 53 patients in the placebo group (73.6%) (P < 0.001).
Figure 1B shows the relationship between changes in the MRI Activity Scores at week 24 and the MRI Activity Scores at baseline in individual patients in the infliximab group. The probability plots of the baseline score and the change score mirror each other in a virtual line close to zero. The interpretation is that the effect of infliximab on AS activity as determined by MRI is almost complete and is present in almost every patient. Figure 1C shows that, unlike the infliximab group, changes observed in the placebo group were sporadic, occurred in both directions (improving and worsening), were independent of baseline activity, and, for the most part, did not reflect an elimination of inflammation.
The results of this study show that treatment with the anti-TNF monoclonal antibody infliximab leads to an almost complete elimination of spinal inflammation in patients with active AS. This finding is consistent with the clinical data previously reported (20) and provides additional support for the role of TNFα in the pathogenesis of AS as well as for the role of anti-TNF therapy in the treatment of the disease. Our study also confirms the feasibility of using MRI to visualize inflammation in the axial skeleton of patients with AS. The results of the MRI assessments contribute additional objective information that is not obtainable through clinical assessments or laboratory parameters.
This was the largest study of any kind, including studies of anti-TNF therapy, in which MRI was used to evaluate spinal inflammation in patients with AS. Both the application of MRI and the treatment with TNF-blocking drugs have had a great impact on the management of AS. The potential of MRI to visualize spinal inflammation in AS was first reported in 1994 in a study of the sacroiliac joints (13) and in 1998 in a study of the spine (14), while the first pilot study reporting the efficacy of infliximab in AS was published in 2000 (25). The scoring system we used was shown to be reliable in this study as well as in previous studies (19, 21–23).
The treatment effect observed in this study was convincing in all aspects. Conventional statistical analysis demonstrated a highly statistically significant contrast between infliximab and placebo treatment. But the probability plots of individual patient data truly unveiled how profound and complete the effect of infliximab on spinal inflammation actually is. The plots show that almost all inflammation was abolished regardless of the level of inflammation at baseline, leaving only marginal residual inflammation. The pattern was clearly different in the placebo group, in which improvement was sporadic and occurred as frequently as worsening, without any relationship between the baseline MRI Activity Score and the change in the MRI Activity Score from baseline to week 24 (Figure 1C). Changes observed in the placebo group are consistent with measurement error, and the curve for the data from the placebo group suggests that regression toward the mean was not a major factor here.
The improvement in clinical disease activity and spinal inflammation observed in patients who received infliximab in this study was similar to that reported in previous studies (19). However, because the previous MRI studies were comparatively small, this large study formally confirms the efficacy of infliximab in improving the spinal inflammation detected by MRI in patients with AS. Although a similarly large study has not been performed with the TNF receptor fusion protein etanercept, the available data suggest that this agent may also reduce spinal inflammation (26, 27), indicating a class effect for the TNF-blocking agents. The assessment time points of baseline and 6 months were chosen to show a treatment effect. However, there is some evidence from smaller studies that improvement in spinal inflammation may be detected as early as 6 or 12 weeks (26, 27).
Our knowledge about the natural course of the spinal inflammation detected by MRI in patients with AS is still rather limited. The main open question concerns how inflammation is linked to the characteristic osteoproliferative changes (e.g., syndesmophytes and ankylosis) that occur in the majority of patients with AS (21). There is a preliminary report suggesting that there was almost no radiographic progression during infliximab therapy in a small number of patients (28), but more data are needed. The challenges in evaluating radiographic progression in AS are that measurable progression may be slow, comparisons are limited to historical controls (29), and NSAIDs may also have an impact (30).
This study proves that anti-TNF therapy with infliximab decreases spinal inflammation detected by MRI in patients with AS. Further investigation is necessary to examine possible correlations between spinal inflammation, osteodestruction, and osteoproliferation and to examine the effect of anti-TNF therapy on radiographic structural changes in patients with AS.
We thank the investigators, patients, and study personnel who made the study possible. We acknowledge the writing support of Scott Newcomer (Centocor, Inc.), who assisted in the preparation of the manuscript.
- 12Ankylosing spondylitis: an overview. Ann Rheum Dis 2002; 61 Suppl III: iii8–18., , , , .
- 19Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis, before and after successful therapy with infliximab: evaluation of a new scoring system. Arthritis Rheum 2003; 48: 1126–36., , , , , , et al.
- 20Ankylosing Spondylitis Study for the Evaluation of Recombinant Infliximab Therapy Study Group. Efficacy and safety of infliximab in patients with ankylosing spondylitis: results of a randomized, placebo-controlled trial (ASSERT). Arthritis Rheum 2005; 52: 582–91., , , , , , et al, and the
- 26Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis before and after therapy with the tumor necrosis factor α receptor fusion protein etanercept. Arthritis Rheum 2005; 52: 1216–23., , , .
- 29ASsessment in Ankylosing Spondylitis International Working Group. ASsessment in Ankylosing Spondylitis International Working Group/Spondylitis Association of America recommendations for conducting clinical trials in ankylosing spondylitis. Arthritis Rheum 2005: 52; 386–94., , , , , , et al, and the