Abstract
Objective. Serum C-reactive protein (CRP) associates with radiographic progression in patients with ankylosing spondylitis (AS) untreated with tumor necrosis factor (TNF) antagonists. We assessed correlations between serum CRP and radiographic progression/magnetic resonance imaging (MRI)-detected inflammation after 2 years of anti-TNF therapy.
Methods. Patients with active AS receiving golimumab (GOL)/placebo through Week 16 (early escape) or Week 24 (crossover by design), followed by GOL through 4 years, had sera/images obtained through Week 208. Lateral spinal radiographs and spinal MRI were scored with the modified Stoke AS Spine Score (mSASSS) and the AS spine MRI activity (ASspiMRI-a) score, respectively. ANOVA assessed differences based on CRP levels and mSASSS progression. The relationships between CRP levels and mSASSS/ASspiMRI-a were assessed by Spearman correlation and logistic regression.
Results. Of the randomized GO-RAISE patients, 299 (84.0%) had pre- and posttreatment spinal radiographs. Larger proportions of patients with Week 104 CRP ≥ 0.5 mg/dl (n = 47) versus < 0.5 mg/dl (n = 236, 40.4% vs 22.9%, p = 0.0121) had mSASSS changes ≥ 2 at Week 104. Across several visits, serum CRP demonstrated weak associations with mSASSS change (rs ≤ 0.21, p < 0.05, n = 262–293) and moderate associations with ASspiMRI-a change (rs = −0.33 to 0.54, p < 0.05, n = 65–89). Higher baseline CRP was associated with increased risk for syndesmophytes at Week 104/Week 208, and large, short-term decreases in CRP from baseline to Week 14/Week 24 also yielded increased syndesmophyte formation risk.
Conclusion. Elevated CRP after 2 years of anti-TNF treatment correlated with greater radiographic progression risk at 4 years. Elevated CRP at baseline or Week 14/Week 24 of anti-TNF treatment weakly predicted subsequent radiographic progression and modestly predicted residual spinal inflammation in patients with AS treated with anti-TNF. Findings are useful regarding new treatment options in patients treated with anti-TNF. ClinicalTrials.gov: NCT00265083.
- C-REACTIVE PROTEIN
- RADIOGRAPH
- MAGNETIC RESONANCE IMAGING
- ANKYLOSING SPONDYLITIS
- TUMOR NECROSIS FACTOR
- GOLIMUMAB
Ankylosing spondylitis (AS) is an immune-mediated inflammatory disease mainly affecting the axial skeleton. The initial sacroiliac spinal inflammation that characterizes AS is typically followed by new bone formation, e.g., syndesmophytes and ankylosis, which can markedly diminish physical function and quality of life.
While the introduction of antitumor necrosis factor (anti-TNF) biologic agents can significantly improve the signs and symptoms of AS1 and reduce magnetic resonance imaging (MRI)-detected spinal inflammation2,3,4, the effect of such therapy on future radiographic progression is less clear. While studies evaluating etanercept (ETN), infliximab (IFX), and adalimumab did not demonstrate inhibition of radiographic progression after 2 years of therapy5,6,7, radiographic progression appeared to have slowed after 4 years of IFX anti-TNF therapy8. Also, results through 4 years of the phase III, randomized, placebo-controlled GO-RAISE trial of golimumab (GOL) in AS indicated that the radiographic progression rate remained stable at years 2 and 4 of anti-TNF treatment, with no acceleration of new bone formation9. In longer-term assessments of patients with AS, the total duration of anti-TNF treatment was inversely associated with progression; patients who received anti-TNF agents for > 50% of their disease duration had significantly less risk of progression than patients who did not10.
Given the ultimate goals of AS treatment, i.e., control of disease activity, prevention of radiographic progression, and maintenance of physical function, it would be valuable to have a biomarker that is measured pretreatment or after short-term treatment — one that can either identify patients likely to progress rapidly or reliably predict longterm outcomes. In a cohort of patients with AS not treated with anti-TNF agents, the inflammatory marker C-reactive protein (CRP) has been shown to predict radiographic progression11,12. At the same time, serum CRP levels are known to decrease in patients with AS receiving TNF antagonists and demonstrating clinical improvement13,14. In a prospective study of MRI-detected inflammation, radiographic progression, and inflammatory biomarkers in patients with axial spondyloarthritis (axSpA) beginning anti-TNF therapy, development of new syndesmophytes and progression in the modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) were both associated with larger decreases in, and normalization of, CRP levels15. To further assess such relationships, we used data from the GO-RAISE trial of GOL in patients with active AS (NCT00265083) to analyze correlations between serum CRP levels during therapy, radiographic progression, and spinal inflammation as detected by MRI.
MATERIALS AND METHODS
Study design and patients
The GO-RAISE protocol was reviewed and approved by each site’s institutional review board or independent ethics committee (see Appendix 1), and all patients provided written informed consent. Details of the GO-RAISE patient selection criteria and study design have been described16,17. Eligible patients had definite AS according to the modified New York criteria18 and active disease, as defined by a Bath Ankylosing Spondylitis Disease Activity Index19 score ≥ 4 and total back pain visual analog scale score ≥ 4 (both on a scale of 0–10 cm).
The GO-RAISE study design is depicted in Figure 1. Patients with active AS enrolled into the phase III, multicenter, randomized, placebo-controlled, double-blind GO-RAISE trial were randomly assigned (1:1.8:1.8) to receive subcutaneous doses of placebo, GOL 50 mg, or GOL 100 mg at baseline and every 4 weeks. Randomization was stratified by investigational study site and screening CRP level (≤ 1.5 mg/dl, > 1.5 mg/dl). Patients in the placebo group who had < 20% improvement in total back pain and morning stiffness received double-blind early escape treatment at Week 16; thus, the study was fully placebo-controlled from weeks 0 to 16. At Week 24, all patients still receiving placebo crossed over to receive GOL 50 mg. All patients continued double-blind treatment through Week 100.
Design of the GO-RAISE study of golimumab in ankylosing spondylitis. AS: ankylosing spondylitis; MRI: magnetic resonance imaging.
The GO-RAISE longterm extension started with the Week 104 administration of GOL. At the investigator’s discretion, the GOL dose could be increased from 50 mg to 100 mg every 4 weeks or decreased from 100 mg to 50 mg every 4 weeks during the longterm extension20.
Biomarker assessments
Serum samples collected at weeks 0, 4, 14, 24, and 104 of the GO-RAISE trial were tested for selected biomarkers, including CRP, using an ELISA platform by Quintiles Laboratories. The reference range for this assay was 0.0–0.6 mg/dl, and the lower limit of quantitation was 0.3 mg/dl.
Imaging assessments
Lateral view radiographs of the cervical and lumbar spine were performed at weeks 0, 104, and 208. Radiographs were scored using the mSASSS method21, whereby scores of 0, 1, 2, and 3 indicated normal vertebral unit (VU); VU with erosion, sclerosis, or squaring; VU with syndesmophyte; and VU with bridging syndesmophyte, respectively. The total mSASSS ranged from 0 to 72.
Serial spine MRI scans of the cervical, thoracic, and lumbar spine in the sagittal plane were acquired with the patient in the supine position using 1.5 Tesla scanners and phase array spine or quadrature coils at weeks 0, 14, and 104. Image sequences were scored using the AS spine MRI-activity (ASspiMRI-a) score (range 0–138)2.
Radiographs and MRI were read by 2 qualified, experienced, and independent readers who were blinded to treatment information, patient identity, and chronology of the images. Further details of radiographic9 and MRI3 protocols and scoring have been described.
Statistical analysis
Analyses of imaging data collected through Week 208 used observed data; missing data were not imputed. ANOVA using the van der Waerden ranking methodology were used to assess differences in CRP levels at Week 0 and changes at weeks 14 and 24 between patients with mSASSS change ≥ 2 and those with mSASSS change < 2 at Week 104 and Week 208. Reciprocal analyses assessed the differences in the proportions of patients with mSASSS change ≥ 2 between patients with CRP levels ≤ 0.3 mg/dl versus > 0.3 mg/dl and < 0.5 mg/dl versus ≥ 0.5 mg/dl at Week 104.
The relationships between CRP levels and ASspiMRI-a scores were assessed by the generation of Spearman correlation coefficients (rs). P values were adjusted for multiplicity of testing using the Bonferroni methods.
Logistic regression analyses were conducted to assess whether CRP levels at various timepoints and baseline factors (age, sex, HLA-B27 status, mSASSS, and smoking status) conferred an increased risk of syndesmophyte formation or radiographic progression from baseline to Week 104 or Week 208. Syndesmophyte formation was defined as having at least 1 vertebral level on radiograph that changed from a score < 2 at baseline to an mSASSS of 2 or 3 at Week 104 or 208, according to at least 1 reader. Radiographic progression was defined as a change of 2 or more units in the mSASSS from baseline to Week 104 or 208.
RESULTS
Analysis groups
The vast majority of randomized patients (299/356, 84.0%) had pre- and posttreatment spine radiographs scored by the mSASSS. Patients with data available for Spearman correlation analysis included 262–293 patients with CRP and mSASSS data and 65–89 patients with CRP and ASspiMRI-a data across the timepoints assessed.
About 20%–25% of the patients in each group were initially assigned to placebo. As reported, the demographic and baseline characteristics for the MRI substudy patients were generally consistent with those of the overall GO-RAISE patient population3. Assessment of baseline patient and disease characteristics by CRP level at Week 104 indicated more severe and active disease at baseline among patients with CRP levels remaining ≥ 0.5 mg/dl after up to 2 years of GOL therapy (Table 1).
Baseline patient and disease characteristics among randomized patients by CRP level at Week 104. For the variables, n/N refers to number of patients with that data/total number of patients.
Serum CRP levels and mSASSS
Although statistically significant, Spearman correlation coefficients indicated only weak associations between serum CRP levels (baseline, Week 14, Week 24) and mSASSS (Week 104, Week 208) across the assessed timepoints (rs ≤ 0.21). However, median baseline CRP levels were higher and median decreases from baseline to Week 14 and/or Week 24 were greater among patients with mSASSS change scores ≥ 2 at Week 104 and Week 208 (p < 0.05; Table 2). Consistent with these findings, the median mSASSS change from baseline to Week 104 was significantly greater in patients with CRP levels ≥ 0.5 mg/dl than in those with CRP < 0.5 mg/dl at Week 104 (1.09 vs 0.00, p = 0.0369; Figure 2A). Additionally, larger proportions of patients with CRP > 0.3 mg/dl versus ≤ 0.3 mg/dl (34.4% vs 23.4%, p = 0.0819) and CRP ≥ 0.5 mg/dl versus < 0.5 mg/dl (40.4% vs 22.9%, p = 0.0121) at Week 104 had mSASSS change ≥ 2 at Week 104 (Table 3). We observed no consistent difference between the 50-mg and 100-mg doses of GOL (data not shown).
Relationships between mSASSS and serum CRP levels, including mSASSS changes by CRP levels at (A) Week 104 and (B) median mSASSS over time. CRP: C-reactive protein; mSASSS: modified Stoke Ankylosing Spondylitis Spine Score; IQ range: interquartile range.
Median serum CRP levels (mg/dl) through Week 24 and subsequent radiographic progression.
Patients with mSASSS progression by CRP level, mg/dl, at Week 104. Values are n/N (%) unless otherwise specified.
The relationship between serum CRP levels and mSASSS over time is further detailed in Figure 2B. Significant differences in the mSASSS were observed between patient subgroups defined by CRP levels ≤ 0.3 mg/dl versus ≥ 0.5 mg/dl at baseline (Week 0, p = 0.0008) and 4 years (Week 208, p = 0.0027).
Further assessment through logistic regression analysis showed a mildly increased risk of syndesmophyte formation at weeks 104 and/or 208 to be associated with baseline mSASSS, age, sex, and greater short-term decreases in serum CRP levels from baseline to Week 14 and Week 24 (Table 4). The association of change in CRP from baseline to Week 14 or 24 with syndesmophyte formation confirmed the finding shown in Table 2, and suggested that a large decrease in CRP with treatment, rather than the actual baseline or posttreatment CRP values, was predictive of future new syndesmophyte formation. In terms of radiographic progression at weeks 104 and 208, only baseline mSASSS was a significant predictor of subsequent damage (Table 4).
Logistic regression analysis of syndesmophyte formation/radiographic progression adjusted by serum CRP levels and other patient and disease factors.
Serum CRP levels and ASspiMRI-a
Results of the Spearman correlation analyses indicated a moderate degree of association between serum CRP levels and ASspiMRI-a scores at baseline and Week 14 (rs = −0.33 to 0.54, p < 0.05). Baseline CRP levels (rs = −0.40, p = 0.0177) and changes in CRP levels at Week 104 (rs = 0.37, p = 0.0452) also demonstrated moderate associations with MRI-detected change in disease activity at Week 104 (Table 5).
Spearman correlation analysis of serum CRP mg/dl levels and ASspiMRI-a scores.
DISCUSSION
In our investigation, serum biomarker and imaging data from a large trial of patients with AS receiving anti-TNF therapy have been evaluated in a comprehensive and systematic manner. Our data show that serum CRP levels at different timepoints are associated with radiographic progression in patients treated with anti-TNF agents. This could have important implications, since it may suggest that some patients are in need of more or stronger antiinflammatory treatment with TNF antagonists.
Prior to our report, elevated serum CRP levels have been shown to predict radiographic progression in patients with AS who had been conventionally treated11,12,22. Our present analysis shows that changes in serum CRP levels can predict and correlate with radiographic progression in anti-TNF–treated patients, and that the reduction of CRP in such patients correlates with a decrease in spinal inflammation as detected by MRI. In addition, patients whose CRP level was reduced to < 0.5 mg/dl at Week 104 had far less radiographic progression from baseline to Week 104 than did patients whose CRP levels remained ≥ 0.5 mg/dl.
Interestingly, results of logistic regression analysis indicated an increased risk of syndesmophyte formation at weeks 104 and/or 208 to be associated with baseline mSASSS, age, sex, and greater short-term decreases in serum CRP levels from baseline to Week 14 and Week 24. Elevated CRP levels at Week 14 or 24 were not significantly associated with subsequent new syndesmophyte formation. In this respect, our results are similar to a smaller study in patients with axSpA beginning anti-TNF therapy, in which larger, short-term decreases in, and normalization of, CRP levels were associated with development of new syndesmophytes and progression in mSASSS15. On the other hand, in terms of radiographic progression at weeks 104 and 208, only baseline mSASSS was a significant predictor of subsequent increase in mSASSS by ≥ 2 units.
Overall, these findings are consistent with the assumption that, despite the robust reduction of serum CRP levels in anti-TNF–treated patients, there can remain inflammation that appears to contribute to new bone formation. Residual inflammation in anti-TNF–treated patients has also been reported for other studies. In the ESTHER trial, for example, only 15% of patients treated with ETN achieved clinical remission accompanied by no spinal inflammation at 1 year23. Our findings are also consistent with data from the OASIS study24, the certolizumab axSpA trial25, and the treat-to-target recommendations for AS26.
Based on the AS imaging data available in the literature, it is likely that spinal inflammation precedes bone formation in AS4. However, this sequence does not seem to be the only germane one, because the regression of inflammation induced by TNF antagonists, as shown by clinical signs and symptoms, MRI, and biomarkers, alone does not appear to be strong enough to inhibit radiographic progression in the first years of treatment5,6,7. The 2 MRI signals that have been shown to be associated with new bone formation are bone marrow edema (BME) and fat4,27,28,29,30,31. While the former likely represents osteitis, it remains unclear what the latter represents because granulation tissue is usually reported in these areas32. In any case, the most probable sequence of events in this scenario is that the initial inflammation, which may be induced by mechanical stress on a genetic background33, at least in some cases followed by erosion2, is then transformed into a tissue that is both inflammatory and degenerative, with the MRI signal demonstrating both BME and fat. From this tissue, syndesmophytes are more likely to develop than from any other MRI signal27,28. Importantly, at this stage in the disease process, anti-TNF therapies are not effective — showing why it appears to take years until the effect of anti-TNF therapy on bone formation becomes measurable8,9.
A limitation of our study is the relatively small sample size. However, serial measurements within the same patients were available in our study while not available in others11,22. Additionally, given the length of time it appears to take for anti-TNF treatment to influence radiographic progression/bone formation, evaluations of biomarkers and images beyond 4 years should be performed in future clinical trials. Of note, we did not perform dose-response analyses in our posthoc study because no difference in mSASSS change was observed between GOL 50 mg and 100 mg in a previous analysis of GO-RAISE radiographic data9.
Results presented herein suggest that incomplete suppression of inflammation in patients with AS may be a relevant factor for new bone formation. Additionally, serum CRP appears to be a pertinent predictive biomarker for longterm AS progression among patients receiving TNF antagonists.
Acknowledgment
The authors thank the patients, the investigators, and the study personnel who made this trial possible. The authors also thank Michelle Perate, MS, and Mary Whitman, PhD, of Janssen Scientific Affairs LLC, medical writers who helped draft, collate author comments for, collect approvals for, and submit the manuscript.
APPENDIX 1
APPENDIX 1.
The following institutional review boards/ethics committees reviewed and approved this study: China Medical University Hospital Instituted Review of Board 9F, Taichung, Taiwan; Comité de Protection des Personnes Ile-de-France III, Hôpital Tarnier-Cochin 89, Paris, France; Commissie Medische ethiek van de Universitaire Ziekenhuizen KU Leuven/U.Z. Gasthuisberg, Leuven, Belgium; Ethikkommission der ärztekammer Westfalen-Lippe und der Medizinischen Fakultät der WWU Münster, Münster, Germany; Human Experiment and Ethics Committee, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan; HUS Helsingin ja Uudenmaan sairaanhoitopiiri, Medisiininen eettinen toimikunta, Biomedicum Helsinki, Finland; Institution Review Board, Biomedical Research Institute, Seoul National University Hospital, Jongro-gu, Seoul, South Korea; Institution Review Board, Dong-A University Hospital Clinical Research Center, Seo-Gu, Busan, South Korea; Institutional Review Board, Guro Hospital, Korea University Medical Center, Guro-Gu, Seoul, South Korea; Institution Review Board, Hanyang University Hospital, Sungdong-Ku, Seoul, South Korea; Institution Review Board, Pusan National University Hospital, Seo-Gu, Busan, South Korea; Institutional Review Board, Seoul St. Mary’s Hospital/The Catholic University of Korea, Seocho-gu, Seoul, South Korea; Joint Institution Review Board No. 201, Taipei, Taiwan; METC azM/UM Maastricht, Maastricht, the Netherlands; and Western Institutional Review Board, Olympia, Washington, USA.
Footnotes
Supported by Janssen Research and Development LLC, and Merck/Schering-Plough Research Institute Inc.
- Accepted for publication May 31, 2016.
