Background Vascular endothelial growth factor (VEGF) and C-reactive protein (CRP) levels, although having been shown to predict radiographic progression, are known to decrease in patients (pts) with ankylosing spondylitis (AS) receiving tumor necrosis factor (TNF)-antagonists.
Objectives Using data from the study GO-RAISE that demonstrated efficacy of golimumab (GLM) in pts with active AS, we analyzed correlations between serum VEGF and CRP levels, radiographic progression and inflammation as detected by magnetic resonance imaging (MRI).
Methods In GO-RAISE, pts with definite AS, BASDAI ≥4, and back pain score ≥4 were randomized to SC GLM 50mg, 100mg, or PBO q4wks. PBO-treated pts crossed over to receive GLM at wk16/24. All pts had lateral view radiographs of the cervical and lumbar spine performed at baseline (BL), wk104, and wk208, all scored using the modified Stoke AS Spine Score (mSASSS). Spinal MRIs in the sagittal plane were acquired using 1.5T scanners with T1 and short tau inversion recovery (STIR) sequences at BL, wk14, and wk104 for 98 pts; images were scored using the AS spinal MRI score for activity (ASspiMRI-a) score. Radiographs and MRIs were assessed by two readers who were blinded to treatment and image time order, the mean scores were used for analysis. Sera were collected from 140 pts for analysis of CRP and VEGF at BL and multiple visits. Spearman correlation analyses were conducted to assess the relationship between VEGF or CRP levels and both mSASSS and MRI-a score at various time points through wk 208. In addition, logistic regression analyses were conducted to assess if VEGF levels conferred an increased risk of a change in mSASSS ≥2 or of syndesmophyte formation from baseline to wk104 or wk208. Syndesmophyte formation was defined as a change from a mSASSS score <2 at baseline to a score of ≥2 (syndesmophyte, bridging syndesmophyte) at wk104 or 208 identified in ≥1 vertebral corner by both readers.
Results CRP serum levels correlated with mSASSS scores at baseline, but not with radiographic progression or changes in MRI-a scores over time. No significant correlations were observed between VEGF serum levels and mSASSS at any time point (Table). Logistic regression analyses failed to show an increased risk of changes in mSASSS >2 or of syndesmophyte formation at wk104 and wk208 associated with VEGF (odds ratio, range: 0.990-1.006, all p=n.s). While a good correlation was observed between changes in ASspiMRI-a and VEGF level at wk14 (p=0.0008), the analysis showed that baseline and wk14 VEGF levels were not predictive of MRI-a scores including change scores at wk104 (Table).
Conclusions Confirming earlier data, CRP serum levels correlated with baseline mSASSS scores. However, they did not predict radiographic progression or spinal inflammation after anti-TNF treatment. When comparing this data with previous studies with historical cohorts, it needs to be taken into account that these and our pt population clearly differed. Similarly, both VEGF and CRP serum levels at baseline were not predictive of either radiographic progression or spinal inflammation in these anti-TNF treated patients. Overall, our data suggest that suppression of VEGF and CRP is not sufficient to halt new bone formation in AS.
Disclosure of Interest X. Baraliakos Grant/research support: Janssen R & D, LLC, K.-G. Hermann Grant/research support: Janssen R & D, LLC, S. Xu Employee of: Janssen R & D, LLC, B. Hsu Employee of: Janssen R & D, LLC, J. Braun Grant/research support: Janssen R & D, LLC