Article Text

THU0235 The Role of C-Reactive Protein as a Predictor of Treatment Response in Patients with Ankylosing Spondylitis
  1. X. Baraliakos1,
  2. A. Szumski2,
  3. A. Koenig2,
  4. H. Jones2
  1. 1Rheumazentrum Ruhrgebiet, Ruhr-University Bochum, Herne, Germany
  2. 2Specialty Care, Pfizer Inc., Collegeville, United States


Background High baseline C-reactive protein (CRP) levels have been shown to be indicative of treatment response in patients with ankylosing spondylitis (AS).1,2 A range of other baseline AS-specific demographics and characteristics have also been associated with treatment response.1,2

Objectives The objective of this post hoc analysis was to determine if CRP levels correlate with demographics and disease characteristics and to compare CRP to other AS-specific measures as predictors of treatment response in patients with active AS.

Methods Responses to etanercept (ETN), sulfasalazine (SSZ), or placebo plus non-steroidal anti-inflammatory drugs (PBO) were analyzed from four pooled large randomized controlled AS studies. Baseline CRP ≤upper limit of normal (ULN) versus CRP >ULN was compared across a number of AS-specific baseline measures. Baseline predictors, including CRP (≤ULN versus >ULN), were analyzed using odds ratios (ORs) in logistic models with treatment response defined as ASAS20 at Week 12 (Assessment of SpondyloArthritis International Society 20 criteria).

Results In total, data from 1283 patients were used for this analysis: ETN n=867, SSZ n=187, and PBO n=229. Across all treatment groups, baseline CRP >ULN was significantly associated with lower age at diagnosis, higher Ankylosing Spondylitis Disease Activity Score (ASDAS-CRP) score, more males, and a higher proportion with positive Human Leukocyte Antigen B27 (HLA-B27) status (P<0.05 for all). With ETN treatment, the following baseline parameters were significantly predictive of ASAS20 response after 12 weeks (OR, 95% CI): CRP >ULN (1.7, 1.3–2.3), ASDAS-CRP >3.5 (1.7, 1.3–2.3), age of diagnosis ≤40 (1.8, 1.3–2.5), positive HLA-B27 (2.0, 1.4–3.0), and higher nocturnal back pain based on quartiles (1.2, 1.1–1.4), with all predictors appearing to have similar effects on ASAS20 based on similar odds ratios. A Bath Ankylosing Spondylitis Disease Activity Index score of ≤40 was significantly predictive of response to SSZ (OR 3.0, 95% CI 1.1–8.0), as was ASDAS-CRP >3.5 (OR 0.51, 95% CI 0.28–0.92) for PBO.

Conclusions Baseline CRP was significantly associated with typical parameters of the disease (age of diagnosis, ASDAS-CRP, gender, HLA-B27) and significantly predicted short-term treatment response in patients receiving ETN but not other active compounds such as SSZ or NSAIDs.


  1. Arends S et al. Arthritis Res Ther 2011;13:R94.

  2. Vastesaeger N et al. Ann Rheum Dis 2011;70:973-81

Acknowledgements We wish to thank all patients who participated in the trials and medical staff of all participating centres. Medical writing support was provided by John Bilbruck of Engage and was funded by Pfizer Inc.

Disclosure of Interest X. Baraliakos Grant/research support from: MSD, Janssen, Pfizer, Abbott, Novartis, Chugai, UCB, Celgene, Consultant for: MSD, Janssen, Pfizer, Abbott, Novartis, Chugai, UCB, Celgene, Speakers bureau: MSD, Janssen, Pfizer, Abbott, Novartis, Chugai, UCB, Celgene, A. Szumski Consultant for: Employee of inVentiv Health and was contracted by Pfizer Inc. to provide statistical support, A. Koenig Employee of: Employee of Pfizer Inc., H. Jones Employee of: Employee of Pfizer Inc.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.