Article Text

THU0207 Comparative Analysis of the Effectiveness of anti-TNF Therapies in Non-Radiographic Axial Spondyloarthritis Using Novel Statistical Techniques
  1. A. Benedict1,
  2. J. Ishak2,
  3. P. Gal1,
  4. I. Proskorovsky2,
  5. J.C. Cappelleri3,
  6. T. Everiss4,
  7. H. Jones5,
  8. R. Hughes4,
  9. A. Jenkins4
  1. 1Evidera, Budapest, Hungary
  2. 2Evidera, St-Laurent, Quebec, Canada
  3. 3Pfizer Inc, Groton, CT, United States
  4. 4Pfizer Ltd, Tadworth, Surrey, United Kingdom
  5. 5Pfizer Inc, Collegeville, PA, United States


Background The anti-tumor necrosis factor (TNF) agents adalimumab (ADA), certolizumab pegol (CZP), and etanercept (ETN) are approved in Europe for the treatment of severe non-radiographic axial spondyloarthritis (nr-axSpA). The efficacy of these agents has been demonstrated against placebo in randomised controlled trials,1-3 but no head-to-head studies have been conducted. Moreover, the results from a typical network meta-analysis may be misleading in nr-axSpA due to possible heterogeneity between studies caused by differences in populations.

Objectives To assess the relative effectiveness of ETN vs ADA and CZP in axSpA (primarily nr-axSpA) populations using the novel matching-adjusted indirect comparison (MAIC) and simulated treatment comparison (STC) statistical techniques to adjust for between-population differences in clinical characteristics across studies.

Methods Pooled index trials of ETN (ongoing EMBARK3 [nr-axSpA] + 314-EU4 [ankylosing spondylitis]) were compared with the ABILITY-11 and RAPID-axSpA2 trials of ADA and CZP. Differences in changes in BASDAI and BASFI from baseline to week 12 for ETN vs ADA and CZP were derived adjusting for baseline characteristics (via weighting in MAIC and prediction in STC). To adjust for differences in characteristics that could not be matched and missing variables, differences between placebo arms were also measured to derive the study effect; study effect was subsequently subtracted from the adjusted estimate of relative treatment effect to account for possible residual confounding or study effects.

Results Incremental changes in BASDAI and BASFI for ETN compared with ADA and CZP are shown (table). The analyses show similar and not statistically significantly different changes in BASDAI and BASFI for ETN and comparators at week 12. Similar or identical treatment results were observed using MAIC and STC.

Table 1.

Comparison of outcomes with ETN vs ADA and ETN vs CZP using MAIC and STC methods

Conclusions Based on analyses using MAIC and STC methods, etanercept was not different in efficacy from ADA and CZP in patients with axSpA.


  1. Sieper J, et al. Ann Rheum Dis. 2013;72:815-22.

  2. Landewe R, et al. Ann Rheum Dis. 2014;73:39-47.

  3. Dougados M, et al. Arthritis Rheumatol. 2014;66:2091-102.

  4. van der Heijde D, et al. Ann Rheum Dis. 2006;65:1572-7.

Disclosure of Interest A. Benedict Consultant for: Pfizer Ltd, J. Ishak Consultant for: Pfizer Ltd, P. Gal Consultant for: Pfizer Ltd, I. Proskorovsky Consultant for: Pfizer Ltd, J. C. Cappelleri Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, T. Everiss Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, H. Jones Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, R. Hughes Employee of: Pfizer Inc, A. Jenkins Shareholder of: Pfizer Inc, 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.