Skip to content


Open Access

Predictors of response and anti-TNF drugs comparative efficacy

  • Helena Canhão1,
  • Ana F Mourão1,
  • Fernando Martins1,
  • Maria J Santos1,
  • Canas Silva1,
  • Joaquim Pereira1,
  • JA Pereira Silva1,
  • José Costa1,
  • Domingos Araujo1,
  • Candida Silva1,
  • Eugenia Simões1,
  • Catia Duarte1,
  • José Antonio P Silva1,
  • Fernando Pimentel1,
  • Jaime Branco1 and
  • João E Fonseca1
Journal of Translational Medicine20119(Suppl 2):P55

Published: 23 November 2011


InfliximabPropensity ScoreEtanerceptAdalimumabMatching Propensity Score


Tumor necrosis factor (TNF) antagonist therapies changed the rheumatoid arthritis treatment paradigm but arguments to choose among anti-TNF are lacking.

The aim of this work was to evaluate the comparative effectiveness of adalimumab, etanercept and infliximab over 1 year in clinical practice.

Secondary aims were to look for clinical predictors of response.


Analyses were performed upon Response to therapy was defined according to EULAR criteria. Probability of response was modeled. Multivariate logistic regression model predicting response over 1 year with all variables and automated stepwise selection models were built. In addition, we performed analyses using propensity score 1:1:1 nearest neighbor matching algorithms to obtain comparable groups regarding baseline features.


632 RA patients, 250 treated with etanercept, 206 with infliximab and 161 with adalimumab were evaluated.

At baseline visit, infliximab treated patients were significantly older, had lower education level, more comorbidities and more concomitant therapies.

We have found significant negative association with anti-TNF response and smoking, presence of RF and ACPA and higher physician VAS at baseline.

In unadjusted analysis, proportion of good response over 1 year was 59.2% in etanercept, 52% in infliximab and 59.6% in adalimumab group (p=0.21). GLM showed no significant differences between the 3. Multivariate logistic regression analyses demonstrated similar effectiveness. Also, no significant differences in treatment response were detected after applying the matching propensity score analysis.


Smoking, RF, ACPA and physician VAS were identified as predictors of response. In our study using different statistic techniques, we observed similar efficacy over 1 year for the 3 anti-TNF agents studied.

Authors’ Affiliations

Rheumatology Research Unit, Instituto de Medicina Molecular, Lisbon, Portugal


© Canhão et al; licensee BioMed Central Ltd. 2011

This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.