Abstract
OBJECTIVE: To examine the baseline demographic and disease characteristics that might influence improvement as measured by the Assessment in Ankylosing Spondylitis Response Criteria (ASAS 20) in patients with ankylosing spondylitis (AS). METHODS: A multicenter Phase 3 study was performed to compare the safety and efficacy of 24 weeks of etanercept 25 mg subcutaneous injection twice weekly (n = 138) and placebo (n = 139) in patients with AS. The ASAS 20 was measured at multiple time points. Using a significance level of 0.05, a repeated measures logistic regression model was used to determine which baseline factors influenced response in the etanercept-treated patients during the 24-week double blind portion of the trial. The following baseline factors were used in the model: demographic and disease severity variables, concomitant medications, extra-articular manifestations, and HLA-B27 status. The predictive capability of the model was then tested on the patients receiving placebo after they had received open-label etanercept treatment. RESULTS: Baseline factors that were significant predictors of an ASAS 20 response in etanercept-treated patients were C-reactive protein (CRP), back pain score, and Bath Ankylosing Spondylitis Functional Index (BASFI) score. Although clinical response to etanercept was seen at all levels of baseline disease activity, responses were consistently more likely with higher CRP levels or back pain scores and less likely with increased BASFI scores at baseline. CONCLUSIONS: Higher CRP values and back pain scores and lower BASFI scores at baseline were significant predictors of a higher ASAS 20 response in patients with AS receiving etanercept but predictive value was of insufficient magnitude to determine treatment in individual patients.
