Measurement properties of disease activity instruments in peripheral spondyloarthritis: a post-hoc analysis of the CRESPA trial.

Abstract

Background: Unravelling the performance of disease activity measures in peripheral spondyloarthritis (pSpA) is crucial for the development of clinical studies. We aimed to evaluate the construct validity and discriminatory capacity of various instruments assessing disease activity and response criteria in patients with pSpA.

Methods: Post-hoc analysis of the CRESPA randomised controlled trial including patients with early active pSpA. Patients were randomised to golimumab (GOL) or placebo (PBO). Data of the placebo-controlled part until week 12 were used. Construct validity (known group discrimination assessed with standardised mean difference, SMD) in the 12-week data, longitudinal construct validity (ie, standardised response mean and effect size) and trial discrimination (SMD) of several disease activity instruments were assessed. As part of trial discrimination, a χ² test was performed for binary outcomes.

Results: A total of 60 patients (40 GOL, 20 PBO) were included. Construct validity was better for composite outcomes (ie, Disease Activity in Psoriatic Arthritis (DAPSA), Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and Axial Spondyloarthritis Disease Activity Score (ASDAS)-higher SMD between active and inactive patients). Longitudinal construct validity was consistently good for composite outcomes (eg, ASDAS, DAPSA) and global assessments (Patient Global Assessment (PGA), Physician Global Assessment (PhGA)). Clinical trial discrimination was good for composites (BASDAI, ASDAS, DAPSA), global assessments (PGA, PhGA) and joint counts (swollen joint count (SJC66) and tender joint count (TJC68)).Among binary outcomes, trial discrimination was strongest for clinical remission (ie, absence of arthritis, enthesitis and dactylitis), BASDAI50, DAPSA-Low Disease Activity (LDA) and ASDAS-LDA.

Conclusion: While both composite and global outcome measurement instruments performed well in pSpA, composite scores like DAPSA, ASDAS and BASDAI showed better construct validity. The clinical remission definition was the most discriminatory response criterion.