Background In the absence of a psoriatic arthritis (PsA)-specific composite disease activity score, rheumatoid arthritis composite scores have predominantly been utilized although they do not incorporate outcomes for accurate PsA assessment. Several PsA-specific composite measures are in development.
Objectives The purpose of this analysis is to assess the performance and relationship between existing and novel PsA composite measures for evidence of validation.
Methods Observed data from the PRESTA trial, a randomized study evaluating 2 etanercept (ETN) dose regimens in psoriasis patients with active PsA, were analyzed. Patients with values to 4 composite indices: the modified composite psoriatic disease activity index (mCPDAI), disease activity index for PsA (DAPSA), PsA disease activity score (PASDAS), and arithmetic mean desirability function (AMDF) at all 3 time points: baseline (BL) and Week 12 and 24 were included. Correlations among endpoints were evaluated using Spearman correlations and dosing regimens were compared using ANCOVA. Parameters influencing each composite index were determined using forward stepwise regression analyses. mCPDAI uses 4 domains: joints (66/68 SJC/TJC, HAQ), skin (PASI andDLQI), dactylitis, and enthesitis for a score of 0–12. DAPSA sums patient global health and joint pain VAS assessments, 66/68 SJC/TJC, and CRP for a score of 0–160. PASDAS utilizes patient global health and physician global arthritis VAS, 66/68 SJC/TJC, CRP, predicted physical component score, dactylitis, and enthesitis. AMDF is calculated from patient psoriasis, joints, and global health VAS assessments, PASI, 66/68 SJC/TJC and HAQ for a score 0–1.
Results All indices could distinguish response to treatment comparing BL and 12/24-week values. Of the 4 measures, mCPDAI and AMDF were able to differentiate between the dosing regimens at Week 12 (P<0.05). Although all 4 measures were correlated at BL, they were more highly correlated at Weeks 12/24. The strongest correlations at Weeks 12/24 were between AMDF versus PASDAS and DAPSA (absolute r>0.86) and PASDAS versus DAPSA (r>0.84). mCPDAI was correlated with all indices (absolute r>0.72) at Weeks 12/24. AMDF, CPDAI, DAPSA, and PASDAS were largely influenced by patient VAS assessments and joint scores (r2>0.90), dactylitis and enthesitis (r2>0.69), TJC and CRP (r2>0.94) and physician and patient VAS (r2>0.73), respectively. Interestingly, PASI was found to be the least influential on AMDF (r2=0.01) and CPDAI (r2=0.08).
Conclusions In addition to distinguishing treatment responses, both AMDF and mCPDAI which better reflect most PsA disease domains, strongly correlate with other composite measures. Further clinical trial analysis and validation of measures is required.
Acknowledgements The PRESTA trial was funded by Pfizer Inc. Medical writing support was provided by Stephanie Eide of UBC Scientific Solutions and funded by Pfizer Inc.
Disclosure of Interest P. Helliwell Speakers bureau: Pfizer Inc, O. FitzGerald Speakers bureau: Pfizer Inc, L. Marshall Employee of: Pfizer Inc, R. Pedersen Employee of: Pfizer Inc, E. Bananis Employee of: Pfizer Inc