Patient Profiles in Randomized Controlled Trials Versus a Real-World Study in Psoriatic Arthritis: Scoping Review and Metaanalysis.

Abstract

Patients with psoriatic arthritis (PsA) in randomized controlled trials (RCTs) may not reflect patients with PsA in clinical practice. Our objective was to perform a metaanalysis comparing the characteristics of patients with PsA in RCTs of biologic disease-modifying antirheumatic drugs (bDMARDs) to patient profiles in a real-world study. Data sources included (1) a scoping literature review of phase III RCTs of bDMARDs in PsA published between 2015 and 2020, and (2) an international observational study of patients with PsA starting a bDMARD enrolled between 2015 and 2018 (PsABio; ClinicalTrials.gov: NCT02627768). Data collected at baseline included swollen and tender joint counts (SJC/TJC), presence of enthesitis, skin involvement (body surface area [BSA]), C-reactive protein (CRP), physician global assessment (PGA), and patient-reported outcomes (PROs; Health Assessment Questionnaire [HAQ], pain). Univariate random effects metaanalysis was conducted to calculate pooled means and proportions. Overall, 5654 patients from 10 RCTs were compared to 930 PsABio patients. Demographic data were similar. SJC/TJC were higher in RCTs than in PsABio (pooled means: 11.8/21.5 vs 5.7/11.9), and enthesitis was more frequent in RCTs (64.7% vs 48.2%), as were patients with a BSA ≥ 3% (62.2% vs 54%). PGA was higher in RCTs (59.7 vs 54.1). In contrast, PROs were similar, whereas CRP was significantly higher in PsABio (1.4 vs 1.1 mg/dL). Patients with PsA starting a bDMARD in RCTs had highly active disease and a high patient-reported disease burden. In contrast, PsABio real-world patients starting a bDMARD had lower SJC/TJC, skin involvement, and PGA, but presented with similar patient-reported disease burden. The extrapolation of RCT data in clinical practice should take these elements into account.