Association between MRI findings and patient‐reported outcomes in patients with rheumatoid arthritis in clinical remission and at relapse

Abstract

To investigate whether magnetic resonance imaging (MRI) pathologies in the wrist/hand of rheumatoid arthritis (RA) patients are associated with patient-reported outcomes (PROs) at clinical remission and relapse. Wrist/hand MRIs and wrists/hands/feet radiographs were obtained in 114 established RA patients in clinical remission, before tapering their biologic disease-modifying antirheumatic drugs. MRIs were assessed according to the Outcome Measures in Rheumatology (OMERACT) RA MRI score (RAMRIS) for inflammation (synovitis/tenosynovitis/bone marrow edema) and damage (bone erosion/joint space narrowing) at baseline (ie remission) and in case of a relapse (n=70). Radiographs were assessed according to the Sharp/van der Heijde (SvH) method at baseline. These scores were assessed for associations with health assessment questionnaires (HAQ), visual analog scales (VAS global/pain), EuroQol-5 dimensions and Short-Form 36 physical and mental component summary (SF-36 PCS/MCS) using Spearman correlations, univariate/multivariable linear regression analyses and generalized estimating equations. Furthermore, MRI pathologies were assessed for association with specific hand-related HAQ items using Jonckheere trend tests. Magnetic resonance imaging-assessed damage was associated with impaired HAQ and SF-36 PCS at remission and relapse (P<.01), independent of clinical and radiographic measures, and was also associated with most of the hand-related HAQ items (P<.03). In multivariate models including MRI, SvH scores were not associated with PROs. MRI-assessed inflammation was not associated with PROs at remission or relapse. Magnetic resonance imaging-assessed wrist/hand damage, but not inflammation, in patients with established RA is associated with patient-reported physical impairment at remission and relapse. The amount of damage in the wrist/hand is associated with reduced hand function.