Open Access Open Badges Research article

MRI osteitis predicts cartilage damage at the wrist in RA: a three-year prospective 3T MRI study examining cartilage damage

Fiona M McQueen15*, Alexandra McHaffie2, Andrew Clarke2, Arier C Lee3, Quentin Reeves2, Barbara Curteis1 and Nicola Dalbeth145

Author Affiliations

1 Department of Molecular Medicine and Pathology, Faculty of Medical and Health Sciences, University of Auckland, 85 Park Rd, Grafton, Auckland, New Zealand

2 Department of Radiology, Auckland City Hospital, Auckland District Health Board, Auckland, New Zealand

3 Department of Epidemiology and Biostatistics, University of Auckland, Auckland, New Zealand

4 Bone and Joint Research Group, Department of Medicine, University of Auckland, Auckland, New Zealand

5 Department of Rheumatology, Greenlane Clinical Centre, Auckland District Health board, Auckland, New Zealand

For all author emails, please log on.

Arthritis Research & Therapy 2014, 16:R33  doi:10.1186/ar4462

Published: 30 January 2014



Cartilage damage impacts on patient disability in rheumatoid arthritis (RA). The aims of this magnetic resonance imaging (MRI) study were to investigate cartilage damage over three years and determine predictive factors.


A total of 38 RA patients and 22 controls were enrolled at t = 0 (2009). After 3 years, clinical and MRI data were available in 28 patients and 15 controls. 3T MRI scans were scored for cartilage damage, bone erosion, synovitis and osteitis. A model was developed to predict cartilage damage from baseline parameters.


Inter-reader reliability for the Auckland MRI cartilage score (AMRICS) was high for status scores; intraclass correlation coefficient (ICC), 0.90 (0.81 to 0.95) and moderate for change scores (ICC 0.58 (0.24 to 0.77)). AMRICS scores correlated with the Outcome MEasures in Rheumatoid Arthritis Clinical Trials (OMERACT) MRI joint space narrowing (jsn) and X-Ray (XR) jsn scores (r =0.96, P < 0.0001 and 0.80, P < 0.0001, respectively). AMRICS change scores were greater for RA patients than controls (P = 0.06 and P = 0.04 for the two readers). Using linear regression, baseline MRI cartilage, synovitis and osteitis scores predicted the three-year AMRICS (R2 = 0.67, 0.37 and 0.39, respectively). A multiple linear regression model predicted the three-year AMRICS (R2 = 0.78). Baseline radial osteitis predicted increased cartilage scores at the radiolunate and radioscaphoid joints, P = 0.0001 and 0.0012, respectively and synovitis at radioulnar, radiocarpal and intercarpal-carpometacarpal joints also influenced three-year cartilage scores (P-values of 0.001, 0.04 and 0.01, respectively).


MRI cartilage damage progression is preceded by osteitis and synovitis but is most influenced by pre-existing cartilage damage suggesting primacy of the cartilage damage pathway in certain patients.