Table 3

Results of Cox regression analyses of MRI-detected subclinical inflammation at baseline in relation to arthritis development

HR (95% CI)p Value
Continuous RAMRIS scores
Univariable
 Total inflammation score, per unit1.14 (1.08 to 1.20)<0.001
 Total synovitis score, per unit1.29 (1.14 to 1.47)<0.001
 Total BME score, per unit1.28 (1.13 to 1.46)<0.001
 Total tenosynovitis score, per unit1.25 (1.11 to 1.39)<0.001
Multivariable
 Total synovitis score, per unit1.09 (0.86 to 1.38)0.47
 Total BME score, per unit1.20 (1.03 to 1.38)0.016
 Total tenosynovitis score, per unit1.15 (0.94 to 1.41)0.17
MRI dichotomised for the presence of any subclinical inflammation and for specific inflammatory features
Univariable
 Presence of any MRI-detected inflammation6.12 (2.32 to 16.19)<0.001
 Presence of MRI-detected synovitis2.22 (1.03 to 4.78)0.042
 Presence of MRI-detected BME2.39 (1.04 to 5.46)0.039
 Presence of MRI-detected tenosynovitis7.56 (3.30 to 17.32)<0.001
Multivariable
 Presence of MRI-detected synovitis0.72 (0.31 to 1.69)0.45
 Presence of MRI-detected BME2.09 (0.91 to 4.81)0.084
 Presence of MRI-detected tenosynovitis8.39 (3.38 to 20.81)<0.001
  • Presented are the HRs of analyses including 143 patients with clinically suspect arthralgia that underwent MRI of which 27 developed clinical arthritis. The HR of 1.14 for the total inflammation score indicates that when the total MRI inflammation score increased with 1 unit the hazard on arthritis development increased with a factor 1.14.

  • BME, bone marrow oedema; RAMRIS, rheumatoid arthritis MRI scoring system.