Table 5

Baseline MRI findings as predictors for development of radiographic erosions in the same joint

Presence of MRI featuresOR (95% CI) of radiographic erosive evolution
(in joints with no radiographic erosions at baseline)
n (%)Crude associationsAdjusted associations*
Synovitis
 Grade 1170/386 (44.0)1.66 (0.58 to 4.77); p=0.351.24 (0.35 to 4.45); p=0.74
 Grades 2 and 3†88/386 (22.9)5.48 (1.71 to 17.62); p=0.0044.83 (1.35 to 17.24); p=0.02
Flexor tenosynovitis66/378 (17.5)2.06 (1.04 to 4.11); p=0.041.92 (0.87 to 4.27); p=0.11
Focal erosions
 Grade 1124/435 (28.5)1.35 (0.73 to 2.49); p=0.341.13 (0.53 to 2.40); p=0.75
 Grades 2 and 3‡51/435 (11.7)3.45 (1.68 to 7.09); p=0.0012.55 (1.11 to 5.83); p=0.03
Subchondral bone attrition21/436 (4.8)4.18 (1.29 to 13.57); p=0.025.58 (1.46 to 21.36); p=0.01
Cysts18/435 (4.1)3.03 (0.97 to 9.41); p=0.063.19 (0.88 to 11.60); p=0.08
Osteophytes§
 Grade 252/437 (11.9)3.09 (1.53 to 6.23); p=0.0024.11 (1.84 to 9.18); p=0.001
 Grade 321/437 (4.8)4.85 (2.04 to 11.54); p<0.0014.78 (2.04 to 11.18); p<0.001
Joint space narrowing¶
 Grades 2–3115/437 (26.3)9.13 (4.08 to 20.39); p<0.0017.78 (3.19 to 18.95); p<0.001
Bone marrow lesions37/438 (8.4)4.93 (2.04 to 11.95); p<0.0015.61 (2.05 to 15.38); p=0.001
Malalignment12/438 (2.7)2.30 (0.54 to 9.86); p=0.2610.18 (2.01 to 51.64); p=0.005
Collateral ligament discontinuity209/437 (47.8)1.41 (0.73 to 2.74); p=0.301.43 (0.64 to 3.22); p=0.38
  • Generalised Estimating Equations with separate models for each MRI feature. Statistically significant associations are shown in italics.

  • *Adjusted for age, sex, BMI, follow-up time and other radiographic erosive joints at baseline.

  • †Few joints with synovitis grade 3 (n=8).

  • ‡Few joints with focal erosions grade 3 (n=6).

  • §Osteophyte grade 0 (n=67) and grade 1 (n=297) as reference.

  • ¶Joint space narrowing grade 0 (n=170) and grade 1 (n=152) as reference (erosive development occurred only in joints with joint space narrowing grades 1–3 at baseline).

  • BMI, Body Mass Index.