Radiographic joint space in rheumatoid glenohumeral joints. A 15-year prospective follow-up study in 74 patients.

OBJECTIVE To evaluate radiographically the glenohumeral (GH) joint space in patients with long-term rheumatoid arthritis (RA). METHODS A cohort of 74 patients with RA were followed prospectively for 15 yr. At the end point, 148 shoulders were radiographed using a standard method. The GH joint space was examined from the radiographs using a method developed previously for population studies; the joint space was measured at three different sites and the average of the three measurements, the integral space, was calculated. Destruction of the GH joints was assessed with the Larsen method on a scale of 0-5 and compared with the joint space measurements. RESULTS The mean GH joint space in RA patients was 3. 1 (S.D. 3.3), range -17.3 to 5.7 mm; 2.7 mm (S.D. 4.5) in men and 3. 2 mm (S.D. 2.8) in women. The mean of the affected joints (Larsen grades 2-5), 1.7 mm (S.D. 4.5), was notably narrower than the mean 4. 4 mm (S.D. 0.6) of the non-affected (Larsen grades 0-1) joints. Pathological GH joint space, less than 2 mm, was found in five (15%) of 36 joints in men and in 14 (13%) of 112 joints in women. All the joints graded as Larsen 4 and 5 (n = 17) fulfilled this pathological criterion. Joint space narrowing was associated [r = - 0.66, 95% confidence interval (CI): -0.56 to -0.75] with increasing destruction (Larsen grading) of the joint. The narrowing was significant between non- (Larsen 0, 1), moderately (Larsen 2, 3) and severely (Larsen 4, 5) affected joints (P < 0.001). However, a remarkable step in this process occurred between Larsen grades 3 and 4 when the mean joint space diminished from 3.1 to 0.3 mm. CONCLUSIONS Joint space narrowing is a frequent consequence of GH joint rheumatoid affection. However, joint space narrowing is a late phenomenon occurring not until after marked erosive destruction, which should be noted when using the Larsen method for GH joints.