Ultrasound evaluation of muscle quality in rheumatoid arthritis: a reliability study of the Spanish Ultrasound Muscle Assessment in Rheumatoid Arthritis (SpUMAR) score.

OBJECTIVE: This study aimed to 1) develop and test the reliability of a semiquantitative ultrasound score to assess muscle quality in patients with rheumatoid arthritis (RA), and 2) evaluate how incorrect adjustments of ultrasound settings affect muscle quality assessment. METHODS: The Spanish Ultrasound Muscle Assessment in RA (SpUMAR) Score consisted of a semiquantitative 0–3 score based on the degree of loss of the normal muscle ultrasound pattern that was agreed upon by 15 rheumatologists expert in musculoskeletal ultrasound. A total of 164 static ultrasound images of the four bellies of the quadriceps muscle from 57 patients with RA were scored by the experts in two rounds. Additionally, the experts scored 20 ultrasound images that were acquired with incorrectly adjusted settings, which were interposed among the other images. Gwet’s AC2, Fleiss' kappa (κ), Cohen’s κ, applying linear weighting, were used for agreement analysis. RESULTS: The interobserver reliability of SpUMAR was moderate in both rounds (Gwet’s AC2 [95% CI] 0.57 [0.54–0.60] and 0.54 [0.51–0.57], percent agreement 81.76% and 80.28%, respectively). Intraobserver reliability between the two rounds was good (Gwet’s AC2 [95% CI] 0.66 [0.64–0.68]). The agreement between the scores given to the images with incorrect settings and the original scores for those images was poor in both rounds (κ [95% CI]: 0.27 [0.21–0.33] and 0.30 [0.24–0.36], respectively). CONCLUSION: Multiobserver assessment of muscle quality using the SpUMAR score demonstrated acceptable reliability in patients with RA. Incorrect adjustment of ultrasound settings greatly affects muscle quality scoring; therefore, their standardised optimisation is necessary for multicentre studies.