Metrologic properties of ultrasound versus clinical evaluation of synovitis in rheumatoid arthritis: results of a multicenter, randomized study

Objective: To evaluate the intraobserver reliability, face validity, and discriminant capacity of different global ultrasound (US) scoring systems for measuring synovitis in rheumatoid arthritis (RA).

Methods: This study was ancillary to a 52-week, multicenter, prospective, randomized, open-label, parallel-group outpatient study conducted in patients with moderate RA who were randomized to receive either etanercept combined with methotrexate or various disease-modifying antirheumatic drugs. A total of 66 different synovitis scoring systems were constructed and evaluated, including 11 different joint combinations; data derived from clinical findings, gray-scale US, and power Doppler US (PDUS); and both binary counts and semiquantitative scores.

Results: Due to discontinuation of the trial, only 62 patients, a subset of the initially planned number of patients, were included in this study. Reliability was found to be better for gray-scale US and PDUS than for clinical evaluation of synovitis in patients with stable disease between the screening and baseline visits (range for intraclass correlation coefficient 0.6, 0.95 for gray-scale US and 0.56, 0.93 for PDUS versus 0.31, 0.75 for clinical indices). The median (range) difference in the discriminant capacities of clinical indices versus gray-scale US and versus PDUS was 0.25 (-0.64, 0.96) and -0.025 (-0.59, 0.53), respectively, in the period from baseline to 12 weeks. No relevant differences in metrologic properties were observed regarding the number and composition of joints between the different scoring systems. Our findings suggested that a simplified scoring system referring to gray-scale US and PDUS findings might be sufficient.

Conclusion: Our findings indicate that gray-scale US and PDUS have better reliability than generally used clinical indices for evaluating synovitis in RA. PDUS has at least as good discriminant capacity as clinical assessment of synovitis for distinguishing between treatment arms.