Background/Purpose: Currently there are several outcome metrics used for the assessment and quantification of musculoskeletal (MSK) ultrasound inflammatory burden in rheumatoid arthritis (RA). This study compares the effect sizes of MSK ultrasound metrics in RA. Effect size is an important factor that judges the robustness of response in a standardized fashion and impacts power and sample size calculations in clinical trials.

Methods:

This study was an open-label 24 week trial of 54 RA patients starting tocilizumab IV recruited from two rheumatology clinic sites. Inclusion criteria required DAS28-ESR>4.4 and power Doppler (PDUS) >10. At baseline patients started 4mg/kg dosing and were dose escalated to 8mg/kg at 12 weeks if the DAS28-ESR>3.2. Joints examined by ultrasound included bilateral MCPs, PIPs, wrist midline/radioulnar, medial/lateral parapatellar knee recesses, and MTP 2-5. Joint images of PDUS and grey-scale synovial hypertrophy (GSUS) were assessed based on a semiquantitative scale (0-3). Ultrasound measure means were calculated: maximum PDUS by joint, PDUS by view, German US7 PDUS (wrist, MCP/PIP 2/3, MTP2/5), modified Global OMERACT-EULAR Synovitis Score (GLOESS) 22 (all except shoulder, elbow, ankle, hindfoot), modified GLOESS 9 (excludes elbow, shoulder), GLOESS 4 (MCP2-5), PDUS dichotomous (0 vs 1,2,3), PDUS dichotomous (0,1 vs 2,3), GSUS US7, GSUS max joint, and GSUS by view. Cohen’s d effect sizes were computed as the mean change in the measure from baseline to 4, 12, and 24 weeks of the study divided by the standard deviation of the changes.

Results: 54 RA patients enrolled in the trial and 44 patients completed the study. Of all the MSK ultrasound measures calculated, the PDUS dichotomous 0 vs 1,2,3 had the largest effect size of 0.92 at 24 weeks and the GLOESS 4 demonstrated the smallest effect size at 0.55 (Figure 1). The US7 PDUS measure also performed well with the second best effect size of 0.85 at 24 weeks.

Conclusion: Effect sizes are important in clinical trials to judge if the response was robust and meaningful. It has important implications for calculating the power and sample size in randomized clinical trials. For MSK ultrasound outcome measures, the PDUS dichotomous measure of 0 vs 1,2,3 had a large effect size and GLOESS measures had moderate effects sizes. Based on this study, a simple PDUS dichotomous measure of 34 joints will decrease the sample size by approximately 60% compared to utilizing the GLOESS 4.

« Back to 2017 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/effect-size-comparison-of-ultrasound-measures-in-rheumatoid-arthritis/