Ultrasonographic analysis in knee osteoarthritis: evaluation of inter-observer reliability

Evidence for the validity of US in detecting structural joint pathology in OA is increasing. However, despite the rapidly emerging field of US in OA, few studies have reported on the inter-observer reliability of US to date. The objective of this study was to assess inter-observer reliability of ultrasonography (US) in the evaluation of specifically defined features in osteoarthritis (OA) of the knee.

US was performed independently by two rheumatologists in 60 outpatients fulfilling the American College of Rheumatology clinical criteria for knee OA. The acquisition protocol comprised medial meniscus protrusion, synovial hypertrophy, effusion, infrapatellar bursitis and cartilage thickness. Cartilage thickness and meniscal protrusion (if >3 mm) were measured on a continuous scale, all other variables were scored dichotomously.

Inter-observer agreement (κ-value) was moderate for protrusion of the medial meniscus (0.54), good for infrapatellar bursitis (0.66) and effusion (0.74), excellent for Bakers' cyst (0.85) and poor for the detection of synovial hypertrophy (-0.08). Inter-observer reliability was good for the measurement of medial meniscus protrusion (correlation coefficient 0.80, 95% limits of agreement -1.93 to 1.94 mm) and cartilage thickness (correlation coefficient 0.62 and 0.68, 95% limits of agreement -0.87 to 0.84 mm and -0.77 to 0.96 mm at the medial and lateral condyle respectively).

This study demonstrated good reproducibility of US in the assessment of the majority of the investigated mechanical, inflammatory and degenerative features of knee OA, and contributes to exploring the use of US in knee OA as a useful tool in research as well as in clinical practice.