Inter-observer reliability of ultrasound detection of tendon abnormalities at the wrist and ankle in patients with rheumatoid arthritis

OBJECTIVE: To assess inter-observer reliability in US detection of tendon inflammatory and structural changes at wrists and ankles in RA patients. METHODS: Fourteen consecutive RA patients underwent bilateral US assessment of the extensor carpi ulnaris (ECUT) and tibialis posterior tendons (TPTs) by two blinded rheumatologists, with different level of experience in musculoskeletal (MS) US. Grey scale and power Doppler (PD) US assessment was focused on detection of tenosynovitis, tenosynovial and intra-tendon PD signal and structural lesions (i.e. tendinosis, tendon erosion, partial or total rupture). RESULTS: The frequency of US findings detected by Investigator 1 was 28.6% for inflammatory changes and 51.8% for structural damage changes while Investigator 2 detected 34 and 53.6% for the corresponding abnormalities. A high overall agreement (82.7%) was found for inflammatory pathology and 89.7% for structural lesions in all tendons. Mean kappa (κ) values for all tendons and pathology was moderate (κ = 0.42), with fair level of agreement for the wrist region (0.27-0.34) and moderate to good values for the ankle region (κ = 0.47-0.62). Subclinical abnormalities were detected in 37.5% of the tendons by Investigator 1 and 28.6% of the tendons by Investigator 2. CONCLUSIONS: MSUS showed high overall agreement and fair to moderate inter-observer κ-values between investigators with different levels of experience in detection of tendon pathology at the wrist and ankle in RA patients. Further standardization of scanning method and pathology definitions may improve MSUS reproducibility