Purpose: To assess whether ultrasonographic (US) findings associate with clinical findings in severe knee osteoarthritis (OA). Association of US findings with side-of-knee pain and inter-reader agreement of knee US were also evaluated.
Methods: One-hundred-two patients (in total 123 knees) with severe knee OA were recruited for this cross-sectional study. US was performed by a single observer, and on 53 knees by two independent observers to assess inter-reader reliability. Preoperative clinical data was available for 69 knees. Cutoff values were applied to dichotomize US and clinical findings. The Chi-square test, Mann-Whitney test, and prevalence- and bias-adjusted kappa (PABAK) were applied for statistical analyses.
Results: Seven of 99 associations tested were statistically significant. Associations were observed between range of flexion and lateral femoral (P = .009) and tibial (P = .001) osteophytes, mediolateral instability and damage to the lateral femoral cartilage (P = .014) and damage to the lateral meniscus (P = .031), and alignment and damage to the lateral femoral cartilage (P < .001), lateral tibial osteophytes (P = .037), and damage to the lateral meniscus (P < .001). A strong association was observed between medial-sided pain and same-sided cartilage damage and osteophytes (P < .001). That inter-reader agreement was excellent on the medial side of the knee joint (PABAK = 0.811-0.887).
Conclusions: US findings show a rather poor association with clinical OA findings. Inter-reader agreement of knee US is excellent on the medial side.
Keywords: Knee Society Score; knee; musculoskeletal system; osteoarthritis; ultrasonography.
© 2019 Wiley Periodicals, Inc.