EULAR Recommendations for Abnormalities in Knee Osteoarthritis Need an Update

The EULAR thresholds of an ultrasound for knee osteoarthritis (OA) were found to have high misclassification rates with poor sensitivity for both synovial effusion and hypertrophy. These were the findings of a new study that has recommended updated cutoffs for synovial effusion and hypertrophy to help screen for patients with symptomatic radiographic knee osteoarthritis (RKOA).

Knee OA is a joint disorder that can be detected through ultrasound. Even though ultrasound is more sensitive than clinical examination at detecting synovial effusion and hypertrophy in knees, values for effusion and hypertrophy are rather limited. Current EULAR cutoff thresholds are not sensitive, leading to patients being misclassified. In the current study, investigators established optimal cutoffs for symptomatic knee OA and analyzed the normal ranges of ultrasound synovial changes in healthy patients, stratified by age, gender, and laterality.

Of the total 299 patients, both male and female, 163 were categorized as healthy controls that had no knee pain or RKOA and 44 patients had symptomatic RKOA. Power doppler signaling was 0.65% in healthy patients and 7% in symptomatic RKOA (P = .0083). Because the frequency was very close to 0 in healthy patients, normal ranges and cutoffs were not needed.

For healthy individuals, age and laterality did not affect synovial effusion and hypertrophy, but gender did. On average, men had greater synovial effusion and hypertrophy than women (median synovial effusion, 4.7 mm versus 3.4 mm, P = .0035; median synovial hypertrophy 2.0 mm vs 0 mm, P = .0012). Synovial hypertrophy differences in gender were still significant after adjusting for height (P = .019); however, effusion was not (P >.05). In most healthy patients, the normal ranges for effusion were 0 mm-10.3 mm in men and 0 mm-9.8 mm in women; the normal ranges for hypertrophy were 0 mm-6.8 mm in men and 0 mm-5.4 mm in women.

After finding the maximum sensitivity and specificity possible, investigators determined that the optimal threshold for effusion in men and women were 7.4 mm and 5.3 mm, respectively, and for synovial hypertrophy, 3.7 mm for men and 1.6 mm for women. Therefore, optimal cutoffs to screen for people with abnormal synovial effusion changes in symptomatic RKOA would be approximately 7 mm in men and 5 mm in women, while abnormal synovial hypertrophy would be 4 mm in men and 2 mm in women. Investigators also determined a threshold for high specificity that would be useful if a more stringent cutoff is required such as in randomized clinical trials. Thresholds with a specificity of 90% was defined as 8.9 mm (~9 mm) in men and 7.8 mm (~8 mm) in women for synovial effusion and 5.8 mm (~6 mm) in men and 4.2 mm (~4 mm) in women for synovial hypertrophy.

The increased use of ultrasound for knee OA requires more specific classifications for categorizing patients that are healthy and patients that have symptomatic RKOA. The authors recommend that the different thresholds found for both synovial effusion and hypertrophy can be used to revise previous EULAR guidelines for ultrasound abnormalities in knee OA.

Reference

Sarmanova A, Hall M, Fernandes GS, et al. Thresholds of ultrasound synovial abnormalities for knee osteoarthritis — a cross sectional study in the general population [published online November 16, 2018]. Osteoarthritis Cartilage. doi: 10.1016/j.joca.2018.09.018.