Six trials (n=378) were included in the study. Most had adequate completeness of data and outcome reporting, but were inadequate allocation concealment, randomisation and blinding. As a result, five studies were considered to have a high risk of bias and one study provided insufficient information to allow risk of bias to be estimated.
Ultrasound was associated with reduced pain compared with control at the end of the intervention (SMD -0.49 on VAS, 95% CI -0.79 to -0.18, I2=51%) and 10 months following completion of ultrasound (SMD -0.77 on VAS, 95% CI -1.15 to -0.39, I2=0%; two studies). Ultrasound was associated with increase in self-reported physical function and walking performance. Only one study reported what the review authors considered an indirect measure of cartilage repair, which showed a positive effect in the index of arthritis severity measured at eight weeks in patients in the lowest third (mean difference 0.80, 95 % CI 0.32 to 1.28) and middle third (mean difference 1.8, 95% CI 0.85 to 2.75), but not the highest third.
No studies were found of patient perception.
Two studies reported no adverse events; the other studies did not report adverse events. A range of clinical subgroup analyses was performed and reported.