Objectives To define the correlation between joint ultrasonography and clinical examination in patients with juvenile idiopathic arthritis (JIA) and to assess whether synovitis detected by ultrasonography in clinically inactive patients predicts arthritis flares.

Methods 88 consecutive patients with JIA—46 (52%) with persistent oligoarthritis, 15 (17%) with extended oligoarthritis, 15 (17%) with rheumatoid factor-negative polyarthritis and 12 (14%) with other forms of JIA, all clinically inactive for a minimum of 3 months—underwent ultrasound (US) assessment of 44 joints. Joints were scanned at study entry for synovial hyperplasia, joint effusion and power Doppler (PD) signal. Patients were followed clinically for 4 years.

Results US was abnormal in 20/88 (22.7%) patients and in 38/3872 (0.98%) joints. Extended oligoarthritis and rheumatoid factor-negative polyarthritis were more frequent in US-positive than in US-negative patients (35.0% vs 11.8% and 30.0% vs 13.2%, respectively; P=0.005). During 4 years of follow-up, 41/88 (46.6%) patients displayed a flare; 26/68 (38.2%) were US-negative and 15/20 (75%) were US-positive at baseline. Abnormality on US examination, after correction for therapy modification, significantly increased the risk of flare (OR=3.8, 95% CI 1.2 to 11.5). The combination of grey scale and PD abnormalities displayed a much higher predictive value of relapse (65%, 13/20) than grey scale alone (33%, 6/18).

Conclusions US abnormalities are a strong predictor of relapse at individual patient level. Irrespective of treatment, the risk of flare in US-positive versus US-negative patients was almost four times higher. In case of US abnormalities, patients should be carefully followed regardless of both the International League of Associations for Rheumatology and Wallace categories.