Objective. Musculoskeletal US can be useful in monitoring RA. It can be time-consuming and there is no consensus in defining the joints to evaluate. We assessed the validity, sensitivity to change and feasibility of a reduced 6-joint US score in patients with RA starting therapy with an anti-TNF agent. Methods. A group of consecutive RA patients starting etanercept were investigated. The patients underwent clinical evaluation, laboratory tests and US assessment at baseline and 3 months. A semi-quantitative score (0-3) was used to evaluate synovial effusion (SE), synovial proliferation (SP) and power Doppler (PD) signal in 12 joints. A process of data reduction, based on the frequency of synovial site involvement by US-SE, US-SP and US-PD signal, was conducted to investigate the validity of a 6-joint US assessment. Results. Forty-five RA patients were evaluated. A significant decrease in all clinical, serological and 12-joint US parameters was found at follow-up. A significant correlation between changes in the DAS-28 and changes in the US scores in the 12-joint assessment was observed at follow-up (P