Background The pathophysiological processes leading from musculoskeletal (MSK) complaints to clinically manifest rheumatoid arthritis (RA) are not fully understood. The urgency for imaging and serological markers that predict arthritis development in individuals at risk of RA is of clinical importance. Objectives To identify ultrasound (US) markers that can predict arthritis development. Methods Patients presenting with MSK complaints and a positive Anti-Citrullinated Protein Antibody (ACPA) test were referred from primary care units to the Rheumatology Unit. Those lacking clinical signs of arthritis, confirmed by absence of synovial hypertrophy with Doppler activity on US examination, were recruited into the Risk-RA prospective program. A total of 66 patients with complete US records were included between years 2015 up to December 2016. Hands and feet, including symptomatic joints were US-evaluated for synovitis, hyperemia and bone erosions. The presence of wrist (compartments 1–6) and finger (flexor and extensor) tenosynovitis, according to OMERACT guidelines were also recorded in all patients. Serum samples from inclusion were analysed on a multiplex immunoassay Results 66 Risk-RA patients (85% female, median age 50 years, range 22–82) were included and followed up to arthritis onset (median 8 months, range 1–27), or to the end of year 2017 (median 25 months, range 11–43). 27 patients (41%, 86% female, median age 52 years, range 22–74) developed arthritis. Of these 7 had tenosynovitis detected by US at inclusion and 7 more developed tenosynovitis at follow-up visits (in total n=14). At the time of diagnosis, 20 out of 27 patients presented with both tenosynovitis and synovitis. A large majority of patients with tenosynovitis (12 out of 14, 86%) and a minority without tenosynovitis (15 out of 52, 29%) developed arthritis, resulting in an increased relative risk of 3.0 (95% CI 1.8–4.8) to develop arthritis for patients presenting with tenosynovitis at baseline or follow-up visits (p=0.001). Concentrations of the anti-CCP antibodies, anti-CEP antibodies and anti-citrullinated vimentin 60–75 antibodies tended to be higher in patients with tenosynovitis developing arthritis (n=12, median of 70 AU/ml, range 2–175 for anti-CCP, median of 68 AU/ml, range 0–673 for anti-CEP, median of 53, range 0–644 for anti-vim) than those without tenosynovitis developing arthritis (n=15, median of 35 AU/ml, range 1–100 for anti CCP, median of 12, range of 0–1179 for anti-CEP, median of 29, range 0–332 for anti-vimentin). Same trend was observed when comparing patients with tenosynovitis developing arthritis to those without-tenosynovitis not-developing arthritis. The 2 patients with tenosynovitis not developing arthritis, had lower levels of the antibodies as compared to those with tenosynovitis developing arthritis. No significant differences in other patient baseline characteristics were seen between those with, and those without tenosynovitis (86 vs 85% female, median (range) 54 years 29–71 vs 50 years, 22–82 mean visual analogue scale pain 34 vs 31, mean c-reactive protein 2.7 vs 3.2; tender joint count 1.2 vs 0.7). Conclusions Ultrasound detected tenosynovitis in the context of ACPA positivity is a good clinical predictor of rapid arthritis onset in individuals at risk of developing RA. Disclosure of Interest None declared