Background Some studies revealed an association of Power Doppler (PD) ultrasound (US)-detected residual synovitis (PDUSS) and risk of relapse and radiographic progression (RP), in individual patients in rheumatoid arthritis (RA). However, the longitudinal relationship between clinical remission and repeated US residual lesions during follow-up is not so well-known.

Objectives The aim of this study was to evaluate the ability of PDUSS in consecutive examinations to predict unfavorable outcome (i.e. clinical relapse or RP) at 1 year.

Methods RA patients ≥18 years fulfilling 2010 ACR-EULAR criteria, treated with synthetic or biologic (b) DMARDs and in clinical remission (DAS28-ESR<2.6 and no clinically active synovitis) for less than 6 months, were included in the longitudinal prospective SONORE study (ClinicalTrials.gov identifier: NCT02618954). Clinical and biological characteristics of patients were collected at baseline, and every 3 months during 1 year. RA treatment had to be stable during follow-up. A standard US examination on 40 joints for the presence of synovial hypertrophy and PD signal was performed by an independent investigator blinded to clinical and radiographic data at each visit during 1 year. Presence of US synovitis was defined by a PD signal≥1 in at least one joint. Radiographs of hands, wrists and feet were scored at baseline and 1 year. Outcome measures: RP was defined by an increase ≥1 point of the modified total Sharp score. A relapse was defined by a DAS28>3.2 at ≥1 follow-up visits AND a change of DMARDs, excluding change due to safety issues; or an increase in the DMARD or Corticosteroid (CS) dosage (≥5mg/d). Baseline variables, including PDUSS and its persistence during the follow-up, were assessed for their association with time to progression to unfavorable outcome using univariate then stepwise multivariate Cox regression analyses to obtain adjusted HRs.

Results The 115 included patients had a mean (SD) age of 58.9 (±12.8) years, mean disease duration of 9.3 (±9.3) years, a mean duration of remission of 2.1 (±2.3) months. 74.8% were female, 79.1% of the patients were anti-CCP positive, 51.4% had erosive disease. The mean DAS28-ESR was 2.03 (±0.63). 59.2% received methotrexate, 59.9% bDMARD and 11.7% CS. PDUSS was detected in ≥1 joint in 75 patients (72.1%) at baseline. 41/75 (54.7%) had persistence of at least one PDUSS during the follow-up. 26 (23.2%) had a relapse (after a mean duration of 9.1 (±2.6) months) or a RP at 1 year. In multivariate analysis, persistence of at least one PDUSS during the follow-up (HR=5.24 [1.74–22.5], p=0.009) and baseline number of tender joints (HR=1.32 [0.95–1.68], p=0.052) were predictors of relapse or RP at 1 year. Duration of remission, other baseline US findings including baseline PDUSS, autoantibodies, and erosive disease had no additional predictive value.

Conclusions Persistence of a PDUSS during the follow-up, rather than baseline PDUSS, predicts unfavorable outcome at 1 year in RA patients in clinical remission. This suggests that initial US findings are not sufficient to justify therapeutic change, but that the persistence of a residual PDUSS requires careful follow-up, and might even potentially merit strategy adaptation.

Disclosure of Interest None declared