Abstract
Abstract Background In anti-cyclic citrullinated peptide antibody-positive (ACPA+) individuals without clinical synovitis (at-risk), to define the critical ultrasound (US) features sufficiently predictive for inflammatory arthritis (IA) to enable logical initiation of therapy. Methods In a single centre prospective cohort, at risk ACPA+ individuals with a new musculoskeletal symptoms underwent an US scan of 38 joints and 18 tendons at first visit. The predictive value of US abnormalities (Power Doppler (PD), Grey Scale (GS), erosion or tenosynovitis (TSV)) for progression to IA was analysed and the best predictive joints determined by Multivariable Cox Regression, adjusted for confounders. The US results were combined with clinical symptoms/findings to produce predictive models. Results Consecutive at-risk ACPA+ individuals (n = 457, mean age 50.3 years old, 74.2% women) were followed up for median of 15.4 months (range 0.1-127.4), a complete dataset with follow-up of at least 6 months was available for 319 of them. 135 (29.5%) developed IA after a median of 11.3 months (range 0.1-111.7). The negative predictive value of a US scan without any abnormality was 82%. In multivariable Cox regression, both PD and TSV were predictive of progression, with respectively hazard ratios of 1.2 (9=0.026) and 1.13 (p = 0.025). All US abnormalities had a high specificity (spec) but only moderate sensitivity (sens), PD was the most specific with a spec/sens of 0.94/0.23, followed by TSV with a spec/sens of 0.91/0.26 but the best area under the curve (AUC) of 0.599 (P = 0.0015). The addition ACPA titre (high compared to low), but not GS, improved spec/sens up to 0.92/0.34 and AUC to 0.964 (p < 0.001). A selection of US and clinical data of 14 joints also improved prediction, with an AUC of 0.670 (p < 0.001) and a spec/sens of 0.65/0.62. A selection of the 34 most predictive features reached the same sens/spec as the ACR/EULAR 2010 classification criteria for RA, showing a spec/sens of 0.80/0.56. Conclusion In at-risk ACPA+ individuals, the presence of sub-clinical US abnormalities are highly specific for progression to IA. The only moderate sensitivity can be improved by using joints or features selection in combination with clinical examination. These results are the first step in providing guidance for which at-risk ACPA+ individuals to treat. Disclosures L.M. Duquenne None. K. Mankia None. L. Garcia Montoya None. A. Di Matteo None. J. Nam None. P. Emery None.
Published Version (Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have