Background Ultrasound-detected synovitis improves the prediction of RA in early disease. Although ultrasound-detected tenosynovitis in common in early RA, its relative utility in the prediction of RA alongside ultrasound-detected synovitis and conventional serological and clinical variables is unclear.
Objectives To identify a minimal core set of ultrasound, clinical and serological variables predicting RA in a cohort of patients with early arthritis using a data-driven approach.
Methods 107 patients [female n=60, median age 51] with ≥1 joint with clinically apparent synovitis and a symptom duration ≤3 months underwent clinical, laboratory and ultrasound assessments. Final diagnoses were determined at 18 month follow-up. Ultrasound assessment determined the presence of synovitis at 28 joints (bilateral MCPs 1–5, PIPs 1–5, wrists, and MTPs 2–5) and tenosynovitis at four tendon sites (bilateral finger flexor and extensor carpi ulnaris tendon). First, principal component analysis (PCA) was performed on 1) clinical and serological variables (age, gender, symptom duration, ESR, CRP, RF, ACPA, duration of early morning stiffness, tender joint and swollen joint count), and 2) joint and tendon ultrasound variables. Then the variable with the highest loading factor from each component was extracted and made available in a forward step-wise multivariate logistic regression analysis.
Results 46 patients developed RA, 17 developed non-RA persistent inflammatory arthritis and 44 patients had resolving disease. Seven components were identified on clinical and serological PCA and four components were identified within ultrasound synovitis and tenosynovitis variables (Table 1). The final multivariate logistic regression model included ACPA positivity (OR=12.61, CI: 3.50–45.52, p<0.001), digit flexor ultrasound tenosynovitis (OR=4.58, CI: 1.67–12.56, p=0.003) and MCP 1 ultrasound synovitis positivity (OR=4.75, CI: 1.78–12.64, p=0.002).
Conclusions Both ultrasound-detected synovitis and tenosynovitis provide independent data in addition to clinical and serological variables in the prediction of RA in early disease. MCP 1 and digit flexor scanning provide the optimal predictive data in this cohort.
Disclosure of Interest None declared