OBJECTIVE: This study aimed to determine whether ultrasound measurements of the hands could predict treatment failure in treatment-naive women with early rheumatoid arthritis. METHOD: In a prospective case-control study, 48 women underwent blind assessments four times over 48-weeks, considering three failure stages: failure 1 (methotrexate), failure 2 (leflunomide), and failure 3 (adalimumab). Bilateral ultrasound exams evaluated wrist, 2nd, and 3rd Metacarpophalangeal Joints (MCPs), and Proximal Interphalangeal Joints (PIPs) for inflammatory indicators (synovial and tenosynovial proliferation using grayscale and Power Doppler [PD]) and joint damage (bone erosion and cartilage damage). RESULTS: The study involved 48 women, aged 47.7 ± 11.6 years, with an average disease duration of 7.5 ± 3.5 months. Of these, 41 (85.41 %) experienced failure 1, 25 (52 %) experienced failure 2, and 5 (10.5%) experienced failure 3. Predictors for failure 1 included PD/Q10 total score >
2.5 (OR = 18.00), PD/SQ10 total score >
5.0 (OR = 23.2), PD/Q MCP score >
1.5 (OR = 14.58), and PD/SQ MCP score >
3.0 (OR = 35). For failure 2, predictors encompassed PD/Q10 total score >
4.5 (OR = 4.81), PD/SQ10 total score >
9.5 (OR = 4.81), PD/Q MCP score >
2.5 (OR = 4.92), PD/SQ MCP score >
5.0 (OR = 6.22), and PD/Q PIP score >
1.5 (OR = 6.66). In relation to failure 3, a PD/Q wrist score >
2.5 (AUC = 0.79
p = 0.035) was indicative. CONCLUSIONS: Power Doppler proved to be a predictive indicator for treatment failure in early rheumatoid arthritis among treatment-naive women. It emerged as a predictor for both the initial and 2nd DMARD treatments, as well as the 1st immunobiological treatment, based on hand joint assessments. TRIAL REGISTRATION: Clinical trials.gov NCT04752748.