MP06: Use of ultrasound and x-ray to predict improvement in hip osteoarthritis symptoms following intra-articular steroid injection

Abstract

Introduction: Intra-articular steroid injection (IASI) is commonly used in the emergency department for management of osteoarthritis (OA) symptoms. Hip IASI carries risks, such as avascular necrosis, and there is currently no reliable way to predict long-term response of a patient’s OA to IASI. Ultrasound (US) conveniently assesses for active arthropathy by detecting effusion-synovitis, and x-ray (XR) is useful for visualizing bone-related changes. We investigated the extent that a response to hip IASI could be predicted from baseline OA patient clinical and physical features alongside US and XR imaging features. Methods: 97 consenting patients with symptomatic hip OA presenting for hip IASI were evaluated at baseline (XR and US) and again 8-weeks after IASI (US only). Self-reported pain (WOMAC), hip range of motion (ROM) were measured at baseline and follow up. On US images we quantified joint effusion and synovial thickening, i.e., “effusion-synovitis”, by the bone-capsule distance (BCD) at the apex of the femoral head from outer femoral cortex to outer synovium. On XR, we measured minimum joint space width (cm) and Kellgren-Lawrence (K-L) Grade for osteophytes and sclerotic changes. Results: In our 97 patients (43 female) aged 28-87 years (mean 59+/-13 years, K-L grades averaged 2.5+/-1.5, and US BCD averaged 5.9+/-2.0 mm. We performed multiple linear regression using age, sex, BMI, ROM of hip flexion, US BCD, radiographic joint space width and K-L grade against the dependent variable, change in WOMAC pain subscore (R=0.587, P=0.002). We compared the response predicted by this model to the actual change in WOMAC pain. At a threshold value of -20% for minimal clinically important difference, 35/97 patients were responders, and a 2x2 table gave 67% overall model predictive accuracy, 61% sensitivity, and 71% specificity. Likelihood ratio for a positive response (LR+) was 2.13. Conclusion: Combining radiographic information on structural damage, US information on active arthropathy, and demographics correctly predicted about two-thirds of the patients that would benefit from IASI after 8 weeks. A patient with hip OA that met our model criteria was more than twice as likely to respond to IASI. With further model refinement, effective, personalized evidence-based management of symptomatic hip OA is possible using XR and hip US, which could both be performed during an ER visit.