BackgroundIntra-articular corticosteroid injections (IACI) are effective treatments for pain in knee osteoarthritis (OA), but responses to treatment vary. Uncertainty remains as to whether structural factors including accurate intra-articular injection placement influence outcome.ObjectivesTo determine whether structural factors including sonographically determined accuracy of IACI placement influence pain reduction in knee OA.MethodsIn a pragmatic prospective observational study, 141 subjects with knee osteoarthritis (mean age 63.8 years, 62% female) routinely referred for IACI underwent assessment of personal and demographic factors, x-rays (AP, lateral and sky-line patellofemoral) and ultrasound (US) before aspiration and injection of the joint, guided by clinical examination. US demonstration of an air-arthrosonogram was used to determine whether injections entered the joint cavity. Pain severity at baseline and at three and nine weeks post injection was assessed using the 500mm WOMAC pain subscale. Mean baseline pain was 271 (SD 96.7).Response to injection was defined as 40% reduction in pain from baseline. Characteristics of responders and non responders and mean change in pain for “accurate” intra-articular vs “inaccurate” extra-articular injections were compared using univariate statistics with the aim of creating logistic regression models. Inter-observer reliability of the air-arthrosonogram assessment was calculated as mean kappa of 0.79 for agreement between three raters.ResultsEighty-three subjects (53%) were classified responders at three weeks and fifty-six (44%)at nine weeks. Ninety-eight subjects (70%) had a positive arthrosonogram.Accurate injection neither resulted in a higher rate of response to treatment than inaccurate injection (57.7% vs 63.4% p0.355 at three weeks,39.3% vs 51.4% p 0.148 at nine weeks) nor greater mean pain reduction (-110.7 vs -116.9 mm VAS, p 0.781 at three weeks, and -65.2 vs -92.8, p 0.247 at nine). There were no differences in mean measurements of sonographic effusion and synovial hypertrophy, or scores for power Doppler signal and individual radiographic features of osteoarthritis (joint space narrowing and osteophytes) between responders and non responders at three or nine weeks. Those who had received previous injection were less likely to respond at nine weeks (p 0.026, OR 0.41, CI 0.20-0.86) and responders at nine weeks had lower mean baseline pain score compared to non-responders (250 vs 288 mm respectively, p 0.031)ConclusionsWe found no evidence that accurate injection of corticosteroid results in superior outcome to inaccurate injection in knee OA, nor that physical factors including sonographic features of synovial inflammation or radiographic severity influence outcome. These results raise potential questions about the routine use of US to enhance or predict response to IACI in knee OA.Disclosure of InterestNone declared