POS0612 HOW DO WE PREDICT TREATMENT RESPONSE IN OSTEOARTHRITIS ANALYSIS OF A COHORT STUDY IN PEOPLE WITH KNEE OSTEOARTHRITIS

Background: Osteoarthritis (OA) is the most common arthritis worldwide, yet no disease-modifying treatments are available. The European Alliance of Associations for Rheumatology (EULAR) advise combined pharmacological and non-pharmacological (exercise and physiotherapy) management. EULAR advise paracetamol first line with non-steroidal anti-inflammatory drugs (NSAIDs) considered if unresponsive. Opioids are commonly prescribed in clinical practice if NSAIDs are ineffective but are not recommended by EULAR [1]. Meta-analysis shows that opioids do not improve pain or function compared with placebo with more adverse effects [2]. Lack of opioid efficacy raises concerns given the opioid abuse prevalence in chronic pain patients [3]. Intra-articular corticosteroid injections (IACI) are recommended in acute pain flare-ups with joint replacement considered in advanced disease. Although several risk factors for OA pain and functional impairment are well recognised, little information is available to assess factors influencing treatment response. Objectives: We aimed to evaluate factors influencing response to knee OA medical management in a real-world UK cohort study (ARPAIN IRAS number: 234659) to: 1.Identify pain characteristics. 2.Assess risk factors influencing response. 3.Evaluate the efficacy of pharmacological treatments. 4.Develop a stratification tool to aid treatment decisions for patients and clinicians. Methods: We recruited participants with knee OA in this multi-centre prospective trial who met American College of Rheumatology (ACR) eligibility criteria for painful symptomatic knee OA on usual National Health Service (NHS) care. During baseline assessment, we collected demographics (age, sex, ethnicity, BMI, analgesia history), Visual Analogue Scale (VAS), painDETECT and Western Ontario and McMasters Universities Arthritis Index (WOMAC). A painDETECT score ≥19 categorised participants as "sensitised" to pain. Appropriate participants underwent clinical or ultrasound-guided IACI in addition to usual care (paracetamol, NSAIDs or opioids). The presence/absence of synovitis/effusion was noted. Knee radiographs were assessed by two Musculoskeletal Radiology Consultants and Kellgren-Lawrence graded. At 3 months VAS, PainDETECT and WOMAC pain questionnaires were repeated. We defined treatment response as at least a 20% improved baseline WOMAC pain score. Univariate and multivariate logistic regression was conducted using R version 4.2. Clinical experience and statistical theory aided variable selection for the global model. We utilised the average area under the receiver operating characteristic (AUROC) across leave-one-out cross-validation (LOOCV) to determine our model's predictive accuracy. Results: We recruited 92 participants with knee OA, with 28(42%) meeting our response criteria. 88(96%) received IACI. The variables included in our final multivariate model were age and baseline WOMAC pain score (LOOCV AUROC 0.72) (Figure 1). Age (per 10-year increase) significantly predicted response in univariate (OR 2.06, p=0.01) and multivariate analysis (OR 1.91, p=0.02). Increased baseline WOMAC pain score (median 53.3, range 13.8-94.8) significantly predicted negative response in univariate (OR 0.68, p=0.01) and multivariate analysis (OR 0.71, p=0.02). Pain sensitisation predicted a negative response in univariate analysis (OR 0.23, p=0.03). Baseline opioid prescription was a significant negative response predictor in univariate analysis (OR 0.29, p=0.05). Figure 2 shows changes in WOMAC pain score over 3 months: response rates are displayed for the whole study group on usual care including paracetamol, NSAIDs or opioids with 96% receiving IACI (Figure 2a) and in a subgroup receiving opioids at baseline (Figure 2b). Figure 1 Figure 2 Conclusion: Our study highlights that the parameters selected in our model can predict knee OA treatment response. We demonstrated improved response in older participants, without pain sensitisation with lower baseline WOMAC pain scores, suggesting lower response with pain catastrophisation. Participants taking baseline paracetamol/NSAIDs demonstrated improved response compared with those taking additional opioids. Our finding of baseline opioid use being a negative response predictor is consistent with opioids' lack of efficacy in knee OA in the literature. Validation of our predictive model could allow for more personalised OA management with improved resource allocation. REFERENCES: [1] A. Pendleton et al, Ann Rheum Dis, vol. 59, no. 12, pp. 936-44, Dec 2000, doi: 10.1136/ard.59.12.936. [2] M. C. Osani, L. S. Lohmander, and R. R. Bannuru, Arthritis Care Res (Hoboken), vol. 73, no. 10, pp. 1413-1424, Oct 2021, doi: 10.1002/acr.24363. [3] L. A. Deveza, D. J. Hunter, and W. E. Van Spil, Osteoarthritis Cartilage, vol. 26, no. 3, pp. 293-295, Mar 2018, doi: 10.1016/j.joca.2017.12.003. Acknowledgements: We acknowledge funding from The Rosetrees' Trust (Grant number M11-F3). Disclosure of Interests: None declared . © The Authors 2025. This abstract is an open access article published in Annals of Rheumatic Diseases under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Neither EULAR nor the publisher make any representation as to the accuracy of the content. The authors are solely responsible for the content in their abstract including accuracy of the facts, statements, results, conclusion, citing resources etc.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.