Patterns of intra-articular injection use after initiation of treatment in patients with knee osteoarthritis: data from the osteoarthritis initiative

Abstract

Purpose: Despite that the efficacy is lacking and recommendations from guidelines are conflicting, intra-articular injections such as hyaluronic acid and corticosteroid injections are increasingly used in patients with osteoarthritis (OA). Studies on patterns of injection use and factors associated with these patterns are scarce. The aim of this study was to describe and evaluate longitudinal use of intra-articular injections after treatment initiation among adults with radiographically confirmed knee OA. Methods: Using publicly available data from the Osteoarthritis Initiative, we used 9 years of data and included participants with at least one knee with radiographically confirmed OA (Kellgren-Lawrence grade (K-L) ≥2) at baseline. We identified 412 participants had newly initiated hyaluronic acid or corticosteroid injections. Among those, 96 initiated hyaluronic acid and 316 initiated corticosteroid injections. After treatment initiation, 79 participants switched the injection type, 90 participants continued the treatment use, and 243 participants did not receive any injections use during the follow-up. Switching and continued users were then matched to 243 one-time users by the frequency of distribution of follow-up times between the first initiation and switching/continued. According to types of injections initiated, sociodemographic and clinical characteristics were calculated by patterns of treatments use at the index visit. Average yearly changes in symptoms and clinical assessments between treatment initiation and switching/continued treatment use were also examined. Multinomial logistic regression models were used to identify factors associated with patterns of injection use including sociodemographics and clinical/functional factors. We estimated adjusted odds ratios (aOR) and 95% confidence intervals (CI) for each group compared with a common reference group (one-time users) after adjusting for sociodemographics and clinical/functional factors. Results: Approximately 4 in 5 of those initiating injections users switched injection type or did not receive any injections after their index injection use. Among hyaluronic acid initiators, 24.0% switched to corticosteroid injections. Among corticosteroid initiators, 17.8% switched to hyaluronate injections. One-time users were common in both types of index injections (60.4%: hyaluronate; 58.5%: corticosteroid). Information to identify factors associated with patterns of injection use among participants initiating hyaluronic acid use were limited due to the small sample size. Among participants initiating corticosteroid injection use, switching and continued injection use was less likely to be associated with increasing age (aOR switching: 0.97 (95% CI: 0.93-0.1.02); aOR continued: 0.96 (95% CI: 0.92-0.99)). Greater comorbidity status (grade 1 versus 0) was associated with increased odds of switching injections use (aOR: 3.35; 95% CI: 1.30 to 8.60), continued users (aOR: 2.16; 95% CI: 0.94 to 4.97). Symptoms such as WOMAC pain and physical function at index visit was correlated with both switching and continued injections use: pain (aOR switching: 0.95 (95% CI: 0.86 to 1.06); aOR continued: 0.89 (95% CI: 0.81 to 0.97)); physical function (aOR switching: 0.96 (95% CI: 0.93 to 1.00); aOR continued: 0.96 (95% CI: 0.93 to 1.01)). SF-12 PCS was associated with decreased odds of both switching injections use (aOR: 0.97; 95% CI: 0.92 to 1.02) and continued users (aOR: 0.99; 95% CI: 0.94 to 1.05) compared to one-time users. Conclusions: Among individuals with radiographically confirmed OA initiating corticosteroid injection use, report of switching or continued treatment use is associated with indicators of greater comorbidity status and functional status. However, the proportion of switching injection use and one-time users after treatment initiation was substantial in both types of commonly used treatment agents. This may indicate that longer-term efficacy regarding symptom relief and/or slowing disease progression of these agents remains in question among patients with OA in the real-world setting.