Authors' objectives: Intra-articular corticosteroid injections are an adjunct to core treatments for osteoarthritis. The National Institute for Health and Care Research Health Technology Assessment programme commissioned this research to address uncertainty around the long-term benefits and potential risks associated with recurrent intra-articular corticosteroid injections. Characterise current intra-articular corticosteroid injection practice. Establish longer-term effects and safety of single and recurrent intra-articular corticosteroid injections. Explore views and experiences of patients and clinicians. Assess the priorities/feasibility for future research. Osteoarthritis (OA) is the most common musculoskeletal condition worldwide and it is a global public health burden. It is an irreversible and progressive disease, associated with pain, morbidity, functional decline and loss in quality of life. Intra-articular corticosteroid injections (IACIs) are an adjunct to core treatments for osteoarthritic joint pain such as weight loss and exercise, and are recommended by the UK’s National Institute for Health and Care Excellence (NICE) for this purpose. Evidence for this recommendation was limited and based on data indicating a short-term benefit of repeated IACIs for pain relief in knee OA. The benefits and risks of single and recurrent use of IACIs in different joints affected by OA remain unknown. Furthermore, there is concern that IACI use could become more common in future for reasons such as the predicted increasing prevalence of OA, to bridge longer waiting times before undergoing joint replacement, or as an alternative treatment for people who do not wish to undergo surgery or in those where surgery is not an option due to significant comorbidities. The National Institute for Health and Care Research (NIHR) Health Technology Assessment (HTA) programme therefore commissioned this research to explore the use of IACIs in OA. To characterise the current practice of IACIs for OA. To establish the longer-term effects and safety of single and recurrent IACIs. To explore the views and experiences of patients and clinicians for IACIs. To identify priorities and associated feasibility considerations for future primary research.

Authors' results and conclusions: There were 23,899 (10.8%) osteoarthritis patients receiving intra-articular corticosteroid injections (40% received > 1 injection). Incidence of intra-articular corticosteroid injection at 5-year follow-up was lowest for elbow (5.2%) and highest for shoulder (13.6%). Incidences remained stable for all joints between 2005 and 2019 but varied between regions {3.8 [95% confidence interval 3.4 to 4.1] to 1.4 [95% confidence interval 1.3 to 1.5] injections per 100 patient-years}. Intra-articular corticosteroid injection for knee osteoarthritis was associated with lower incident use of several pain medications at 5-year follow-up; recurrent knee intra-articular corticosteroid injections were associated with greater risk reduction. In primary analyses intra-articular corticosteroid injection was associated with a lower 5-year cumulative incidence of knee replacement (number needed to treat 17, 95% confidence interval 12 to 40), but not hip replacement. In certain analyses, incidences of diabetes, ischaemic heart disease and myocardial infarction were numerically higher with intra-articular corticosteroid injections but 95% confidence interval spanned the null; there were no significant associations with other outcomes. Qualitative interviews demonstrated clinicians were more cautious about administering intra-articular corticosteroid injections compared to patients considering receiving intra-articular corticosteroid injections. Patients valued intra-articular corticosteroid injections; however, access was variable, with contributory factors including clinician concerns about the evidence base and their individual competence and confidence performing intra-articular corticosteroid injections. The Delphi research priority list included 14 questions covering long-term effects, clinical and cost-effectiveness, outcomes measurement, comparison to other treatments, provision, safety, identifying responders, maximising benefits, patient experience, delaying joint replacement and dosage. Overall, in the UK, intra-articular corticosteroid injection use is unchanged over 15 years; however, there was wide regional heterogeneity. We observed a sustained reduction in usage of certain pain medications after intra-articular corticosteroid injections. Intra-articular corticosteroid injections appear generally safe, although more research is needed on potential safety signals, particularly diabetes and cardiovascular events. Observational work Of the 220,125 OA patients identified (58.8% women; mean age of 65 years), 23,899 (10.8%) patients were injected on or after diagnosis. Patients with shoulder (15%) and knee (13%) OA had a higher prevalence of injection at any time after diagnosis. The least injected joints were the elbow and wrist (6% and 9%, respectively). Of those injected, 40% had more than one injection. The joints with most injections per patient were the knee and hand (mean of two injections per injected patient). The earliest injected joints were the shoulder (median time of 94 days after OA diagnosis) and knee (median 295 days). Incidence of first injection after OA diagnosis was higher in women. Incidence of first injection in any joint was higher in Yorkshire and The Humber, and East Midlands with incidences of 3.8 (95% confidence interval 3.4 to 4.1) and 3.4 (3.0 to 3.7) injections per 100 person-years. The regions with the lowest injection usage were South East with 1.2 (1.1–1.3) injections per 100 person-years, and South West with 1.4 (1.3–1.5). Incidences of injections were stable for all joints during the study period, except for a significant decline for all joints in 2020 during the COVID-19 pandemic. With regard to examining longer-term effects, the number of patients included in analyses of OA at each joint was up to 5180 for ankle, 16,775 for hand, 35,159 for hip, 83,188 for knee and 3007 for shoulder (elbow and wrist not analysed due to insufficient sample size). In the main IV analyses, single IACI for hand OA was associated with lower incident use of several pain medications at 1-year follow-up: oral non-steroidal anti-inflammatory drug [NNT = 14 (11–45)], oral glucocorticosteroids [NNT = 31 (27–135)], paracetamol [NNT = 10 (9–114)] and partial opioids [NNT = 5 (4–12)]. Incident or repeated use of most pain medications studied were similarly reduced over 5-year follow-up after a single IACI for knee OA. For patients with knee OA, generally larger reductions in incident pain medication use were observed after recurrent IACI use than were observed after single IACI use: for example repeat IACI relative risk (RR) = 0.39 (0.23–0.65) for initiating partial opioids during 5-year follow-up, while single IACI RR = 0.62 (0.37–0.87). IV analyses likewise found single IACI associated with lower 5-year cumulative incidence of knee replacement: NNT = 17 (12–40), but no association with hip replacement. In secondary analyses, PS matching suggested that recurrent use of IACI for hand, hip or knee OA was associated with lower rates of incident use of partial opioids at 5-year follow-up. However, rates of hip and knee replacement at 5-year follow-up were higher in the PS-matched IACI group. IACI was found to have no adverse impact on secondary outcomes after joint replacement. With regard to examining safety, 197,301 patients remained after exclusions (58.6% women; mean age 65.5 years). 8241 (9.35%) patients were injected. When stratifying by joint, there were no significant differences between crude incidence rates of adverse events following OA diagnosis (excluding those who received an IACI) compared to following an injection. After PS matching (1 : 2), we arrived at a sample size of 8327 injected patients matched to 16,654 non-injected patients. PS-matched analysis with Cox regression showed no statistically significant results for all outcomes after a single IACI: a hazard ratio (HR) of 0.97 (0.48–1.95) for mortality, 1.22 (0.26–5.78) for infection, 1.50 (0.63–3.60) for diabetes, 0.81 (0.32–2.10) for cerebrovascular disease, 6.21 (0.94–40.92) for MI and 2.49 (0.96–6.48) for IHD. Results were similar for repeated IACI use, as they were in Fine-Gray regression analyses controlling for the competing risk of death. Self-controlled cohort analyses pointed towards the null effect for all outcomes with a similar risk in 6 months after OA diagnosis as in 6 months after injection. Intra-articular corticosteroid injections are commonly used for treating patients with OA with around one out of nine patients diagnosed with OA receiving an IACI. Of those, 40% received more than one IACI. The most injected joints were the shoulder and knee, and median time from OA to first injection for those joints was ˂ 1 year. The use of injections has been unchanged over the past 15 years. However, there was wide heterogeneity in use of injections between UK regions, with qualitative work suggesting that this may relate to the skills, confidence and training of primary care healthcare professionals to perform IACIs. There was evidence that IACIs were associated with a sustained reduction in usage of several pain medications (up to 5 years post-injection), and a reduction in partial opioids at 1 year after single IACI for OA in any of the joint sites studied. For knee OA, larger reductions in incident pain medication use were observed after recurrent IACI use compared with single IACI use. A single IACI for knee OA may delay knee replacement with no adverse effect on postsurgical outcomes, although this finding requires confirmation in future studies. The safety of IACIs appears acceptable overall, with no increase in incidence of mortality, infection or bleeding at 6 months compared to after OA diagnosis. However, in certain analyses the incidences of diabetes, IHD and MI were numerically higher in the IACI group and approached borderline statistical significance, which require further investigation. This may be due to residual confounding and low number of outcomes, with these findings not present when using self-controlled methods. There was no dose–response observed in these analyses for single compared to recurrent IACIs. Both patients and clinicians valued the potential of IACIs to relieve OA symptoms and improve quality of life. Variability in patients’ access to treatment appeared to be related to clinicians’ confidence in delivering injections and their concerns about the evidence base. Variation in dose frequency and timing may reflect clinicians’ uncertainty about their familiarity with and the evidence base contained within current guidance. Despite variation in effectiveness, patients preferred IACIs to other forms of pain medication and to delay or avoid surgery. IACIs were used across the whole OA pathway, but most often as an adjunct treatment before surgery was offered. A robust consensus technique with key stakeholders identified 14 priority research questions to guide future research into IACIs for OA, covering long-term effects, clinical and cost-effectiveness, outcome measurement, comparison to other treatments, provision, safety, identifying responders, maximising benefits, patient experience, delaying the need for joint replacement, and dosage. Additionally, work exploring regional variation in IACI access would be beneficial to help ensure patients with OA have equal access to IACIs across the UK.

Authors' methods: A cohort study of incident osteoarthritis patients (2005–20) was performed using United Kingdom primary care data (Clinical Practice Research Datalink) linked to hospital data (Hospital Episode Statistics). Incidence of first intra-articular corticosteroid injection was stratified by age, calendar year, gender and geographical region. Longer-term outcomes included incident pain medication and joint replacement. Instrumental variables based on practice preference for intra-articular corticosteroid injection were used in primary analyses. Safety was assessed with propensity score matching and a self-controlled cohort, with outcomes (mortality, bleeding, hemarthrosis, wound infection, diabetes, stroke, ischaemic heart disease, myocardial infarction) assessed at 6 months. Semistructured telephone/videocall interviews were conducted (patients = 38, primary care clinicians = 19), with inductive thematic analysis used to investigate views and experiences of intra-articular corticosteroid injections. A three-round modified Delphi study with patients (n = 41), healthcare professionals (n = 25) and academics/researchers (n = 25) was performed to identify future primary research priorities and feasibility. Observational analyses: Possibility of residual confounding. Qualitative research: Self-selecting nature of participation; and while the funding remit focussed on primary care, qualitative findings suggest orthopaedic consultants are key stakeholders. Delphi study: Unable to engage commissioners. Observational work An observational cohort study was performed using UK primary care data (Clinical Practice Research Datalink GOLD) linked to hospital data [Hospital Episode Statistics (HES)]. Patients ≥ 20 years old with a first diagnosis of ankle/foot, elbow, hand, hip, knee, shoulder or wrist OA between 2005 and 2019 were eligible for inclusion. Those receiving IACIs were identified. To establish the current practice of IACI use, the incidence of first injection per 100 person-years and prevalence were estimated for each joint, and stratified by calendar year, gender and geographical region. Injections per OA patient and mean number of injections per injected patient’s joint were calculated. Median time from OA diagnosis to injection was calculated per each joint. To establish the longer-term effects of IACIs, we analysed patients who had not already been referred to orthopaedic surgery. Primary outcomes were incident pain medication prescriptions and joint replacement. Among those who received joint replacement, secondary outcomes included: complications [stroke, myocardial infarction (MI), thrombosis, joint infection], reoperation and patient-reported outcome measures. Instrumental variable (IV) analyses were conducted using two-step Poisson regression. Index date for the IV was 6 months after OA diagnosis date and IV was defined as practice preference for IACI use in the year prior to the index date (single and recurrent use considered separately, compared to non-use). Outcomes were measured at 1 and 5 years from the index date. For primary outcomes, propensity score (PS) matching was carried out in secondary analyses. To establish the safety of IACIs, primary outcomes assessed were mortality, bleeding, hemarthrosis, infection, diabetes, stroke, ischaemic heart disease (IHD) and MI. We calculated incidence of outcomes per 100 person-years during 6 months after the first IACI (in patients with OA who received an injection) and after OA diagnosis (in patients with OA who did not receive an injection). Follow-up was censored at subsequent OA diagnosis, IACI, joint replacement or death. Incidence was further stratified by joint. To avoid duplicate events we excluded patients with previous diabetes, stroke, IHD and MI for those events. To minimise confounding we applied PS matching 2 : 1 by several confounders. We also performed a self-controlled cohort analysis using only IACI patients and comparing 6 months after OA versus 6 months after first injection for the same patient, adjusting by age.

Project Status: Completed

URL for project: https://www.journalslibrary.nihr.ac.uk/programmes/hta/NIHR129011

Year Published: 2025

URL for published report: https://www.journalslibrary.nihr.ac.uk/hta/LFAJ9337

URL for additional information: English

English language abstract: An English language summary is available

Publication Type: Full HTA

Country: England, United Kingdom

DOI: 10.3310/LFAJ9337

Organisation Name: NIHR Health Technology Assessment programme

Contact Address: NIHR Journals Library, National Institute for Health and Care Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK

Contact Name: journals.library@nihr.ac.uk

Contact Email: journals.library@nihr.ac.uk