Intra-Articular Corticosteroid Injections for Symptomatic Knee Osteoarthritis

What the Orthopaedic Provider Needs to Know

Cody L. Martin, MD; James A. Browne, MD


J Am Acad Orthop Surg. 2019;27(17):e758-e766. 

In This Article


Corticosteroid injections into the knee for management of osteoarthritis have been shown to be statistically and clinically significant at reducing pain in the short term.[14,37–39] However, the exact duration of pain relief remains a controversy. Hepper et al[14] performed a systematic review published in 2009 of level I studies and found that corticosteroid injections for management of pain related to osteoarthritis of knee showed statistically and clinically significant pain relief at 1 week after injection. There seemed to be some benefit in the short term after 1 week, but the data were not statistically significant compared with placebo. A Cochrane Review update in 2015 supported this result by concluding that the clinical benefits of intra-articular steroid knee injections remained unclear at 1 to 6 weeks after injection because of low quality of evidence.[37] No evidence of clinical benefit was found 6 months after patients received an injection. Studies on this topic are variable, and a lack of high-quality, high-powered, placebo-controlled prospective randomized trials exist; the 2015 Cochrane Review graded the quality of evidence as "low," which was defined as having little confidence in the results due to the discordant studies based on small studies.[37]

Individual patient variables and severity of arthritis may affect the efficacy of injections. Recently, Matzkin et al[38] studied a cohort of 100 patients who received a steroid injection for management of symptomatic knee osteoarthritis, and they found improved WOMAC (P < 0.001) and Visual Numeric Scale (P < 0.05) scores at 3, 6, 12, and 24 weeks after injection, except the Visual Numeric Scale score at 24 weeks (P = 0.114), compared with scores at baseline. However, obese patients had statistically significant worse WOMAC scores compared with nonobese patients at baseline (P = 0.003), 6 weeks after injection (P = 0.010), and 3 months after injection (P = 0.009), although they showed improvement from baseline. Those with worse arthritis (Kellgren-Lawrence grade 3 or 4) had statistically significant poorer scores at 6 weeks after injection (P = 0.028) and 3 months after injection (P = 0.004), although they showed improvement from baseline. Maricar et al[39] followed up a cohort of 207 patients and characterized structural predictors of clinical response to intra-articular steroid injections in symptomatic osteoarthritis of the knee. They found that 73.4% of patients responded to treatment in the short term of 2 weeks and 20.1% in the long term at 6 months after injection. Structural predictors of response were not found in the short term, but they were in the long term. Worsening disease as seen on radiographs including increased overall joint space narrowing (P = 0.047) and overall Kellgren-Lawrence score (P = 0.010) were associated with lack of response long term.[39] Maricar et al[40] also performed a systematic review of predictors of response to steroid injections in osteoarthritic knees and found increased likelihood of response in patients who had an effusion, absence of synovitis, worse baseline symptoms, and less severe disease on radiographs and in those who received an injection with ultrasonography guidance.

As mentioned previously, one concern with repeat injections is chondrotoxicity. Mixed results exist in the literature on whether intra-articular steroid injections may affect the natural history of knee osteoarthritis by accelerating cartilage loss. In 2017, McAlindon et al[35] published a randomized placebo-controlled double-blind trial of 140 patients who received either a steroid injection or saline injection every 3 months for 2 years and showed increased knee cartilage volume loss evaluated on MRI and equivocal pain scores at the 2-year follow-up period in patients who received steroid injections compared with those who received placebo. The cartilage loss measured on MRI in this study has unknown clinical consequences, particularly in patients with advanced preexisting chondral damage. The estimated cartilage loss corresponding to the used MRI index was 0.46 mm, and the value for a minimally clinically important difference has not yet been established.[35]

The study by McAlindon et al can be contrasted by a randomized, double-blind placebo-controlled trial by Raynauld et al[41] in which osteoarthritic knees were injected with steroids every 3 months for 2 years and joint space narrowing was measured on radiographs along with patient outcomes. At 1-year and 2-year follow-up periods, no significant difference was found in joint space loss over time as seen on radiographs. At the 2-year follow-up period, no significant difference was found in pain relief, but by using the area under the curve analysis to describe the changes in outcome scores throughout the time course of the study, statistically significantly improved pain (P = 0.0047) and improved stiffness (P = 0.0511) scores were seen among those who received steroid injections.[41]

Intra-articular injections are rarely effective as a long-term strategy in the management of symptomatic osteoarthritis of the knee. Subsequent injections are often seen to have less associated pain relief or decreased duration of efficacy. Hirsch et al[27] found that patients who had previous intra-articular steroid injections were less likely to respond at 9 weeks (P = 0.021) but not at 3 weeks (P = 0.314) after injection than patients who underwent their initial first-time injection. The study by Raynauld et al also showed this as the 2-year follow-up pain scores were not statistically different between those who received placebo versus steroid; however, if the area under the curve was analyzed, the overall pain scores throughout the 2 years were statistically lower in those who received a steroid injection (P = 0.0047).

The mind-body connection and the placebo affect may also play an important role in the management of chronic painful conditions.[5] A meta-analysis of oral and intra-articular treatments for knee osteoarthritis analyzed 137 studies with 33,243 patients found that intra-articular placebo had a statistically significantly better pain relief than oral placebo (effect size, 0.29 [95% confidence interval, 0.04 to 0.54]) and that none of the oral nonsteroidal anti-inflammatory drugs were significantly superior to intra-articular placebo.[42]