Evidence suggests that intraarticular corticosteroids are effective (short term) and safe (long term)

In a meta-analysis recently published in Osteoarthritis and Cartilage, Donovan et al.¹ stated that ‘recurrent IACS (intraarticular corticosteroids) often provide inferior (or non-superior) longer-term symptom relief compared with other injectables’. However, after carrying out meta-analyses of treatments for osteoarthritis (OA) or reviewing comprehensive ones, many guidelines committees have recommended IACS, at least conditionally, for the treatment of OA. What is the evidence supporting the efficacy and safety of intraarticular corticosteroids, and how can we make sense of these conflicting findings/recommendations?

In their commentary, Richette and Latourte² raise concerns about examining long-term outcomes for treatments that have only short-term effects. We support this view. Our goals in this commentary are three-fold:

1. To expand upon the concerns of Richette and Latourte, with specifics from the Donovan et al. meta-analysis.

2. To revisit a trial that provided the best data on whether repeated steroid injections cause cartilage loss.

3. To review recently published large-scale observational data on the long-term outcomes of IACS injections.

In their meta-analysis, Donovan et al. reported that, at 6 and 9 months, groups receiving repeated steroid injections had worse pain scores than comparators. In the two largest trials contributing most to this conclusion^3,4, patients received steroid injections at baseline and no more than 4 weeks later, with pain assessed at 6 months and beyond. The effects of steroids on pain are transient and would not be expected to last this long⁵. Donovan et al. tested for an effect of IACS that they did not provide.

Incidentally, there are strategies that can prolong the efficacy of steroids. Among these are the use of a higher than usual dose of steroids, such as 80 mg of injected methylprednisolone⁶, and the use of extended-release triamcinolone⁷.

There is a concern that IACS injections may cause cartilage loss. In the most widely cited and definitive trial examining repeated steroid injections, McAlindon et al.⁸ reported not only that 40 mg of injected triamcinolone every 3 months did not improve symptoms compared with placebo injections, but also that after 2 years of this treatment, steroid-treated patients had lost more cartilage than the placebo group. However, this dose would not necessarily be expected to reduce pain 3 months after each injection. In terms of cartilage thickness, the difference between the steroid and placebo groups at 2 years was 0.11 mm, representing an annual loss caused by steroids of 0.055 mm. Given that tibiofemoral cartilage is 3–5 mm thick (assume 4 mm), it would take 10 years of injections every 3 months to thin cartilage by 12.5%. Such tiny effects are likely not clinically important.

Ultimately, long-term observational studies with follow-up sufficient to evaluate disease progression and rates of knee replacement are the optimal approach to address the long-term consequences of IACS injections. Combining data from the MOST and OAI studies, Bucci et al. reported that patients undergoing steroid injections had no greater risk of OA progression nor of knee replacement than those who received hyaluronic acid (HA) injections⁹. Latourte et al. went a step further. They showed not only no increase in progression or knee replacement compared with HA–treated patients, but also no increased risk of these outcomes compared with OA patients not treated with any injections¹⁰.

In conclusion, the efficacy of IACS injections is substantial but transitory⁵. Trials have not shown severe cartilage damage in people receiving injections, and large-scale observational studies have provided evidence that long-term consequences of corticosteroid injections, including repeated ones, are negligible.