Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Steroid injections for OA knee


Many older people have osteoarthritis (OA) of the knee that gives them pain and discomfort. Some of those consult a primary care physician, and some a specialist. As well as oral and topical analgesics, injection of corticosteroid into the knee is often carried out with the aim of producing pain relief. Two systematic reviews [1,2] confirm the efficacy of this intervention, at least over about a month.

The aim is to inject the steroid into the joint space, but not all injections are correctly placed, though it is not clear whether or how much this matters. There are issues, about which corticosteroid to use, which dose of which corticosteroid to use, and whether a long acting is better than a short acting corticosteroid. Steroid injections are often accompanied by use of local anaesthetics. Again, benefits of using local anaesthetic with steroid are argued, dose is an issue, and even whether local anaesthetic can be used alone. Then there is the issue of how long the injections are effective, and how often they can be repeated. So while we have two systematic reviews, we would be fortunate to find all our questions answered.


The reviews were similar. The first [1] sought randomised trials where intra-articular long acting corticosteroids (triamcinolone, methylprednisolone, betamethasone, cortivazol) were compared with placebo. Searching was up to December 2002. The second [2] sought randomised trials where any formulation of steroid was compared with placebo, with searching into 2003. Both studies included patients with osteoarthritis of the knee, and required pain outcomes.


The second review [2], because it included any formulation of corticosteroid, and ended searching later, had more studies (10) than the former [1], which had five. Two of the additional five studies used long acting corticosteroids, and were published in 2003. All of the trials except one compared corticosteroid to saline injection, the other using a sham injection. Several of the studies had a crossover design. Dose of corticosteroid varied widely, between 6 mg and 80 mg prednisone equivalent, though most were between 25 and 50 mg prednisone equivalent.

Figure 1 shows the six studies with outcomes of improvement up to two weeks. This was not a clearly defined term in many of the studies. In these six studies with 317 patients, five used long acting corticosteroids. Improvement up to two weeks occurred in 74% of patients with a steroid injection and 45% given placebo. The relative benefit was 1.7 (95% CI 1.4 to 2.0), and the number needed to treat for one patient to have improvement was 3.4 (2.5 to 5.1). The weighted mean reduction in visual analogue pain scores was 17 mm on a 100 mm scale.

Figure 1: Improvement after steroid injection up to two weeks (dark symbol short-acting steroid)

Three studies with 192 patients had results at 16-24 weeks after the injection (Figure 2). Two used long acting corticosteroid, and the other used hydrocortisone (and was an older study with a low quality score). Overall, 33% of steroid treated patients had improvement at 16-24 weeks compared, and 16% of those given placebo. The relative benefit was 2.1 (1.2 to 3.5), and the NNT 5.8 (3.4 to 19).

Figure 2: Improvement after steroid injection up to 16-24 weeks (dark symbol short-acting steroid)

Adverse consequences of the intra-articular injections were not reported.


What we have here is some very limited data with implications for clinical practice. It is likely that intra-articular corticosteroids produce some pain relief, perhaps for some weeks. There are many problems, though:

Clinical practice and experience suggests that intra-articular steroid injections are helpful for painful knees in osteoarthritis. The trouble is that half of the patients improved with saline alone, and the additional benefits of adding steroid were moderate. Some will say that this is the power of psychiatry with needles, but it is equally possible that the improvement would have come about anyway, because of the ups and downs of symptoms. In many ways the situation resembles that of many alternative therapies, though those usually have less evidence to support them.


  1. M Godwin, M Dawes. Intra-articular steroid injections for painful knees: systematic review with meta-analysis. Canadian Family Physician 2004 50: 241-248.
  2. B Arroll, F Goodyear-Smith. Corticosteroid injections for osteoarthritis of the knee: meta-analysis. British Medical Journal 2004 328: 869-870.

previous or next story