Injections and surgery for knee and hip osteoarthritis: what the trials found

When knee or hip osteoarthritis stops responding to exercise and simple pain relievers, the next questions are usually about injections and procedures. The marketing around some of these options runs well ahead of the evidence. Here is what randomized trials, Cochrane reviews, and the major osteoarthritis guidelines report for each one.

Corticosteroid injections: real but short relief, with a cartilage question

Intra-articular corticosteroid (steroid) injections are the best-supported injection for short-term flares. A 2015 Cochrane review pooled 27 trials with 1,767 participants and found a benefit on pain and function roughly one to six weeks after injection. The reviewers cautioned that trial quality was low and that it was unclear whether any clinically meaningful benefit lasts beyond six weeks, with no evidence of an effect at six months (Juni et al., Cochrane, 2015). In plain terms, the relief is genuine but brief.

A widely cited randomized trial complicated the picture. McAlindon and colleagues gave 140 people with knee osteoarthritis either triamcinolone or saline every three months for two years. The steroid group lost more cartilage than the saline group (mean cartilage thickness change of -0.21 mm vs -0.10 mm) and showed no advantage in pain (McAlindon et al., JAMA, 2017). The cartilage difference was not linked to worse symptoms during the study, so the long-term meaning is uncertain, but the result raised a caution about repeated, scheduled injections. Guidelines still support steroid injections: the 2019 American College of Rheumatology (ACR) guideline strongly recommends them for knee and hip OA, and OARSI conditionally recommends them for short-term pain relief (Bannuru et al., OARSI, 2019).

Hyaluronic acid: mixed evidence and split guidelines

Hyaluronic acid injection, also called viscosupplementation, is where reputable guidelines openly disagree, which tells you the underlying evidence is mixed. OARSI 2019 conditionally recommends it for the knee, citing symptom improvement that may extend beyond 12 weeks and a favorable safety profile. The ACR 2019 guideline reaches a different conclusion: it conditionally recommends against hyaluronic acid for the knee and strongly recommends against it for the hip, judging the benefit over placebo to be small, inconsistent, and prone to bias (ACR press release, 2019; Bannuru et al., OARSI, 2019). Neither group is hiding data from the other. They weighed the same noisy trials and landed in different places, which is a signal to treat any single strong claim about hyaluronic acid with skepticism.

Platelet-rich plasma: popular, but the best trial found nothing

Platelet-rich plasma (PRP) is marketed heavily, yet the evidence is low-certainty and highly heterogeneous, partly because preparations are not standardized from clinic to clinic. The most rigorous test to date, the RESTORE trial, randomized people with mild to moderate knee osteoarthritis to PRP or saline placebo injections. At 12 months, PRP did not improve pain and did not slow loss of medial tibial cartilage volume compared with placebo (Bennell et al., JAMA, 2021). Guidelines reflect that uncertainty. The ACR recommends against PRP for knee and hip OA, and OARSI strongly recommends against it, noting the very low quality of evidence and the lack of a standardized product (ACR press release, 2019; Bannuru et al., OARSI, 2019). PRP may yet find a defined role, but the current high-quality data do not support it.

Arthroscopic surgery: no better than a sham operation

Arthroscopic debridement and lavage (cleaning out and rinsing the joint) was once common for knee osteoarthritis. Two landmark trials undercut it. Moseley and colleagues randomized 180 patients to arthroscopic debridement, lavage, or a placebo (sham) operation, and outcomes after the real surgery were no better than after the fake one (Moseley et al., NEJM, 2002). Kirkley and colleagues then found that arthroscopy added no benefit beyond optimized physical and medical therapy (Kirkley et al., NEJM, 2008). A separate sham-controlled trial, FIDELITY, tested arthroscopic partial meniscectomy for degenerative meniscal tears (without established OA) and again found no benefit over sham surgery (Sihvonen et al., NEJM, 2013). Guidelines now advise against arthroscopic debridement or lavage for osteoarthritis. This does not mean arthroscopy is useless for every knee problem, only that it does not help degenerative osteoarthritis.

Total joint replacement: effective for end-stage disease

For advanced, disabling disease, total joint replacement is the intervention with the strongest outcome data. In a randomized trial, Skou and colleagues compared total knee replacement plus nonsurgical care against nonsurgical care alone in people with moderate-to-severe knee OA. The surgery group improved more on the KOOS4 score (32.5 vs 16.0 points), a clear and clinically meaningful gap. The tradeoff was real: the surgery group had more serious adverse events (24 vs 6), including infections and blood clots (Skou et al., NEJM, 2015). Notably, 26% (13 of 50) of the nonsurgical group had crossed over to knee replacement within 12 months, a sign of how many did not get enough relief from nonsurgical care alone. That is why replacement is generally reserved for end-stage disease after conservative options have been tried, rather than offered early.

Where the guidelines line up

Across OARSI and the ACR, a few points are consistent. Exercise and weight management are first-line for everyone. Steroid injections are accepted for short-term relief. Arthroscopy for osteoarthritis is discouraged. PRP and stem cell injections are advised against. Joint replacement is validated for end-stage disease. The clearest disagreement is on hyaluronic acid, and even there both bodies describe the benefit as small or uncertain.

This article is evidence reporting, not medical advice. Treatment decisions depend on your specific joint, disease stage, and health history, so discuss any injection or procedure with a qualified clinician.

Common questions

How long does a cortisone shot last for knee arthritis?

Trial evidence points to short-term relief. A Cochrane review of 27 trials found benefit mainly at one to six weeks after injection, with no clear evidence of effect by six months (Juni et al., Cochrane, 2015). Individual experiences vary, and this is a description of averages from studies, not a prediction for any one person.

Do cortisone injections damage cartilage?

One randomized trial found that triamcinolone injections given every three months for two years led to more cartilage thinning than saline (-0.21 mm vs -0.10 mm), without better pain control (McAlindon et al., JAMA, 2017). The long-term clinical meaning is uncertain, but the finding raised caution about repeated scheduled injections. This is a reported research finding, not treatment guidance.

Does platelet-rich plasma (PRP) work for knee osteoarthritis?

The highest-quality trial, RESTORE, found PRP no better than saline placebo for pain or cartilage loss at 12 months (Bennell et al., JAMA, 2021). Overall evidence is low-certainty and inconsistent, and both the ACR and OARSI guidelines recommend against PRP for knee and hip OA.

Is arthroscopic knee surgery helpful for arthritis?

Randomized trials found arthroscopic debridement and lavage no better than sham surgery or supervised physical therapy for knee osteoarthritis (Moseley et al., NEJM, 2002; Kirkley et al., NEJM, 2008). Guidelines advise against it for osteoarthritis, though arthroscopy can still be appropriate for some non-arthritic knee problems.

When is a knee or hip replacement considered?

Joint replacement is generally reserved for end-stage disease with significant pain and disability after nonsurgical options. A randomized trial showed total knee replacement improved symptoms more than nonsurgical care but carried more serious adverse events (Skou et al., NEJM, 2015). Timing is an individual decision made with a clinician.