Short answer. Probably not. In the best study so far, PRP injections eased knee pain no more than injections of salt water.
Evidence grade. Probably does not help.
What it is
PRP stands for platelet rich plasma. A small amount of your own blood is taken, spun in a machine to concentrate the platelets, and injected into the knee. Platelets carry growth factors, and the idea is that these might calm the joint or help the cartilage repair. A course is usually one to three injections.
What the research shows
The largest and best designed trial was done in Melbourne and Sydney and published in 2021. It followed 288 people with mild to moderate knee arthritis for a year. Half had three PRP injections and half had three salt water injections. Neither the patients nor the people giving the injections or measuring the results knew which was which. At 12 months both groups had less pain, and the difference between them was too small to matter. MRI scans showed no difference in the cartilage either.1
Some earlier studies found PRP worked, and reviews that pool them still rank PRP above other injections.2 Most of those studies were small, compared PRP with gel injections rather than a placebo, and the patients or doctors often knew who had PRP. When a treatment is tested against a placebo in a large trial where nobody knows who had what, the benefit disappears. That is the pattern seen with PRP.
PRP is also used for other joints and for tendon problems. The evidence for those is thinner again, and this page is about the knee only.
Risks and effort
The injections themselves are low risk. Pain and swelling for a few days is common, and infection is rare. The main drawback is time spent on a treatment that is unlikely to help, and time not spent on treatments that do.
What usually works instead
Exercise and strength work for the leg muscles, weight loss where that applies, and simple pain relief have the best evidence for knee arthritis. A cortisone injection can give a few weeks of relief for a painful flare. When pain limits daily life despite these, knee replacement is the treatment with the most reliable results. See Osteoarthritis of the Knee.
Questions people ask
My physio said PRP helped their other patients. Why does the trial say no?
Most people with knee arthritis feel better for a while after any injection, including salt water, because the knee is rested, attention is paid to it and pain naturally rises and falls. A trial with a placebo group is the only way to tell how much of the improvement came from the PRP itself. In this case, almost none did.
Does it matter which PRP system is used?
Systems differ in how many platelets and white cells they contain, and some clinics say their system is the one that works. No type of PRP has been shown to beat a placebo in a large blinded trial.
Is PRP better than a cortisone injection?
Cortisone gives quicker relief but it wears off within weeks to months and repeated injections may harm cartilage. PRP has fewer short term side effects but no proven benefit. Neither changes the course of the arthritis.
Would PRP help my meniscus tear or my tendon instead?
The evidence for tendons such as the patellar and Achilles tendon is mixed, and for meniscus tears it is very limited. Ask what the injection is expected to do for your specific problem and what the alternative is.
References
- Bennell KL, Paterson KL, Metcalf BR, Duong V, Eyles J, Kasza J, Wang Y, Cicuttini F, Buchbinder R, Forbes A, Harris A, Yu SP, Connell D, Linklater J, Wang BH, Oo WM, Hunter DJ. Effect of intra-articular platelet-rich plasma vs placebo injection on pain and medial tibial cartilage volume in patients with knee osteoarthritis: the RESTORE randomized clinical trial. JAMA. 2021;326(20):2021-2030. https://doi.org/10.1001/jama.2021.19415
- Qiao X, Yan L, Feng Y, Li X, Zhang K, Lv Z, Xu C, Zhao S, Liu F, Yang X, Tian Z. Efficacy and safety of corticosteroids, hyaluronic acid, and PRP and combination therapy for knee osteoarthritis: a systematic review and network meta-analysis. BMC Musculoskelet Disord. 2023;24(1):926. https://doi.org/10.1186/s12891-023-06925-6
Last reviewed 18 October 2026. Next review due October 2028.