Quick answer: PRP is one of the most polarised topics in knee osteoarthritis. Trials comparing it to hyaluronic acid often favour PRP — but the highest-quality placebo-controlled trial, RESTORE, published in JAMA in 2021, found PRP was no better than saline injections for pain or cartilage volume at 12 months. The ACR and OARSI recommend against it, and AAOS makes no recommendation for routine use. The honest summary: PRP may be better than a treatment that itself works poorly, which is not the same as working.
This article is general information, not medical advice about your knee. Use it to have a more informed conversation with your orthopaedic surgeon.
The Comparison Problem, and Why It Matters
This is the most important thing on the page, and almost no PRP article makes the point.
Most positive PRP research compares it against hyaluronic acid. AAOS’s own technology overview found PRP showed more improvement than HA in nine of twelve high-quality studies reporting patient-reported outcomes, with three finding no difference. On the face of it, that looks encouraging.
The difficulty is that hyaluronic acid itself performs poorly against placebo. Most major guidelines now advise against routine HA use, and the largest analysis found its benefit falls below the level patients can perceive. Beating a weak comparator does not establish that a treatment works — see our guide to hyaluronic acid knee injections for that evidence.
The test that matters is PRP against a genuine placebo, and that is what RESTORE did.
The RESTORE Trial
Published in JAMA in 2021 by Bennell and colleagues, this was a properly blinded, placebo-controlled trial in 288 people over 50 with painful knee OA and mild to moderate radiographic change. Participants received three weekly injections of either PRP or normal saline. Neither the patients, the injectors nor the assessors knew which.
- Both groups improved substantially over 12 months — and by a similar amount.
- No difference in medial tibial cartilage volume on MRI, so no evidence of the structural regeneration PRP is marketed on.
- The PRP group reported more post-injection pain, swelling and stiffness.
An accompanying editorial noted that similar results had emerged from high-quality trials of PRP for ankle osteoarthritis and Achilles tendinopathy: participants improved substantially, but not meaningfully more than those given placebo.
That “both groups improved” finding is worth sitting with. People genuinely do feel better after these injections. The question RESTORE answers is whether the platelets are responsible, and its answer was no — the improvement came from the passage of time, regression to the mean, the procedure itself and expectation.
The trial’s own stated limitation is important and cuts both ways: PRP preparations are not standardised, so the findings may not generalise to every product.
Where the Guidelines Stand
| Body | Position on PRP for knee OA |
|---|---|
| ACR / Arthritis Foundation | Recommends against, citing lack of standardisation and insufficient high-quality efficacy data |
| OARSI | Recommends against, on the same grounds |
| AAOS, 2021 | No recommendation for routine use. Maintains a separate technology overview reflecting the mixed literature |
Note what the objection actually is. Guidelines are not saying PRP has been proven useless — they are saying the evidence is not good enough, and that “PRP” describes a category of wildly varying preparations rather than a single defined treatment. Those are different criticisms, and the second one is the deeper problem.
Why “PRP” Is Not One Thing
The standardisation issue is the root of the entire controversy, and it is the most useful thing to understand if you are considering paying for it.
Two clinics can both offer “PRP” and inject substantially different products, varying in:
- Platelet concentration, from barely above baseline to eight times or more
- Leukocyte content. Leukocyte-poor (LP-PRP) preparations have generally shown better tolerance and results in osteoarthritis than leukocyte-rich (LR-PRP), which may provoke more inflammation in a joint
- Whether the platelets are activated before injection
- Centrifuge system and spin protocol, which change the final product substantially
- Volume, number of injections and spacing
This means trials are not testing the same intervention, that clinic results cannot be compared reliably, and that “the research shows PRP works” is not a well-formed statement — it depends which PRP.
If you proceed, ask specifically: is this leukocyte-poor or leukocyte-rich, what platelet concentration does your system produce, and is it activated? A clinic that cannot answer is not measuring what it is selling you.
What PRP Is and How It Is Given
Platelet-rich plasma is made from your own blood. A sample of roughly 20 to 60 ml is drawn and spun in a centrifuge to concentrate the platelets. Platelets carry growth factors — PDGF, TGF-beta, IGF-1, VEGF and EGF among them — which are involved in tissue repair and inflammation regulation. The concentrate is injected into the knee, usually under ultrasound guidance.
Because it comes from your own blood, immune rejection is not a concern. The biological rationale is genuinely plausible. The difficulty, as with hyaluronic acid, is that plausibility has not reliably translated into results in blinded trials.
The procedure
- Blood draw from the arm, much like a standard blood test
- Centrifugation, typically 10 to 15 minutes in the clinic
- Injection into the joint under ultrasound guidance. About 30 minutes in total
- Afterwards, expect soreness, stiffness and sometimes swelling for one to three days. This was more common in the PRP arm of RESTORE than with saline
- Activity: avoid strenuous exertion for around 48 hours; light walking is usually encouraged
Protocols commonly involve three injections spaced one to four weeks apart, though single-injection protocols exist. Benefit, where reported, tends to appear over four to eight weeks rather than immediately.
Who It Might Suit
Where positive findings exist, they cluster in a fairly consistent group:
- Mild to moderate osteoarthritis, roughly Kellgren-Lawrence grade I to II, with cartilage remaining
- Younger patients
- Leukocyte-poor preparations
- People who have exhausted conservative care and want to try something before considering surgery
PRP is unlikely to help meaningfully in grade III to IV disease with bone-on-bone contact. Note that RESTORE recruited exactly the mild-to-moderate group PRP is supposed to suit, and still found no benefit over placebo — so this should be read as “where benefit is most plausible” rather than “where it is established”.
Cost, Safety and Making the Decision
PRP is typically self-funded, since most insurers classify it as investigational. Costs vary widely by country, clinic and protocol, and a course of three injections costs considerably more than one — ask for the full-course price up front.
Safety is reasonably good. Because it uses your own blood there is no rejection risk, and adverse events in RESTORE were minor and transient. The realistic risks are post-injection pain, swelling and stiffness, and — as with any joint injection — a small risk of infection. A hot, swollen, very painful knee with fever after an injection needs same-day medical assessment.
One claim worth correcting: PRP is sometimes promoted as safer than steroids because repeated corticosteroid injections have been associated with cartilage concerns. That comparison is fair as far as it goes, but “no known risk” is not the same as “proven safe long term”, and long-term PRP data remain limited.
Questions worth asking
- Given that ACR and OARSI advise against PRP, what makes me a candidate?
- Which preparation do you use — leukocyte-poor or leukocyte-rich, and at what concentration?
- What is the total cost of the full course?
- What result, by when, would tell us it has not worked?
- Have I actually completed a supervised strengthening programme yet?
That last question deserves weight. Exercise and weight management remain first-line in every guideline, have far stronger evidence than any injection, and cost a fraction as much — yet are consistently under-prescribed. See our guides to knee osteoarthritis treatment and knee strengthening exercises.
None of this means PRP is a scam, and a reasonable person can decide it is worth trying after conservative care has failed and before considering knee replacement. It means going in with accurate expectations rather than the marketing version.
Frequently Asked Questions
The evidence is polarised. Trials comparing PRP with hyaluronic acid often favour PRP, but the highest-quality placebo-controlled trial, RESTORE, published in JAMA in 2021, found PRP was no better than saline injections for either pain or cartilage volume at 12 months. Both groups improved substantially and similarly. The ACR and OARSI recommend against PRP, and AAOS makes no recommendation for routine use.
Not on its own, and this is the key point. Hyaluronic acid itself performs poorly against placebo — most major guidelines now advise against routine HA use because its benefit falls below the level patients can perceive. Outperforming a weak comparator does not establish that a treatment works. The test that matters is PRP against genuine placebo, and in RESTORE it showed no advantage.
Because PRP describes a category of widely varying preparations rather than one defined treatment. Two clinics can both offer PRP while injecting substantially different products, differing in platelet concentration, leukocyte content, whether platelets are activated, centrifuge system and spin protocol, and injection volume and spacing. This means trials are not testing the same intervention, and clinic results cannot be compared reliably.
Leukocyte-poor preparations have generally shown better tolerance and results in osteoarthritis, since leukocyte-rich formulations may provoke more inflammation within a joint. If you are considering PRP, ask specifically which type the clinic uses, what platelet concentration their system produces, and whether the platelets are activated. A clinic that cannot answer is not measuring what it is selling you.
Reasonably so. It uses your own blood, so there is no rejection risk, and adverse events in the RESTORE trial were minor and transient. The realistic risks are post-injection pain, swelling and stiffness for a few days — more common with PRP than saline in that trial — plus the small infection risk of any joint injection. A hot, swollen, very painful knee with fever after an injection needs same-day assessment. Note that no known risk is not the same as proven safe long term, as long-term data remain limited.
There is no good evidence that it does. RESTORE measured medial tibial cartilage volume on MRI at 12 months and found no difference between PRP and saline. Structural regeneration is central to how PRP is marketed, so the absence of any measurable cartilage effect in the best-designed trial is significant.
Because they genuinely did improve — that part is real. In RESTORE both groups improved substantially over 12 months. What the trial showed is that the platelets were not responsible; improvement came from the passage of time, regression to the mean, the procedure itself and expectation. Personal testimony cannot separate those from a drug effect, which is exactly why blinded placebo-controlled trials exist.
It is a defensible choice after conservative care has failed, provided you go in with accurate expectations and have genuinely completed a supervised strengthening programme first. Exercise and weight management remain first-line in every guideline, have far stronger evidence than any injection and cost a fraction as much, yet are consistently under-prescribed. Ask what result by what date would tell you PRP has not worked, so the decision does not drift into repeated courses.
Sources
- Bennell KL, Paterson KL, Metcalf BR, et al. 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.
- AAOS — Platelet-rich plasma for knee osteoarthritis technology overview.
- AAOS — Management of Osteoarthritis of the Knee (Non-Arthroplasty), 3rd edition, 2021.
- Nonoperative management recommendations for knee osteoarthritis: a review of clinical guidelines and treatment alternatives. PMC.
- National Institute of Arthritis and Musculoskeletal and Skin Diseases — Osteoarthritis.
- NICE guideline NG226 — Osteoarthritis in over 16s.
Last updated: August 2026. Written by Aisha Desai, who covers musculoskeletal topics and is not a clinician. Reviewed for accuracy by our editorial team; this article has not been reviewed by a named clinician. It is general information, not medical advice, and not a recommendation for or against treatment in your case. Evidence on PRP continues to evolve and preparations vary considerably between clinics. Decisions belong with the clinician who knows your imaging and history.
