The short answer
Yes — for the right knee, at the right stage
Yes, PRP works for knee arthritis in the right knee at the right stage, and knee osteoarthritis is the most heavily studied use of platelet-rich plasma anywhere in medicine, and the weight of the evidence is positive. A large pooled analysis, 42 randomized controlled trials covering 3,696 patients, published in Annals of Medicine and Surgery, found PRP outperforming hyaluronic acid, corticosteroid and placebo on pain, and on function in that analysis; stricter analyses, covered below, are less certain about function.
Against placebo, that analysis reported a mean difference of −1.74 on the VAS pain scale (95% CI −2.68 to −0.80, p = 0.0003) and −7.01 on total WOMAC (95% CI −11.26 to −2.76, p = 0.001).
Talk to a physiatrist about your kneeWhat it is
Concentrating your own repair signals
PRP works by concentrating your own repair signals: we draw a small amount of your blood, spin it in a centrifuge to separate the platelet fraction, and inject that concentrate into the joint. Platelets carry growth factors — the signaling proteins your body already uses to repair tissue. Nothing synthetic is added.
That autologous quality is why PRP carries essentially no allergy risk and why it is permitted under current anti-doping rules, which matters to competitive athletes weighing their options. The full mechanism and what a visit involves is covered on our PRP therapy page. In short, PRP concentrates repair signals your body already makes and delivers them to the joint.
The evidence
What the studies actually found
The studies found a consistent pain benefit from PRP, but the three pieces of recent evidence worth knowing about do not all say the same thing. That disagreement is the most useful part.
The large pooled analysis — favorable
The large pooled analysis was favorable: the 42-trial, 3,696-patient meta-analysis found PRP beating hyaluronic acid on WOMAC pain (MD −0.74, 95% CI −1.02 to −0.46, p < 0.00001) and beating corticosteroid on WOMAC pain (MD −8.06, 95% CI −13.62 to −2.51, p = 0.004). Benefit versus hyaluronic acid was greatest at 12 months; versus corticosteroid it peaked around 6 months.
The umbrella review — favorable, with conditions
An umbrella review of 7 meta-analyses published in Clinics and Practice in 2026 concluded PRP delivers clinically significant improvement, specifically in Kellgren-Lawrence grades I–III (mild to moderate disease), sustained through 6 to 12 months. It also noted a trend toward better results with leukocyte-poor PRP and three-injection regimens than with leukocyte-rich preparations and single injections, though inconsistent reporting kept that from being firmly tested.
A stricter 2026 analysis — more cautious
A stricter 2026 analysis was more cautious: a systematic review of 5 placebo-controlled trials in 761 patients, published in the International Journal of Rheumatic Diseases in August 2026, found a statistically significant VAS pain benefit at 6 months (MD 0.53, 95% CI 0.39–0.68) and a WOMAC function difference whose confidence interval touched zero (MD 0.14, 95% CI −0.00 to 0.28), although the authors describe it as significant. They state plainly that long-term effectiveness remains uncertain.
Read together: PRP reliably reduces pain in mild-to-moderate knee osteoarthritis, the effect is most convincing between six and twelve months, and the functional benefit is less consistent than the pain benefit. Anyone telling you it is settled science in both directions has not read the 2026 papers.
One more paper is worth knowing about because it governs how treatment should actually be delivered. A 2025 review in the Journal of Clinical Medicine reports that PRP typically concentrates platelets to 3–5 times baseline blood levels, that leukocyte-poor preparations produce better outcomes and fewer adverse reactions than leukocyte-rich ones, and that triple injections were superior to single injections over 12 months in one study it cites — with two- or three-injection series the most common in trials. Taken together, the studies found real pain relief that depends on disease stage and on how PRP is prepared and delivered.
Comparison
PRP versus the alternatives, with numbers
Measured in numbers, PRP outperformed the alternatives (placebo, hyaluronic acid and corticosteroid) on knee pain in pooled trials, as the table below shows.
| Comparison | Pain outcome | When it peaked |
|---|---|---|
| PRP vs placebo | VAS MD −1.74 (95% CI −2.68 to −0.80) | — |
| PRP vs hyaluronic acid | WOMAC pain MD −0.74 (95% CI −1.02 to −0.46) | 12 months |
| PRP vs corticosteroid | WOMAC pain MD −8.06 (95% CI −13.62 to −2.51) | 6 months |
Corticosteroid still has a role — it works faster, and for an acutely inflamed joint that speed matters. The tradeoffs between the two are worth understanding before you choose, and we have written that comparison up separately in PRP vs cortisone.
Good candidates
Who tends to do well
Patients who tend to do well with PRP have Kellgren-Lawrence grade I–III knee osteoarthritis, and the umbrella review's finding on that is the most practically useful number in this whole article, because it tells you where the ceiling is.
- Kellgren-Lawrence grade I–III osteoarthritis — mild to moderate, not bone-on-bone
- Pain that has not settled with physical therapy and load management
- People wanting to delay or avoid joint replacement
- Patients who have had corticosteroid before and do not want to keep repeating it
- Active adults whose symptoms limit activity rather than daily function entirely
Poor candidates
When we will tell you not to bother
We will tell you not to bother with PRP when it is unlikely to help, because part of the value of seeing a physiatrist rather than an injection service is being told when the answer is no. Dr. Dentico is board certified in Physical Medicine and Rehabilitation with a sports medicine subspecialty — you can read his background on our team page.
- End-stage, bone-on-bone arthritis — a surgical opinion is usually more useful
- Knee pain that is actually referred from the hip or lower back, which an examination can distinguish
- Mechanical symptoms (true locking or giving way) pointing to a structural problem
- Active infection, or certain blood and platelet disorders
Where imaging shows genuine structural damage, other options in our pain and injury recovery program (including shockwave therapy and peptides for soft-tissue recovery) may fit better, alone or alongside PRP.
Practicalities
What a course of treatment looks like
A course of PRP treatment at Protocol Health usually looks like a series of injections rather than one shot, and on dosing, the umbrella review noted a trend toward better results with three-injection regimens than with single injections, and one study cited in the 2025 review found triple injections superior over 12 months. If someone quotes you one shot as a complete course, that is not what most trial protocols used.
Examination and imaging review
Examination and imaging review establish what is generating the pain and how advanced the arthritis is — the Kellgren-Lawrence grade genuinely changes the recommendation.
Draw and preparation
Blood is drawn in the office and centrifuged while you wait. Leukocyte-poor preparation, per the comparative evidence.
Injection
The concentrate is injected into the joint, with guidance where accuracy matters. Local anesthetic is used.
The next few days
In the next few days, expect increased soreness. That inflammatory response is part of the mechanism, not a sign of failure.
Rehabilitation
Rehabilitation means loading and strength work in the following weeks. NSAIDs are usually paused around treatment because they may blunt the signaling PRP relies on.
Review
A review assesses response before deciding on further injections. A series is typical, not a single shot.
The full pathway from first consultation to ongoing care is set out on our how it works page, and the joint-by-joint detail lives on PRP for joint, tendon and cartilage restoration. A typical course, then, is an examination, a series of injections and rehabilitation, with a review between them.
Verdict
Does PRP work for knee arthritis? The verdict
PRP for knee arthritis is a reasonable option in mild to moderate osteoarthritis (Kellgren-Lawrence grades I–III). Pooled trials show more pain relief than hyaluronic acid, cortisone or placebo, clearest at six to twelve months. The benefit for function is less certain, PRP does not regrow cartilage, and a bone-on-bone knee rarely justifies it. The decision starts with an examination and imaging review, not with the injection.
FAQ
Frequently Asked Questions
How long does PRP last in the knee?
The published benefit is clearest between about six and twelve months, and longer-term effectiveness is still uncertain. Many people plan for repeat treatment rather than expecting one injection to be permanent.
How many injections will I need?
A series is common for knee osteoarthritis rather than a single treatment. The number depends on your response, and you should be told the expected schedule before you start rather than discovering it as you go.
Is PRP better than a cortisone shot?
For short-term relief, cortisone often works faster. For chronic problems where you would otherwise be repeating steroid injections, PRP avoids the tissue-thinning concern associated with repeated corticosteroid. In mild-to-moderate knee osteoarthritis, comparative analyses generally favor PRP over both cortisone and hyaluronic acid over six to twelve months.
Does it hurt?
There is discomfort during the injection and typically several days of increased ache afterwards. Local anesthetic is used, and most patients drive themselves home.
What is the recovery time after a PRP knee injection?
Most people have increased soreness for several days after a PRP knee injection and return to normal daily activity within a few days. Your physician will give specific instructions on activity, and on which pain relievers to avoid, for the first one to two weeks. Any benefit builds gradually over weeks rather than appearing straight away.
Explore
Related at Protocol Health
Talk to a physiatrist about your kneeThis article is educational and is not medical advice. It does not diagnose, prevent, treat or cure any condition. Talk with a licensed clinician about your own situation.