Platelet-rich plasma is the most common regenerative injection offered for knee osteoarthritis in the United States, and it is also one of the harder treatments to evaluate as a patient. The research is genuinely mixed, the preparations vary wildly from clinic to clinic, and almost nobody's insurance pays for it. This guide walks through what a course of injections actually involves, what the trials show, and how PRP compares with the two injections your orthopedist is more likely to offer first.

This is information, not medical advice. Decisions about your knee belong to you and a clinician who has examined it and looked at your imaging.

What PRP is, in one paragraph

Blood is drawn from your arm, spun in a centrifuge to concentrate the platelets, and the resulting plasma is injected into the knee joint. Platelets carry growth factors, and the theory is that concentrating them modulates inflammation inside the joint and possibly slows cartilage breakdown. The American Academy of Orthopaedic Surgeons' patient explainer on PRP is a reasonable neutral starting point.

One thing to hold onto: "PRP" is not a standardized drug. Platelet concentration can differ by a factor of five or more between systems. Some preparations are leukocyte-rich (white cells included), some leukocyte-poor. Some are activated with calcium chloride, some aren't. When you read that "PRP works" or "PRP doesn't work," the underlying studies were often testing meaningfully different products.

What the evidence shows

The honest summary: for knee osteoarthritis specifically, PRP has more supportive data than for most other regenerative applications, but the quality of that data is uneven and the highest-quality trials have been the least encouraging.

Pooled analyses lean positive. Multiple meta-analyses of randomized trials have found that intra-articular PRP reduces pain and improves function more than saline or hyaluronic acid at 6 and 12 months, with effect sizes that often exceed the threshold usually considered clinically meaningful. That's the case for the treatment, and it is not trivial.

The best-blinded trials lean negative. The RESTORE trial, published in JAMA in 2021, randomized 288 people with mild-to-moderate knee osteoarthritis to three weekly injections of leukocyte-poor PRP or saline, followed them for 12 months, and found no significant difference in knee pain and no difference in medial tibial cartilage volume on MRI. Both groups improved substantially. That last detail matters: an injection into a painful knee produces a large placebo response, and trials without a真 sham arm systematically overstate benefit.

So the field is split between a large volume of small, often unblinded, industry- or clinic-run trials showing benefit, and a smaller number of rigorous placebo-controlled trials showing little to none. Cochrane's review of platelet-rich therapies across musculoskeletal conditions concluded that the evidence base was insufficient to support routine use, largely because of small samples, inconsistent preparations, and poor blinding.

Professional bodies have been cautious. The American College of Rheumatology's 2019 osteoarthritis guideline recommends against PRP for knee and hip osteoarthritis, citing the heterogeneity of preparations and protocols as much as the efficacy data. The AAOS knee osteoarthritis guideline has treated the evidence as limited rather than conclusive in either direction. Neither position means "PRP never helps anyone." It means the field cannot yet say who it helps, with which formulation, and how reliably.

What a course actually looks like

Most protocols use one to three injections, spaced one to four weeks apart. There is some trial evidence that two or three injections outperform a single one for knee osteoarthritis, though this is not settled. Be skeptical of packages of six or more; the data thin out fast beyond three.

A realistic timeline, based on how trials report outcomes:

  • Days 1–5: Often worse. A flare of soreness, stiffness, and swelling is common because you have injected an inflammatory stimulus into an already irritated joint. Ice and acetaminophen are the usual advice; many protocols ask you to avoid NSAIDs around the injection window, on the theory that they blunt the platelet response, though that rationale is mechanistic rather than proven.
  • Weeks 2–6: Gradual change, if it comes. Almost nobody has a dramatic overnight response, and a clinician promising one is overselling.
  • Months 2–3: Where trials typically see peak effect.
  • Months 6–12: Effects tend to persist longer than a cortisone shot does, but usually taper. Many patients who respond return for repeat treatment at 9–12 months.

If you have had no meaningful change by three months after completing a course, further injections of the same preparation are unlikely to help.

Who tends not to respond

The pattern across trials and observational series is fairly consistent, even if none of it is a guarantee:

  • Severe radiographic arthritis. People with Kellgren-Lawrence grade 3–4 disease — substantial joint space loss, "bone on bone" — respond less well than those with grade 1–2. PRP is not a substitute for a joint replacement in an end-stage knee.
  • Large, recurring effusions or a knee that is mechanically unstable.
  • Significant malalignment (marked varus or valgus), which keeps loading the same damaged compartment regardless of what is injected.
  • Higher BMI and older age, both associated with weaker responses in subgroup analyses, though these findings are inconsistent.
  • Inflammatory arthritis rather than osteoarthritis — rheumatoid or crystal arthropathy is a different disease and PRP is not the intervention studied for it.
  • Mechanical symptoms such as true locking from an unstable meniscal flap. An injection does not fix a mechanical problem.

PRP versus cortisone versus hyaluronic acid

Corticosteroid is the fastest and cheapest option and is usually covered by insurance. Cochrane's review found short-term pain relief that was clear at around one to two weeks but faded by roughly six months. The trade-off: a randomized trial in JAMA found that repeated triamcinolone injections every three months over two years produced greater cartilage volume loss than saline, with no pain benefit — which is why most clinicians now limit how often they repeat them. Steroid is a good tool for a flare, or for getting through an event, or when you need to know whether the pain is coming from inside the joint at all.

Hyaluronic acid (viscosupplementation) occupies a strange middle position. Some professional guidelines recommend against it; some payers cover it; large network meta-analyses have suggested that the average benefit over placebo is small and may not cross the threshold patients would notice. Relief, when it happens, tends to build over weeks and last a few months.

PRP costs the most, is essentially never covered, has the longest onset, and has a plausible case for longer duration than steroid — but the least regulatory validation and the widest variation in what you actually receive. Head-to-head trials against hyaluronic acid have generally favored PRP, though many of those trials were not blinded.

None of the three has convincingly been shown to change the structural course of the disease. Weight management, targeted strengthening, and load modification remain the interventions with the strongest evidence for knee osteoarthritis, per the National Institute of Arthritis and Musculoskeletal and Skin Diseases. PRP is an adjunct to that work, not a replacement for it.

The FDA question

PRP for knee osteoarthritis is not an FDA-approved treatment. The centrifuge systems used to prepare it are generally cleared as devices for producing platelet concentrate from your own blood, but no agency has reviewed and approved PRP as a therapy for arthritis. Because PRP is autologous and minimally manipulated, it is largely practiced under the physician's clinical judgment rather than through the drug-approval pathway.

What that does and doesn't mean: it does not mean the treatment is illegal, fringe, or unstudied. It does mean no regulator has vetted efficacy claims, no standardized dosing exists, and marketing language is not policed the way a pharmaceutical label is. The FDA maintains consumer guidance on regenerative medicine claims that is worth reading, particularly if a clinic starts talking about stem cells or amniotic products rather than plain PRP — those raise separate regulatory issues.

Cost, and what to ask

Expect roughly $500 to $1,500 per knee injection, with three-injection courses commonly landing between about $1,200 and $3,500. Bilateral treatment costs more. Insurance coverage is rare; some plans will cover the office visit and imaging but not the injection itself. Ask for the total course price in writing before the first draw.

Useful questions: What is my Kellgren-Lawrence grade? Is the injection done under ultrasound guidance? Which preparation system, and is it leukocyte-rich or leukocyte-poor? How many injections, and what will you consider a non-response? What is the plan if this doesn't work?

A clinician who answers those precisely, and who tells you plainly that this may not work, is giving you better information than one who promises to "regenerate" your cartilage.