Stem Cells vs PRP for Knee Arthritis: Which Has Better Evidence?

If you have knee osteoarthritis and have started looking at injections, you have probably run into two options that are marketed as "regenerative": platelet-rich plasma (PRP), made from your own blood, and stem cell therapy, made from bone marrow, fat, or donor tissue. One is usually a fraction of the price of the other. The obvious question is whether the more expensive option actually does more.

This article sets out what the randomized trials show for each, what happened in the trials that compared them directly, and where the evidence for both is weaker than clinics tend to admit. It is written by a clinic that offers cell therapy, so we have tried to be especially careful to report the findings that do not favour it.

The quick answer

PRP has the larger evidence base. Pooled trials show it outperforms hyaluronic acid injections on pain and function scores,[2] but the largest placebo-controlled trial — 288 patients, published in JAMA — found no meaningful difference between PRP and a saltwater injection at twelve months.[1]

Stem cell therapy has a smaller and more varied evidence base. In head-to-head trials, bone marrow aspirate concentrate (BMAC) was equivalent to PRP at two years in one study[3] and somewhat better on several scores at one year in another.[4] A network meta-analysis found both beat hyaluronic acid, with no significant difference between them.[5] A 2024 meta-analysis of MSC trials against placebo or usual care concluded they probably make little to no difference to pain.[7]

The short version: neither is a proven disease-modifying treatment. PRP is the reasonable first biologic for most people with mild-to-moderate knee arthritis because it costs less and performs similarly in direct comparisons. Cell therapy is a defensible second step for selected patients — not a guaranteed upgrade.

What each option actually is

Platelet-rich plasma (PRP) is your own blood, spun in a centrifuge to concentrate the platelets, then injected into the knee. Platelets carry growth factors and signalling proteins; the theory is that concentrating them shifts the joint towards a less inflammatory state. PRP is not a cell therapy in the stem cell sense and contains no stem cells. Preparations vary a great deal — in particular, "leukocyte-poor" PRP filters out most white cells, and "leukocyte-rich" PRP keeps them — and that turns out to matter.

"Stem cell therapy" is a label that covers at least three quite different products, which is one of the main reasons this comparison gets muddled:

• Bone marrow aspirate concentrate (BMAC). Marrow drawn from the pelvis, concentrated in a centrifuge the same day, and injected. It is mostly blood and marrow cells; genuine mesenchymal stromal cells are a very small fraction. Most of the head-to-head trials against PRP used this.
• Adipose-derived preparations. Fat harvested by mini-liposuction and processed, either same-day or in a laboratory.
• Culture-expanded mesenchymal stromal cells (MSCs). Cells isolated from marrow, fat, or umbilical cord tissue and grown in a laboratory to a defined, counted dose before injection. Our guide to what happens to cells in the lab explains why a counted, characterized dose is a different product from a same-day concentrate.

This matters because a trial result for BMAC does not automatically transfer to culture-expanded umbilical cord MSCs, or the reverse. When you read "stem cells vs PRP," always ask: which stem cell product?

What the PRP evidence shows

PRP has dozens of randomized trials, most of them comparing it with hyaluronic acid (a lubricating gel injection). A 2021 meta-analysis in the American Journal of Sports Medicine by Belk and colleagues pooled 18 level-one trials with more than 1,600 patients and found that PRP produced substantially larger improvements in WOMAC pain-and-function scores than hyaluronic acid over an average of about eleven months. In a sub-analysis, leukocyte-poor PRP was associated with better outcomes than leukocyte-rich PRP on one of the scores measured.[2]

The difficulty is that hyaluronic acid is itself only a modestly effective comparator. Beating it does not prove PRP beats placebo. That question was tested directly in the RESTORE trial, published in JAMA in 2021 by Bennell and colleagues. It randomized 288 people aged 50 or older with mild-to-moderate medial knee osteoarthritis to three weekly injections of leukocyte-poor PRP or saline. At twelve months, pain improved in both groups — by 2.1 points with PRP and 1.8 points with saline on a ten-point scale — a difference that was not statistically significant. There was no difference in cartilage volume on MRI, and 29 of 31 secondary outcomes showed no difference either. The authors concluded their findings did not support PRP for knee osteoarthritis.[1]

RESTORE did not end the debate — it tested one commercial PRP product and one injection schedule, and other trials have reported more positive results — but it is the most rigorous single study in the field, and it is the one to weigh most heavily.

What the stem cell evidence shows

The cell therapy literature for the knee includes several small randomized trials with encouraging results. In one, Matas and colleagues randomized 26 patients to umbilical cord-derived MSCs (one or two doses) or hyaluronic acid; at twelve months, the two-dose group had significantly lower pain than the hyaluronic acid group, with no severe adverse events and no difference in MRI scores.[6] The result is interesting, particularly on dosing — but eight or nine patients per arm cannot carry much weight.

When the trials are pooled against placebo or usual care, the picture is sobering. A 2024 systematic review in Osteoarthritis and Cartilage by Sadeghirad and colleagues included 16 randomized trials with 807 participants. It concluded, with moderate certainty, that intra-articular MSCs probably result in little to no difference in pain at three to six months and at twelve months. There was low-certainty evidence of a possible improvement in physical function at twelve months, and low-certainty evidence of more adverse events — mainly knee pain and swelling after injection.[7]

We think that is the most important paper for any patient considering cell therapy for the knee to know about, and we would rather you hear it from us than discover it after treatment.

What happened when they were compared head-to-head

A handful of randomized trials have put a stem cell product directly against PRP. All of them used bone marrow concentrate rather than culture-expanded cells.

Anz and colleagues (2022). Ninety patients with mild-to-moderate knee osteoarthritis (Kellgren-Lawrence grades 1–3) received a single injection of either leukocyte-rich PRP or BMAC. Both groups improved significantly, the improvement levelled off by three months, and it was sustained to 24 months — with no difference between PRP and BMAC at any time point. There was no placebo arm, and a meaningful number of patients were lost to follow-up by two years. The authors' conclusion was blunt: BMAC was not superior to PRP.[3]

Dulic and colleagues (2021). 175 patients with grade 2–4 osteoarthritis were randomized to BMAC, PRP, or hyaluronic acid. All three groups improved over twelve months with no serious side effects. BMAC scored better than the other groups on several measures, though not on WOMAC or one IKDC measure when compared with PRP. The groups were very unequal in size (111 BMAC, 34 PRP, 30 hyaluronic acid) and the authors themselves called for more high-quality trials.[4]

Belk and colleagues (2023). A meta-analysis in Arthroscopy pooled 27 level-one trials — about 1,000 PRP patients, 226 BMAC patients, and 1,100 hyaluronic acid patients. Both PRP and BMAC outperformed hyaluronic acid on pain and function scores. Between PRP and BMAC, there was no significant difference.[5]

Why the results are so hard to interpret

Both literatures share the same weaknesses, and it helps to know them before you read any clinic's claims — including ours:

• Placebo response in knee pain is large. In RESTORE, saline alone improved pain by almost two points on a ten-point scale. Any trial without a placebo arm will make its treatment look better than it is.
• Products are not standardized. PRP preparations differ in platelet concentration and white cell content; "stem cells" range from a same-day marrow concentrate to a counted dose of cultured cells. Pooling them together blurs real differences.
• Small, short trials. Most cell therapy trials enrol tens of patients and follow them for one to two years.
• Industry involvement. Many trials in both fields are funded or supplied by device and biologics companies. That does not make them wrong, but it is worth noticing.
• Structural change is unproven. Neither PRP nor any cell product has reliably been shown to regrow cartilage or change the long-term course of the disease in humans.

Who might reasonably try which

With those caveats, here is how we think about the choice in practice. None of it replaces an assessment of your own imaging.

PRP is a reasonable first biologic if:

• You have mild-to-moderate osteoarthritis (Kellgren-Lawrence grade 1–3) that has not responded to exercise therapy, weight management, and simple pain relief
• You want the lower-cost, lower-effort option first, knowing the benefit may be modest
• You can have it done close to home — PRP generally does not justify international travel on its own

Cell therapy may be worth discussing if:

• You have moderate osteoarthritis with cartilage still present on imaging, and a genuine trial of PRP or other injections gave little or short-lived benefit
• You are trying to delay a knee replacement for age or personal reasons, and understand that delay rather than repair is the realistic goal
• You accept, in writing, that the pooled evidence against placebo is weak and that non-response is a real possibility

The conditions we treat and how we assess knee cases are set out on our page on stem cell therapy for knee and joint conditions in Colombia. Our longer guide to stem cells for knee osteoarthritis covers what a course of treatment involves.

When this may not be right for you

Neither injection is likely to help — and we would advise against paying for either — in these situations:

• End-stage, bone-on-bone arthritis (grade 4) with no measurable joint space. For a joint that has structurally collapsed, the evidence points to replacement; we compare the two honestly in stem cells vs knee replacement.
• Severe malalignment or an unstable knee, which no injection corrects.
• Pain that does not match the imaging — for example, pain that is actually coming from the hip or spine.
• Inflammatory arthritis such as rheumatoid or psoriatic arthritis that is not controlled.
• Active infection, active cancer, pregnancy, or a blood disorder that makes a harvest or injection unsafe.
• An expectation of a cure or a rebuilt knee. No injection currently available can promise that.

The bottom line

PRP and stem cell therapy for knee arthritis are closer in their evidence than their price tags suggest. Both have outperformed hyaluronic acid in pooled trials. Both have struggled to separate from placebo in the most rigorous studies. And in the trials that compared them directly, bone marrow concentrate has mostly performed about the same as PRP.

That means the sensible order for most people is to try the simpler, cheaper option first, and to consider cell therapy only as a considered second step with realistic expectations. If a clinic tells you stem cells are simply "stronger PRP," or that either treatment regrows cartilage, it is overstating what the research has shown. What we can do is review your X-rays or MRI and tell you plainly whether either injection makes sense for your knee — including when the answer is that neither does.