PRP vs. Hyaluronic Acid Injections for Knee Osteoarthritis: What Does Research Show?
- Updated November 14, 2026
- Stop 3 of 4 on the PRP injections line
In head-to-head trials, platelet-rich plasma (PRP) has tended to score better than hyaluronic acid (HA) on some pain and function measures. A 2026 meta-analysis of 28 randomized trials (2,685 participants) found that advantage, but the trials disagreed with each other to an unusually high degree. Guidelines read the wider evidence differently: the American College of Rheumatology’s 2026 update strongly recommends against PRP for knee and hip osteoarthritis (low certainty of evidence) and conditionally recommends against HA, while the May 2026 VA/DoD guideline finds insufficient evidence to recommend for or against either injection in the knee. Beating HA does not make PRP a recommended treatment, and neither injection has been shown to regrow cartilage.
What are PRP and hyaluronic acid injections?
PRP is made from your own blood. A sample is drawn and spun in a centrifuge to concentrate platelets, which carry growth factors and other signaling proteins. The resulting plasma is injected into the knee. Preparations differ in their platelet and white blood cell concentrations, including whether they are leukocyte-rich or leukocyte-poor (white blood cells reduced to a lower concentration), and in the number of injections.
HA, sometimes called viscosupplementation or “gel” injection, is a manufactured form of a substance naturally present in joint fluid. It is intended to improve the fluid’s lubricating and cushioning properties.
Neither is a stem cell product. Cell-based injections and the question of cartilage regrowth are covered separately in whether stem cell therapy can regrow knee cartilage.
What outcomes are compared in clinical trials?
The trials in this comparison measure symptoms, not structure:
- Pain scores, often a 0 to 10 or 0 to 100 visual analog scale (VAS)
- WOMAC, a questionnaire covering pain, stiffness and physical function in knee and hip osteoarthritis
- KOOS and IKDC, other knee function questionnaires
- Follow-up timing, for example 3, 6 and 12 months
A result can be statistically significant without being large enough for a person to notice. When you read a study, look for the size of the difference on the scale used, not only whether it was significant.
What do randomized trials and meta-analyses show?
A 2026 pooled analysis is a systematic review and meta-analysis in the Journal of ISAKOS, published online September 15, 2026 (volume 21, article 101218). It included 28 randomized trials with 2,685 participants comparing intra-articular PRP with intra-articular HA in adults with symptomatic knee osteoarthritis. Its findings:
- PRP improved pain and function more than HA on pooled estimates. According to the abstract, the WOMAC total score was 8.14 points lower with PRP (mean difference −8.14, 95% CI −10.79 to −5.48) and pain VAS was 7.89 points lower (mean difference −7.89, 95% CI −12.90 to −2.89). The abstract does not state the scale ranges used, so these differences should not be read as a percentage of any particular scale.
- Heterogeneity was very high: I² of about 96% for WOMAC and 97% for pain VAS. An I² this high means the individual trials reported substantially different results, so a single pooled number summarizes a wide spread rather than a consistent effect.
- Two study features explained part of that spread. Osteoarthritis severity and the number of injections were statistically significant effect modifiers, together explaining about 24% of the between-study variance. The larger benefits were seen in mild-to-moderate osteoarthritis.
- In a subgroup analysis, treatment effects were larger with leukocyte-poor PRP (interaction P = 0.04). That is a subgroup signal within this review, not a head-to-head trial result, and it does not mean leukocyte-rich PRP failed: an earlier meta-analysis of randomized trials in Archives of Orthopaedic and Trauma Surgery found PRP better than HA whether the preparation was leukocyte-rich or leukocyte-poor and whether one or several injections were given. Subgroup comparisons like these do not establish which preparation or schedule is best.
- Using the GRADE approach, the authors rated the certainty of evidence as moderate for WOMAC and low for pain VAS. That rating applies to this PRP-versus-HA question; it is a different question from the ACR’s assessment of whether PRP should be used at all.
Three limits follow from the design. First, the comparison is PRP against HA, not PRP against placebo or exercise; an advantage over HA does not establish benefit over doing nothing different. Second, protocols varied widely, so “PRP” in one trial is not the same preparation as in another. Third, the outcomes are symptoms, not cartilage.
How do 2021 and 2026 guidelines differ?
Guidelines answer a different question from a head-to-head meta-analysis. They ask whether a treatment should be used at all, weighing the full body of evidence, including placebo-controlled trials, harms and certainty.
| Guideline | Year | On PRP | On HA |
|---|---|---|---|
| American College of Rheumatology recommendation summary | Approved September 8, 2026 | Strongly recommends against for knee and hip OA; certainty of evidence low | Conditionally recommends against for knee, hip and hand OA; certainty moderate for knee and hip, very low for hand |
| VA/DoD clinical practice guideline for non-surgical management of hip and knee OA, Version 3.0 | May 2026 | Recommendation 20 (knee): insufficient evidence to recommend for or against PRP; strength and direction "Neither for nor against" | Recommendation 17 (knee): insufficient evidence to recommend for or against viscosupplementation; "Neither for nor against." For the hip, Recommendation 18 suggests against it (weak) |
| VA/DoD patient companion, Orthobiologics at a Glance | 2026 | A plain-language handout, not a graded recommendation; explains that some studies found improvements and others did not, and more research is needed on who may benefit and whether results last | Not addressed in the companion |
| AAPM&R consensus guidance on PRP for knee OA | First published online April 16, 2026 | A consensus practice document with five evidence-based recommendations and 11 consensus best practices; this article does not summarize its individual recommendations | Not its subject |
| AAOS clinical practice guideline on knee OA (nonarthroplasty), third edition | Published August 31, 2021 | PRP may reduce pain and improve function in symptomatic knee OA; strength of recommendation "Limited" (downgraded) | Not recommended for routine use in symptomatic knee OA |
Two distinctions matter when reading this table.
Strength of recommendation is not certainty of evidence. The ACR’s recommendation against PRP is strong, but the evidence certainty behind it is rated low. That combination means the panel was confident enough to advise against routine use, not that PRP has been proven to have zero effect. You can read the wording in the ACR 2026 summary.
Different documents serve different purposes. The ACR summary gives graded recommendations on whether to use treatments. The <a href="https://www.healthquality.va.gov/HEALTHQUALITY/guidelines/CD/OA/Osteoarthritis-CPG2026-Guidelinefinal_20260618.pdf”>VA/DoD guideline, Version 3.0 also grades its recommendations, and for knee PRP and knee viscosupplementation it takes no position either way because it judged the evidence insufficient. The VA/DoD companion is a plain-language patient handout tied to the 2026 VA/DoD guideline. The AAPM&R guidance is a specialty society consensus on PRP practice. The AAOS guideline page reflects its own earlier review; its 2021 PRP and HA wording and strength ratings are listed in the guideline summary on Guideline Central. Read side by side, the 2021 AAOS document leans cautiously toward PRP, the 2026 ACR summary advises against it, and the 2026 VA/DoD guideline takes no position, which is why the dates matter. They should not be treated as identical endorsements or as opposing trials.
For context, the same ACR summary strongly recommends exercise for knee and hip osteoarthritis and strongly recommends intra-articular glucocorticoid (steroid) injections, each with stated certainty ratings.
Side effects, logistics, and patient-specific considerations
Both carry the general risks of putting a needle into a joint. A 2026 knee-specific review in PM&R described the adverse events seen after PRP as mostly mild and short-lived, such as knee pain and swelling, and found that PRP had higher rates of mild knee pain and swelling than HA, a difference seen with leukocyte-rich PRP but not leukocyte-poor PRP. Those safety details, and when to call a clinician after an injection, are covered in the risks and side effects of PRP injections.
Practical differences include cost, since PRP may be paid out of pocket, the number of visits a protocol requires, and whether your insurer covers either option.
Questions to ask when comparing options
- Which preparation is being proposed: leukocyte-rich or leukocyte-poor PRP, and how many injections?
- How severe is my osteoarthritis on imaging and examination, and does the evidence apply to that severity?
- What improvement, on what scale, should I realistically expect, and for how long?
- How does this option compare with exercise therapy and steroid injection, which the 2026 ACR summary recommends?
- What will it cost, and does my plan cover it?
Is PRP always more effective than gel injections? No. Pooled results favor PRP on some symptom measures, but the trials vary so widely that the advantage is not consistent across studies, preparations or levels of arthritis.
Do either of these injections regrow cartilage? The trials summarized here measured pain and function. They do not establish that either injection rebuilds cartilage.
Every stop on the PRP injections line
- Does PRP Help Tennis Elbow? What Clinical Trials Show
- PRP for Achilles Tendinopathy: What Does the Evidence Say?
- PRP vs. Hyaluronic Acid Injections for Knee Osteoarthritis: What Does Research Show?
- What Are the Risks and Side Effects of PRP Injections?
Leave a note on this stop