Read the literature in the right order

Two filters do most of the work, and neither is special pleading.

Does PRP work for knee osteoarthritis?

The consensus layer came first. ESSKA-ICRS and ESSKA-ORBIT both published consensus statements in 2024; AAPM&R issued a guidance statement; AAOS produced a technology overview and a clinical practice guideline on nonarthroplasty management.

The recent pooled data has moved with it.

• Liu and colleagues (2026) found PRP significantly reduced pain and improved function against placebo across short, mid and long-term follow-up, with the magnitude varying but remaining clinically relevant.
• Bensa and colleagues (2025) concluded the improvement with PRP in knee osteoarthritis reaches clinical significance, not merely statistical significance.
• Jawanda and colleagues (2024) found PRP, bone marrow aspirate concentrate and hyaluronic acid all outperformed corticosteroid — the comparator that is routinely covered.
• Belk and colleagues (2021) pooled randomized trials of PRP against hyaluronic acid.
• Patel and colleagues (2024), in a triple-blind randomized trial, showed a higher platelet dose beat a conventional dose on both pain and function.

The honest qualifier belongs here rather than in a footnote. Han and colleagues (2026) reported that intra-articular orthobiologics, including PRP, did not exceed the minimal clinically important difference when compared with hyaluronic acid at up to twelve months. That is a comparison against another active injection, not against placebo, and it sits alongside Liu’s placebo-controlled finding rather than canceling it. Both are real; they answer different questions.

Does PRP work for tendon injuries like tennis elbow?

Lateral epicondylitis has the cleanest tendon result. Xu and colleagues (2024) found PRP produced better long-term functional improvement and pain relief than the comparators studied.

Rotator cuff tendinopathy has changed. Feltri and colleagues concluded in 2023 that PRP did not improve clinical results in rotator cuff disorders. In 2026 Yuwarungsikul and colleagues, at Level I, found PRP produced statistically significant if clinically modest improvements in pain and shoulder function over corticosteroid, with fewer adverse events at six months, and read the findings as suggesting PRP may be the more durable option. The newer, higher-level analysis is the one to quote, and the older one should be acknowledged rather than hidden.

Achilles tendinopathy is genuinely mixed. Barreto and colleagues (2025) pooled randomized trials favorably; Assi and colleagues (2026) found no significant clinical or radiological benefit against placebo at three or six months. We state both. This practice does not present Achilles tendinopathy as a settled indication.

Does PRP work for low back pain?

ASIPP’s 2025 update is the document to quote for spine, and it is worth quoting carefully because it is candid about its own limits: 119 pages, 35 authors, 33 in the formal consensus, appraised with GRADE, 19 recommendations agreed at 100% consensus.

Quote them at that level and nowhere above it. ASIPP’s own background says regenerative medicine remains in early stages of clinical validation and calls for careful optimism, and names a scarcity of high-quality studies as its limitation. A society that publishes unanimous recommendations while naming the weakness of its own evidence base is doing evidence appraisal properly — and it is the opposite of what a payer’s medical policy does.

Since that guideline, Wang and colleagues (2025) reported a randomized trial in discogenic low back pain in which PRP significantly reduced pain and improved lumbar function against a methylene blue comparator, with imaging changes on Pfirrmann grade and diffusion. Pei and colleagues (2026) found PRP reduced low back pain and pain-related disability to six months, with consistent improvement in non-discogenic pain and inconsistent longer-term discogenic results, and cautioned that heterogeneity across studies limits the reading — which is the preparation problem again.

Where does the evidence not support PRP?

These belong on the page. Leaving them off would make everything above less credible, and a reviewing expert will find them anyway.

A practice that recommends orthobiologics for every joint is not reading this literature. We do not offer them for hip osteoarthritis on the strength of the knee data.

How to use this in a case

The move that holds up is narrow: establish the indication, cite only the evidence for that indication, state its level honestly, and ask what the opposing evidence actually measured. A moderate recommendation at Level III, accurately quoted, survives cross-examination. A confident claim built on the wrong body region does not.

This practice is not aligned to a side. Its obligation runs to the patient’s health and safety, which is the same reason it is not beholden to any position that would keep it from seeing what is actually there. We do not offer opinions on liability, we do not accept a referral or a review conditioned on reaching a particular conclusion, and we will say plainly when the evidence does not support the claim — whichever party was hoping otherwise.

Frequently asked questions

Is the evidence strong enough to call this standard care?

For knee osteoarthritis it is supported by multiple 2024 society consensus statements and by placebo-controlled pooled data through 2026 finding clinically relevant benefit. For the lumbar spine ASIPP grades it Level III to IV with moderate to low consensus recommendations. Those are different claims and are made separately.

Why do some meta-analyses still find no benefit?

Usually because of what they pooled. Dose and leukocyte content change outcomes and much of the older literature did not report either — see dose and preparation — and because a null against hyaluronic acid is not a null against placebo.

Should the negative studies be volunteered?

Yes. Naming the hip and Achilles findings before the responding party does is what makes the favorable citations believable — see records that survive cross.

Does any of this establish causation in a case?

No. Whether a treatment works is separate from whether a collision caused the condition being treated. That is causation and apportionment.

Does PRP work?

It depends on what it is used for. The evidence is strongest in knee osteoarthritis, where 2024 consensus statements and placebo-controlled pooled data favor PRP, and lateral epicondylitis (tennis elbow) has the cleanest tendon result. The lumbar spine carries a real recommendation at a modest level. Hip osteoarthritis and Achilles tendinopathy are where the evidence is unfavorable or unsettled, and this practice says so.

Is PRP better than a cortisone shot?

In the comparisons on this page, yes. Jawanda and colleagues (2024) found PRP, bone marrow aspirate concentrate and hyaluronic acid all outperformed corticosteroid in knee osteoarthritis. In rotator cuff tendinopathy, a 2026 Level I analysis found PRP produced statistically significant, if modest, gains in pain and shoulder function over corticosteroid, with fewer adverse events at six months.

Related reading

• Dose and preparation
• Orthobiologics: PRP and BMAC
• A denial is a budgeting decision
• Future care and damages
• Expert testimony