Does Platelet-Rich Plasma Injection Work?

Give your blood … to yourself!

Blood therapy, anyone? You can inject your own blood into an injury to “increase circulation” in an almost brutal way, delivering everything that blood has to offer at high concentrations right where healing needs to happen. This is the injection of “autologous whole blood” (AWB). Platelet-rich plasma (PRP) gets a little fancier: the blood is processed to make a concentrated mixture dominated by platelets, a blood ingredient that releases more growth factors. Either way, the hope is that a bloody injection will stimulate healing where it is otherwise failing — one of the simplest kinds of “regenerative” medicine, especially for stubborn, slow-motion injuries like tendinitis.1

It’s easy to pay someone to do this for you these days: extract some of your blood, spin it in a centrifuge to get the platelets, and then pump them back into you. PRP is not cheap, but PRP injections have become popular, particularly with elite athletes — ever the guinea pigs for unproven, expensive new treatments for musculoskeletal injuries. It seems perfect for injuries like patellofemoral pain, an extremely common pseudo-arthritis of the knee in runners,2 or IT band syndrome, another kind of common runner’s knee, or incredibly common conditions like plantar fasciitis or frozen shoulder — so the market is immense. In 2009, ScienceBasedMedicine.org scathingly criticized the marketing of PRP:3

Without any clear evidence of benefit beyond placebo, PRP is now being marketed aggressively as a cure-all for sports injuries. And at about $300 per injection (the NYT reports $2000/treatment), there’s plenty of money to be made. … a nation-wide marketing initiative has begun, using sports celebrities as guinea pigs.

A Case Study In Aggressive Quackery Marketing, Jones (ScienceBasedMedicine.org)

At that time, the problem was that the marketing was irresponsible in light of the lack of evidence. It was a short wait for more. Alas, in the 2020s, the marketing is more at odds with the unimpressive evidence we now do now have …

Cynics can stop reading here — you know how this ends! Despite a lot of allegedly “promising” evidence over the years, both AWB and PRP remain experimental, supported almost exclusively by low quality trials with mixed and generally unimpressive results. I will summarize all the highlights from the literature, but the bottom line is clear: if PRP or AWB work at all, it’s tricky and unreliable, probably diluted and polluted by a lot of exotic variables we don’t understand and cannot control — as with most chronic painful conditions.

And these injections can also hurt a lot! One reader described it as “among the worst experiences of my life, definitely surpassing childbirth.”

Relationship to stem cell therapy

You can also inject injuries with stem cells, which are generic cells that do not yet have a job. They produce a lot of chemical signals related to growth and development and healing, and they could even become what we need them to be, a potentially powerful tool in medicine. Stem cell therapy is a broad concept in regenerative medicine, and it is a hot hot hot topic right now. It’s identical in spirit to the other two main regenerative therapies in musculoskeletal and pain medicine: platelet-rich plasma and autologous chondrocyte implantation. But neither PRP or ACI is technically a stem cell therapy — they use mature, specialized cells, so they are just cell therapies.

The theme of regeneration unites them, and the topic of stem cell therapy overlaps so much with PRP and ACI that they are practically the same thing with regards to safety, efficacy, and the concerns of skeptical consumers and regulators. They are all biologically intriguing treatments that might amount to something someday — after all, we know regeneration is possible, thanks to salamanders! — but it’s a depressing mess so far, instead of being inspiring and promising. These treatments are all being rushed to market in the same way, all sold as high-tech medicine to desperate consumers long before the science is done.

Why platelets?

Meet the Clotters! Platelets are the major clotting tool in your blood, and they are curious critters, neither cells nor molecules, but a strange hybrid often called “cell fragments”: platelets are to blood cells what wood chips are to a log … if the chips were extremely clever. Platelets definitely do have a bunch of fascinating biological features, but they are best known for their work in clotting — and that’s mainly what gives them that healing mystique.

There are countless biochemical factors that regulate healing — it’s complex, to say the least. Platelets are part of that equation, playing “a critical role in tissue repair and regeneration”; specifically they “regulate fundamental mechanisms involved in the healing process including cellular migration, proliferation, and angiogenesis.”4

Since they are involved in healing, more of them must be better, right?

That is the rationale for PRP. That’s how sophisticated it is: more is better.

The ruffled white one in the middle is a platelet — an “activated” platelet, specifically. When calm, they are smaller and smoother.

In fact, PRP is often called “regenerative medicine,” because the idea of genuinely accelerated healing is so tantalizing, happy science fiction. But it’s more marketing than advanced biomedical science, surprise surprise — we don’t really have a good understanding of the relationship between platelets and healing.5 You could probably talk people into drinking a platelet smoothie if you told them it would “regenerate” them.

But it is not safe to assume a soup of platelets is regenerative. In fact, it’s not even safe to assume it’s safe …

Myotoxic? Myo-maybe! Safety issues with PRP

Health Canada isn’t a fan, and notably they have safety concerns (as of 2019):6

Many emerging autologous cell therapy products may eventually prove to be safe and effective. However, most of these products are currently at the investigational stage of development with an on-going need to gather supporting scientific evidence.

Injecting medications into muscles might not be harmless. (No one’s surprised by that, right? Good!) Anaesthetics and NSAIDs probably are a little myotoxic — poisonous to muscles — and there’s “conflicting evidence” about PRP.7 It might be fine, but it’s important to bear in mind that faddish new injection treatments are never risk free.

Note that autologous whole blood injection is probably safer (but the evidence is too thin to say).

Like a salad spinner

To make PRP, blood is spun in a centrifuge. Different blood components separate into layers.

Why would PRP by myotoxic? The possibility highlights our near perfect ignorance of the principle of PRP therapy. Who says more platelets are going to stimulate healing? Is that in the Platelet User Guide? “For extra healing, generously apply platelets to wound.” Dosage is critical with many medicines. More is not only not always better, it’s often worse. Do other cells like being bathed in ten times the normal number of platelets? Or is it a suffocating mess that throws everything off kilter?

Or is it just kind of ho hum?

In PRP marketing and hype, it’s common to see claims that it’s a “natural” treatment — because it’s your own blood being returned to you, see? — and what could be safer and healthier than you-stuff? But this is bio-illogical: there’s lots of stuff inside of me that I do not want to be extracted, concentrated, and returned! Pick any hormone, for instance: many of those are just as involved in healing as platelets, but too much of most of them is just a disaster.

For instance, there is a disease of excessive iron, hemochromatosis — a major component of red blood cells, essential to life, something you could easily think you want a lot of for vitality and healing. And indeed you do, if you’re anemic. But chronically absorb too much, and it’s a serious disease.

What you want in biology is just-right amounts of everything, rarely lots of extra anything. It’s really quite odd to assume that a platelet-rich sauce o’ blood is natural and safe and helpful just because the stuff came from you. Which is why this treatment needs to be tested, not assumed — like every treatment.

The overly optimistic expert

In a 2018 podcast, I listened to a really credible expert guest boldly state that “there’s very strong evidence for platelet-rich plasma in osteoarthritis.” Thinking I must have missed something, I checked the primary reference in the show notes, which is also still the most recent review of PRP for osteoarthritis. Did it back up the expert?

Not even close!8 As far as I can tell, there is just zero justification for what she said: it’s not “very strong,” it’s actually very weak. It exists, and it’s technically positive, but it’s just not compelling.

I have come across many others like her over the years. Now, let's look at that evidence…

Science says “probably not but it’s hard to tell”

PRP fans and purveyors will tell you there is good evidence that PRP works, but they are cherry picking from a few studies that worked out in their favour one way or another. A few positive studies never mean much; indeed, many “positive” study results are … not.9

Taken as a whole, the PRP evidence is perpetually adrift somewhere between inconclusive and discouraging (unless you sell the stuff) . Although more research is needed (of course!) enough decent studies have now been done that the evidence reviews have started to come out. They all warn that most of the evidence is poor quality, and they are all basing their conclusions on just not nearly enough good data. They all emphasize that PRP methods are not standardized — there are many versions of PRP, all based on speculation, not data.

The bad news got rolling in 2010, a few years after it started to get popular. Gina Kolata (a trusted science journalist) for The New York Times:10

Now, though, the first rigorous study asking whether the platelet injections actually work finds they are no more effective than saltwater.

Nothing has improved significantly since. Highlights from the more recent literature, including noteworthy individual trials and some major reviews, for a variety of conditions:

• Moraes et al. found “insufficient evidence to support the use of PRT for treating musculoskeletal soft tissue injuries.”11
• Bell et al. concluded that PRP had “no additional benefit in the treatment of mid-portion Achilles tendinopathy.”12 And de Vos et al was super negative regarding tendinopathy:13

  “Three high-quality studies (75%) and two low-quality studies showed no significant benefit at the final follow-up measurement or predefined primary outcome score when compared with a control group. One high-quality study (25%) showed a beneficial effect of a PRP injection when compared with a corticosteroid injection (corticosteroid injections are harmful in tendinopathy). Based on the best evidence synthesis, there is strong evidence that PRP injections are not efficacious in chronic lateral epicondylar tendinopathy.”

• One of the only really good trials done so far — just a straight up randomized controlled trial, adequately powered, published in the New England Journal of Medicine in 2014 — was just completely negative. PRP for muscle strains did not beat a placebo, period.14
• In 2014, Sandrey found “strong evidence” that PRP does not improve plantar fasciitis when combined with several other therapies, and limited evidence that it might be beneficial on its own.15

• The two least lousy trials of PRP for plantar fasciitis in recent history … have positive results and seem well-designed. And that's where most people would stop! Unfortunately, one of them is a weirdo, with baffling flaws and too-good-to-be-true results I just don’t believe,16 and the other is less flawed but also delivers less to believe, merely mediocre results even if true.17

  On paper, these studies from 2016 and 2019 are as good as it gets, and (as of 2025) they are still the only placebo-controlled trials of PRP for plantar fasciitis. And that is the only kind of trial that matters if you really want to know what works. Big scientific reviews are relatively useless, always crippled by “garbage in, garbage out.” These two studies are the cream of the crop. Which ain’t saying much, unfortunately. It’s important to get this right — I cannot scoff at the results of well-designed trials without carefully "showing my work" — so I’ve explained the problems in great detail in the huge footnotes here (basically whole blog posts in each of those last two footnotes; PainSci members got a podcast episode that also went all the way into those weeds).

  Bottom line for PRP for plantar fasciitis: neither study delivers what PRP fans want to see, and there’s nothing better for plantar fasciitis. PRP for plantar fasciitis is not based on good evidence — because there simply is no good evidence to base it on.

• A 2018 review in Current Reviews In Musculoskeletal Medicine is both positive and negative, emphasizing how variable results are depending on the condition: some mildly encouraging results for some things (tennis elbow, patellar tendinopathy, plantar fasciitis), but definitely not others (Achilles tendinitis, muscle strain, fractures, and ACL tears).18

• In early 2018, the journal Sports Medicine piled on with a review of six (crappy) studies of PRP for muscle injury (“pulled” muscles, strains).19 Even the optimistic expert I mentioned above expressed her doubts about PRP for muscle injury:

  “The promising biological rationale, the positive preclinical findings, and the successful early clinical experience of PRP injections are not confirmed by the recent high-level RCTs.”

• A 2018 meta-analysis of PRP for tendon and ligament healing20 was technically positive, but with small effect sizes and this weak-sauce conclusion: “may reduce the pain.” And the authors (predictably) emphasize the “significance” (statistical) of their results rather than the effect size.
• A 2019 meta-analysis of PRP for knee arthritis is underwhelming in exactly the same ways:21 technically positive, but not really trustworthy, and unimpressive in any case: PRP “might be more effective” than hyaluronic acid injections.22 Be still my beating heart.

Any good news? Any hope of any kind?

Most of the good news is coming from isolated or flawed studies. Isolated positive evidence about over-hyped treatments is a huge red flag, which usually means “researchers made errors in their favour.” It’s the pattern of evidence that counts, and so far the pattern is distinctly bad.

Any hope? Maybe a little. There are different ways of doing PRP, and there are different conditions in different stages may respond better or worse. It’s biologically plausible that PRP could fail with chronic tendinitis but still succeed with an acute muscle strain, for instance, or even fail with one kind of muscle strain and succeed with another. Hammond et al., an experiment on rats — rats were harmed and treated for our edification — reported a difference between two kinds of muscle strain. It worked better on a more serious injury, where regeneration of muscle tissue was part of the healing process. PRP might assist with that regenerative process, but have no effect on a less serious strain where no regeneration is occurring.23

But these are faint hopes. In general, one would hope that the methods and conditions tested so far are at least in shouting distance of being the right formula — close enough to be at least a little more encouraging.

Initially promising in principle, I predict that PRP will now be mired in trumped-up controversy for years. It will die a slow death, only beaten into submission over many years by a growing pile of underwhelming evidence, while its proponents continue to overconfidently sell the service and defend it from detractors, mainly by betting — with dwindling odds — that just the right formula can still be proven effective for just the right kind of patient. If so, great: I will be pleased to admit that my prediction was wrong! But I’m betting against them for now.

After the centrifuge treatment, platelets are separated from the other components of blood.

Platelets give good placebo

My final word on this topic has to be “placebo” — PRP is a perfect storm for it. It’s got everything! Bearing in mind that it’s been thoroughly demonstrated that people get stronger placebo effects from treatment features as trivial as a more potent pill colour …

• People expect injected medicines to be more powerful.
• People are also strongly influenced by what elite athletes are doing.
• And of course it’s both high-tech and “natural” — a rare combination. Usually these qualities are contradictory, but PRP is blessed by both.

I can hardly imagine a better formula for a powerful expectation effect or “relief from belief.” Unfortunately, despite placebo’s weirdly good reputation, its powers are quite limited.24 The next time you hear a positive anecdote about PRP, remember: it’s probably the placebo talking.

Is it worth a try anyway?

The bar for “worth a try” is fairly high for normal folks. No invasive treatment can qualify for it without being proven at least safe. And you really need clear, consistent evidence of non-trivial benefit across several good trials before anything injected is “worth a try.” Before that it’s more like “hey, it’s your knee, don’t stab it”!

The equation is always different for elite athletes, of course: the slightest edge could be a big deal. But that sword cuts both ways! It might help just a little, and that might matter a lot… or it might hurt just a little, and that might matter a lot. Every athlete and coach is going to have a different opinion on whether that risk is worth it.

Did you find this article useful? Interesting? Maybe notice how there’s not much content like this on the Internet? That’s because it’s crazy hard to make it pay. Please support (very) independent science journalism with a donation. See the donation page for more information & options.

About Paul Ingraham

I am a science writer in Vancouver, Canada. I was a Registered Massage Therapist for a decade and the assistant editor of ScienceBasedMedicine.org for several years. I’ve had many injuries as a runner and ultimate player, and I’ve been a chronic pain patient myself since 2015. Full bio. See you on Facebook or Twitter., or subscribe:

SUBSCRIBE

Get PainSci updates in your inbox, mostly small, 0–2/week depending on coffee dosage. More info.

Notes

1. “Tendinitis” versus “tendonitis”: Both spellings are acceptable these days, but the first is the more legitimate, while the second is just an old misspelling that has become acceptable only through popular use, which is a thing that happens in English. The word is based on the Latin “tendo” which has a genitive singular form of tendinis, and a combining form that is therefore tendin. (Source: Stedmans Electronic Medical Dictionary.)

   “Tendinitis” vs “tendinopathy”: Both are acceptable labels for ticked off tendons. Tendinopathy (and tendinosis) are often used to avoid the implication of inflammation that is baked into the term tendinitis, because the condition involves no signs of gross, acute inflammation. However, recent research has shown that inflammation is actually there, it’s just not obvious. So tendinitis remains a fair label, and much more familiar to patients to boot.

2. It’s not really an arthritis at all, but it is often perceived that way because of the way it feels — a nagging ache — and a loose correlation with a slight degeneration of the kneecap cartilage (chondromalacia patellae). But, because it is perceived as being arthritis-y, it is a popular target for PRP.
3. ScienceBasedMedicine.org [Internet]. Jones V. A Case Study In Aggressive Quackery Marketing; 2009 Oct 22 [cited 12 Mar 9]. PainSci Bibliography 55587 ❐
4. Gawaz M, Vogel S. Platelets in tissue repair: control of apoptosis and interactions with regenerative cells. Blood. 2013 Oct;122(15):2550–4. PubMed 23963043 ❐
5. Gawaz et al. writes (emphasis mine): “Control of apoptosis/cell survival and interaction with progenitor cells, which are clinically relevant but poorly understood aspects of platelets in tissue repair, will be highlighted in this review. Gaining deeper insight into the less well-characterized molecular mechanisms is necessary to develop new therapeutic platelet-based options.” Deeper insight is necessary? An understatement!
6. www.canada.ca [Internet]. Health Canada. Health Canada Policy Position Paper: Autologous Cell Therapy Products; 2019 May 15 [cited 19 May 31]. PainSci Bibliography 52439 ❐
7. Reurink G, Goudswaard GJ, Moen MH, et al. Platelet-rich plasma injections in acute muscle injury. N Engl J Med. 2014 Jun;370(26):2546–7. PubMed 24963588 ❐

   No one knows for sure exactly how toxic, but permanent/serious harm is unlikely. Complete recovery should be possible in most cases. That was my reasoning before I did some reading, and then my reading confirmed it. Reurink et al.:

   Although the extent of muscular damage differs between specific agents and is dose dependent, the time course and the histological changes that are observed appear to be rather uniform. Within minutes after injection, hypercontraction of the fibres occurs, followed by degeneration of fibres, oedema and infiltration of inflammatory cells and necrosis over the following days. The pathogenesis of the myotoxicity is complex and not completely understood. The pathological mechanism is thought to be an interaction of local anaesthetics with the sarcoplasmic reticulum channels, initiating increased intracellular Ca2? levels and subsequent Ca2?-activated pathways of cell death. Initially, the histological damage appears severe, but the necrosis of fibres is reversible, and muscle regeneration usually occurs within 2–5 weeks.

   So it’s counterproductive for sure, but lasting damage would probably only occur in patients who got an unusually large dose and/or who had some special biological vulnerability.

8. Mascarenhas R, Saltzman BM, Fortier LA, Cole BJ. Role of platelet-rich plasma in articular cartilage injury and disease. J Knee Surg. 2015 Feb;28(1):3–10. PubMed 25068847 ❐

   The full text of this paper concludes:

   Recent systematic reviews on the topic conclude that there is still a paucity of high-quality data providing sufficient evidence to support or disprove the clinical utility of PRP in symptomatic osteoarthritis of the knee. There is even less clinical evidence supporting its use in other joints or in the treatment of focal osteochondral defects despite the basic science evidence in favor of its use. In addition, not all basic science and clinical studies on PRP have concluded it has positive effects.

   So garbage in, garbage out, no real conclusions possible: not enough good data even for the knee, even less for other joints. And there’s contradictory evidence.

9. Ingraham. Most Pain Treatments Damned With Faint Praise: Most controversial and alternative therapies are fighting over scraps of “positive” scientific evidence that damn them with the faint praise of small effect sizes that cannot impress. PainScience.com. 4417 words. Many seemingly evidence-based treatments for pain and injury are actually supported by evidence that damns with faint praise: results showing only a tiny benefit, which is usually not even real. Frustratingly, unimpressive evidence like this is routinely integrated into scientific reviews and clinical guidelines, which carry more weight with the uninformed. This pattern is practically standard in alternative medicine, sports medicine, and physical therapy.
10. New York Times [Internet]. Kolata G. Popular Blood Therapy May Not Work; 2010 Jan 12 [cited 20 Jan 25]. PainSci Bibliography 55584 ❐
11. Moraes VY, Lenza M, Tamaoki MJ, Faloppa F, Belloti JC. Platelet-rich therapies for musculoskeletal soft tissue injuries. Cochrane Database Syst Rev. 2014;4:CD010071. PubMed 24782334 ❐
12. Bell KJ, Fulcher ML, Rowlands DS, Kerse N. Impact of autologous blood injections in treatment of mid-portion Achilles tendinopathy: double blind randomised controlled trial. BMJ. 2013;346:f2310. PubMed 23599320 ❐ PainSci Bibliography 54578 ❐
13. de Vos RJ, Windt J, Weir A. Strong evidence against platelet-rich plasma injections for chronic lateral epicondylar tendinopathy: a systematic review. Br J Sports Med. 2014 Jun;48(12):952–6. PubMed 24563387 ❐
14. Reurink 2014, op. cit. In this randomized trial involving athletes with acute hamstring muscle injuries, injection of platelet-rich plasma did not result in shortening of the time until patients could resume their sports activity or in a reduction in reinjury rates, as compared with placebo.
15. Sandrey MA. Autologous growth factor injections in chronic tendinopathy. J Athl Train. 2014 Jun;49(3):428–30. PubMed 24840581 ❐ PainSci Bibliography 53813 ❐
16. Mahindra P, Yamin M, Selhi HS, Singla S, Soni A. Chronic Plantar Fasciitis: Effect of Platelet-Rich Plasma, Corticosteroid, and Placebo. Orthopedics. 2016;39(2):e285–9. PubMed 26913766 ❐

    This 2016 trial seems like a big win for PRP from the right kind of science: a double-blind, randomized, and placebo-controlled design. But it is too seriously flawed to take seriously. Indeed, I was seriously irritated by the time I was done with it.

    Mahindra et al. sloppily tested PRP, corticosteroids, and placebo injections for seventy-five chronic plantar fasciitis patients. The authors present their work as a landmark study proving that PRP works much better than placebo, with mighty effect sizes that beat even steroids. They did design a great looking trial, but, alas, they did not conduct one. Bizarrely, they did not (statistically) test PRP versus placebo! This is such a major malfunction that it casts doubt on the whole thing. Maybe the paper should even have been retracted (although that normally takes more than methodological bloopers, not matter how big). Regardless, the peer reviewers sure seemed to be asleep on the job.

    Ironically, if they had done the right stats, it likely would have made PRP look great. But they didn’t, and it truly is a deal-breaker, a textbook example of the “within-group significance fallacy.”

    Without replication, this trial is best viewed as an outlier. It’s the worst kind of science in my opinion: the kind that looks way better than it actually is. Shiny junk. And I bet it has been extensively cited by people selling PRP.

    The details…

    The big news reported by Mahindra et al. is that both PRP and corticosteroids produced allegedly “large” (huge!) improvements at three months for both pain and function (using the visual analog score or VAS, and the American Orthopaedic Foot and Ankle Society Ankle-Hindfoot Score).

    The effect sizes reported here are striking! Both PRP and corticosteroid injections reduced pain from about 7.5/10 at baseline to ~2.5–3.5 at 3 months, and the function scores improved from ~50 to over 80. Those are impressive changes: a 4–5 point drop on VAS is well beyond the minimal clinically important reduction of at least 1.5 points often cited for body pain. Similarly, a ~30-point AOFAS improvement (on a 100-point scale) is clearly substantial.

    But they are only impressive if they’re real! At first glance, the results make PRP look not just “promising,” but hugely effective, actually superior to steroids (at least by one functional measure), but without steroid's notorious side effects. Unfortunately, there are several serious caveats here:

    • First, the difference between PRP and steroids at three months was marginal: just a few points on the function score, and not mirrored by the pain scores. Does this justify touting PRP as just as effective “or more effective”? Only barely.
    • Follow-up was quite short, just three months — and that’s not long enough to judge the true value of a treatment for a chronic condition well-known for coming and going in frustrating waves. Without six- or twelve-month data for PRP, we can’t say it’s an effective treatment — even if the short-term result here was believable.
    • The placebo group’s complete lack of improvement is surprising, if not downright suspicious. Placebo groups usually improve, and chronic plantar fasciitis is notorious for fluctuating, often improving spontaneously even after long symptomatic periods. Why would the natural history of such a condition entirely stall in 25 patients in the same three-month period? In pain studies, regression to the mean, patient expectations, and even subtle unblinding normally generate some placebo responsiveness. The flatlined placebo here is quite a strong hint of methodological or reporting artifacts.
    • The blinding is another concern. PRP required drawing nearly 30 mL of blood, versus just 5 mL in the other groups. Even if patients were shielded from seeing syringes, venipuncture differences are difficult to disguise. Was the promise of a “special” injection implicitly communicated? If blinding was compromised, expectancy effects could account for much of the benefit. Few things can corrupt trial results like a blinding flaw.
    • But the real whopper and deal-breaker is the statistical methods …

    A bizarre statistical glitch

    Mahindra et al. only ran between-group statistical tests on PRP vs steroids. They never did the most basic statistical test: does PRP outperform placebo? Of course it probably does, with effect sizes like that! But we don’t “eyeball” things in science, and the entire point of doing the stats properly is to tame the many ways that happy results can be an illusion.

    More thoroughly and technically: only simple t-tests were applied to repeated measures, inflating the chance of spurious significance (a good example of “p hacking”). No correction for multiple comparisons is reported. More technically stated: they failed to compare treatment, nothing, and placebo with a between-group omnibus analysis (e.g. ANOVA or mixed-effects model across all timepoints). This is a stark statistical error.

    And so perhaps it’s not surprising that the study claimed unbelievable p-values of “zero” (reported as P = 0.00, rounded down from 0.001) — an unheard of statistical significance given the noisy data of pain outcomes. That’s really only possible by declaring statistical significance of the wrong comparison.

    Ironically, if they had done the right comparison, it almost certainly would have worked in PRP’s favour, based on the data they’d collected. But that data also includes that strangely flat placebo effect — and a weaker placebo response makes PRP look relatively stronger. So even if they hadn’t stunk up the stats, the result may well have still have been inflated by another mechanism.

    •

    The authors frame their work as an important placebo-controlled trial, but the implausible lack of placebo effect, fragile blinding, badly broken statistical analysis, minimal sample, and short follow-up all cast serious doubts on the results. This is way more than the usual methodological nitpicking that can be done to any study. When PRP looks “highly effective” in such a small and badly flawed study, the much more parsimonious explanation is bias and methodological artifact.

17. Shetty SH, Dhond A, Arora M, Deore S. Platelet-Rich Plasma Has Better Long-Term Results Than Corticosteroids or Placebo for Chronic Plantar Fasciitis: Randomized Control Trial. J Foot Ankle Surg. 2019 Jan;58(1):42–46. PubMed 30448183 ❐

    In 2019, Shetty et al. compared platelet-rich plasma (PRP), corticosteroid, and placebo injections for 90 patients with chronic plantar fasciitis — one of the few three-arm RCTs of its kind, and with unusually long follow-up (18 months). On paper, that makes it a more ambitious and credible effort than most studies of PRP (or much of anything else in this business) — better in both principle and execution than Mahindra et al., which was aiming almost as high but missed by a lot more.

    And yet this superior study still has three real strikes against it:

    • At least one non-trivial statistical flaw.
    • A high risk of bias.
    • Results that — although undeniably good news — don’t exactly thrill even if they can be believed.

    So this is quite a bit more credible than Mahindra et al., but still has one clear blooper, and its results are less promising! It’s a study that puts a skeptic in an awkward position: reluctant to blow off a perfectly good positive result, but still thinking, “So … meh?” This is hardly the validation PRP fans want to see.

    Now, more detail…

    The headline results were that all three groups improved, but PRP won in the long term by a respectable amount.

    Steroids gave the biggest early boost (within weeks), but the benefit plateaued while PRP continued to show slightly better results through 6, 12, and 18 months. By the end, PRP had the lowest pain scores: 3 points more improvement than placebo, which is enough to matter. Function improved too (Roles & Maudsley score), though only 1.7 points better than placebo. Quality of life (SF-12) improved in all groups, but the most in PRP.

    Re-injection and surgery rates were lower for PRP than for steroids, and no complications were reported.

    The strengths here were good randomization with three arms, including placebo — rare in this field! The long-term follow-up is also quite unusual. And there aren’t any major red flags in the statistics.

    Despite all that goodness — and the quality really does stand out — the sample is still small (30 per arm), which makes it fragile to outliers or imbalance. And although blinding was decent, it wasn’t bulletproof: all injections contained lidocaine and were delivered by the same surgeon, but both injectors and patients could have developed hunches about whether they were delivering or “getting the good stuff.”

    And there are orange flags on the stats: the authors failed to mathematically ask whether there was there an overall difference in the course of outcomes between groups over 18 months. This is a substantial limitation, and arguably an “error”; it makes the analysis weaker than it should be, and the apparent “wins” at certain timepoints could well be mirages, thanks to a greater chance of spurious “significant” differences. So that's really not great. (But it also isn't anywhere near as glaring as the similar error made by Mahindra et al.)

    And of course hidden flaws are always a possibility, and more likely with a high risk of bias — which this study has. The team acknowledges “self-bias” (authors measuring their own outcomes) and “institutional bias” (they produce PRP in-house). While not necessarily a deal-breaker, that is exactly the kind of conflict of interest that lifts the eyebrows of any skeptic.

    No good skeptic should refuse to accept the results of a study just because of a risk of bias and some methodological weaknesses. The design and reporting here are good enough to take seriously, but the evidence is still limited by small numbers, bias and the possibility of hidden flaws — so skepticism remains warranted but outright dismissal would be unfair.

    My final caveat: it’s not the size of positive result that we really want to see. Yes, it does seem to be genuinely positive — but slow, modest improvement barely even registers as “promising.” Showing that PRP may outperform steroids over the long run (a year and a half!), while actually confirming that steroids can give quicker relief (albeit less durable), is not going to excite many plantar fasciitis patients.

18. Le ADK, Enweze L, DeBaun MR, Dragoo JL. Current Clinical Recommendations for Use of Platelet-Rich Plasma. Curr Rev Musculoskelet Med. 2018 Dec;11(4):624–634. PubMed 30353479 ❐ PainSci Bibliography 52517 ❐
19. Grassi A, Napoli F, Romandini I, et al. Is Platelet-Rich Plasma (PRP) Effective in the Treatment of Acute Muscle Injuries? A Systematic Review and Meta-Analysis. Sports Med. 2018 Jan. PubMed 29363053 ❐
20. Chen X, Jones IA, Park C, Vangsness Jr. The Efficacy of Platelet-Rich Plasma on Tendon and Ligament Healing: A Systematic Review and Meta-analysis With Bias Assessment. Am J Sports Med. 2018 07;46(8):2020–2032. PubMed 29268037 ❐ PainSci Bibliography 52535 ❐
21. Han Y, Huang H, Pan J, et al. Meta-analysis Comparing Platelet-Rich Plasma vs Hyaluronic Acid Injection in Patients with Knee Osteoarthritis. Pain Med. 2019 07;20(7):1418–1429. PubMed 30849177 ❐ PainSci Bibliography 52532 ❐
22. “Lube” injections — a treatment which used to be associated with all the same hype and hope as PRP, but which is now widely regarded as ineffective (though it is still being sold to patients!), and therefore a decent sham to compare something else to. See Should You Get A Lube Job for Your Arthritic Knee? Reviewing the science of injecting artificial synovial fluid, especially for patellofemoral pain.
23. Hammond JW, Hinton RY, Curl LA, Muriel JM, Lovering RM. Use of autologous platelet-rich plasma to treat muscle strain injuries. Am J Sports Med. 2009 Jun;37(6):1135–42. PubMed 19282509 ❐
24. Placebo is fascinating, but its “power” isn’t all it’s cracked up to be, not even for relief of purely subjective symptoms like pain: the power of belief is strictly limited and accounts for only a little of what we think of as “the” placebo effect, which is actually a collection of diverse nonspecific effects and research artifacts. For more information, see Placebo Power Hype: The placebo effect is fascinating, but its “power” isn’t all it’s cracked up to be.