PRP for Thumb Arthritis: A Clear Pain Signal, and One Result That Did Not Move

Arthritis at the base of the thumb is one of the most disruptive forms of osteoarthritis a person can have, and one of the least discussed. It is the joint you use to turn a key, open a jar, hold a pen, pinch a zipper. When it hurts, the tasks it interferes with are constant and small and hard to avoid. Patients often tell me they can tolerate the pain itself but not how often it interrupts them.

The joint is called the carpometacarpal joint, usually shortened to CMC, and you will also see it called the basal joint. A 2025 meta-analysis looked at what platelet-rich plasma does for it. This post reports what that paper found, including one result that did not improve, which I think is the most useful thing in it. If you want the underlying mechanism explained first, our general guide to PRP for joint pain covers how it is drawn, spun and delivered.

What the paper is, and how small it is

The review is "Safety and efficacy of platelet-rich plasma injections in basal thumb osteoarthritis; should we offer it or not?" by El Sewify and colleagues, published in the Journal of Hand and Microsurgery in February 2025. It is a systematic review with a meta-analysis, pooling seven articles covering 115 patients and 129 affected joints.

I want to be direct about the size of that. Individual studies ranged from 2 to 24 patients. A two-patient study is close to a case report. In my own view, pooling seven small studies does not produce the evidentiary weight of one large randomized trial, and the authors do note that the low patient numbers prevented them from running subgroup analysis. Mean follow-up was 14.1 months, which the authors say does not support long-term recommendations.

The treatment itself, as pooled across those studies, was modest: an average of 1.3 mL per injection, an average of 1.4 injections per joint, given about 3.2 weeks apart. That is a small volume into a small joint, and typically one or two sessions rather than a long course.

What improved

Pain improved, and it was the strongest signal in the paper. Across five studies covering 98 patients, the meta-analysis found a statistically significant pain reduction, with a standardized mean difference of 2.2 and a 95% confidence interval of 2.0 to 2.3. That is a large effect with a narrow interval.

Hand function improved as well. Across four studies covering 64 patients, DASH and QuickDASH scores decreased significantly, with a standardized mean difference of 0.7 and a confidence interval of 0.1 to 1.2. DASH stands for Disabilities of the Arm, Shoulder and Hand, and it is a questionnaire rather than a measurement, which matters when we get to the next section. Overall patient satisfaction across the pooled studies was 73.7 percent, or 56 of 76 patients.

The result that did not move

Grip strength did not improve. In the 39 patients where it was measured, the standardized mean difference was 0.1 with a confidence interval running from negative 1.1 to positive 1.2. That interval crosses zero, which means the data cannot distinguish the result from no effect at all.

Pinch strength did improve, with a standardized mean difference of 0.3 and a confidence interval of 0.1 to 0.5. Significant, but a small effect, and noticeably smaller than the pain effect.

I am drawing this out because it is the honest shape of the finding and it is what a patient actually needs to plan around. The measures that improved most are the ones a patient reports: how much it hurts, how much it interferes. The measure that did not improve is the one an examiner records with an instrument. That does not make the pain relief unreal. Pain is the reason people come in, and relieving it is a legitimate goal on its own. But if your hope is that an injection will restore the strength of the hand, this paper does not support that, and I would rather say so before a treatment than after one.

Against cortisone, in one study

The comparison most patients ask about is cortisone, because it is what is usually offered first. I have written more broadly about how PRP and cortisone differ.

Two of the included studies compared PRP against corticosteroid, and they did not agree. I am going to give you both, because giving you only the favourable one would be the easiest way to mislead you in this entire post.

Malahias and colleagues, in 2021, reported significantly better results for PRP than for corticosteroid on both VAS pain score, at a p value of 0.015, and Q-DASH score, at 0.025. The satisfaction difference at twelve months in that study was the striking part: 69 percent for PRP against 12.5 percent for corticosteroid, at a p value of 0.002.

Sabaah and colleagues, in 2020, found the opposite pattern. All groups improved at four weeks, but by twelve weeks the improvement was sustained only in the hyaluronic acid group. It was not sustained in the PRP group or in the corticosteroid group. In that trial PRP faded on the same timescale cortisone did, and the arm that held was the third one, hyaluronic acid.

There is a third result that belongs here for the same reason. Winter and colleagues, in 2023, found the largest pain reduction with combined fat and PRP against a saline group, at a p value of 0.003. But in that same trial PRP on its own did less well than saline on pain, and no group improved in pinch or grip strength.

So the honest reading is that one trial favoured PRP strongly over cortisone, one found PRP fading like cortisone while hyaluronic acid held, and one found PRP alone underperforming saline. The review's own closing line asks for exactly this to be tested properly, and that request is the correct response to a picture this mixed.

Safety, as reported

No adverse events occurred across the five studies that reported safety. There was one complication, a palmar wrist ganglion, which the authors put at a 1.0 percent rate across 98 joints.

As with any small evidence base, that should be read for what it is. A pooled population of roughly one hundred joints can tell you that nothing common and serious happened. It cannot characterise rare events, and the 14.1 month mean follow-up does not speak to the long term.

Why this is not settled

The authors list their own limitations plainly, and they are worth repeating because they explain why two people can read this literature and come away with different impressions. Patient numbers were too low to permit subgroup analysis. The reported outcomes were heterogeneous, meaning the studies did not all measure the same things the same way. There was no standardized PRP preparation protocol across studies. Imaging guidance varied between ultrasound, fluoroscopy and X-ray. And the mean follow-up, by their assessment, does not support long-term recommendations.

The preparation point is the same one that runs through the whole PRP literature, in the knee and the hip as much as the thumb. PRP is not a single standardized product. Two preparations can carry the same name and differ in what they contain, which is a fair thing to ask any provider about directly.

Their conclusion, in full, is that "Intra-articular PRP injections in thumb CMC OA yields favorable outcomes for pain relief and hand function without major complications. However, procedural and data heterogeneity affect reliability. Further randomized controlled trials comparing PRP and cortisone injections are needed."

How I read this

Conservative care comes first, as it does everywhere in this field. For the thumb that usually means activity modification, a thumb spica orthosis, hand therapy and appropriate use of oral medication. Those approaches change how the joint is loaded, which an injection does not do. Osteopathic manipulative treatment belongs in that same first tier, because the hand rarely works in isolation from the wrist, forearm and shoulder girdle above it.

For someone who has worked through those and still has pain that interrupts daily tasks, PRP is a reasonable thing to raise. What this paper supports is a discussion about pain and function, held alongside the fact that two of its seven studies did not favour PRP. What it does not support is a promise about grip strength, a claim that the effect is durable past a bit over a year, or a confident statement that PRP is the better choice than cortisone or hyaluronic acid, because the included trials disagree on exactly that point.

Whether it makes sense in your case depends on the stage of the arthritis, what your hand is asked to do, what has already been tried, and an examination. This post cannot make that call and you should not make it from what you read here.