Does PRP work for shoulder pain? Better than cortisone for a few months

Whether PRP works for shoulder pain has been tested twice against the standard corticosteroid injection, the anti-inflammatory steroid used as routine care, both times double-blind, meaning neither the patients nor the staff scoring them knew which injection had been given. Both trials favoured PRP in the first months. In 100 adults with an irritated supraspinatus tendon, one of the four rotator cuff tendons that hold the shoulder in its socket, the PRP group reported less pain at one year than the corticosteroid group - 1.68 against 2.3 on the trial's pain rating, where lower means less pain [2].

That trial also counted how often each injection simply did not hold, which it defined as pain still bad enough at three months to need another one. It happened to 12% of the PRP group and 30% of the steroid group - about 1 in 8 against nearly 1 in 3 [2]. Shoulder function, scored on a questionnaire where higher is better, finished at 89.8 after PRP against 78.0 after the steroid [2]. These were young shoulders: everyone was between 18 and 50, averaging about 28.

The second trial complicates the one-year picture. It followed 99 people with tendon irritation or a partial-thickness rotator cuff tear, meaning a tear through part of the tendon's thickness rather than all of it. At three months pain had fallen by 13.6 points in the PRP group while the steroid group's rose by 0.4, on the scale that had started them at 46 and 35 [3]. There was no difference at six weeks, before that gap opened, and none at twelve months, after it closed - and the same share of each group went on to agree to surgery [3].

Pooling more trials keeps the direction and blurs the detail. A meta-analysis, which combines the results of earlier trials into one estimate, covered 17 randomised trials of PRP for common shoulder conditions, randomised meaning a coin toss decided who got which treatment and found better pain relief and function than the comparison treatments from 8 to 12 weeks out to a year or more, against placebo injections and corticosteroids alike [4]. A review of 8 trials in rotator cuff tendinopathy, the irritated-tendon condition, reached the same conclusion for long-term pain and function, though its trials used different PRP recipes and different comparisons - four against a saline injection, which is salt water with nothing active in it, and four against rehabilitation or dry needling, where a needle goes in but nothing is injected [1]. For early frozen shoulder, where the capsule around the joint becomes stiff and painful, 3 trials in 177 people put PRP ahead of physiotherapy at twelve weeks, and PRP was the only one of four drug treatments tested that improved how far the shoulder could actually move [15].

Clinic records are less flattering than trials. A review looking back at 400 PRP treatments found that of the common reasons for treatment, rotator cuff injuries satisfied the fewest patients at two months and shoulder osteoarthritis the fewest at six [12].

PRP is made in the room. Blood is drawn, spun in a centrifuge until the platelets settle into a concentrated layer, and that layer is injected into the painful tissue [10]. Platelets do more than help blood clot: as they break open they release growth factors, the signalling proteins that tell nearby cells to multiply and build tissue, among them vascular endothelial growth factor and platelet-derived growth factor, both of which drive the growth of new blood vessels [10].

Blood supply is the plausible reason this would matter in a shoulder. Tendon tissue carries little of it, so repair signals and nutrients arrive slowly. In the surgical trial, ultrasound of the repaired tendon showed more blood flow at the PRP repair site three months after the operation, and in the surrounding tissue for up to a year [9].

The same biology explains why the trials disagree. There is no single PRP: how concentrated the platelets are, how many times the blood is spun, whether white blood cells are kept in or filtered out, and whether the mixture is chemically activated before injection all differ between clinics [12]. The 17-trial analysis found its strongest pain relief at six to seven months in the trials that spun the blood twice, concentrated platelets at least fivefold, kept the white cells in and activated the mixture first [4].

The strongest head-to-head trial does not describe most people with shoulder pain. Its participants were 18 to 50 years old, averaged about 28, had their tendon problem confirmed on a scan, and had already failed to improve with conservative treatment [2]. That is a young athletic shoulder, not the worn cuff of a 70-year-old, and nothing here tests PRP in older or arthritic shoulders.

Where these studies are weak:

• In the partial-tear trial, randomisation left the PRP group worse off at the start on all three of its scores, which flatters any improvement it later showed [3].
• The 8-trial review rated one of its studies at high risk of bias and another unclear, and its trials compared PRP against three different things - saline, rehabilitation, dry needling - so better than usual care means something different in each [1].
• The frozen shoulder finding carries its own authors' warning about risk of bias, variation between studies and inconsistent results [15].
• The 400-treatment review looked back at records with no comparison group and measured only patient satisfaction, so it cannot separate the injection from time passing [12].

Two things this evidence does not say. The surgery trial used PRP as an add-on during keyhole repair of medium-to-large tears, where the tendon tore again within two years in 2 of 52 PRP patients against 10 of 50 without it - that is PRP alongside surgery, not instead of it [9]. And the most complete tendon-to-bone regrowth anywhere in this research was produced in rats, with a PRP-loaded gel that is not a treatment anyone can book [11].

Longevity and regenerative-medicine clinics offer PRP among their injection services, usually filed under regenerative medicine or orthobiologics, the clinic label for treatments made from a patient's own blood or tissue. For a shoulder, a first visit tends to start with imaging or a look at an existing scan, because both trials selected people whose tendon problem had been confirmed on ultrasound or magnetic resonance imaging rather than by examination alone [2][3]. The procedure is one appointment: blood taken from the arm, spun for a few minutes, then injected into the shoulder, in the trials under ultrasound guidance so the needle is placed by sight rather than by feel [3]. Two things are worth asking before booking - how the clinic prepares its plasma, since that varied between trials and tracked with how well it worked [4], and what the plan is if nothing has changed by three months, the point where both trials took their reading [2][3]. Whether PRP makes sense for a particular shoulder is a judgement for a doctor who has examined it and seen the images.