What each injection actually does

Three injections, three entirely different mechanisms — and that difference explains why disease stage drives the choice.

PRP (platelet-rich plasma) is drawn from the patient's own blood, concentrated to raise the platelet count, and injected into the joint. The platelets release growth factors that modulate the joint environment — dampening inflammation and signalling tissue repair — rather than replacing or rebuilding what is lost. Because it works through the body's own biology, PRP is best understood as a biological modulator.

Hyaluronic acid (HA) is a viscosupplement. The knee joint normally contains hyaluronic acid as a core component of synovial fluid; in osteoarthritis, this fluid thins and loses its lubricating quality. HA injections aim to restore that viscosity and cushioning effect. The mechanism is physical, not regenerative — HA does not stimulate repair or modify the underlying disease process.

Polyacrylamide hydrogel (iPAAG, sold as Arthrosamid®) takes a third approach entirely. A single 6 ml injection of this non-resorbable synthetic gel integrates into the synovial membrane, where it remains as a permanent mechanical cushion. It is not absorbed, not metabolised, and not regenerative — it augments joint structure rather than altering biology or supplementing fluid.

Thinking of these as interchangeable options misses the point. A biological modulator, a viscosupplement, and a structural implant target different problems at different stages of the same disease.

PRP versus hyaluronic acid: what the trials show

The most-cited data point in this comparison comes from Belk et al. 2021 — a meta-analysis now carrying over 500 citations — which found that PRP produced a mean WOMAC total score improvement of 44.7% against 12.6% for HA (P<0.01). That gap held across both the six- and twelve-month follow-up points, covering pain, stiffness, and physical function domains.

This was not an isolated result. Belk et al. 2023 and Li et al. 2025 (Arthroscopy Journal) independently confirmed consistently better outcomes with PRP compared with HA for knee OA patients, reinforcing the signal across separate research groups and time points.

HA, for its part, reliably reduces pain and improves function in mild-to-moderate disease — its evidence base is the most mature of the three options considered here. The practical limitation is duration: for most patients, relief tends to tail off before the six-month mark, and a second course becomes necessary. HA also carries a well-established safety record, broad insurance coverage in many health systems, and flexibility in dosing format — from multi-injection regimens to single high-molecular-weight preparations.

One complication to the straightforward PRP-versus-HA framing is worth noting: Du et al. 2025 meta-analysis found that combining the two — injecting PRP and HA together — achieves better pain relief and functional improvement than either agent used alone. This is an emerging rather than established standard-of-care finding, but it introduces a clinically actionable option for patients where both biological modulation and lubrication goals are in play simultaneously.

Arthrosamid's durability case and its trade-offs

For patients who find repeat clinic visits difficult — whether because of geography, time, or the cumulative cost of annual courses — a single injection that may last three to five years changes the calculation considerably.

Arthrosamid® (iPAAG) has been directly compared with HA in the Bliddal et al. 2024 randomised controlled trial. The headline finding was non-inferiority: WOMAC pain improved by 18.5 points in the iPAAG arm versus 14.8 points in the HA arm over one year. iPAAG did not outperform HA at twelve months — but it achieved comparable outcomes from one injection rather than a course of repeat doses. The trade-off that demands transparent reporting is the adverse device effect rate: 28.9% of iPAAG recipients experienced such effects compared with 7.6% in the HA group. Critically, all were described as mild and transient, and no serious device-related adverse events emerged in either the trial or the subsequent TACIT trial with its five-year extension data. That long-term dataset is iPAAG's primary durability differentiator — sustained symptom improvement across three to five years from a single procedure, a timeframe no HA course consistently matches.

Real-world data add breadth to the picture. The LUNA trial, NHS cohort studies, and a 2025 retrospective cohort by Aykaç et al. — comparing iPAAG directly against HA and corticosteroids — collectively extend the evidence beyond the controlled trial setting, though this remains a smaller and younger body of literature than HA or PRP have accumulated over decades.

The practical fit is moderate-to-severe established OA, where the non-resorbable, long-lasting profile of the hydrogel outweighs the preference for repeat injectable therapy. One regulatory note for completeness: iPAAG is approved in the UK and multiple other markets, but does not currently hold FDA approval in the United States — UK patients are not affected by this distinction.

Matching the injection to your OA stage

Stage determines fit more reliably than brand preference or clinical fashion. For patients with early or mild knee OA whose primary goal is supporting the joint's biological environment, PRP has the best evidence — Belk et al. 2021's 44.7% WOMAC improvement against 12.6% for HA reflects a consistent signal across multiple meta-analyses. Leukocyte-poor formulations with an optimised platelet count tend to perform best, and Du et al. 2025 adds an actionable refinement: combining PRP with HA outperforms either alone, making PRP+HA a reasonable consideration when both biological modulation and lubrication are goals simultaneously.

Where cost or insurance coverage is the deciding factor, HA remains the most accessible first step for mild-to-moderate disease. Its evidence base is mature, its safety record long-established, and coverage in NHS-adjacent and private insurance pathways is broader than for either PRP or iPAAG. The shorter duration — typically four to six months — is a known limitation, not a disqualifier, for patients who are well placed to return for repeat courses.

For established moderate-to-severe OA, and particularly where repeat clinic visits are a practical burden, iPAAG (Arthrosamid®) offers the strongest durability data. Bliddal et al. 2024 showed it non-inferior to HA at one year; TACIT trial five-year extension data extend that relief considerably further from a single injection. The higher adverse device effect rate (28.9% versus 7.6% for HA) remained mild and transient, but it belongs in any shared decision-making conversation.

Specialist grading of OA severity — not self-reported stage — is what determines which of these pathways applies. The three options are not interchangeable across disease progression.

Why PRP results vary so widely

The label 'PRP' covers a wider range of preparations than most patients realise — and that variability explains much of the inconsistency seen across published trials.

Three variables matter most. First, platelet concentration: the absolute count in the final product influences how much growth-factor activity reaches the joint, and getting this wrong in either direction affects outcome. Second, leukocyte content: preparations are broadly split into leukocyte-poor (LP-PRP, low white-cell) and leukocyte-rich (LR-PRP, high white-cell). In knee OA specifically, LP-PRP is generally preferred — high white-cell formulations may increase post-injection inflammation in an already sensitised joint. Third, injection frequency: a single-dose protocol and a course of three injections are both described as 'PRP' in the literature, yet they are not equivalent, and meta-analyses that pool across these approaches can obscure what any individual formulation actually achieves.

Belk et al. 2021 and Belk et al. 2023 both reflect this challenge: the strong aggregate results they report come from trials using varied protocols, so no single 'standard' PRP product can claim that evidence by default.

The practical implication is straightforward rather than alarming. PRP's evidence base is genuine and consistent — the variability is a quality-assurance and informed-consent question, not a reason to dismiss the therapy.

Three questions worth asking any clinic offering PRP

• What is the target platelet count in the final injection, and how is it verified?
• Is the formulation leukocyte-poor or leukocyte-rich, and why has that choice been made for knee OA?
• How many injections are proposed, and which trial protocols does the regimen most closely follow?

What the evidence still cannot tell you

Several comparisons in this article rest on indirect evidence — that is worth stating plainly. No single trial has run PRP, HA, and iPAAG against each other in the same patient population at the same time. The rankings outlined in the staging section above come from pooling separate trials with different patient groups, different timepoints, and different outcome measures. That is not a flaw in the evidence so much as a reflection of where clinical research currently stands.

Two further limits deserve naming. First, no injection therapy has demonstrated proven structural cartilage repair in rigorous long-term imaging trials. Symptom improvement and tissue regeneration are distinct outcomes — trials have measured the first consistently; the second remains unestablished for all three agents. Improved pain scores do not confirm a rebuilt joint surface. Second, iPAAG's evidence base, though growing steadily since the TACIT trial and Bliddal et al. 2024, is younger and smaller in volume than HA's decades of accumulated trial data; comparative safety beyond five years remains limited.

None of this unsettles the staging logic above. What it does confirm is that OA severity, injection history, and how much a patient values single-visit convenience over a longer-established evidence record all pull in different directions — and those trade-offs cannot be resolved by a meta-analysis average. That is where individual assessment earns its place.