Why these four injections are not interchangeable

'I've been offered an injection for my knee — but which one?' is often the question patients arrive with after a referral that mentioned several options almost interchangeably. Corticosteroids, hyaluronic acid (HA), platelet-rich plasma (PRP), and polyacrylamide hydrogel (PAAG) are routinely grouped together as 'knee injections', but they work by entirely different mechanisms, carry different durations of effect, and sit in different regulatory categories.

• Corticosteroids suppress joint inflammation directly. The effect can onset within days, but is transient — lasting weeks rather than months.
• Hyaluronic acid supplements the viscosity of synovial fluid, restoring mechanical lubrication. Some formulations may provide benefit for several months.
• PRP delivers a concentrated mix of growth factors derived from the patient's own blood, aiming to modify the joint environment biologically. In some patients the effect may persist for months, though evidence is mixed.
• PAAG (Arthrosamid®) functions as a mechanical spacer that integrates into the synovial tissue. Crucially, it is regulated as a medical device rather than a pharmaceutical — which affects how its safety and efficacy data are generated and assessed.

No single randomised trial has yet placed all four options head-to-head in humans; the comparisons available in 2026 are built from separate trial programmes and network meta-analyses. Patient factors — including disease severity (Kellgren–Lawrence grade), age, and comorbidities such as diabetes — are increasingly shaping which option is considered, rather than any single universal first-line recommendation.

Corticosteroids and the speed-versus-duration trade-off

Speed matters acutely when a patient has significant knee swelling limiting daily function — and on that measure, corticosteroids have no equal among the available intra-articular options. A 2025 meta-analysis of nine RCTs confirmed that corticosteroid injection outperformed biological therapies at three months, establishing them as the fastest-acting option in the current comparison landscape.

That advantage narrows quickly. Pain relief typically peaks within the first few weeks and fades before the three-month mark for most patients. The field now treats this as a meaningful clinical limitation rather than a minor caveat: the MOTION RCT, a multicentre international trial running across at least 45 centres with a target of 264 participants, was powered specifically to establish whether a rival approach can achieve superior sustained benefit over corticosteroid at six months.

Extended-release formulations: a pharmacokinetic step forward

Standard triamcinolone acetonide (TA) reaches a peak plasma concentration of around 9,629 pg/mL within hours of injection, and synovial fluid levels become undetectable in most patients by Week 6. An extended-release microsphere formulation (FX006, 32 mg) changes that profile substantially: synovial drug levels remain quantifiable through Week 12, and peak plasma concentration falls to 836 pg/mL — roughly one-tenth the systemic exposure — from a single dose. Whether the prolonged joint residency translates into meaningfully superior clinical outcomes requires further validation, but the pharmacokinetic improvement is well-documented.

Cartilage safety and injection frequency

The question of how repeated corticosteroid injections affect articular cartilage over time is contested but not without signal. Some imaging studies have raised concern about cartilage volume loss associated with more frequent dosing, though trials have produced inconsistent findings and no reliable frequency threshold has been established. What clinical practice has broadly converged on is treating this as a frequency-dependent consideration: injection interval requires careful thought, particularly in younger patients or those with less advanced disease where cartilage preservation is a longer-term priority. A single well-timed injection for acute symptom burden sits in a different risk category from repeated courses over months or years.

Hyaluronic acid and the debate over how well it works

Hyaluronic acid has accumulated more clinical trial data than any of the other three options covered here — a track record built over decades of use in rheumatology and orthopaedics. Expert consensus statements, including Henrotin et al. (2015) and the Bannuru network meta-analysis published in the Annals of Internal Medicine the same year, continue to position viscosupplementation as an appropriate option for suitable patients with knee OA.

The debate over its placebo-adjusted effect is real, however, and worth understanding. Trials consistently show patients improving after HA injections — but when pooled across systematic reviews, the gap between HA and well-designed placebo (saline or sham) narrows to modest effect sizes. Those improvements are not artefacts; they are genuine. The question is whether they are large enough to be consistently clinically meaningful, and that is where reviewers part ways. Guideline bodies that have examined this data include HA as an option rather than removing it, reflecting both the evidence of benefit and recognition that response varies across patient groups.

One injection or five?

Patients are sometimes offered single, three-, or five-injection courses — which can be confusing when the rationale is not explained. A systematic review of 11 studies found no consistent difference in patient-reported outcomes between single- and multiple-injection formulations, and five-injection regimens were not demonstrably superior to three-injection courses. Earlier pooled data did suggest that single injections may be less effective than two-to-five injection courses over saline, so this question is not fully settled. In practice, most guidance defaults to three-injection protocols while the evidence on single-injection approaches matures — a position reinforced by the growing interest in reticulated HA derivatives engineered for longer joint residence.

Safety at scale

The safety record is reassuring. An analysis of nearly 700,000 real-world HA knee injections found severe acute localised reactions within three days in no more than 0.03% of cases, with neither avian origin nor cross-linking status confirmed as a consistent driver of that risk.

Duration of benefit typically falls in the range of several weeks to up to six months — placing HA in the middle ground between the shorter-acting corticosteroids and the longer-horizon options discussed in the sections that follow.

PRP and the gap between trial results and clinical enthusiasm

Evidence on PRP sits in an uncomfortable position: a high-quality blinded trial says it does not work, while a body of network meta-analyses says it may be the most effective option available. Both claims deserve scrutiny rather than a tidy resolution that does not exist.

The RESTORE trial, published in JAMA in 2021, is the most rigorous test of PRP to date. It enrolled 288 patients aged 50 and over with KL grade 2–3 knee OA and randomised them to three weekly injections of leukocyte-poor PRP or saline placebo, with participants, injectors, and outcome assessors all blinded. At 12 months, mean pain reduction was -2.1 in the PRP group and -1.8 in the placebo group — a difference that did not reach the minimum clinically important threshold of 1.8 points. MRI-assessed cartilage volume showed no significant difference between groups either.

Set against that, a 2025 Bayesian network meta-analysis of 14 RCTs (934 patients) comparing ultrasound-guided intra-articular injections ranked PRP highest across total WOMAC score (SUCRA 85.86%), stiffness (93.24%), and function (90.9%). A separate 2025 dosing NMA of 10 RCTs found three-injection PRP protocols outperformed single- and double-injection regimens at one, three, and six months for both VAS pain and WOMAC scores, with no major adverse events leading to discontinuation in any group.

The discrepancy is real, and formulation is the most likely explanation. RESTORE used leukocyte-poor PRP — a preparation that removes white blood cells to reduce inflammatory potential. Many trials in the NMAs used leukocyte-rich formulations, which may behave differently in synovial tissue. Platelet concentration, injection volume, patient age, KL grade at baseline, and whether ultrasound guidance was used also vary across studies, making direct comparison unreliable. In effect, 'PRP' is not a single treatment — it is a category covering biologically distinct preparations.

For patients, this means PRP cannot yet be described as a proven universal step in knee OA management. It may offer meaningful benefit in the right patient, with the right formulation, delivered under imaging guidance — but those conditions are not yet standardised enough to predict with confidence who will respond.

PAAG hydrogel as a single-injection option

Of the four options compared here, PAAG (Arthrosamid®) makes the most distinctive practical claim: a single injection, no repeat dosing, sustained relief over months or years. Its classification as a medical device — rather than a drug or biologic — reflects that it acts as a mechanical spacer within the joint space, a mechanism already set apart from the pharmacological options covered in earlier sections.

The largest available human dataset is a 24-month PROMs-based cohort study of 314 knees, which found significant improvements in VAS pain scores, Oxford Knee Score, and Lysholm scores from a single injection, sustained across the full follow-up period. Outcomes were consistently better in older patients, those with lower Kellgren-Lawrence grade, those without diabetes, and those with bilateral OA. A 12-month open-label study by Bliddal et al. (J Orthop Surg Res, 2024) adds supporting signals on efficacy and safety, though open-label design limits certainty about the size of the true treatment effect.

The cohort data also carry figures that require stating clearly. During follow-up, 49 of the 314 knees (roughly 16%) ultimately proceeded to total knee replacement, with higher KL grade the strongest predictor — a reminder that a single injection does not halt disease progression in every patient. Complications were recorded in 155 of the 314 knees, approaching half the cohort. The nature of those events is not fully characterised in the available data. Whether they predominantly reflect injection-site reactions and transient pain flares — the kinds of events commonly logged in post-injection cohorts — or more clinically significant occurrences is not specified in the published abstract, and that gap matters for how the figure should be read.

Most critically, PAAG has not yet been tested against a placebo in a blinded human trial, which means its true benefit-versus-placebo effect remains uncertain. An equine head-to-head trial showed a meaningfully higher lameness-free rate versus triamcinolone acetonide and sodium hyaluronate at six and twelve weeks, but findings from equine carpal joints do not translate directly to human knee OA. A placebo-controlled RCT in humans remains the evidence gap.

Which injection suits which patient

No single option leads for every patient. The evidence points to four practical variables that shape which injection is most appropriate: how quickly symptoms need to improve, what the Kellgren-Lawrence grade shows on imaging, age and comorbidities, and whether ultrasound-guided delivery is available.

Kellgren-Lawrence grade shapes the likely ceiling for benefit. Both the PAAG 24-month cohort data and PRP network meta-analyses report better outcomes at lower KL grades (2–3); patients with more advanced structural change are less likely to see durable results from any injection option. Corticosteroids remain a reasonable choice across the full grade range when rapid symptom relief is the clinical priority.

Symptom urgency is a legitimate clinical variable. Patients managing an acute flare, or needing meaningful pain reduction before a specific event, are better served by corticosteroids — the fastest-acting option in the comparison. Those seeking longer-term symptom management without moving to surgery may find HA, PRP, or PAAG more relevant, depending on their imaging and comorbidity profile.

Age and comorbidity add further nuance. The PAAG cohort found consistently better outcomes in older, non-diabetic patients; repeated corticosteroid use warrants particular care in anyone with diabetes or systemic steroid sensitivity.

Injection guidance also matters. Ultrasound-guided delivery is associated with better outcomes for PRP in particular, and is increasingly used across all injectables. It is reasonable to ask whether image-guided injection is available before committing to a treatment course.

What the evidence does not yet support is a universal first-choice recommendation. Patient stratification by KL grade, age, comorbidity, and symptom profile is now the central framework for injection selection — and that assessment requires a clinical review of your specific history and imaging rather than a generalised ranking.

1. [1] Polyacrylamide hydrogel injections in knee osteoarthritis: A PROMs-based 24 month cohort study. (2025). https://doi.org/10.1016/j.jcot.2025.103136 https://doi.org/10.1016/j.jcot.2025.103136
2. [2] Efficacy of ultrasound-guided intra-articular injection in the treatment of knee osteoarthritis in early and middle stages: a network meta-analysis. (2025). https://doi.org/10.3389/fmed.2025.1700950 https://doi.org/10.3389/fmed.2025.1700950
3. [3] A Double-Blinded Positive Control Study Comparing the Relative Efficacy of 2.5% Polyacrylamide Hydrogel (PAAG) Against Triamcinolone Acetonide (TA) And Sodium Hyaluronate (HA) in the Management of Middle Carpal Joint Lameness in Racing Thoroughbreds. (2021). https://doi.org/10.1016/j.jevs.2021.103780 https://doi.org/10.1016/j.jevs.2021.103780
4. [4] Risk of Severe Acute Localized Reactions for Different Intraarticular Hyaluronic Acid Knee Injections in a Real-World Setting. (2021). https://doi.org/10.1177/19476035211025815 https://doi.org/10.1177/19476035211025815
5. [5] Synovial and systemic pharmacokinetics of triamcinolone acetonide following intra-articular injection of an extended-release microsphere-based formulation (FX006) or standard crystalline suspension in patients with knee OA. (2018). https://doi.org/10.1016/j.joca.2017.10.003 https://doi.org/10.1016/j.joca.2017.10.003
6. [6] Comparative efficacy of different doses of platelet-rich plasma injection in the treatment of knee osteoarthritis: a systematic review and network meta-analysis. (2025). https://doi.org/10.1186/s13018-025-05650-1 https://doi.org/10.1186/s13018-025-05650-1
7. [7] Intra-Articular Stromal Vascular Fraction and Mesenchymal Stem Cell Injections Show Variable Efficacy and Higher Potential Complications Compared to Corticosteroid and Hyaluronic Acid in Treatment of Knee Osteoarthritis: Meta-Analysis of RCTs. (2025). https://doi.org/10.1016/j.arthro.2025.01.050 https://doi.org/10.1016/j.arthro.2025.01.050
8. [8] Effect of Intra-articular Platelet-Rich Plasma vs Placebo Injection on Pain and Medial Tibial Cartilage Volume in Patients With Knee Osteoarthritis: The RESTORE Randomized Clinical Trial. (2021). https://doi.org/10.1001/jama.2021.19415 https://doi.org/10.1001/jama.2021.19415
9. [9] MOTION Study Protocol: Genicular Artery Embolization vs Corticosteroid Injections for Symptomatic Knee Osteoarthritis. (2025). https://doi.org/10.1007/s00270-025-03994-z https://doi.org/10.1007/s00270-025-03994-z