What deferral actually means — and what injections can realistically deliver

For patients who have been told they will eventually need a knee replacement, the practical question is rarely "can I avoid surgery forever?" — it is more often "how long can I reasonably put this off?" That is a question the evidence can actually address, at least in part.

Injections can delay the need for total knee replacement (TKR) in some patients. That outcome — deferral — is documented in real-world data across more than one treatment type. But the word "delay" is doing serious work in that sentence. No injection changes the underlying structural damage of osteoarthritis: cartilage does not regenerate, bone-on-bone narrowing does not reverse. The mechanism is symptom control and functional improvement sustained long enough to push back the point at which surgery becomes unavoidable.

Patient selection shapes the outcome as much as the choice of injection. In mild-to-moderate OA, where some joint space and cartilage remain, injections have a meaningful biological target. In end-stage disease — severe joint-space loss, bone-on-bone contact — the evidence for any meaningful response is weak across all modalities.

The quality of that evidence also differs sharply between agents, and understanding those differences matters as much as the deferral figures themselves.

Hyaluronic acid: the clearest evidence for delay

Among all the injectable options, hyaluronic acid (HA) — also called viscosupplementation — carries the most direct published data on time to surgery.

The clearest single figure comes from a US insurance claims database of 182,022 knee OA patients, all of whom eventually had a total knee replacement. Those who had received at least one course of HA injections waited a median of 484 days before surgery; those who received no HA waited a median of 114 days — a gap of roughly one year (Altman 2015). A dose-response relationship was also documented in the same dataset: patients who completed multiple HA courses waited longer than those who received only one, with some population-level data from UK and French health databases pointing to average delays of one to three years overall, and up to five to seven years in specific patient subgroups.

In practical terms, a repeat-injection plan — typically one course every six to twelve months — is a viable treatment strategy rather than a one-off attempt.

A Cochrane systematic review (CD005321) and subsequent network meta-analyses support HA for symptomatic relief in mild-to-moderate osteoarthritis, reinforcing its role as a tool for managing the knee while deferral remains realistic.

The important caveat is that every piece of TKR-delay data for HA is observational. Confounding by indication — the likelihood that patients offered injections tended to have less severe disease to begin with — cannot be excluded without a randomised trial specifically designed to measure delay as its primary outcome. The figures are meaningful, but they are not proof of a guaranteed deferral for any individual patient.

PRP: the most striking deferral claims — and why RCTs complicate them

Platelet-rich plasma (PRP) produces the most striking deferral numbers in the published literature — and also the most contested.

A retrospective survival analysis by Sánchez et al. (2021) followed knee OA patients treated with PRP and reported that roughly 74% delayed total knee replacement by more than 1.5 years. The median deferral was 5.3 years, and more than 85% of patients had avoided surgery altogether during a five-year follow-up period. Those figures are the largest deferral claims attached to any injection modality.

The difficulty is that they come from retrospective data, not a randomised trial. A large JAMA-published RCT (Bennell 2021, n=288) tested PRP against saline placebo and found no statistically significant difference in structural cartilage improvement — specifically, medial cartilage volume — at 12 months. This is a serious challenge to any claim that PRP slows structural deterioration. One plausible explanation for the gap is patient selection bias: patients who choose PRP in observational studies may be earlier in their disease course, more physically active, or more engaged with their own care than those who do not — all factors that independently predict slower progression toward surgery.

PRP is not, however, without supportive evidence. Comparative studies suggest it outperforms HA for pain relief at six and twelve months (Murali 2024), which is clinically meaningful even if structural modification remains unproven. Health-economic modelling at the Cleveland Clinic, using a Markov decision model, indicates PRP can be cost-effective as a deferral strategy — but only when the delay is actually achieved, making the economic case contingent on the same uncertain benefit.

The honest summary is that PRP remains under active investigation: promising for symptom control, intriguing in observational delay data, but not yet supported by randomised trial evidence of structural benefit.

Corticosteroids: effective pain bridge, not a delay strategy

Corticosteroids sit in a different category from both HA and PRP. They are not deferral tools — and the evidence is unambiguous on this point.

Among all injectable treatments for knee OA, corticosteroids have the strongest randomised controlled trial evidence for short-term pain relief. The relief is real but brief, typically lasting four to six weeks. No clinical evidence shows that corticosteroid injections delay the eventual need for total knee replacement; they reduce symptoms while the underlying joint remains in the same structural state. Their role is bridging — managing acute pain spikes or buying time before a more sustained treatment can take effect.

Where corticosteroids raise a distinct and practical concern is surgical timing. Injections administered within 90 days of a planned knee replacement are associated with up to a 40% increase in periprosthetic joint infection (PJI) risk, with the highest exposure window at zero to four weeks before the procedure. Major orthopaedic centres, including Hospital for Special Surgery, mandate a minimum 90-day gap between a corticosteroid injection and planned TKR as a result.

For patients already on a surgical waiting list — or who expect to be listed soon — this has a direct scheduling consequence: a routine injection for pain relief could require postponing a procedure that is already booked. Discussing recent injection history with the operating surgeon before any new injection, or before surgery is confirmed, is a step that is easily overlooked and worth raising proactively.

Arthrosamid and longer-duration options: what early evidence suggests

Arthrosamid® (polyacrylamide hydrogel) works on a different principle from every other option covered so far. Rather than introducing a biological agent or lubricant that the body metabolises, it acts as a physical scaffold — a space-filling cushion that integrates with synovial tissue and modifies the joint's mechanical environment.

Published open-label data report approximately 2–3 years of symptom relief from a single injection in suitable patients — a longer reported interval than typical HA or PRP courses. The important caveat is that this figure comes from uncontrolled, open-label studies; no randomised controlled trial has yet evaluated Arthrosamid against a comparator, and no head-to-head data against HA or PRP currently exist. The 2–3 year figure is a useful signal, not a confirmed benchmark.

In clinical practice, it tends to be considered for patients who have not achieved adequate or lasting relief from HA, or those looking to extend the gap between active treatments while they weigh their surgical options. The single-injection format is also relevant for patients who find repeated injection courses burdensome. Suitability still depends on OA severity and overall joint status — a specialist assessment is needed to establish whether it is appropriate.

Who is most likely to benefit — and how to think about treatment planning

Mapping these options onto a patient's situation is less about choosing a single "best" injection and more about matching modality to purpose at a given point in the disease course.

For patients at the mild-to-moderate stage primarily trying to push the surgical date out, the evidence hierarchy is reasonably clear: hyaluronic acid carries the only large-scale data directly measuring time to surgery, and repeated courses appear to extend that window beyond what a single injection achieves. PRP shows superior pain reduction at 6–12 months compared with HA (Murali 2024), which may make it a stronger choice where symptom control is the priority — though the absence of RCT-level structural evidence means that retrospective deferral figures should not be taken at face value. Arthrosamid represents a longer-duration option for patients who have gained only short-lived relief from earlier agents; published open-label data suggest roughly 2–3 years of benefit per injection, though randomised evidence remains pending.

Corticosteroids sit outside this deferral logic entirely. Their relevance here is principally the scheduling implication: a patient approaching a surgical waiting list must account for the 90-day pre-surgery gap before receiving one.

None of these treatments operates in isolation. Physiotherapy, weight management, and activity adaptation all affect how long an injection's symptomatic benefit holds — and how much functional runway it actually buys. A specialist assessment covering imaging, functional status, and symptom trajectory is the step that establishes which combination is appropriate for a given patient's joint condition and timeline.