The question most patients bring to clinic

Most patients arrive in clinic having already done their research. They come in holding a question that feels simple: should they have a steroid injection or a PRP jab? They have read blogs, watched videos, and received conflicting advice — and they want a straight answer, not another brochure.

The difficulty is that 'which injection' is the wrong starting question. The right one is: what does this joint actually need, and what is still biologically possible before we talk about replacement?

Joint preservation is the clinical frame that matters. Every intra-articular decision — what to inject, when to inject it, and how many times — has to be weighed against the question of whether it is buying meaningful function, slowing degeneration, or simply deferring an inevitable conversation. Choosing a treatment because it is cheaper, more familiar, or more heavily marketed is not a clinical strategy; it is a default.

Both corticosteroids and PRP are legitimate tools in the right clinical context. Neither is universally superior. The decision that determines which is appropriate — if either is — rests on a structured assessment of OA stage, symptom acuity, cartilage status, and what the patient realistically needs the joint to do over the coming months and years.

What corticosteroid injections actually do — and where they fall short

Corticosteroids earn their place in the intra-articular toolkit precisely because they work quickly and reliably. Triamcinolone, betamethasone, and methylprednisolone all act by broadly suppressing the inflammatory cascade — cytokine release slows, synovial swelling reduces, and pain often diminishes within days. That speed, combined with low cost and routine NHS availability, makes them a reasonable first-line choice for specific clinical scenarios: an acute inflammatory flare that is limiting function, a patient who needs rapid mobility to engage with physiotherapy, or a short-term bridge when surgery is planned but some weeks away.

The limitations are real, however, and the evidence is specific enough to merit precision rather than alarm. ESSKA has graded the evidence for chondrotoxicity as Grade A — systematic review data from Wernecke et al. and single-dose experimental work from Dragoo et al. both document adverse effects on articular cartilage. With repeated injections, some evidence suggests the risk of accelerating cartilage degeneration rises further. That does not make a single, well-timed corticosteroid injection categorically contraindicated — it makes repeated or poorly sequenced use harder to justify when joint preservation is the goal.

The preoperative timing issue carries its own clinical weight. A meta-analysis of 28 studies involving 480,532 patients found a significantly raised periprosthetic joint infection risk (OR 1.55) when corticosteroid injection preceded hip arthroplasty within three months. Separately, injections within two weeks of shoulder arthroscopy have been associated with elevated postoperative infection rates. When surgery is on the horizon, that sequencing constraint must be factored into the decision — the injection that provided welcome short-term relief could complicate the operation it was meant to precede.

In tendinopathy, the position is clearer still: the current evidence does not support corticosteroid use. Atrophy rates of approximately 50% have been documented in Achilles tendinopathy, and DaCruz et al. found that 39% of patients eventually required operative intervention — a rate that argues against routine use in any tendon condition.

How PRP works and why the evidence is promising but not yet definitive

Unlike corticosteroids, which suppress inflammation broadly, PRP works by concentrating the patient's own growth factors — principally PDGF, TGF-β, and VEGF — and delivering them into the joint at therapeutic levels. The mechanism is modulatory rather than suppressive: the aim is to shift the biological environment towards repair and tissue homeostasis, not simply to damp down the inflammatory response. Because the product is autologous, the systemic risk profile is low; the clinical questions are about efficacy and durability rather than systemic safety.

The trial evidence broadly supports a two-phase picture. In the short term — typically the first five to six weeks — PRP and corticosteroid produce comparable outcomes on pain and function measures. A prospective RCT of 40 patients with Kellgren–Lawrence grade II–III knee OA found no significant difference between PRP (8 mL intra-articular) and triamcinolone acetonide on VAS, IKDC, or KSS scores up to week five. After that point, the trajectories diverge: PRP-treated patients continued to improve, while those who received corticosteroid began to show recurrence of symptoms. By twelve months, PRP showed sustained superiority across pain and functional outcomes. A 2025 RCT corroborated this pattern, reporting lower VAS scores in the PRP group at every measured interval from one month to twelve, along with greater WOMAC improvement at six and twelve months. A 2024 triple-blinded RCT in chronic zygapophyseal joint OA followed the same arc — corticosteroid led at one month, parity at three, PRP superior at six.

The clinical implication is straightforward: for patients who need durability rather than immediate relief, the evidence favours PRP across the medium term. A 2025 systematic review of six RCTs published between 2019 and 2024 confirmed that both treatments are safe and effective, but reached no consensus on definitive superiority — a distinction worth preserving, because the evidence is real but not yet settled.

The principal reason blanket claims remain premature is formulation variability. PRP is not a single standardised product: leucocyte content, platelet concentration, activation method, and dosing schedule differ across studies and clinical settings, and each variable affects cytokine-release kinetics and, potentially, clinical outcomes. Leukocyte-rich and leukocyte-poor preparations show broadly comparable benefit in knee OA trials, but the heterogeneity across protocols limits the conclusions that can be drawn from any pooled analysis. Most of the comparative trial data also comes from KL grade II–III populations; the evidence thins considerably for more advanced disease.

Combination strategies — PRP alongside hyaluronic acid, or with an initial corticosteroid dose — are under active evaluation. A 2026 RCT in KL grade II–III patients found that PRP plasma gel with HA and CS produced the most consistent six-month outcomes across the arms tested, though no statistically significant differences remained after adjustment. That result warrants attention in selected patients, but the evidence is early and the approach is not yet standard practice.

Matching injection choice to disease stage

The clinical decision starts with staging. Kellgren–Lawrence grading and cartilage status are the primary variables that determine which injection, if any, is appropriate — the same treatment does not suit every disease stage.

For KL grade II–III (mild-to-moderate OA), the evidence supports PRP as the structurally appropriate choice when durable relief is the goal. A prospective RCT of 40 patients with KL II–III knee OA found equivalent outcomes between PRP and triamcinolone acetonide up to five weeks, after which PRP showed diverging superiority through twelve months on VAS, IKDC, and KSS — a pattern consistent across multiple subsequent trials. Corticosteroid remains a legitimate option within this group when rapid pain control is the immediate priority: for a patient who needs short-term comfort to engage with a rehabilitation programme, a well-timed CS injection can be exactly the right tool. The clinical question is always what the injection is being asked to achieve.

An acute inflammatory flare shifts the calculus regardless of OA grade. CS's speed of action — meaningful symptom reduction within days — makes it a useful bridge towards rehabilitation or further workup, not a substitute for either.

At KL grade 3 and beyond, the picture changes materially. PRP evidence thins considerably at this stage; most comparative trial data comes from grade II–III populations, leaving advanced disease poorly evidenced for either treatment. When articular damage is severe, neither CS nor PRP is likely to provide durable relief, and surgical timing becomes the honest conversation. For patients who are not yet candidates for replacement or are actively bridging the gap, cell-based options such as autologous adipose-derived mesenchymal stem cells (ADMSC) represent a distinct next tier beyond PRP — not a direct alternative to CS or PRP within the same treatment sequence, but a separate specialist pathway for a different point on the disease curve. A phase III RCT of 261 patients with KL grade 3 OA found ADMSC produced significantly greater VAS improvement (25.2 vs 15.5, p=0.004) and WOMAC improvement (21.7 vs 14.3, p=0.002) over placebo at six months. These treatments remain specialist-access and are not yet standard practice.

In tendinopathy, the position is cleaner: corticosteroid is not supported by current evidence for most tendon pathologies. PRP combined with structured load-based rehabilitation is the appropriate pairing.

No injection buys unlimited time. When structural degeneration has advanced beyond what any injectable can meaningfully address, the right discussion is about surgical timing — and deferring that conversation serves neither the joint nor the patient's long-term outcome.

Timing, sequencing, and what injection choice means if surgery is ahead

When joint replacement is under active consideration, the injection decision becomes a sequencing problem as much as a clinical one.

The elevated infection risk associated with corticosteroid injections before arthroplasty — the evidence for which was set out earlier — is not a caveat to weigh against short-term benefit. Within the three months before hip replacement, it functions as a hard timing constraint that must enter the discussion as soon as surgery becomes a realistic possibility. Repeated CS injections in a joint already heading towards replacement carry a second cost: the cumulative chondrotoxicity documented by Wernecke et al. and Dragoo et al. is not theoretical. There is limited clinical logic in accelerating the cartilage damage that replacement is intended to address.

PRP does not carry the same preoperative infection signal. No documented elevation in perioperative risk corresponds to PRP in the way the meta-analysis data does for CS, which makes it a more defensible option in the months approaching a potential surgical window. That said, formal guidance on how close to surgery PRP can be administered — and whether it meaningfully alters the operative environment — has not been established by dedicated study. The inference that it is safer in this window follows from its autologous, non-immunosuppressive mechanism; it is not yet supported by the same evidence depth as the restriction on CS.

For any patient where replacement is a realistic possibility within twelve months, the injection conversation needs reframing. The relevant question is not which treatment performs best in isolation, but which — if either — can provide meaningful symptom control without narrowing the surgical options or creating avoidable operative risk. On occasion the most responsible answer is that injection is not the right next step at all: a course of targeted analgesia and rehabilitation, followed by timely surgical planning, may serve the joint better than an intervention chosen to postpone a necessary decision.

Sequencing — not product selection — is the discipline that matters here.

What realistic outcomes look like — and when to stop injecting

Four to eight weeks of meaningful pain relief is what a well-executed corticosteroid injection reliably delivers in appropriately selected patients — no more should routinely be expected, and for a specific subset, that window is precisely what they need. A patient completing a supervised rehabilitation programme needs pain control for twelve to sixteen weeks; CS may not bridge the full course. A patient facing a significant life event needs six weeks of reduced symptoms; it very likely will.

PRP's timeline is different, and this matters for the consent conversation. The first four to six weeks after injection are often indistinguishable from CS in terms of how the patient experiences them — the growth-factor cascade takes time to modify the synovial environment. Patients who expect immediate dramatic relief from PRP are frequently disappointed at week three and quietly reassured by week twelve. Setting that expectation beforehand, not after, is part of what an honest consultation looks like.

On repeat cycles: the evidence base for sequential PRP courses across multiple years is genuinely thin, and this is not a caveat to hide. Similarly, repeated corticosteroid injections carry accumulating chondrotoxic risk that is not theoretical — the Wernecke et al. systematic review and Grade A ESSKA consensus evidence make this position defensible. Neither agent is a long-term maintenance strategy for a progressively deteriorating joint, and presenting either as one does the patient a disservice.

Perhaps the most underused diagnostic signal in injection medicine is the non-response. A patient who receives little benefit from a correctly placed, imaging-guided CS injection — in a joint that on clinical grounds ought to have responded — is communicating something important about the underlying structural state. That information deserves more clinical weight than it typically receives. It is not automatically a reason to try a different injection; it may be a reason to revisit the imaging, reassess disease stage, and ask whether continuing intra-articular treatment in any form is serving the joint or simply deferring a decision the evidence is already pointing towards.

1. [1] Comparison of Corticosteroid and PRP Injections for Symptomatic Relief in Knee OA: A Randomized Controlled Study (IJSR, 2025). (2025). https://doi.org/10.36106/ijsr/2500538 https://doi.org/10.36106/ijsr/2500538
2. [2] PRP Plasma Gel versus Hyaluronic Acid versus Corticosteroid for Knee OA: A Prospective Randomized Clinical Trial (2026). (2026). https://doi.org/10.63403/re.v33i2.480 https://doi.org/10.63403/re.v33i2.480
3. [3] Clinical Efficacy and Safety of Intra-articular Autologous Adipose-Derived MSCs for Knee OA: Phase III RCT (Am J Sports Med, 2023). (2023). https://doi.org/10.1177/03635465231179223 https://doi.org/10.1177/03635465231179223
4. [4] Efficacy and Safety of Intra-articular PRP vs Corticosteroid for Knee OA: Systematic Review of RCTs (Cureus, 2025). (2025). https://doi.org/10.7759/cureus.80948 https://doi.org/10.7759/cureus.80948