When the knee starts to fail — and what can still be done

The knee rarely fails all at once. Pain arrives first — often after a long walk, a flight of stairs, or simply getting up in the morning. Stiffness follows. Range of motion narrows. Activities that once required no thought begin to require planning. Over months and years, what started as a nuisance becomes a persistent constraint on how a person moves through their life.

For the roughly 240 million people living with knee osteoarthritis worldwide, the underlying process is a gradual breakdown of cartilage — the smooth tissue that allows the joint to bear load and move freely. Once lost, cartilage does not rebuild itself. That biological reality has historically steered clinical conversations towards replacement. An experienced clinician's first obligation, however, is to ask what the joint still has to offer before that step is taken.

The organising logic of sound clinical practice here is a hierarchy: preserve, then repair, then regenerate, then replace. Every decision along that sequence is a question of timing and patient selection. Replacement surgery remains an important and sometimes right answer — but positioning it as the default bypasses options that, in appropriate patients at the right stage of disease, can meaningfully extend the life and function of the original joint.

Biological treatments, including platelet-rich plasma injections, sit within that hierarchy. Their clinical value depends entirely on where in the disease course they are used, and for whom. The sections that follow examine the evidence for PRP against that framework.

How PRP works and why platelet concentration matters

Preparing PRP begins with a routine blood draw from the patient. That blood is centrifuged to separate its components: red cells are removed, and what remains is a plasma fraction concentrated in platelets — typically four to five times the level found in whole blood. When those platelets are activated at the injection site, they release a range of growth factors, including platelet-derived growth factor (PDGF), transforming growth factor-beta (TGF-β), insulin-like growth factor-1 (IGF-1), and vascular endothelial growth factor (VEGF). These proteins modulate the local inflammatory environment, influence cell behaviour, and may support tissue repair signalling. More recently, research has identified variable concentrations of specialised pro-resolving mediators — including Maresin 1 and Resolvin D1 — within PRP preparations; higher levels of these mediators correlate with stronger anti-inflammatory effects, offering one explanation for why individual patient responses differ.

The preparation process is where clinical outcomes diverge. A 2025 Level I meta-analysis of 18 RCTs found that only high-platelet formulations — defined as at least 1,000,000 platelets per microlitre — consistently exceeded the minimal clinically important difference for pain relief. Lower-concentration preparations did not. A 2024 systematic review of 29 studies reinforced this dose-response pattern: arms achieving positive outcomes at six months used a mean platelet dose of approximately 5,500 × 10⁶, roughly double that of non-positive arms. Not all PRP is equivalent.

Leukocyte content adds a further variable. Leukocyte-rich and leukocyte-poor preparations produce broadly similar outcomes at twelve months, but leukocyte-rich PRP is associated with more post-injection pain and swelling — making leukocyte-poor formulations the more practical choice for intra-articular use.

It is also important to be clear about what PRP does not reliably do. It is a biological signal delivered to a damaged environment — primarily symptom-modulating rather than cartilage-regenerating. No current evidence establishes PRP as a means of reliably rebuilding lost cartilage tissue, and the absence of standardised preparation protocols across clinical settings means results vary in ways that population-level averages cannot fully capture.

What the trials actually show

The strongest available clinical evidence comes from a 2025 Level I meta-analysis of 18 randomised controlled trials involving nearly 2,000 patients. PRP was statistically superior to placebo on both pain (measured by VAS) and physical function (measured by WOMAC) at every follow-up point up to twelve months. The clinical significance threshold matters here: improvements exceeded the minimal clinically important difference — the margin below which a patient is unlikely to notice a meaningful change — for pain at three and six months, and for function at all timepoints tested. These are not marginal signals; they are also not uniform ones, and the formulation caveats established in the preceding section apply in full.

Against corticosteroids, PRP shows broadly comparable short-term efficacy, with some evidence of more sustained benefit at longer follow-up. A 2025 systematic review of six RCTs found no serious adverse events for PRP and no consensus on superiority over corticosteroids, but documented the possibility of more prolonged pain relief.

Imaging adds both promise and caution. A 2025 retrospective MRI cohort study of 161 patients and 221 knees found improved cartilage integrity scores and a significant reduction in synovial fluid volume at six months — but that reduction in synovial fluid correlated with functional improvement, not with pain relief, suggesting the two response mechanisms are distinct. A separate 3D-MRI study in medial knee OA found fewer than 20% of knees showed cartilage thickness gains in the zones most affected by disease, reinforcing that the primary benefit is symptomatic rather than structural.

A 2025 rat-model study found PRP appeared to aggravate synovial fibrosis; this has not been replicated in human tissue, but it is a reason to hold mechanistic confidence lightly.

The larger evidential problem is standardisation. A 2025 critical review identified PRP preparation heterogeneity and inadequate OA phenotyping as the principal reasons trial results contradict one another — and why population-level averages cannot be read as individual predictions for any particular patient.

Which patients are most likely to benefit

Patient selection is where clinical judgement matters most — and where honest conversations about expectations begin.

PRP is most likely to help patients with early-to-moderate knee osteoarthritis who retain meaningful joint space. Bone-on-bone disease — where cartilage loss is advanced and the mechanical environment is severely compromised — is unlikely to respond meaningfully. The biological signals PRP delivers need a viable tissue environment in which to act; when that environment has largely gone, there is too little to work with.

Age matters in the trial data. Younger patients consistently outperform older cohorts across multiple study populations, a pattern that most likely reflects residual biological capacity rather than age as a rigid exclusion criterion. This does not close the door on older patients, but it is a realistic input into any honest assessment.

Obesity complicates outcomes in a way worth stating plainly. A study of 91 patients with mild-to-moderate OA found that those with a BMI above 30 improved after three PRP injections — but at final follow-up their WOMAC functional scores were significantly worse than those of non-obese patients (17.8 versus 10.5, p=0.023). Obese patients also carried a 3.5 times higher likelihood of progressing to total knee arthroplasty. Short-term pain relief does not translate into equivalent medium-term function when mechanical load remains a persistent, unresolved problem.

On injection protocol, evidence consistently favours three injections over one. A single injection produces a more limited and less durable response; most structured clinical programmes space a course of three over several weeks, and that pattern is reflected in the better-performing arms of the comparative literature.

Where PRP sits within joint preservation — and what comes next

PRP sits at the earlier end of the joint preservation spectrum — appropriate when cartilage damage is present and the inflammatory environment is active, but the joint has not yet crossed into irreparable structural failure. At that stage, the clinical aim is to reduce inflammation, slow deterioration, and restore enough function to preserve the patient's options. It does not reconstruct; it does not replace lost cartilage. That distinction matters for managing expectations honestly.

When PRP delivers insufficient benefit — or when the clinical picture suggests that symptom management alone will not hold — the next question is whether more targeted biological repair is appropriate before replacement is considered. This is where the field has developed meaningfully in recent years, and where clinical judgement about the individual patient determines which rung of the ladder applies.

One development at this level is NanoACi (Non-Arthroscopic Needle-Delivered One-Stage Autologous Chondrogenic Injection), a surgeon-led technique developed by Professor Paul Lee. It combines autologous platelet-rich fibrin (PRF)with cartilage cells from a small ear-concha biopsy, delivered by needle in a single clinic session without theatre or general anaesthetic — addressing directly the patient burden associated with two-stage procedures such as ACI, which require two hospital admissions, two anaesthetics, and an extended interval while cells are cultured externally. The biological components individually carry an established evidence base; the full combined protocol is emerging rather than settled, and Professor Lee makes that distinction explicitly rather than eliding it. NanoACi is not a step beyond PRP for everyone; it is a different intervention for a different clinical indication, appropriate for selected patients where repair rather than symptom management is the objective.

None of this repositions joint replacement as an outcome to be avoided. Replacement, performed at the right time for the right patient, remains an excellent operation. The purpose of the preservation-first pathway is not to defer replacement indefinitely — it is to ensure that when replacement is chosen, it is chosen deliberately.

UK costs, NICE guidance, and what to ask before committing

NICE's position sets the context for every private-sector decision about PRP. Under guidance HTG497 (last reviewed January 2019, next review due 2026), PRP for knee osteoarthritis is not a standard recommended treatment — it is permissible only where the clinician operates under special clinical governance, consent, and audit arrangements. That regulatory framing explains why the procedure sits almost entirely in the private sector: it is not routinely commissioned on the NHS, and standard private health insurance policies rarely cover it.

The financial picture is straightforward. A single PRP injection costs £450–£3800 in the UK private market, with London specialist centres typically at the upper end of that range and regional centres tending to sit lower. A three-injection course — the protocol supported by the better-performing arms of the comparative literature — is usually packaged at £1,500–£2,800. For PRF is ~£2000-3000 Initial consultation fees (£20”–£350) and ultrasound guidance or pre-treatment imaging (£300–£700) may be additional if not bundled.

Before committing, the right questions to ask are:

• What platelet concentration will be used, and does it meet the ≥1,000,000/µL threshold linked to clinically meaningful pain relief?
• How many injections are planned, and over what schedule?
• What does success look like in measurable terms, and over what timeframe?
• If the course produces insufficient benefit, what is the next clinical step?

How a clinician answers those questions reveals as much as any price list. The specifics of platelet dose and injection count are not administrative details — they are the core variables on which the evidence turns, and the NICE review due in 2026 is unlikely to change that.

1. [1] PRP Injections for the Treatment of Knee Osteoarthritis: The Improvement Is Clinically Significant and Influenced by Platelet Concentration: A Meta-analysis of RCTs. (2025). https://doi.org/10.1177/03635465241246524 https://doi.org/10.1177/03635465241246524
2. [2] Efficacy and Safety of Intra-articular PRP Versus Corticosteroid Injections in Knee OA: Systematic Review of RCTs. (2025). https://doi.org/10.7759/cureus.80948 https://doi.org/10.7759/cureus.80948
3. [3] MRI-Based Evaluation of PRP Therapy in Knee Osteoarthritis: WORMS and Synovial Changes at 6 Months. (2025). https://doi.org/10.3390/jcm14186408 https://doi.org/10.3390/jcm14186408
4. [4] 3D-MRI analysis of cartilage thickness changes after PRP injection in medial knee osteoarthritis. (2025). https://doi.org/10.1371/journal.pone.0321067 https://doi.org/10.1371/journal.pone.0321067
5. [5] Platelet-rich-plasma and knee osteoarthritis: a critical overview of the multiple biases and methodological complexity. (2025). https://doi.org/10.1016/j.jbspin.2025.106014 https://doi.org/10.1016/j.jbspin.2025.106014