The knee is where this treatment has the most to stand on.

PRP FOR THE KNEE

Knee osteoarthritis is the indication people most often arrive asking about, and it is the one where the honest answer depends most sharply on what stage you are at. A knee with preserved joint space and a knee that is bone-on-bone are two different problems wearing the same diagnosis.

What is actually happening in an arthritic knee

The familiar description — cartilage wears out — is not wrong, but it is the least interesting part. Articular cartilage has no blood supply and almost no capacity for self-repair, so once it thins it does not thicken again. What changes is everything around it. The synovium becomes inflamed and starts producing catabolic cytokines, principally IL-1β and TNF-α, which drive matrix metalloproteinases that degrade what cartilage remains. Subchondral bone remodels and becomes a pain generator in its own right. The joint capsule stiffens.

This is why osteoarthritis stopped being described as a wear-and-tear disease and started being described as an inflammatory one. It is an active biological process, not an erosion.

Why some knees degrade faster than others, with the numbers

Mechanical load explains part of the variation and nothing like all of it. The rest is systemic, and it has been measured at a scale that is hard to argue with.

A prospective cohort followed 256,364 UK Biobank participants who had no osteoarthritis at baseline for a median of eleven years, staging each person on cardiovascular-kidney-metabolic health from 0 to 4. Compared with stage 0, people at stage 4 developed knee osteoarthritis at a hazard ratio of 2.67 (95% CI 2.44 to 2.93). Two details in that paper matter more than the headline: the association was stronger in people under 60, and in participants who carried both a high genetic risk score and stage 4 metabolic disease the hazard ratio reached 4.00, with a positive additive interaction between the two.

You did not pick your genes. You have some say over the metabolic stage, and the two multiply rather than simply add. That is the argument for treating the terrain in a forty-five-year-old knee, not a seventy-five-year-old one.

The mechanism is not mysterious. Adipose tissue is endocrine tissue: in a person carrying visceral fat it secretes leptin, resistin and IL-6 into general circulation, and those reach the synovium regardless of what the knee is doing. This is why osteoarthritis appears in non-weight-bearing joints in metabolically unwell people, which a purely mechanical model cannot explain. Alongside it, insulin resistance impairs chondrocyte function while advanced glycation end-products stiffen the collagen network, making cartilage more brittle under the same load.

The blood-pressure finding that is not about weight

The obvious objection to all of this is that metabolic illness travels with body weight, and body weight loads the knee. So the interesting evidence is the evidence that survives adjustment for it.

A meta-analysis of 26 observational studies covering 97,960 people found hypertension associated with knee osteoarthritis at an odds ratio of 1.62 (95% CI 1.32 to 1.98), rising to 1.89 for radiographic knee osteoarthritis specifically. Restricted to the studies that adjusted for body mass index, the association with radiographic disease held at 1.42 (1.13 to 1.78). It was particularly strong in women (2.27).

And the control that makes the point: the same analysis found no significant association at the hand (1.19, 0.92 to 1.53) — a joint that carries no body weight at all. If this were simply load, the hand would not be the joint that comes out clean. The authors read it as pointing toward a vascular contribution to cartilage failure, which is the same subchondral perfusion story that runs through everything else on this site.

What is actually modifiable, and by how much

Association is only useful if something moves. A network meta-analysis of 13 randomized trials in knee osteoarthritis, 2,800 participants, compared seven weight-loss strategies. Every one beat control on weight change. On pain, only diet combined with exercise separated from control — and meta-regression put the threshold plainly: meaningful pain relief was anticipated once a patient achieved at least a 7% reduction in body weight.

Two honest notes on that. The intervention that produced the largest weight loss — psychological support with diet and exercise, at 11.2 kg — was not statistically superior to control for pain, so this is not a simple more-is-better relationship. And 7% is a target, not a promise. What it does is turn a vague instruction into a number you can work toward, which is a different conversation from being told to lose weight.

Then there is the driver that appears on no panel: whether you can afford to offload the joint. A knee belonging to someone who stands on concrete for nine hours has a different prognosis than the same knee in someone who can sit down, and no injection changes that arithmetic. It is worth saying plainly, because patients are routinely blamed for a trajectory their working life determined. What we measure, and what we do with it, is on the metabolic health page.

What PRP is doing in this joint specifically

PRP will not regrow cartilage. Anyone who tells you otherwise is selling something. What it appears to do in a knee is shift the intra-articular environment: dampening the IL-1β-driven catabolic signaling, influencing synovial cell behavior, and improving the viscosity and lubricating properties of the joint fluid. The effect people report is less pain and better function, sometimes for many months, in a joint whose radiograph has not changed at all.

That distinction matters when you are deciding what success would mean. If your criterion is a better X-ray, this is the wrong treatment. If it is climbing stairs without planning the route first, it may not be.

What the evidence supports, and where it thins

Knee osteoarthritis has the largest PRP literature of any indication, and it is broadly favorable for pain and function at six to twelve months, with several comparisons suggesting durability beyond what a corticosteroid injection offers. It is also heterogeneous: preparation protocols differ between trials to the point that they are not really studying the same intervention, platelet concentrations vary several-fold, and leukocyte content — which plausibly matters a great deal in a joint — is often not reported at all.

The consistent signal is that earlier-stage disease responds better. The consistent gap is that trial populations are metabolically healthier than the patients who actually walk in.

That gap is the subject of why your insurer will not pay for this, and it is the single most useful thing to understand before deciding.

Where you sit on the grading scale decides most of this

Radiographic knee arthritis is graded Kellgren–Lawrence 1 through 4, and that number predicts how a joint responds to an intra-articular injection better than pain score, age or duration of symptoms does. It is the first thing worth knowing about your own knee, and most people have never been told it.

• Grade 1–2 — joint space largely intact, osteophytes minor. This is where the trial evidence is strongest and where a good response most often lasts a year or longer.
• Grade 3 — space narrowed but present, cartilage thinned rather than gone. Worth treating, with a smaller expected effect and a plan to repeat it.
• Grade 4 — bone contacting bone, usually with deformity. Structure will not be restored here and nobody should tell you otherwise. What treatment does at this stage is buy time and reduce how much medication it takes to get through a day. A blinded placebo-controlled trial conducted specifically in grade IV knees — in people unwilling to have a replacement or carrying a reason to defer one — found serial PRP reached a clinically meaningful improvement in pain and function at three and six months where saline did not.

That last point is worth separating from the usual advice, because the usual advice is wrong about it. Grade 4 is not an automatic exit from this clinic. Most people at that stage are not choosing between an injection and a replacement; they are choosing between an injection and another year of escalating analgesia while they wait, defer, or are told they are too young. Buying that year back, at a lower medication burden, is a legitimate clinical goal and not a consolation prize. Where a replacement is wanted and appropriate, it remains the definitive operation for the structure and we will say so plainly.

Three things then move the decision either way. A corticosteroid injection that worked and then stopped working is a favorable sign, not a failure — it establishes that the joint is the pain source, which is more than most people arrive knowing. Mechanical locking or a large unstable meniscal tear is not an arthritis problem and needs a different conversation. And an uncontrolled blood sugar gets addressed before the knee does, which is a sequence rather than a refusal, for the reasons set out on metabolic health and tissue repair.

The one person this does not suit at any grade is the person who expects the injection to be the whole treatment. The quadriceps work is not optional and it is not a formality.

What a knee replacement actually costs

The comparison people are usually offered is between an injection that might not work and an operation that will. That is not the real comparison, and the figures below are the reason. They are not arguments against having a knee replaced — they are the arithmetic that should be in the room when you decide when.

What is counted	Reported figure
Moderate to severe pain at 1 year	25%
Moderate to severe pain at ~5 years	28%
Dissatisfied with the result at 5 years	12.7%
Still taking opioids at 12 months	21%
Stiffness needing manipulation under anesthesia	1.3–5.8%
Deep periprosthetic infection	1.08%
Venous thromboembolism at 30 days	1.4%
Returned to sport at 3 months / 6 months	18.7% / 70%
Lifetime revision risk if replaced at age 46–50	22.4%
Implant surviving at 10 years	~96%

Two of those rows do most of the work. Pain does not reliably end — a quarter of replaced knees still hurt at a year, and at roughly five years the figure is 28%, so it is not something that fades with time the way people are led to expect. None of it counts as a surgical complication. No infection, no loosening, nothing a registry records. It is simply what a meaningful minority of patients live with afterward. In fairness the traffic is not all one way: one mixed-methods study found many people with persistent pain at a year did report improvement by five to seven years.

And the implant is most finite in the people who are told to wait. A knee replaced between 46 and 50 carries a 22.4% lifetime risk of revision — about one in four — and revision is a larger operation with worse results than the original. That is the actual content of “you are too young.” It is a real problem, and it is a reason to control the timing rather than a reason to accept whatever timing you are handed. How the two options compare.

How treatment runs here

Most knee protocols are one to three injections spaced three to four weeks apart, placed under ultrasound guidance. Blind intra-articular knee injections miss the joint more often than clinicians like to admit, particularly in a knee with no effusion, and a preparation placed in the fat pad instead of the joint is a wasted draw.

Alongside it we look at your metabolic panel, because the cartilage you are asking to behave better is being fed by the same circulation as everything else, and at your loading pattern, because quadriceps strength is one of the few things that reliably alters the trajectory of an arthritic knee. Neither of those is a consolation prize handed over at discharge. They are the part that determines whether the window the injection opens stays open.

Recovery in a knee

Expect a flare for two to four days — a knee joint is a closed space and the inflammatory response has nowhere to go, so it is often more noticeable here than in a tendon. Walking is fine and encouraged. Impact and deep loaded flexion wait. Most people notice change between weeks four and eight, and the fuller answer arrives at three to six months.

The recovery timeline in detail →

Where this sits against pain medication

A knee that hurts constantly is a common reason people end up on a daily opioid, and it is one of the situations where the drug is doing real work and also steadily losing ground. The aim here is not to take anything away from you. It is to widen the window in which you can move, sleep and load the leg — because quadriceps strength and sleep quality both feed back into how much pain the joint generates, and neither improves while you are immobile. Where that works, dose questions usually answer themselves. Where it does not, we say so.

What people ask about their knee

Is PRP better than a cortisone shot for knee arthritis?

For short-term relief cortisone often works faster. The trade is durability and tissue effect: repeated corticosteroid exposure is associated with cartilage loss over time, which is a poor bargain in a joint you intend to keep. The full comparison is here.

How long does it last?

Reported benefit commonly runs six to twelve months, sometimes longer, and repeat treatment is normal rather than a sign of failure. What to expect from a course of treatment. The platelets are gone in days; what holds is what they signaled.

Will it let me avoid a knee replacement?

Sometimes it defers one. It does not reverse structural disease, and a knee that needs replacing will still need replacing. How PRP and surgery actually compare.

Does my weight matter?

It matters through two separate routes, and the second one surprises people. Load is the obvious one. The other is that adipose tissue behaves as an endocrine organ, sending inflammatory signals to the joint whatever the knee is doing. The useful number is that meaningful pain relief was anticipated at a 7% reduction in body weight in the pooled randomized data, with diet plus exercise the combination that separated from control — a target rather than a lecture. What we measure. Metabolic terrain is the ceiling on the result, and it is measurable.

Related reading

• Why a pain clinic asks to see your bloodwork
• How platelet-rich plasma works
• PRP compared with cortisone
• PRP compared with hyaluronic acid injections
• Bone marrow aspirate concentrate
• Am I a candidate?
• When the knee was injured rather than worn out

There is no magic in a series of three — the number follows the tissue and the first response.