Something can be done: in a 2026 randomized trial of 90 patients with bone-on-bone knees already on the replacement waiting list, two PRP injections improved pain and function at 6 months. They beat cortisone and NSAIDs. PRP does not regrow cartilage. It calms the joint’s inflammatory chemistry and puts the timing of surgery on your calendar.
“You’re bone on bone. There’s nothing left to do but replace it.”
If someone pointed at your X-ray and said that, you were handed a verdict dressed as a fact. The image is real. The meaning pinned to it is where the reasoning breaks down.
Why does a bone-on-bone knee hurt?
Start with something about the body that no one can argue with: cartilage has no nerves. It has no nerve supply. It simply cannot send a pain signal.
That raises an obvious question about the X-ray you were shown. The gap in that image is missing cartilage. Missing cartilage does not make pain. In fact, the gap on your X-ray and the pain you feel are only weakly linked. That is why some people with dramatic images walk with ease. Others with mild findings can barely function.
So what is making your pain? A biological fire inside the joint. It is made of inflammation, enzymes that break down cartilage, and signals that turn up pain. That is an active process, not a fixed gap.
And here is why that difference is the whole argument. A gap in an image cannot be treated without replacing the joint. A fire can be measured. And it can be calmed.
Do PRP injections work on bone-on-bone knees?
A randomized controlled trial published in 2026 studied exactly the people who get told there is nothing left to do. It enrolled 90 patients with advanced, “bone on bone” knees. They were Kellgren–Lawrence grade 3 to 4. All were already on the total knee replacement waiting list.
These were not mild cases being watched. These were people whose surgery was, in principle, already scheduled.
Two platelet-rich plasma injections produced significant, sustained pain and function improvement at 6 months. They also led to lower opioid use. They beat both a cortisone injection and NSAIDs in the same trial.
How does PRP work in an arthritic knee?
What sets this apart from a symptom study is what happened to the joint’s chemistry.
PRP clearly shifted the chemistry inside the joint. Harmful, inflammatory markers went down: COMP, MMP-3, IL-6, IL-18, TNF-α, and the pain signal CGRP. Protective and repair-signal markers went up: sTREM2, sRAGE, and the tissue-building growth factor TGF-β1.
In plain terms, the enzymes and signals that break tissue down went down. The ones tied to repair and protection went up. The therapy did not simply mask pain. It changed how the joint was behaving.
Can PRP regrow cartilage in a bone-on-bone knee?
Here is what most clinics offering this will not say out loud. It needs to be said first, not buried.
PRP does not regrow cartilage. The gap on your X-ray will not close. If someone tells you otherwise, the evidence does not back that claim.
What changes is the joint’s biology and how it behaves. That is exactly why the study’s authors frame PRP as a safe bridge to arthroplasty, meaning joint replacement. It is a way to calm the fire and win back function. It lets you decide when, or even whether, to have a surgery you cannot undo. You decide on your own timeline, not someone else’s.
That matters most if you are younger than the typical replacement patient. It also matters if the timing of surgery is a bad fit rather than urgent. A bridge is not a cure. It is control over the calendar.
Why does PRP work for some people and not others?
Not all PRP is equal. Platelet dose and how carefully it is prepared largely decide whether the treatment does anything at all. The published work on dosing points to roughly 10 billion platelets for a lasting effect. A preparation that falls short of a treatment dose is not a weaker version of the same thing. It is a different treatment with different expectations. That is why “I tried PRP and it didn’t work” is an incomplete sentence until you know what was actually injected.
The injection lands in a body, not just a joint. Your metabolic terrain is the inflammatory and metabolic setting the tissue lives in. It shapes whether the tissue can respond to a repair signal at all. Treating the knee while ignoring the body it belongs to treats half the problem.
Where this leaves you
This is not a universal cure. It is not for everyone. Sometimes joint replacement truly is the right next step, and a good clinician will tell you so.
Here is what it is: evidence-informed, biologically active medicine for people who were handed “the end of the road” and never offered the door marked biology. The evidence is randomized and peer-reviewed. It is shown here with its limits stated, not sanded off.
Whether your knee is a candidate depends on your imaging, your exam, and your metabolic markers. That is a clinical call, and a formal evaluation exists to make it.
Frequently asked questions
If PRP doesn’t regrow cartilage, what is it actually doing?
It changes the joint’s chemistry. In the 2026 trial, harmful markers including MMP-3, IL-6, and TNF-α went down. Repair-signal markers including TGF-β1 went up. Pain and function improved and opioid use fell at 6 months. The gap on the X-ray stays. The fire inside the joint is what changes.
I was told I’m too far gone for anything but replacement. Does this apply to me?
The trial chose grade 3–4, “bone on bone” knees already on the replacement waiting list. So advanced disease was the study group, not a reason to exclude. That said, this is not a universal cure, individual results vary, and for some patients replacement really is the right next step.
Should I stop treatment or cancel surgery based on this?
Do not start, stop, or change any treatment without consulting your physician. If you want to explore this, raise it with the physician managing your care. Or get a formal second evaluation before your surgery date. This page informs the conversation; it does not replace it.
Why would the type of PRP matter?
Because platelet dose and how carefully it is prepared largely decide the effect. Published dosing work points to roughly 10 billion platelets for a lasting effect. Two preparations both called “PRP” can differ so much that only one has a real chance of making the biological shift described here.
Can a bone-on-bone knee be fixed without surgery?
The structure cannot be rebuilt. The gap on the X-ray will not close. The pain is a different story. It comes from inflammation, enzymes that break down cartilage, and signals that turn up pain inside the joint. That fire can be calmed. In the 2026 trial, patients already on the replacement waiting list improved in pain and function at 6 months. That puts the timing of surgery back on your calendar.
What are the best injections for a bone-on-bone knee?
In the 2026 randomized trial of grade 3 to 4 knees, two PRP injections beat both a cortisone injection and NSAIDs at 6 months. The preparation matters as much as the name. Published dosing work points to roughly 10 billion platelets for a lasting effect. A preparation below a treatment dose is a different treatment with different expectations.
How long does PRP last in a bone-on-bone knee?
The 2026 trial measured lasting gains in pain and function at 6 months after two injections. It also found lower opioid use. That is the time point the evidence covers. How long any one knee responds depends on platelet dose, preparation and the metabolic terrain the joint lives in. That is why imaging, exam and metabolic markers decide who is a candidate.
Does PRP reduce the need for pain medication in knee arthritis?
In the 2026 trial it did. Patients with bone-on-bone knees who got two PRP injections used fewer opioids at 6 months. Their pain and function scores were also better. The therapy did not simply mask pain. Harmful markers in the joint fell and repair-signal markers rose. The joint itself was behaving differently.
Key takeaways
- Cartilage has no nerves and cannot make pain. The gap on your X-ray is only weakly linked to what you feel.
- Pain in an arthritic knee comes from an active fire in the joint. That fire is inflammation, enzymes that break down cartilage, and signals that turn up pain.
- In a 2026 randomized trial of 90 grade 3–4 knees on the replacement waiting list, two PRP injections improved pain and function at 6 months. Opioid use fell, and PRP beat cortisone and NSAIDs.
- PRP clearly lowered harmful markers and raised repair markers. It changed the joint’s biology rather than masking pain.
- PRP does not regrow cartilage. The study’s authors frame it as a safe bridge to arthroplasty. It gives you control over when or whether to operate.
Studies referenced
- Lacko M, Awad O, Matúška M, et al. (2026). Intra-articular platelet-rich plasma demonstrates superior clinical and serum biomarker outcomes compared with corticosteroids and NSAIDs in late-stage knee osteoarthritis: a randomised controlled trial. Journal of Orthopaedic Surgery and Research, 21(1). DOI: 10.1186/s13018-026-07013-w
- Bansal H, Leon J, Pont JL, et al. (2021). Platelet-rich plasma (PRP) in osteoarthritis (OA) knee: correct dose critical for long-term clinical efficacy. Scientific Reports, 11:3971.
- Belk JW, et al. (2021). Platelet-Rich Plasma Versus Hyaluronic Acid for Knee Osteoarthritis.
- Bennell KL, et al. (2021). Effect of Intra-articular PRP vs Placebo on Pain and Medial Tibial Cartilage Volume (RESTORE Trial).
- Neogi T, et al. (2016). Association of Joint Inflammation With Pain Sensitization in Knee Osteoarthritis. Arthritis & Rheumatology, 68(3):654–661.
This article is educational and is not a substitute for evaluation, diagnosis, or treatment by a physician. All efficacy described is presented as trial findings, not guaranteed outcomes. Individual results vary, and not every patient is a candidate for the therapies described. Do not start, stop, or change any treatment without consulting your physician.
Find out what is actually driving your pain
Regen.MD begins with a physician-led Clinical Evaluation — a review of your history, imaging, and metabolic data, and a written terrain roadmap. Evaluation is contingent upon review of your data.
Questions? Call or text us.