Where Knee Arthritis Pain Actually Comes From (It Is Not the Cartilage)

What this video covers

• Why cartilage has no nerves, and where knee arthritis pain is actually generated
• What the genicular nerves are, how a diagnostic block helps predict who is likely to respond to ablation, and why a single positive block can still be a false positive
• How radiofrequency ablation interrupts pain signaling without cutting bone or destabilizing the knee — and why the long-term effect of denervating a weight-bearing joint is still unsettled
• What the controlled trials actually measured — 12 weeks in the sham-controlled trial, 12 to 24 weeks in the pooled analysis — and why the commonly quoted three to twelve months of relief is clinical convention rather than a trial result
• The material risks: bleeding or injury to the genicular artery, infection, a numb patch of skin over the knee, painful dysesthesia or neuroma, thermal and grounding-pad burns, a temporary pain flare, radiation exposure from fluoroscopy, and rare reports of osteonecrosis or fracture near the ablation site — plus the honest limits: roughly 4 in 10 carefully selected patients do not reach 50 percent relief, the treated nerves regrow so the procedure is commonly repeated, and ablation does not slow the arthritis itself
• What the pooled placebo-controlled data show about viscosupplementation, its own risks of acute inflammatory or pseudoseptic flare and joint infection, and why the AAOS 2021 and ACR 2019 osteoarthritis guidelines advise against its routine use
• MEDICAL DISCLAIMER: This content is for educational purposes only and is not medical advice. It does not substitute for professional diagnosis or treatment. Always consult a licensed healthcare provider regarding your condition. Viewing this video does not establish a doctor-patient relationship.

You count stairs now. Thirteen down to the garage, and somewhere around the fourth tread the inside of the knee gives that hot, grinding catch — your hand on the railing before you decided to reach. Then there is the part nobody prepared you for: it aches while you lie still. Arthritis was supposed to hurt when you moved. No one said it would wake you at two in the morning. You were handed an X-ray and the words bone on bone, a steroid shot that lasted six weeks and then four and then two, some gel injections, and an appointment in three months. That is a calendar, not a plan. This article explains where knee arthritis pain is actually manufactured, what the small nerves called the genicular nerves do, and what the controlled evidence does — and does not — support.

The tissue on your film cannot generate pain

Start with the fact that reorganizes everything else: articular cartilage has no nerve supply. None. The smooth surface that looks worn on your radiograph contains no pain fibers, so it cannot, by itself, produce a pain signal.

So where is the pain manufactured? In the tissues around the cartilage that are innervated:

• The synovium, the joint’s lining, inflamed and thickened.
• The subchondral bone beneath the cartilage plate — densely innervated, remodeling under load, and prone to microfracture.
• The joint capsule and the periosteum, the sensitive membrane covering bone.
• The infrapatellar fat pad.

This is why knee osteoarthritis is best understood as a whole-joint disease rather than a cartilage disease — and why radiographic severity tracks pain only loosely. People with dramatic films can have modest pain, and people with modest films can be in serious trouble. (Established.)

Layered on top of that is a chemical change. Nerve growth factor, up-regulated in the arthritic knee, sensitizes the peripheral nerve endings in those tissues until ordinary load starts reading as injury. That is peripheral sensitization — the volume knob on the joint’s own reporting turned up.

Why it grinds on the stairs and burns at two in the morning

The stair symptom and the night symptom come from two different places, which is why treating one rarely fixes the other.

The stairs are mechanical. Body mass is multiplied across that joint with every step, and going down stairs is where you feel that arithmetic most. In a sensitized joint, ordinary load reads as injury — so the catch on the fourth tread is real tissue reporting real strain.

The two-in-the-morning ache is not. Sustain that sensitized input long enough and the spinal cord’s dorsal horn begins amplifying it on its own — central sensitization. The afferent traffic keeps firing after the joint has gone still. That is the mechanism behind the symptom that confuses patients most: pain at rest, pain in bed, pain that has stopped waiting for you to move. It is not weakness of character, and it is not “arthritis behaving strangely.” It is a nervous system that has been listening to an inflamed joint for years and has changed how it listens.

The genicular nerves: the wiring that carries it

The front of the knee capsule is supplied by a set of small sensory branches known as the genicular nerves. They are pure messengers — they carry sensation, they do not move the leg.

Three are consistently addressed in interventional practice:

• Superior medial, at the junction of the femoral shaft with the medial epicondyle.
• Superior lateral, at the junction of the femoral shaft with the lateral epicondyle.
• Inferior medial, at the junction of the tibial shaft with the medial epicondyle.

The inferior lateral genicular nerve is deliberately left alone. The common peroneal nerve — which lifts the foot — runs too close to it, and a foot drop is not an acceptable price for a knee.

One more anatomical fact governs technique: these nerves travel alongside the genicular arteries. That is the reason needle position is earned under live imaging — fluoroscopy, ultrasound with color Doppler, or both — rather than estimated by feel.

What the standard approach misses

None of this is a criticism of the physicians who prescribed your anti-inflammatories or gave you a steroid injection. Symptom-control medication and a well-timed intra-articular steroid do something real: they lower inflammation and buy function. What they do not do is answer the question of which structure is generating your pain.

The failure mode is skipping straight to treatment without ever proving the knee is the source. Several conditions produce knee pain that is not coming from the genicular pathway at all:

• Hip osteoarthritis refers into the medial knee with a completely normal knee examination — internal rotation of the hip reproduces it.
• L3 or L4 radiculopathy projects into the anterior thigh from the spine.
• Pes anserine bursitis is tender below the joint line, not inside the joint.
• Subchondral insufficiency fracture and osteonecrosis are MRI diagnoses, not X-ray diagnoses.
• Crystal arthropathy — gout or pseudogout — behaves differently and is treated differently.
• A locked meniscus is a mechanical, surgical problem. Silencing the nerves that report it is a serious error.
• A hot, swollen joint is septic until proven otherwise.

Examination, weight-bearing films, laboratory work, and MRI when indicated exist to clear that list before anything is injected. This part is not optional, and it is the part a fifteen-minute visit tends to lose.

The first procedure is a question, not a treatment

A diagnostic genicular nerve block places a small volume of local anesthetic on those sensory branches under imaging. Then you are sent out to use the knee and record what happens, hour by hour, on paper.

If those nerves are carrying your pain, it falls sharply — and briefly. If it does not fall, there are two possible reasons, and they are not the same finding: either the pain is generated somewhere other than that pathway, or the anesthetic never reached the branch it was aimed at. That difference has to be settled before anyone makes a lesion.

Two honest qualifiers belong here. A block can be falsely positive, and it can be falsely negative. The volume is kept small precisely so the anesthetic cannot spread onto neighboring structures and manufacture a positive. And an equivocal result gets repeated, not rationalized.

Understand what the block does and does not do: it helps predict who is likely to respond to ablation. It does not prove anything.

Practical safety point: the anesthetic can leave the leg temporarily weak. Arrange a ride. You do not drive afterward.

What ablation does — and what the trials actually measured

Genicular radiofrequency ablation follows a clear block. Electrodes are placed at the same targets under fluoroscopy. Before any lesion is made, low-level current is used to stimulate: sensory testing confirms the tip is on the branch carrying your pain, and motor testing confirms it is not on anything that moves the leg. Only then is tissue heated. The axons undergo Wallerian degeneration — an orderly breakdown of the nerve fiber beyond the lesion — and stop transmitting.

No bone is cut. No ligament is divided. The knee is no less stable. But the lesion sits against periosteum and heat is delivered next to bone, which is why the rare bone complications named below are real rather than theoretical. The full procedural detail — how the targets are chosen, what the day itself looks like, what recovery involves — is on the genicular radiofrequency ablation treatment page.

Now the arithmetic, stated the way it deserves.

The sham-controlled trial. Choi and colleagues published a double-blind, sham-controlled randomized trial in Pain in 2011. Thirty-eight elderly patients were randomized — all with knee osteoarthritis pain lasting beyond three months, all having failed conservative care, all with a positive diagnostic genicular block. In the treated arm, 10 of 17 analyzed patients — 59 percent — reached at least 50 percent relief at 12 weeks. No sham patient did. Then hold that number the way its size deserves: 38 randomized, 17 analyzed in the treated arm, and 12 weeks is where that trial ended. It says nothing about six months or a year. (Controlled evidence — but a single small trial, and only out to 12 weeks.)

The pooled analysis. A 2026 systematic review and meta-analysis by Barreto and colleagues in Pain Medicine pooled eight sham-controlled trials, 627 adults with knee osteoarthritis who were ineligible for joint replacement. At 12 weeks, treated groups averaged 1.65 points better on a 0-to-10 pain scale than sham (95% CI −2.57 to −0.74) and about 11.4 points better on WOMAC function (95% CI −18.83 to −3.91). Those are averages of a difference, not a promise to you — and the honesty belongs in the same breath: heterogeneity was very high (I² = 83 percent for pain and 91 percent for function), meaning the trials barely agree with each other, and long-term data are sparse. (Established but low-certainty; do not read it as more than it is.)

A detail worth knowing before you consent. On subgroup analysis in that pooled review, the best-performing technique was conventional monopolar radiofrequency — the off-label one — not the cooled system that carries the regulatory clearance. Cooled radiofrequency systems are FDA-cleared for knee osteoarthritis pain; conventional monopolar and bipolar radiofrequency of these nerves, and chemical neurolysis with phenol or alcohol, are off-label. Regulatory status and evidence are pointing in different directions here. Ask which device and which technique is planned for you.

What the trials do not say. The commonly quoted three to twelve months of relief is clinical convention, not a trial result — a practice observation rather than a controlled finding. Individual results vary. Nerves regenerate, so repeat treatment is ordinary. Roughly four in ten carefully selected patients never reach 50 percent relief. Ablation does not slow the arthritis. And the long-term consequence of denervating a weight-bearing joint remains unsettled. (Contested and unresolved — presented as such deliberately.)

Gel shots: what the pooled placebo-controlled data show

Viscosupplementation — hyaluronic acid injected into the joint and described as lubrication — deserves a plain answer, because most patients are given a warmer one.

In 2022, Pereira and colleagues pooled 15 large placebo-controlled trials in the BMJ, more than 6,000 patients. Pain reduction against placebo was about 2 millimeters on a 100-millimeter scale (SMD −0.08; 95% CI −0.15 to −0.02) — smaller than the review’s prespecified minimal clinically important difference, and far below any threshold the field considers meaningful to a patient. Serious adverse events were more common in the treated groups: relative risk 1.49 (95% CI 1.12 to 1.98). Hyaluronic acid carries its own harms, including acute inflammatory or pseudoseptic flare and joint infection.

The American Academy of Orthopaedic Surgeons (2021) and the American College of Rheumatology (2019) both advise against its routine use in knee osteoarthritis. (Established guideline position.)

Saying this costs practices revenue, including this one. It is still the honest read of the data.

What actually carries weight in an arthritic knee

Two things have the strongest claim on your effort, and neither involves a needle.

Quadriceps and hip abductor strengthening. A weak quadriceps lets the joint absorb load the muscle should have taken. Rebuilding it changes how force crosses the knee on every step.

Weight management. Body mass is multiplied across that joint with each stride, so change here has leverage nothing else matches.

Nutritional strategy, omega-3-derived specialized pro-resolving mediators, and red and near-infrared light are used at this practice around these procedures. For knee osteoarthritis specifically, the mechanisms are largely preclinical and the human trials are small. These are adjunctive and emerging — investigational, not standard of care — and never the reason a procedure worked. Individual results vary.

An honest picture of a good outcome

What follows is a composite — a picture assembled from many patients with this condition, not one person’s chart.

Someone in their late sixties. Arthritis in both knees, worse on the right. A cardiac history and body weight that made a surgeon unwilling to operate. Two steroid injections, each lasting less than the one before. Gel injections, no change. Hip and lumbar examinations normal, no effusion, no fever. A diagnostic block dropped the pain from seven to two for about four hours, with the expected temporary leg weakness — someone else drove home. Then cooled radiofrequency ablation.

The first ten days were worse: a pain flare, ice, a phone call on day six asking whether this had been a mistake. Relief consolidated over about three weeks. Not zero pain — roughly half. Enough to reclaim the stairs, sleep on that side, and load the quadriceps in therapy. The numb patch of skin above the kneecap is permanent. Somewhere around six months the pain began returning as the nerves regrew, and the ablation was repeated after a second conversation about the same modest numbers.

Six months was chosen for that composite deliberately. Some people get longer. Some get considerably less. The sham-controlled evidence stops at 12 to 24 weeks, so nobody — including your physician — can tell you in advance which you will be. That is the shape of a genuinely good result: partial, time-limited, repeatable. Individual results vary.

Risks, and who should not have this

This is not for you if you have an active joint or skin infection, a locked meniscus, a bleeding risk or anticoagulation that cannot safely be held, or untreated inflammatory arthritis.

The material risks of genicular block and ablation: bleeding or injury to a genicular artery, infection, a numb patch of skin over the knee, painful dysesthesia or neuroma formation, thermal and grounding-pad burns, a pain flare lasting days, radiation exposure from fluoroscopy, temporary leg weakness and fall risk after the block, no meaningful benefit in roughly 4 of 10 selected patients, and rare reports of osteonecrosis or periarticular fracture near the treated site.

Frequently asked questions

Key takeaways

• Cartilage has no nerves, so knee arthritis pain is generated in the synovium, subchondral bone, capsule, periosteum and fat pad — which is why the X-ray and the pain so often disagree.
• Ache at rest and pain that wakes you at night reflect sensitization, peripheral and then central, not a knee “behaving strangely.”
• The genicular nerves carry that signal from the front of the capsule; a diagnostic block tests the pathway but only predicts response — it does not prove the diagnosis.
• The controlled evidence for ablation is real but modest and short-horizon: 10 of 17 analyzed patients, 59 percent, reaching at least 50 percent relief at 12 weeks in one small sham-controlled trial, a 1.65-point pooled advantage at 12 weeks with very high heterogeneity, and roughly 4 in 10 selected patients getting no meaningful benefit.
• Pooled placebo-controlled data show viscosupplementation delivers benefit below the minimal clinically important difference with more serious adverse events; AAOS 2021 and ACR 2019 advise against routine use.

Medically reviewed by Dr. Gurpreet Singh Padda, MD, MBA, MHP — Board Certified in Anesthesiology, Pain Medicine, Interventional Pain Management, Addiction Medicine, and Obesity Medicine. Last reviewed July 2026.

This article is educational and is not a substitute for evaluation, diagnosis, or treatment by a physician. Individual results vary. Do not start, stop, or change any medication without consulting your physician. To have the source of your knee pain identified before another injection is scheduled, call (314) 481-5000 or text (314) 886-5902.

References

1. Choi WJ, Hwang SJ, Song JG, Leem JG, Kang YU, Park PH, Shin JW. Radiofrequency treatment relieves chronic knee osteoarthritis pain: a double-blind randomized controlled trial. Pain. 2011;152(3):481-487. PMID 21055873. PubMed
2. Barreto RB, Barreto AJ, do Nascimento ALM, Carvalho DMB, Queiroz B. Efficacy and safety of genicular nerve ablation techniques for knee osteoarthritis: a systematic review and meta-analysis of sham-controlled randomized trials. Pain Med. 2026;27(4):449-461. PMID 41063397. PubMed
3. Pereira TV, Jüni P, Saadat P, Xing D, Yao L, Bobos P, Agarwal A, Hincapié CA, da Costa BR. Viscosupplementation for knee osteoarthritis: systematic review and meta-analysis. BMJ. 2022;378:e069722. PMID 36333100. PubMed

Dr. Gurpreet Singh Padda, MD, MBA, MHP