Genicular Nerve Block: Targets, Evidence and the Steroid Question

The genicular nerves are the sensory supply to the knee capsule, and blocking them is one of the few interventional options for a painful knee that does not involve putting anything inside the joint. That makes the technique attractive in exactly the population where intra-articular options have run out.

It also makes it easy to overstate. The evidence for genicular nerve block is genuinely split, and a practitioner who has read only the favourable half will promise patients more than the literature supports. This reference sets out the anatomy, what the block is actually for, where the evidence is solid and where it is not, and the two questions clinicians most often get wrong: whether to add steroid, and when to move to ablation. It sits under our clinical reference to nerve blocks in outpatient pain practice.

The three nerves, and the fourth question

The classical target set is three articular branches that run against periosteum at predictable points: the superomedial, inferomedial and superolateral genicular nerves. Each is approached where it crosses the junction of the femoral or tibial shaft with the condyle, which is why the technique is taught on bony landmarks and confirmed with imaging rather than by seeking a paraesthesia.

The inferolateral genicular nerve is deliberately left alone in most protocols because of its relationship to the common peroneal nerve. That omission is a safety decision, not an oversight.

Whether three targets are necessary in every knee is an open question with recent evidence attached. A 2026 clinical series compared medial-only blocks (superomedial and inferomedial) against medial plus superolateral, and found that adding the superolateral block was associated with greater short-term walking pain reduction in advanced medial-compartment osteoarthritis. The same work reported that lateral femoral bone marrow lesions were more common in non-responders, which is the more useful finding: it suggests the knees that fail a medial-only block may be identifiable in advance on imaging rather than by trial and error.

Treat that as a direction of travel rather than a settled protocol. It was a non-randomised comparison across two consecutive clinical periods.

Diagnostic, therapeutic, or a gateway

Genicular nerve block is used in three different ways, and conflating them is the most common source of disappointed patients.

• Diagnostic — a local anaesthetic block to establish that the pain is transmitted by the genicular nerves before committing to radiofrequency ablation. Here the block is a test, and a short duration of relief is the expected result, not a failure.
• Therapeutic — a block intended to give a period of useful relief in its own right, usually in a patient who is not a surgical candidate or who is waiting for arthroplasty.
• Perioperative — an adjunct around total knee arthroplasty, which is a different literature with different endpoints.

The reasoning behind selecting patients for non-spinal injection generally, and the steroid pharmacology that applies when steroid is used, is set out in joint and extremity injection.

What the evidence supports, and what it does not

This is where the honest account matters, because the published syntheses do not agree.

Several meta-analyses report clear benefit. A 2024 systematic review and meta-analysis in The Clinical Journal of Pain found a significant pooled effect on pain at one and three months, with the minimal clinically important difference achieved at both time points, and a smaller but significant effect on knee function. A 2025 meta-analysis of six randomised trials in 420 participants reported a standardised mean difference of 1.00 for pain and 1.07 for function, with the pain advantage largest at two weeks and the functional advantage largest at twelve. A separate 2025 meta-analysis of thirteen studies in 731 patients found a more modest pooled effect.

Two cautions travel with those numbers. Heterogeneity was high — I² of 85% for pain in the 2024 review — and the authors of that review declined to recommend routine use on the strength of their own positive result, citing study heterogeneity and limited follow-up. When the people who produced the favourable estimate will not endorse routine use, that is worth repeating to a patient.

The counterweight is the most methodologically demanding synthesis in the set. A 2025 meta-analysis in Osteoarthritis and Cartilage assessed twenty-five randomised trials in 2,049 patients across radiofrequency ablation, genicular nerve block and cryoneurolysis, applying GRADE and the Cochrane RoB-2 tool. It rated the certainty of evidence low to very low across every comparison. For genicular nerve block specifically it concluded the evidence is very uncertain, resting on a single trial suggesting small pain reduction and moderate functional improvement at four weeks. For radiofrequency ablation it found modest short-term relief against sham at four and twelve weeks, no benefit at twenty-four or forty-eight weeks, no functional improvement at any time point, and advised against routine use pending better data.

Both accounts are defensible. The difference is what each review was willing to count. The defensible clinical position is that genicular nerve block has a reasonable short-term signal, an unresolved durability question, and a favourable safety profile — and that patients should be told the durability question is unresolved rather than given a number.

A multicentre pragmatic trial now underway is designed to settle it: roughly 900 knees across thirty United States centres randomised between image-guided intra-articular hyaluronic acid with corticosteroid, genicular nerve block with liposomal bupivacaine and corticosteroid, and radiofrequency ablation. Until it reports, the durability claim stays open.

Whether corticosteroid adds anything

It appears not, and this is one of the cleaner answers in the set.

A placebo-controlled randomised trial allocated 81 patients with knee osteoarthritis to genicular nerve block with local anaesthetic alone, the same block with local anaesthetic plus corticosteroid, or a placebo procedure, with every patient following a standardised daily exercise programme. Both block groups improved on pain, disability and physical capacity. The corticosteroid arm did not separate from the local-anaesthetic-only arm on the outcome measures.

The practical consequence is worth stating plainly: if steroid is not adding benefit at the genicular nerves, then the cumulative steroid exposure it creates — and the counselling, the diabetic considerations and the interval restrictions that come with it — is being spent for nothing. The general principles of steroid selection and cumulative dose are covered in corticosteroid injections.

Note also what that trial controlled for. Every arm did the exercise programme, which is a reminder that the block is an adjunct to load management rather than a replacement for it.

Block versus ablation

A systematic review comparing genicular nerve ablation with genicular nerve block across randomised and retrospective studies found both reduced pain and improved function, with ablation likely producing more substantial and longer-lasting effects than a diagnostic block. Critically, the same review found that the superiority of ablation over a therapeutic block with steroid was not conclusive for pain reduction, and that functional improvements were comparable between the two.

Adverse events were minimal and transient in both, and the 2025 GRADE synthesis found no difference in serious adverse events between these interventions and sham.

So the sequence most often described — diagnostic block, then ablation — is reasonable, but the assumption that ablation is simply the better version of the block is not supported. It is better than a diagnostic block, which is designed to be short-lived. Against a therapeutic block, the case is unproven.

Around knee replacement

The perioperative literature is separate and should not be used to support the chronic-pain claim.

A 2025 meta-analysis of six randomised trials in 335 patients undergoing total knee arthroplasty found genicular nerve block lowered pain scores at four, eight, twelve and twenty-four hours, at rest and on activity. It found no significant effect on knee flexion and none on morphine consumption, and the authors noted the absence of complication assessment across the included trials.

A Bayesian network meta-analysis of 42 studies in 2,857 patients placed genicular nerve block first for movement-pain relief at twenty-four hours, while concluding that adductor canal block combined with infiltration between the popliteal artery and the knee capsule was the better overall regimen once function, opioid use and length of stay were weighed together.

Read together: a real early analgesic effect, no demonstrated opioid-sparing, and not the first-choice technique when recovery rather than pain score is the endpoint.

Safety

The reported safety profile is the strongest part of the case. Across the randomised evidence, adverse events have been minimal and transient, and no difference in serious adverse events has been shown against sham.

That does not make the procedure free of consequence. The general safety architecture — local anaesthetic systemic toxicity, asepsis, and the anticoagulation documents that are routinely confused with one another — applies here as it does to every block, and is set out in the nerve blocks reference. The deliberate avoidance of the inferolateral target because of the common peroneal nerve is the one anatomy-specific caution worth carrying into every case.

For clinicians

Genicular nerve block is a landmark-and-imaging technique across three targets, and accuracy depends on recognising the shaft-condyle junction on a live knee rather than on a diagram. This reference carries no needle depths or trajectories for that reason. What ultrasound guidance does and does not improve — accuracy and safety rather than pain scores — is covered in musculoskeletal ultrasound for pain practice.

Empire teaches the genicular nerve block under ultrasound guidance as part of Advanced Musculoskeletal Ultrasound Guided Injections, alongside the other peripheral blocks in the pain management academy.