Medial Branch Block: A Diagnostic Test, Not a Treatment

The medial branch block is the best-evidenced diagnostic procedure in interventional spine practice, and it is also the one most often performed for the wrong reason. Its purpose is not to treat pain. Its purpose is to answer a question: is this patient’s pain coming from the facet joints, and will denervating them help?

Getting that distinction right is what separates a defensible practice from an indefensible one, because the procedure that follows — radiofrequency denervation — is only as good as the selection that preceded it. This reference sets out what the nerve supplies, why single blocks mislead, what relief threshold to use, and where the evidence is strongest. It sits under our clinical reference to nerve blocks in outpatient pain practice.

What the medial branch supplies, and why two levels are blocked

Each lumbar facet joint receives innervation from the medial branches of the dorsal rami above and below it. A single joint is therefore not served by a single nerve, and blocking one medial branch does not anaesthetise one joint.

The practical consequence is that targeting a suspected joint requires blocking two medial branches, and that the nerves are named for the vertebral level at which they run rather than the joint they supply. This is the most common source of level-numbering errors in referrals and reports, and it is worth confirming which convention a report is using before acting on it.

The nerve courses at a reproducible bony location — the junction of the superior articular process and the transverse process in the lumbar spine, and the articular pillar waist in the cervical spine — which is why these are image-guided procedures rather than landmark ones.

The false-positive problem

This is the central issue in the medial branch literature, and it is not a minor technical caveat.

As reported in ASIPP’s evidence syntheses, lumbar medial branch blocks rest on Level I evidence drawn from eighteen dual-block studies showing 75% to 80% relief, with estimated false-positive rates of 25% to 50%. Cervical blocks are rated Level II with false-positive rates reported between 26% and 63%, and thoracic between 42% and 58%.

A false-positive rate approaching half means that if you proceed to denervation on the strength of one positive block, a substantial proportion of those patients never had facet-mediated pain at all. They will be exposed to a procedure that cannot help them, and their subsequent failure will be attributed to the technique rather than to the selection.

Those figures should be attributed carefully. They trace to ASIPP evidence syntheses and to policy documents built on them, and an independent replication of the estimate is not readily located. Say “as reported in ASIPP’s evidence syntheses” rather than “studies show”.

Controlled blocks and the relief threshold

The response to the false-positive problem is the controlled block: either a second confirmatory block, or blocks performed with two different local anaesthetics of differing duration, with the expectation that the duration of relief tracks the agent used.

A 2025 summary comparing three international consensus guidelines alongside the American Society of Pain and Neuroscience found that all of them converge on the same core: controlled diagnostic blocks, fluoroscopic or CT guidance, and consideration of radiofrequency ablation only after diagnostic confirmation. The relief thresholds used to justify progression cluster between 50% and 80%, paired with functional improvement rather than a pain score alone.

Which end of that range a practice adopts is a real decision with real consequences. A 50% threshold treats more people and accepts more failures; an 80% threshold treats fewer and is harder to defend to a patient who improved by 70%. What is not defensible is having no threshold, or recording one that was never measured.

What the block predicts

Where the diagnostic block is positive under controlled conditions, radiofrequency denervation is expected to relieve roughly 60% to 80% of patients. That expectation is the entire justification for the block, and it is the subject of our post on radiofrequency ablation versus chemical facet neurolysis.

Denervation is not permanent. The medial branch regenerates, and pain commonly returns over months to a couple of years, at which point the question of repeating the procedure arises. A patient told the effect is permanent has been misled.

A German S3 guideline developed by GRADE methodology addressed radiofrequency denervation of the facet joints and sacroiliac joint across twenty key questions, reaching complete consensus on all but one recommendation. Its stated aim is worth repeating: to avoid both overtreatment and undertreatment, a framing that acknowledges the procedure is both overused and underused in different settings.

Why this is not a treatment

A medial branch block performed with local anaesthetic gives relief lasting hours. Adding steroid is common practice and turns a diagnostic test into something patients reasonably interpret as a treatment, which then confuses the interpretation of the next block.

If the intent is diagnostic, the cleanest approach is to keep the block diagnostic. If the intent is therapeutic, that should be a separate, stated decision rather than an incidental addition, and the cumulative steroid exposure it creates should be tracked as set out in corticosteroid injections.

Safety and screening

These are among the safer image-guided spinal procedures, but they are performed near the neuraxis and the usual architecture applies: local anaesthetic systemic toxicity, asepsis, and anticoagulation guidance that differs between societies and is routinely confused. That is set out in the nerve blocks reference.

Pre-procedural contraindication screening is a documented step rather than an assumption, and the International Pain and Spine Intervention Society safety practices are the reference point most commonly used for it.

For clinicians

Medial branch blocks are fluoroscopically guided procedures where the target is a bony landmark seen on an image, and competence is acquired under supervision rather than from a description. This reference carries no needle trajectories or contrast criteria for that reason.

Empire teaches medial branch block alongside the transforaminal approach, nerve root block and the imaging that supports them within Pain Management Training, part of the pain management academy.