The Full Picture — Anatomy, Diagnosis & Evidence
What's Actually Going On
The radial nerve runs down the outside of your arm and dives through the muscles of the upper forearm. The branch that continues from there is almost purely a motor nerve: its job is to straighten your fingers and thumb, not to carry feeling from your skin. The theory behind this condition is that the nerve gets squeezed as it passes under the edge of the supinator muscle, in a region called the radial tunnel, and that the squeeze produces a deep ache without ever producing weakness.
There are three problems with that theory, and a good clinician will tell you all three.
The blamed structure is normal. Pooling twenty anatomical studies, the arcade of Frohse is present in 66% of adults and in none of the fetuses examined. It is a normal band of tissue that most adults develop. When surgeons find one during an operation, they are finding something two out of three people have.
When the whole nerve was scanned properly, the trouble was somewhere else. In a study using high-resolution nerve imaging, only 16% of people had a problem where the nerve enters that forearm muscle. In 84% the abnormality was up in the upper arm, roughly three inches above the elbow. That study looked at people who had actual weakness rather than pain alone, so it does not settle the question. But it is the strongest reason to hold the forearm story loosely, and surgeons can only find problems in the place they open.
It is pain in a nerve that does not carry pain from the skin. This condition is defined as pain with no weakness and no numbness. That definition is exactly why the examination is vague and why electrical nerve testing usually comes back clean. The tests are not failing by accident. They are being asked to detect something the definition rules out.
How to Identify It
The single most consistent finding is where it hurts. Across the published case series, tenderness over the radial tunnel was present in 97% of people diagnosed. That sounds impressive until you realize it is a count of people who already had the label, not a measure of how well the sign separates them from everyone else.
Here is the honest position on every test used for this condition.
Where the tenderness is worst
Convention says 3 to 5 cm below the bony bump, over the radial tunnel. How often this correctly catches the condition or correctly clears it: never measured.
Maudsley's test (resisting your middle finger straightening)
How often this correctly catches the condition or correctly clears it: never measured. A 2023 review described its validity and its anatomical reasoning as unclear, "if at all."
Pain on resisted twisting of the forearm
Used as one of the criteria in the 2025 trial. How often it correctly catches or clears the condition: never measured.
Electrical nerve conduction studies
Abnormal in only 8.9% of people carrying the diagnosis. It cannot rule the condition in or out, though it remains the right test if there is real weakness.
Numbing injection into the radial tunnel
Used to make the diagnosis in 74% of published cases. Since salt water works as well as the drug, a good response identifies who responds to injections, not who has the condition.
Why there are no percentages on this page: no study has ever measured how good any of these tests is. Not one. A dedicated search for those numbers returned nothing at all. Anyone who quotes you a figure for the accuracy of a radial tunnel test should be able to name the study it came from, and they will not be able to. For some conditions that gap does not matter, because the diagnosis is obvious on sight. This is not one of those conditions. Here the tests are the only thing holding the diagnosis up, which is why it should be held loosely.
What it is not. The main alternative is ordinary tennis elbow, and it is around thirty times more common. In tennis elbow the tenderness is on or immediately below the bony bump and the pain comes on when you resist the wrist bending back. In radial tunnel syndrome it is conventionally described as sitting a couple of inches further down, with pain on twisting the forearm instead. Be aware that this distinction is clinical convention rather than a proven discriminator. Other things worth ruling out: a trapped nerve in the neck, a nerve palsy after a broken arm, and, in anyone with rheumatoid arthritis, a nerve problem that mimics a snapped tendon.
The Debate
The older evidence, 2019, 40 patients, no comparison group
A single steroid injection produced large improvements in pain and function, still holding at one year, with 57% reaching a meaningful improvement. This is the study that put the injection into routine practice.
The newer evidence, 2025, 49 patients, placebo-controlled
Steroid versus salt water. No difference at two weeks. No difference at three months. Both groups improved substantially.
Follow the 2025 trial. The 2019 study had no comparison group, so it measured natural recovery plus the effect of being treated, and credited all of it to the drug. The salt-water arm of the newer trial reproduces the same recovery curve. This is the single most useful thing on this page: it means the expected course of this condition is improvement over roughly three months.
The story everyone repeats, dating to the mid-20th century
Radial tunnel syndrome is essentially tennis elbow that would not settle, and the two conditions travel together.
The population data, 2023, 75,459 patients
Only 5.7% of people diagnosed with radial tunnel syndrome also had tennis elbow recorded on the same side within six months.
Hold the story loosely. It came from surgical series of hand-picked stubborn cases, where the overlap is guaranteed by how patients were selected. One honest caveat: this data cannot answer the reverse question, which is what share of stubborn tennis elbow is really this condition. Nobody has studied that.
Honest Limitations
Nearly every treatment success here comes from patients picked because they had already responded.
The research finding: a fluid-injection technique produced complete, lasting relief in 11 out of 11 patients, none of whom needed surgery.
The gap: you only got into that study if a previous injection had already given you more than 80% relief. The study selected the best responders and then treated them. A surgical series reporting good results screened 182 records down to 14 patients. These designs cannot separate the treatment from the selection.
The diagnosis is usually made by an injection, and that injection has now failed its own test.
The research finding: three quarters of published cases were diagnosed by injecting the radial tunnel and seeing what happened.
The gap: if the drug does not beat salt water, then responding to the injection is not evidence of the pathology. The way the condition is identified and the treatment that failed are the same procedure, which is a circle nobody in this field has broken.
The natural course is good, and almost nobody says so.
The research finding: the salt-water group improved significantly on every measure across three months, and in insurance records, fewer than 3 in 100 people with this diagnosis have surgery within a year.
The gap: the overwhelming majority of these patients get no specific treatment at all, and the literature is nearly silent about what happens to them. It is the most useful thing to tell a patient and the thing least often said.
The Nuance
Surgery against conservative care: what the numbers actually are.
Success rate without surgery: no formal figure exists. The closest available: in the placebo group of the only real trial, pain and function both improved significantly over three months with no active treatment at all. In insurance records covering 75,459 people, 97.6% did not have an operation within a year. In one follow-up study, 23% eventually went to surgery.
Success rate with surgery: no controlled figure exists at all. A 2008 review of 21 surgical case series found only 6 of adequate quality, concluded there was a "tendency" toward benefit, and called for a proper randomized trial. A 2026 review of 11 studies covering 401 arms found that one surgical approach produced better reported scores than another, alongside wide variation in technique and, again, no controlled comparison.
Nobody knows whether surgery for this condition works. Every study supporting it lacks a comparison group, in a condition where the placebo arm of the one real trial improved substantially on its own. That combination, an uncontrolled body of evidence plus a genuinely improving natural course, is exactly the setup that produces decades of confident practice on an effect that might not exist.
None of that means surgery never helps. Surgeons find real things when they operate, and patients report real improvement afterward. It means the true size of the benefit is unknown, and anyone consenting to this operation deserves to be told that rather than shown a success rate calculated without a comparison group.
