Category: Working out what it is

WILL IT COME BACK

This is the question underneath most of the others, and it is rarely asked out loud. The honest answer has structure to it: some things recover reliably, some recover slowly, and one thing recovers poorly once it is lost.

Three different injuries wear the word numbness

A nerve can be in trouble at three depths, and they carry three different prognoses. Understanding which one you have is most of the answer.

• Conduction block. The nerve is intact but not transmitting across a compressed segment — the wire is fine, the signal is interrupted. This is the good one. Relieve the compression and function returns over weeks, and follow-up work on ulnar conduction block at the elbow shows recovery occurring far too quickly to be explained by regrowth. Nothing had to regrow. It was never severed.
• Axon loss. Fibers have actually died back and must regrow from the point of injury, at roughly a millimeter a day. From the elbow to the fingertips that is months, not weeks. Recovery is real but slow, and it is incomplete more often as the distance and the delay increase.
• Chronic denervation of muscle. The one that does not wait. Muscle deprived of its nerve supply for long enough undergoes changes that reinnervation cannot fully reverse, which is why visible wasting of the small hand muscles is treated as a different level of urgency than numbness is.

Why sensation usually comes back better than strength

Because the two are asking different things of the repair. Sensory recovery can be partial and still feel like recovery — the brain is good at making use of a degraded signal, and remapping does some of the work. Motor recovery requires the axon to physically reach a muscle fiber and form a working junction with it, and the target has been deteriorating the whole time it waited.

This shows up clearly after decompression surgery: sensation improves ahead of strength, and long-standing wasting recovers slowly and sometimes incompletely, while imaging follow-up shows the nerve itself continuing to change structurally out to two years. Which is worth knowing, because being told to wait a year is not a brush-off.

What actually predicts your outcome

Duration first. Across nerve compression generally, how long the deficit has been present is a more reliable predictor than how severe it currently feels, and it is the one variable still partly under your control.

Then the type of deficit. Numbness and tingling with normal strength is a favorable presentation even when it is deeply unpleasant. Measurable weakness is a different category on a shorter clock, and that distinction is the reason we ask about buttons, keys and dropped mugs rather than asking you to rate pain.

Then whether the compression is at one site or two. A nerve squeezed at two points behaves worse than either site predicts, and treating one while leaving the other is the common reason a technically successful operation leaves someone still numb.

What this means for what you do next

It means the timeline is not indefinite, and it also means it is not tomorrow. Stable numbness in one arm is assessed in weeks. Progressive weakness, clumsiness in both hands, or a change in how you walk is assessed in days — that short list exists precisely so it stays usable.

And it means the first job is still establishing which nerve and where. Which fingers are numb narrows the level, and electrodiagnostic testing separates conduction block from axon loss directly — which is the distinction this whole page turns on, and it is answerable in one appointment rather than inferred over a year.

The fear is not the pain. It is that this might be permanent. For most people reading this, it is not — and the variable that matters most is how long it is left.

IT DID NOTHING

A negative response to a well-placed injection is one of the more useful results in this field. It is routinely reported as a failure, and it is usually a finding.

An injection is two things at once

Every injection carries a therapeutic hope and a diagnostic question, and the two are almost never separated when the result is discussed. If a subacromial injection abolishes your pain for four hours, the shoulder was generating it. If the same injection changes nothing at all, the shoulder probably was not — and that is a genuine narrowing of the field, delivered for the cost of an injection you have already had.

This is a recognized use of local anesthetic rather than an after-the-fact rationalization. In the shoulder specifically, anesthetizing the subacromial space measurably improves the specificity of the cuff examination tests, because pain inhibition is removed and what is left is true weakness. That is a diagnostic maneuver by design.

So what does a genuine negative tell you?

• The target was wrong. The commonest explanation, and the most useful. A shoulder that does not respond to a shoulder injection is pointing somewhere else — frequently at the lower cervical facet joints or a C5 root. The shoulder gets scanned, injected and treated while the source sits six inches away.
• There is more than one generator. Partial relief that fades is not a failed injection; it is a second contributor announcing itself. This is the same logic as double crush in a nerve, applied to joints.
• The structure is involved but is not the driver. Common in long-standing pain, where the peripheral source has been overtaken by central sensitization. That changes the plan rather than ending it.

And what a false negative looks like

Not every non-response is meaningful, and it is worth knowing the ways a test can fail on its own terms rather than on yours.

• Volume and spread. Too much injectate does not make a better block, it makes a less specific one — it spreads to structures you were not asking about, and imaging work on cervical medial branch blocks has mapped exactly how dispersion changes with volume. A block that anesthetizes everything answers nothing.
• Timing of the assessment. A diagnostic block is read during the anesthetic window, which is a matter of hours. Asked about it a week later, most people cannot separate that window from the days around it, and the result is lost to recall rather than to physiology.
• No confirmatory block. One positive result has a meaningful false-positive rate, which is why two blocks are done before ablation is considered rather than one.

The question to ask next

Not “can I have another one.” The useful question is: what did that result rule out, and what is now the leading candidate? If the answer is a category rather than a structure, the diagnosis has not moved.

This is also why we will decline to repeat an injection into a level that nothing has identified. Repeating a test that already answered clearly is not treatment, it is optimism with a needle. Whether you are a candidate depends on which structure is causing it, and a negative result is part of how that gets established.

I have been on the wrong end of this reasoning myself, injecting a shoulder twice before accepting that the first result had already told me the answer. The result was not ambiguous. I was.

LEFT OR RIGHT

More than the intensity does. Sidedness is one of the few things you can report that narrows the structure before anyone has touched you — and one of the few that changes what a procedure would even look like.

One side usually means one joint

The facet joints are paired, two at every level, one on each side. A joint on the right cannot refer pain to the left. So a pain that stays on one side, does not cross the midline, worsens on extension and rotation toward that side, and refers into the shoulder blade or the base of the skull in a consistent map is behaving exactly the way a facet joint behaves.

That is not a small observation. Facet-mediated pain accounts for a substantial share of persistent neck pain, it is invisible on imaging, and its referral maps are segment-specific enough to be useful. Sidedness is the first filter and costs nothing to establish.

Both sides is a different question

Symmetry changes the differential rather than doubling it. Pain across the whole back of the neck is usually muscular and load-related. But symptoms in both arms — numbness, clumsiness, a change in handwriting or in how you walk — are not two pinched nerves that happened at once. Bilateral upper-limb symptoms point toward the central canal and the cord rather than toward a nerve root, and that is a different clock. The short list that does not wait includes this one.

Side plus fingers gives you the level

Side narrows to a column. Fingers narrow to a row. Together they specify a single nerve root, which is the unit everything downstream is actually delivered to. Right-sided neck pain with numbness in the right thumb and index finger is a right C6 story until something contradicts it — and the finger map is worth learning before your appointment.

Where sidedness misleads

Three ways, and all three are common enough to plan around.

• Muscle referral does not respect your expectations. A trigger point in the sternocleidomastoid refers to the same-side ear, jaw and behind the eye, and patients reasonably describe that as a head problem rather than a neck one. One muscle produces a remarkable amount of it, and referral patterns generally travel further than people expect.
• The side may be telling you about your job, not your spine. Which arm reaches for the mouse, which shoulder carries the bag, which way the second monitor sits. A neck that is worse on Thursday and fine by Sunday is reporting a loading pattern, and the side is the clue that identifies it.
• Shoulder pathology is emphatically one-sided too. Unilateral pain is not self-evidently cervical. The shoulder gets scanned, injected and treated in a large number of people whose problem sits six inches away.

Why we ask before we plan anything

Because the diagnostic procedures are sided. A medial branch block anesthetizes the two small nerves supplying one facet joint on one side; there is no such thing as blocking “the neck.” If that block abolishes the pain and a confirmatory block repeats the result, the joint has been identified and ablation becomes a reasonable conversation. If the pain was never one-sided to begin with, the whole sequence was aimed at the wrong target.

Bilateral facet pain does exist and gets treated bilaterally. The point is that it has to be established rather than assumed, because assuming it doubles a procedure for someone who needed half of one.

What this is worth to you

You can supply this yourself, for free, before any appointment: which side, whether it crosses the midline, which fingers if any, and what movement reproduces it. Four answers. In our experience they narrow the field further than the first scan usually does, and they are the difference between an examination that starts from your neck and one that starts from a report about it.

SHARP AND BRIEF

Short, electrical pain behaves differently from an ache, and the difference is diagnostic rather than a matter of severity. Duration tells you which tissue is talking.

Why the length of it matters

A sustained, dull, movement-related ache comes from joint and muscle. Those tissues are densely supplied with slow-conducting fibers, and what they produce is a pain that builds, spreads poorly and outlasts the movement that provoked it.

Pain that arrives fully formed, lasts one to a few seconds and vanishes is generated differently. That is a neural discharge — a nerve firing rather than a tissue reporting — and the list of structures that behave that way is short. Which is good news, because a short list is a workable one.

So the question we ask first is not how bad it was. It is how long it lasted, what set it off, and where it traveled.

The common ones

• Occipital neuralgia. Sharp, shooting pain starting at the base of the skull and running up the back of the head, sometimes finishing behind the eye. Seconds at a time, often triggered by brushing hair or turning the head, with a tender spot over the greater occipital nerve. It is one of the few diagnoses a block confirms and treats at once, and it overlaps considerably with headache that begins in the neck.
• A facet joint catching. A sudden stab on rotation or extension, usually to one side only, gone as soon as you come out of the position. The most common source of lasting neck pain is invisible on every scan, and this catching quality is one of its signatures.
• Face rather than neck. Electric shocks in the cheek, jaw or gum triggered by brushing teeth, cold air or eating belong to a different nerve entirely. That is trigeminal neuralgia, and it is treated nothing like neck pain.

The one that changes the order of the queue

If bending your chin toward your chest produces a brief electric sensation running down your spine, or into the arms and legs, that is Lhermitte’s phenomenon. It is not a muscle and it is not a joint. It is the spinal cord being mechanically provoked, and in this population it most often points to degenerative cervical myelopathy — cord compression from the same wear-and-tear changes that otherwise mean little.

This is not an emergency on its own and it should not be read as one. It is a finding that moves you to the front of the queue rather than to the emergency department, because myelopathy is a condition where the timing of recognition genuinely affects the outcome. Paired with clumsy hands, buttons becoming difficult, or a change in how you walk, it stops being something to think about next month. The findings that need urgent assessment are listed here.

New, severe, and unlike anything you have had

There is one presentation of sudden neck pain that is not in the same category as anything above, and it deserves a clear sentence rather than a hedge. Sudden severe neck pain or headache that is genuinely unlike your usual pattern — particularly with visual change, slurred speech, unsteadiness, a drooping eyelid or a new pulsing noise in one ear — needs assessment the same day.

The concern is cervical artery dissection: a tear in the wall of a vertebral or carotid artery. In population terms it is uncommon, on the order of a few cases per 100,000 people per year. Among adults under fifty who have a stroke, it accounts for a strikingly disproportionate share — which is the population that makes it worth naming. It frequently announces itself as pain days before anything neurological happens, and that interval is the entire opportunity.

That is the whole of the alarming content on this page, and it is deliberately specific so that everything else on the list can be read calmly. Brief electrical neck pain, on its own, in a pattern you have had before, is usually one of the first three things above.

What we do with it

Brief neural pain is one of the more satisfying things to work up, because the diagnostic tests are also the treatments. A block that abolishes the pain for the duration of the anesthetic has told you which nerve was generating it, and told you before anything irreversible was contemplated. Nothing is done on the first visit — the first appointment exists to work out which of these it is.

WHICH NERVE ROOT

Your hand is a map of your neck. Which fingers are numb narrows the level before anyone books imaging, and it is the single most useful sentence you can bring to an appointment.

The map

Each cervical nerve root carries sensation from a defined strip of skin and power to a defined set of muscles. When one root is irritated, the pattern it produces is not random and it is not “the whole arm.”

• C5 — the outer shoulder and upper arm, stopping around the elbow. No fingers. Weakness shows up lifting the arm out to the side. This is the level most often mistaken for a shoulder problem, because in every practical sense it looks like one.
• C6 — down the thumb side of the forearm into the thumb and index finger. Biceps weakness, and a reduced biceps or brachioradialis reflex.
• C7 — the back of the arm into the middle finger. Triceps weakness, a reduced triceps reflex. The most commonly affected root of the four.
• C8 — the little finger side of the forearm into the ring and little fingers. Weak grip and weak finger spread, because the small muscles inside the hand run on this level.

Two fingers and a movement you have lost is a more specific piece of data than most of what gets collected at a first appointment. If you take nothing else from this: write down which fingers, before you go.

Where the map is honest, and where it is not

Dermatome charts are drawn with hard borders. Real people have overlap between adjacent roots, meaningful variation in how the brachial plexus is assembled, and a fair amount of individual difference in where a given root’s territory actually ends. A C6 pattern that spills into the middle finger has not disproved anything.

The provocation tests have the same character. Spurling’s test — extending and rotating the neck toward the painful side under gentle compression — is a genuinely useful confirming test when it reproduces the arm symptom, because it is fairly specific. It is not sensitive. A negative Spurling’s in a patient with a clear dermatomal story does not clear the root, and systematic reviews of the physical examination in cervical radiculopathy have been consistent about this for years.

So the map is a hypothesis generator, not a verdict. Used that way it is excellent. Used as proof it produces confident errors.

What electrodiagnostics add that the map cannot

The dermatome tells you where the symptom is. It cannot tell you where along the nerve the problem sits, and that is the question that changes the treatment. Needle EMG looks for denervation in a myotomal distribution — muscles supplied by one root but by different peripheral nerves. If C7-innervated muscles are affected across two different peripheral nerve territories, the lesion is at the root. If the abnormality follows a single peripheral nerve instead, it is not a neck problem at all.

Two honest limitations. Needle findings take one to three weeks after onset to appear, so an early study can be normal in a genuinely compressed root. And a purely sensory radiculopathy — pain and numbness with no motor involvement — may never produce a positive needle study, because the sensory cell body sits outside the compression. A normal EMG narrows the field; it does not empty it.

Why we care which level it is

Because everything downstream is level-specific. A cervical epidural steroid injection is delivered to a segment, not to a neck. The same is true of a transforaminal approach, of surgical decompression, and of the decision about whether the finding on the scan is the finding that matters. Getting the level wrong is not a small miss — it is a treatment delivered adjacent to the problem.

This is also where the imaging finally earns its place. Once the examination has produced a level, the scan is being asked a narrow question it can answer, rather than an open one it cannot. Read in the other order it is much less useful.

What to bring

Three sentences, prepared in advance, and you will have contributed more to the diagnosis than most first visits manage:

• Which fingers, specifically. Not “my hand” — which fingers.
• What position makes it worse, and what makes it stop. Overhead, bent elbow, at night, turning the head.
• Whether anything has actually become weaker. Dropping things, jars, a key in a lock, buttons.

The fear underneath this is usually not the pain. It is whether the numbness is going to be permanent. That question has an answer, and getting the level right is the first step toward it.

NOT A PINCHED NERVE

The phrase is doing an enormous amount of work. It names a mechanism most people have never had confirmed, and it quietly closes a question that is still open — which is how a predictable number of arms end up being treated at the wrong end.

What the phrase actually claims

A pinched nerve in the neck means one thing specifically: a cervical nerve root is compressed or chemically irritated where it leaves the spine. That is a real diagnosis, it has a name — cervical radiculopathy — and in population studies it runs at roughly 85 cases per 100,000 people per year, weighted toward the fifth and sixth decades.

What the phrase has become is a container for any symptom that travels from the neck into the arm. Numbness, tingling, weakness, a burning line between the shoulder blade and the wrist: all of it gets the same two words, from patients, from search engines, and often enough from the first clinician who hears the story. The label arrives before the examination does.

That matters because the treatments diverge completely. A compressed C7 root and a compressed ulnar nerve at the elbow produce overlapping complaints and share almost nothing in their management.

The four that account for most of it

These are the conditions we see wearing the label most often. Each has a physical sign that separates it, and none of them requires a scan to suspect.

• The ulnar nerve at the elbow. Numbness in the little and ring fingers, worse when the elbow is bent for a while — on the phone, asleep, driving. The nerve is compressed in the cubital tunnel, roughly two feet from your neck. The elbow and the neck produce the same two numb fingers, which is exactly why this one is missed.
• The median nerve at the wrist. Thumb, index and middle finger, worst at night, relieved by shaking the hand. Carpal tunnel gets confused with a C6 root in both directions. A release that worked halfway is the clue, not the failure.
• The shoulder itself. Pain down the outer arm that stops at the elbow, worse reaching overhead or behind the back, no numbness anywhere. Subacromial pathology refers in a pattern that reads as radicular to almost everyone. A sac the thickness of paper causes a surprising share of it, and a reported cuff tear feels like an answer long before it has earned that status.
• Muscle referral. A taut band in the scalenes, levator scapulae or sternocleidomastoid refers pain into the shoulder blade, the arm and sometimes behind the eye, with no nerve involved at all. A knot in one muscle refers pain a long way, and the sternocleidomastoid in particular is a practiced impersonator.

Why the wrong label survives contact with a scan

Here is the uncomfortable part, and it is the reason this correction has to be made deliberately rather than assumed. Nearly every cervical MRI in an adult over forty shows something: disc desiccation, a bulge, foraminal narrowing, osteophytes. The scan almost always contains a finding that will confirm whatever story you brought to it.

So a person with cubital tunnel syndrome gets a neck MRI, the MRI shows C5–C6 degenerative change, and the loop closes. Nobody lied. The imaging simply cannot tell you whether the finding it reported is the one generating the symptom. A normal scan does not close the question either, and for the same reason in reverse.

What settles it

An examination that maps the symptom to a distribution, and electrodiagnostic testing when the distribution is ambiguous. Nerve conduction studies and needle EMG answer a question no image asks: not what the nerve looks like, but where along its course it stopped working. A conduction block at the elbow and a denervation pattern in a C8 myotome are different findings, and they are visible on the same afternoon.

For what it is worth, I spent the first stretch of my career ordering the scan first and reasoning backward from it. It is the faster path and it feels rigorous. It is also how you end up operating on the wrong level, which is a lesson that only has to land once.

And sometimes both are true

A nerve compressed at two points along its length is more symptomatic than either compression would predict alone. That is double crush, and it is the reason “the operation worked but I am still numb” is such a common sentence. Treating one site and declaring the question answered leaves the other one in place.

Which is why the useful posture is not picking a winner between neck and elbow. It is establishing how much each one contributes before anything irreversible happens.