Category: Living with neck pain

POSTURE

Posture is the most blamed and least useful explanation in this whole field. The evidence points somewhere adjacent to it, and the adjacent thing is treatable.

The claim, and what the data actually supports

The story everyone has been told is that bad posture causes neck pain and correcting posture cures it. When posture has been measured properly — with instrumented assessment rather than eyeballing — the relationship between a person’s habitual head position and whether their neck hurts is far weaker than the confidence of the advice implies. Plenty of people with textbook alignment have persistent neck pain, and plenty with forward head posture have none.

What does associate with neck pain in computer users, across systematic review, is closer to duration and organization of work than to geometry: hours at the workstation without a break, high job demands with low control, and limited variation in position. In other words, the problem is how long a posture is held, not which posture it is.

I gave posture advice for years because it is easy to give and patients expect it. It is not that it was wrong so much as that it was aimed at the wrong variable, and it put the responsibility on the person rather than on the pattern.

So do posture correctors work?

As a way of reminding you to change position, briefly, yes. As a treatment, no — and the mechanism is the problem. A brace that holds your shoulders back does the work your postural muscles would otherwise do, which is precisely the capacity you want to build. Worn consistently it substitutes for the thing that needed training.

If you already own one, use it as a prompt rather than a support: put it on for short periods to notice the position, not for the whole working day.

What the trials do support

• Strength and endurance training of the neck and shoulder girdle. This is the intervention with the most consistent support in workplace neck pain, and meta-analysis points to resistance training combined with workplace modification rather than to either alone. The evidence sits with endurance work, not stretching.
• Interrupting the duration. Frequent short changes of position beat one correct position held for hours. This is the single change with the best ratio of effect to effort.
• Exercise delivered at work rather than prescribed for home. Office-worker trials find the programs that actually get done are the ones embedded in the working day, which is an adherence finding rather than a physiological one, and no less real for that.
• Understanding what is happening. Patient education on its own is a weak intervention — the Cochrane review is clear that it does not substitute for active treatment — but it changes what people do with the rest of the plan.

The chair question

A better chair helps some people and it is oversold to all of them. It changes the comfort of a position; it does not change how long you hold one, and duration is the variable doing the damage. Buying the chair and keeping the eleven-hour day is the common version of this, and it does not work.

A neck that is worse on Thursday and fine by Sunday is reporting a loading pattern very precisely, and that pattern is more useful than any assessment of your alignment. It also carries good news: recovering over a weekend means the tissue still has capacity.

And the counterpart claim is worth retiring too. The sixty-pound figure comes from a computer model, not from a measured spine, and the eleven hours are the part that matters.

WHAT NOT TO DO

Most cervical radiculopathy settles without anything being done to it. Whether it settles quickly is decided largely by a list of things not to do, and almost none of them appear on the printout you were handed.

Start with the number, because it changes how you should read everything below. The majority of people with a compressed cervical nerve root improve substantially within weeks to a few months without surgery. That is the baseline. The interventions on offer are competing against a condition that largely resolves, which means the ones that slow resolution down do real damage even when they feel protective.

Do not put it in a collar

A soft collar is the most intuitive thing to reach for and one of the least useful. The clearest data comes from the whiplash literature, where immobilization in a collar was compared directly against acting as usual and against active mobilization — and the collar did not come out ahead.

The mechanism is not mysterious. Deep cervical flexors lose endurance quickly when they stop working. A week in a collar buys you a few days of comfort and hands back a neck with less capacity than it had, which is exactly the state that turns a six-week problem into a six-month one.

Do not buy a traction device on the strength of the marketing

This one deserves a fairer hearing than it usually gets. Cervical traction is not dangerous and it is not nonsense; there is a coherent mechanical rationale, and some people report genuine short-term relief. What the systematic reviews show is that adding traction to a physical therapy program does not produce the improvement the marketing implies, and the trials supporting it are small and inconsistent.

So the honest position is not “traction does not work.” It is that traction is not the lever, and buying a door-mounted unit is often the thing a person does instead of getting the level identified. The device is cheaper than the appointment. It is also not an answer.

Do not have the neck manipulated while the arm is symptomatic

This is our position and we apply it consistently. High-velocity cervical manipulation in the presence of an active radiculopathy is not something we send patients toward. The evidence base for it in degenerative cervical radiculopathy specifically is thin and low in quality, and the argument for accepting an unquantified risk is correspondingly weak when the natural history is already favorable.

Mobilization, range work and graded loading are a different matter entirely and we use them. The objection is to the thrust, not to hands on the neck.

Do not stop using the arm

Protective disuse is the single most common self-inflicted complication we see. The arm stops being raised, the shoulder stiffens, the scapular stabilizers switch off, and within two months there is a second problem sitting on top of the first — one that outlasts the nerve irritation that caused it. Some movement helps and some makes it worse, and the distinction is learnable in one appointment.

The rule we give patients: symptoms in the neck and shoulder blade during exercise are acceptable and informative. Symptoms driven further down the arm are not. That is peripheralization, and it means stop that particular movement, not stop moving.

Do not wait out a weak arm

Pain runs on one clock. Motor loss runs on another. A genuinely weak deltoid, biceps or triceps — not painful-and-guarded, but weak against resistance — is a different conversation and a shorter timeline, because muscle that stays denervated does not recover indefinitely.

The same applies to anything involving both arms, the legs, balance or the hands becoming clumsy. That short list does not wait, and it is short precisely so that it is usable.

Do not accept the scan as the end of the conversation

“Degenerative changes at C5–C6” is a true statement about most necks past forty. It is a description of a population, not an explanation of your symptom, and treating it as a diagnosis closes the file before anyone has correlated the finding to a distribution. Your scan showing wear is not the finding people think it is.

None of this is an argument for doing nothing. It is an argument for spending the first six weeks on the things that change trajectory — load, sleep, movement, and establishing the level — rather than on the things that feel like treatment and are mostly a way of waiting.

CRACKING AND POPPING

A neck that makes noise is almost never the problem. What matters is whether the noise arrives with something else.

What the noise is

Two different sounds get described the same way. The sharp single pop when a joint is taken toward end range is cavitation — a gas bubble forming in the synovial fluid as pressure drops inside the joint capsule. It is not bones grinding, nothing has moved out of place, and the joint cannot repeat it for a while afterward because the gas takes time to redissolve. That refractory period is the giveaway.

The other sound is a finer, repeatable grating or rustling through a range of movement. That is usually tendon or soft tissue passing over bone, or roughened joint surfaces, and it can repeat as often as you move.

The reassuring part

Noise on its own carries almost no diagnostic weight. Painless clicking in a neck with full movement and no arm symptoms is a finding to note and leave alone. There is no good evidence that self-cracking damages a normal cervical spine, and the widely repeated claim that it causes arthritis has never been demonstrated in the neck.

What people are usually reporting when they mention it is not the sound. It is the compulsion — the sense that the neck needs to be cracked, several times an hour, for a relief that lasts minutes.

The compulsion is the actual finding

That urge is worth more attention than the noise. It generally means a segment is stiff and the surrounding muscle is guarding, and taking the joint to end range briefly interrupts the guarding. The relief is real and it is short, because the reason for the guarding has not changed.

So the cycle runs: stiffness, crack, brief relief, stiffness returns slightly sooner. People end up doing it more often over months and reasonably conclude their neck is getting worse. The segment is not degenerating from the cracking; it is that the underlying irritation was never addressed and end-range movement is a poor substitute for load.

The way out is unglamorous and it is endurance work rather than range — which exercises, and which make it worse.

When the noise does matter

• New noise after a collision or a fall, particularly with pain on movement. That warrants assessment rather than reassurance — what whiplash actually injures.
• Noise with arm symptoms — numbness, tingling or weakness traveling below the elbow. The noise is incidental; the arm is not. Cervical radiculopathy.
• Noise with dizziness or visual disturbance on neck movement, which needs looking at rather than dismissing. What vestibular testing shows.
• Any of the urgent features — clumsy hands, changed gait, leg heaviness. Those do not wait.

Having someone else do it

Manipulation delivered by a clinician is a different proposition from self-cracking, and it has a genuine short-term effect — the Cochrane work on manual therapy with exercise finds the combination outperforms either alone. The qualifier is in that sentence: with exercise. Manipulation on its own, repeated indefinitely, is the version that does not hold up, and it should never be the whole plan.

Anyone offering a long pre-paid course of neck manipulation before establishing what is generating the pain is selling a schedule, not a diagnosis.

What is usually driving it

A stiff segment, a thoracic spine that has stopped extending so the neck finds range it does not have, and a loading pattern that has not changed. Those are the three things worth addressing, and none of them is the sound.

ICE OR HEAT

The honest answer is that the evidence is thinner than the confidence with which this is usually answered, and that the choice matters far less than what you do in the twenty-three hours either side of it.

What each one actually does

Cold narrows local blood vessels, slows conduction in small pain fibers and reduces tissue metabolic rate. That is a real physiological effect and it is why an ice pack genuinely takes the edge off a freshly strained neck. Heat does close to the opposite — vasodilation, increased tissue extensibility, and a reduction in the protective muscle guarding that follows an injury by a few days.

Neither of them reaches a cervical facet joint. The joint capsules that generate most persistent neck pain sit under muscle, and surface temperature changes attenuate within the first centimeter or so of tissue. What you are treating with a hot pack is the muscular layer above the problem and, to a considerable degree, the nervous system’s reading of that region.

That is not a criticism. Changing how a segment is perceived is a legitimate therapeutic effect. It is just not the same as treating the structure.

What the trials actually studied

Very little of the thermotherapy literature is about necks. Cryotherapy has been studied most carefully in sports medicine, where recent critical reviews have been blunt about how modest and short-lived the effects are, and about how much of the practice rests on convention rather than data. The best evidence for superficial heat comes from the low back, where a heat wrap combined with exercise has been tested in randomized trials.

So when someone tells you confidently that it is ice for the first forty-eight hours and heat after that, they are extrapolating from ankles and lumbar spines to a cervical facet joint. It might be right. It has not been shown.

How to actually decide

• Something acute, within a day or two, especially with a clear mechanism: cold, ten to fifteen minutes, with a layer of cloth against the skin.
• Stiffness and guarding after the first few days, or a neck that is worst on waking: heat, and use it before you move rather than instead of moving.
• Nerve pain down the arm: neither reliably helps, and chasing the wrong one wastes weeks. That symptom is asking a different question — see cervical radiculopathy.
• If it helps, use it. The physiological argument is weaker than the practical one, and the risk is close to zero at sane durations.

The confession

For years I told patients to alternate, as though the sequence carried the therapy. It does not. What I was really doing was giving people something to do while the tissue got on with it, and that is a defensible thing to prescribe as long as everyone knows that is what it is.

The problem starts when the hot pack becomes the plan. A neck that still needs heat every evening nine months later is not a neck with a temperature problem.

The part that is not about temperature

Two things determine whether a strained neck settles, and neither appears on a heating pad box. The first is load — how many hours a day the segment spends in sustained flexion, and whether the deep cervical flexors are doing any work at all. The second is sleep, because the repair signaling that resolves soft-tissue injury is largely nocturnal and a person sleeping five broken hours is running that process at half rate.

The third is neither biological nor behavioral: it is that most people are managing this alone, at a desk they cannot change, in a job that does not pause. That is a real constraint and a treatment plan that ignores it is a plan for somebody else.

Which is why the durable answers here are changing the loading pattern and the right exercise, not a better pack.

When to stop experimenting

If a neck has not meaningfully improved in six weeks of sensible self-management, the question is no longer hot or cold. It is which structure is generating the pain, and that is answerable — see why a normal scan does not close the question.

And some symptoms should not wait six weeks at all. The findings that need urgent assessment.

WHEN STIFFNESS PERSISTS

Most stiff necks resolve within two weeks without anyone identifying what they were. The ones that do not are a different problem, and waiting longer is not the treatment for them.

What ordinary stiffness is

A neck that seizes after an awkward night or a long drive is usually doing something protective. A segment gets irritated, the surrounding muscle splints it to limit movement, and the splinting itself becomes most of what you feel. That is why the stiffness is often worse than the original injury and why it eases as soon as the guarding lets go.

Guidelines across Europe and North America converge on the same early management for this, and it is unglamorous: keep moving within tolerance, do not immobilize it, and expect improvement over days to a couple of weeks.

Why one side

Almost all of these are one-sided, and people read that as significant. Usually it is not. The facet joints are paired, the muscles that guard are segmental, and a single irritated joint produces a single stiff side with rotation limited toward it.

What does matter is whether the limitation is in one direction or all of them, and whether it is easing. A neck that loses rotation progressively over weeks, in every direction, is not behaving like a strain.

The six-week line

There is nothing magic about six weeks except that it is roughly the point at which continuing to wait stops being conservative and starts being a decision. By then the tissue has done what healing it is going to do largely unaided, and what remains is either a structure that is still generating pain or a nervous system that has learned the pattern.

Both are treatable. Neither improves by waiting longer, and the second gets harder the longer it runs.

What makes a stiff neck not a stiff neck

These are the features that change the question rather than the answer:

• Pain that travels below the elbow, with numbness or weakness in a pattern. That is a nerve root until proven otherwise — cervical radiculopathy.
• Clumsy hands, buttons, a changed gait, or heaviness in the legs. Cervical cord signs are not an outpatient wait. The findings that need urgent assessment.
• Fever, a neck that resists passive flexion, or profound malaise. Different differential entirely, and urgent.
• New headache with scalp tenderness or jaw claudication over fifty — temporal arteritis, which is a same-day problem because of the visual risk.
• A neck that will not straighten and pulls the head to one side over weeks rather than hours. Cervical dystonia is uncommon and frequently mistaken for spasm.

What is actually keeping it going

When a stiff neck does not settle, three things are usually running at once. There is a structure still generating input — most often a facet joint, sometimes a disc. There is a sensitized dorsal horn that has been receiving that input for long enough to amplify it. And there is a loading pattern that has not changed since the day it started, because the desk, the commute and the job did not change.

Treating only the first gives a good result that lasts six weeks. Treating only the third is the advice people have already had and could not follow. It is the combination that holds, and that is the argument for identifying the structure rather than continuing to guess at it.

How the structure gets identified

Not by a scan. Degenerative findings are near universal past middle age and correlate poorly with which segment hurts, which is why a normal MRI settles nothing and an abnormal one settles less than people assume.

It gets identified by examination and, where the pattern fits, by anesthetizing a specific joint’s nerve supply and seeing what happens to the pain. What a diagnostic block proves.

A neck that locked overnight is a different problem with a much better outlook. That one usually loosens inside three weeks.