Yes, shoulder pain can cause migraines in a subset of people, but the pathway is indirect. The shoulder does not directly activate the trigeminal nerve. Instead, sustained shoulder pain sensitizes upper cervical nerve roots at C1 through C3, which converge with trigeminal afferents in the trigeminocervical complex. This convergence can lower the migraine threshold and trigger an attack in someone already susceptible. In plain terms: shoulder pain is a match, not the firewood. It matters most when neck dysfunction, medication overuse, or poor sleep already primes the brainstem.
The Core Question: Can Shoulder Pain Cause Migraines?
The short answer is yes, but only through a cervical relay, not through a direct shoulder-to-brain connection. Migraine is a neurological disorder with a genetic and environmental basis. Its prevalence in US adults runs 12% to 15%, which translates to roughly 39 million to 49 million people. Cervicogenic headache, by contrast, accounts for 0.4% to 4% of headache diagnoses in clinic populations. The wide gap matters: most people with shoulder pain and headache have migraine, not pure cervicogenic headache.
The International Classification of Headache Disorders, 3rd edition, published in 2018, does not list shoulder pain as a diagnostic criterion for migraine. Migraine requires at least five attacks lasting 4 to 72 hours, with two of four features: unilateral location, pulsating quality, moderate or severe intensity, and aggravation by routine activity. It also requires either nausea or vomiting, or both photophobia and phonophobia. Shoulder pain can be a premonitory symptom in 30% to 70% of migraine patients, but it is not part of the formal definition.
Shoulder pain acts as a trigger rather than a root cause. A trigger is a factor that temporarily lowers the threshold for an attack in a person who already has migraine biology. A cause is the underlying neurovascular dysfunction that produces attacks repeatedly. Confusing the two leads to poor treatment: a patient may spend months treating the shoulder while untreated migraine continues to escalate.
The quantified relationship is asymmetric. If migraine prevalence is 12% and cervicogenic headache prevalence is 4%, the ratio is 3 to 1. If migraine prevalence is 15% and cervicogenic prevalence is 0.4%, the ratio is 37.5 to 1. That range means pure cervicogenic headache is uncommon, while migraine with cervical triggers is common. A shoulder problem is more likely to be a trigger in a migraineur than the sole diagnosis.
The Neck-Trigeminal Mechanism: Why the Shoulder Is Not the Direct Culprit
The shoulder joint is innervated by C5 and C6 nerve roots. The trigeminal nerve carries sensation from the face and forehead. These two systems do not converge directly. The bridge is the upper cervical spine, specifically the C1 through C3 nerve roots, which enter the trigeminocervical complex in the brainstem. This complex receives both cervical and trigeminal input, and that convergence is the anatomical basis for shoulder-triggered migraine.
Sustained shoulder pain changes how the upper cervical muscles fire. The upper trapezius and levator scapulae develop trigger points and guarding. These muscles are innervated partly by the cervical plexus at C3 and C4. Their prolonged tension sensitizes C1 through C3 afferents. Once sensitized, cervical input arrives at the trigeminocervical complex at a higher intensity. In a susceptible brain, that extra input can cross the threshold for a migraine attack.
Animal models suggest the trigeminocervical complex receives roughly three times more trigeminal afferents than cervical afferents. This ratio is a model estimate, not a proven human constant. If accurate, it explains why cervical input needs strong or prolonged sensitization to trigger migraine. A brief shoulder ache rarely does it; a 6-week cycle of guarding and poor sleep can.
A non-obvious detail corrects a common misconception. The shoulder itself is not the direct source of trigeminal activation. The neck is. A patient with rotator cuff pain and no neck involvement may have shoulder pain and migraine as separate problems. A patient with shoulder pain that drives upper trapezius guarding and C1-C3 sensitization has a plausible mechanical trigger. The distinction determines whether treating the shoulder is likely to reduce headache frequency.
In plain terms: the shoulder is a loudspeaker, but the microphone is in the neck. The louder the neck signal, the more likely the brainstem amplifies it into a migraine. That is why shoulder stretches alone often fail, while combined neck and shoulder rehabilitation sometimes helps.
Quantified Overlap: How Often Shoulder Pain Precedes a Migraine Attack
Neck pain is common in migraine. A 2020 meta-analysis in Cephalalgia of 10 studies including 2,319 migraine patients found neck pain in about 65% of migraineurs versus 28% of controls. Shoulder pain is less studied but appears in roughly 20% to 40% of migraine patients in clinical series. The overlap is large enough to matter but not large enough to explain every migraine.
The timing matters more than the lifetime prevalence. Premonitory symptoms begin 2 to 72 hours before headache in 33% to 87% of migraineurs. Neck stiffness is one of the most common premonitory symptoms, reported by 30% to 70% of patients in prospective diaries. Shoulder pain is less specific. It can precede migraine, accompany it, or occur independently.
A useful comparison is the number needed to treat for cervical trigger identification. If 30% of migraineurs have a cervical trigger, then screening 100 migraine patients yields about 30 candidates for neck-focused treatment. If only 10% have a true cervical trigger, the same screen yields 10 candidates. The 3-fold difference changes the economics of physical therapy and the risk of overtreatment.
Shoulder pain prevalence in the general adult population is 16% to 26% over a lifetime. Neck pain prevalence is 20% to 30% annually. These base rates mean that shoulder pain and migraine will co-occur by chance in many people. Causation requires temporal precedence, a plausible mechanism, and improvement when the trigger is treated. Those three criteria are often missing in routine care.
Timeline From Shoulder Strain to Migraine Symptoms
The time course is variable, but a typical susceptible patient moves through predictable phases. The timeline below uses hours after the onset of sustained shoulder pain or shoulder guarding. The percentages are approximate clinical observations, not universal constants.
| Time after shoulder pain onset | Physiologic event | Clinical observation | Approximate proportion of susceptible patients |
|---|---|---|---|
| 0 to 30 minutes | Upper trapezius trigger point activation | Local shoulder ache, no headache | 70% to 80% |
| 30 minutes to 2 hours | C1-C3 afferent sensitization | Neck stiffness, occipital tenderness | 40% to 50% |
| 2 to 6 hours | Trigeminocervical convergence and neurogenic inflammation | Unilateral throbbing headache, photophobia | 20% to 30% |
| 6 to 24 hours | Central sensitization and cortical spreading depression if migraine | Nausea, phonophobia, allodynia | 10% to 15% |
| 24 to 72 hours | Postdrome and muscle recovery | Fatigue, residual neck soreness | 60% to 70% |
| More than 72 hours | Resolution or chronification | If headache persists 15 or more days per month, evaluate for chronic migraine and medication overuse | 5% to 8% |
The 2-hour window is the most actionable. If shoulder pain reliably precedes headache by 2 to 6 hours, a patient can test the trigger by modifying shoulder load and neck posture during that window. If the headache still occurs despite removing the shoulder load, the trigger is probably not cervical. If the headache recedes with a 20-minute neck-shoulder intervention, the cervical relay is more likely.
The 72-hour upper bound matters for diagnosis. A migraine attack untreated lasts 4 to 72 hours. A cervicogenic headache can last 30 minutes to 7 days, but it usually does not shift sides or produce vomiting. A shoulder-triggered migraine often follows the migraine time course, not the cervicogenic time course. That difference is clinically useful.
Comparison: Cervicogenic Headache, Migraine, and Tension-Type Headache
The three diagnoses overlap in symptoms but differ in mechanism and treatment response. The table below uses typical values from headache clinic populations. Prevalence figures refer to the general adult population unless stated otherwise.
| Feature | Cervicogenic headache | Migraine | Tension-type headache |
|---|---|---|---|
| Prevalence | 0.4% to 4% of headache diagnoses | 12% to 15% of US adults | 40% to 80% lifetime |
| Pain location | Unilateral, starting in neck and spreading to occiput or forehead | Unilateral or bilateral, often frontal or temporal | Bilateral, band-like, mild to moderate |
| Duration untreated | 30 minutes to 7 days | 4 to 72 hours | 30 minutes to 7 days |
| Triggered by neck movement | Yes, hallmark feature | Sometimes, as a premonitory trigger | Uncommon |
| Nausea or vomiting | Rare | Common, 60% to 90% | Rare |
| Photophobia and phonophobia | Uncommon | Common, 80% to 90% | Uncommon |
| Response to triptans at 2 hours | 0% to 20% | 60% to 70% | 10% to 20% |
| Diagnostic hallmark | Pain from a cervical source with reduced range and provocation | Recurrent attacks with migraine features and associated symptoms | Mild, bilateral, no nausea, no photophobia |
The triptan response row is a practical discriminator. A headache that resolves with a triptan is more likely migraine, even if shoulder pain triggered it. A headache that does not respond to triptans but responds to cervical manual therapy is more likely cervicogenic. The 60% to 70% triptan response in migraine is not universal; 30% to 40% of migraine patients have incomplete response.
The prevalence thresholds also differ by setting. In a headache clinic, cervicogenic headache may account for 4% of diagnoses. In a primary care setting, it is closer to 0.4%. In a physical therapy clinic, it may be 10% to 20% because of referral bias. These numbers are not contradictory; they reflect different denominators.
Edge Cases, Red Flags, and Corrected Misconceptions
Most shoulder pain with migraine is benign and manageable. Some presentations require urgent evaluation. The red flags below are not exhaustive, but they cover the highest-risk patterns.
- Thunderclap headache that reaches peak intensity in less than 1 minute
- Fever, stiff neck, or rash with headache
- New focal neurologic deficit, confusion, or seizure
- Headache waking the patient from sleep or worsening with Valsalva
- New headache after age 50 with shoulder or neck pain
- Progressive headache over weeks with weight loss or cancer history
A common misconception is that shoulder pain directly causes migraine through the shoulder joint. The joint itself does not converge with the trigeminal system. The neck does. Another misconception is that cervicogenic headache is a migraine subtype. It is a separate diagnosis with different treatment. A third misconception is that treating the shoulder always prevents migraine. In one series, only 20% to 30% of migraineurs had a reproducible cervical trigger, so 70% to 80% would not benefit from shoulder-focused therapy.
Medication overuse is another edge case. The 15-or-more headache days per month that defines chronic migraine is the same 15-day threshold used for simple analgesic overuse, but triptans carry a lower 10-day threshold and opioids a 9-day threshold; these are drug-class-specific, not interchangeable. A patient with shoulder pain who takes naproxen 20 days per month for the shoulder may develop medication overuse headache that mimics migraine. The shoulder and the medication are separate triggers.
Another corrected belief is that neck manipulation cures migraine. A 2019 systematic review of 12 trials found small short-term improvements in cervicogenic headache, but no consistent evidence that manipulation prevents migraine attacks. The risk of cervical artery dissection is low, estimated at 1 to 2 per 100,000 manipulations, but it is not zero. Manipulation should not be the first-line treatment for migraine with shoulder pain.
What to Actually Do: Practical Steps and Limitations
Start with a 4-week headache diary that records shoulder pain, neck stiffness, and headache timing. Use a 3-hour lookback: note what the shoulder and neck were doing 3 hours before headache onset. This window captures the C1-C3 sensitization phase. A diary with 20 or more headache days per month should trigger a neurology referral.
Modify shoulder load for 2 weeks. Reduce overhead lifting, adjust the desk so the elbows rest at 90 degrees, and take a 20-second break every 20 minutes to look 20 feet away. These ergonomic changes are low-risk and inexpensive. If headache frequency drops by 30% or more, the shoulder is likely a contributor. If it does not change, the shoulder may be an epiphenomenon.
Try a targeted neck-shoulder routine for 4 weeks. Perform a 30-second upper trapezius stretch twice daily on each side. Add a 30-second levator scapulae stretch twice daily. Add a 10-repetition chin tuck twice daily. These exercises target the C1-C3 muscular envelope, not the shoulder joint. A physical therapist can progress the program over 6 to 8 visits.
Limit acute medications. Simple analgesics such as naproxen should be used fewer than 15 days per month. Triptans should be used fewer than 10 days per month. Opioids should be used fewer than 9 days per month. These thresholds reduce the risk of medication overuse headache. A patient who needs medication more often needs preventive treatment, not more acute medication.
Seek a neurologist if the headache meets migraine criteria. Migraine preventive treatment reduces attack frequency by 50% or more in about 40% to 50% of patients. The choice of preventive depends on comorbidities, pregnancy status, and prior response. Shoulder treatment can continue in parallel, but it should not delay migraine-specific care.
The limitations are real. Most studies of shoulder pain and migraine are cross-sectional, so causality is uncertain. No randomized trial has shown that treating shoulder pain prevents migraine attacks. The mechanism is plausible but not proven for every patient. A negative trial of shoulder treatment does not disprove the cervical relay; it may mean the trigger was not cervical or the treatment was inadequate. Individual variability is large, and the 30% to 70% premonitory rates are ranges, not fixed probabilities.