The 30 most common migraine triggers, how often patients report them, and how strong the evidence really is.
The most common migraine triggers are stress, hormonal changes, skipped meals, weather changes and disturbed sleep. In the largest single survey (Kelman 2007, 1,207 patients), 79.7% named stress, 65.1% of women named hormones, 57.3% named not eating, 53.2% named weather and 49.8% named sleep disturbance. Most people have several triggers that add up, and the mix is personal: in a 90-day diary study, the average patient had about four, and 85% of profiles were unique.
So treat the list below as a map of where to look. The table rates each trigger by how good the evidence is, not just how often people blame it, and says how to test it in your own log.
This is general information, not medical advice. Treat it as a well-sourced starting point for a conversation with your doctor.
The top 10 by how often patients report them
| Rank | Trigger | Patients reporting it (Kelman 2007) |
|---|---|---|
| 1 | Stress | 79.7% |
| 2 | Hormones (women) | 65.1% |
| 3 | Not eating / skipped meals | 57.3% |
| 4 | Weather | 53.2% |
| 5 | Sleep disturbance | 49.8% |
| 6 | Perfume or odours | 43.7% |
| 7 | Neck pain | 38.4% |
| 8 | Lights | 38.1% |
| 9 | Alcohol | 37.8% |
| 10 | Smoke | 35.7% |
Further down Kelman’s list: sleeping late (32.0%), heat (30.3%), food (26.9%), exercise (22.1%) and sexual activity (5.2%). Caffeine and dehydration weren’t on his questionnaire, but they’re on the American Migraine Foundation’s top-10 list and the NHS list.
“Reported” isn’t the same as “proven”. When researchers exposed 27 people with migraine with aura to the lights or exercise they said triggered them, only 3 (11%) got an aura attack (Hougaard 2013). A clinical review by Marmura (2018) lists the triggers that held up in studies, not just surveys: fasting, the premenstrual window, “let-down” after stress, and most likely low barometric pressure.
All 30 triggers at a glance
Evidence strength:
- Strong: confirmed in diary or clinical studies.
- Moderate: consistently reported, with some supporting data.
- Weak: often blamed, but controlled studies haven’t confirmed it, or it may be an early symptom rather than a cause.
| # | Trigger | Category | Evidence | How to test it in your log |
|---|---|---|---|---|
| 1 | Acute stress | Stress | Moderate | Rate stress 0–10 each evening |
| 2 | Stress let-down | Stress | Strong | Look for attacks the day after stress drops |
| 3 | Anxiety / low mood | Stress | Moderate | Weekly mood rating vs attack count |
| 4 | Sudden hard exertion | Lifestyle | Moderate | Note intense workouts; attacks within ~6 h |
| 5 | Neck pain / posture | Lifestyle | Weak (often a symptom) | Note whether neck pain came before other early signs |
| 6 | Travel and time-zone change | Lifestyle | Moderate | Tag travel days; attacks in the next 1–3 days |
| 7 | Disturbed / poor-quality sleep | Sleep | Moderate | Log how rested you felt, not just hours |
| 8 | Sleeping late / oversleeping | Sleep | Moderate | Tag lie-ins, weekends especially |
| 9 | Irregular schedule / shift work | Sleep | Moderate | Log bedtime; watch for big swings |
| 10 | Falling estrogen before a period | Hormonal | Strong | Log cycle day with every attack |
| 11 | Ovulation | Hormonal | Weak | Look for a mid-cycle cluster |
| 12 | Perimenopause | Hormonal | Moderate | Compare monthly attack counts over a year |
| 13 | Hormonal contraception / HRT | Hormonal | Moderate | Compare before and after any change |
| 14 | Falling barometric pressure | Weather | Moderate | Compare attacks with local pressure history |
| 15 | Heat and temperature swings | Weather | Moderate | Day-to-day temperature change |
| 16 | High humidity | Weather | Weak | Log alongside pressure |
| 17 | Dry, warm winds (Foehn, Chinook) | Weather | Weak | Tag wind events if you live where they occur |
| 18 | Alcohol | Diet | Moderate | Drink type and amount; attacks within 10 h |
| 19 | Skipped meals / fasting | Diet | Strong | Log meal times; watch gaps over ~5 h |
| 20 | Dehydration | Diet | Weak | Rough daily water intake |
| 21 | Aged cheese, cured meat | Diet | Weak | Name the food, not just “cheese” |
| 22 | Chocolate | Diet | Weak (often a craving) | Note whether a craving came first |
| 23 | MSG, nitrates, aspartame | Diet | Weak | Only a planned elimination and re-test tells you |
| 24 | Bright or flickering light | Sensory | Weak to moderate | Log light environment in the hour before |
| 25 | Strong smells | Sensory | Moderate | Name the smell |
| 26 | Smoke | Sensory | Moderate | Tag smoke exposure |
| 27 | Loud noise | Sensory | Weak | Tag noisy environments |
| 28 | Caffeine withdrawal | Medication | Strong | Log daily caffeine, not just on attack days |
| 29 | Medication overuse | Medication | Strong | Count days per month you take acute medication |
| 30 | Vasodilating medicines | Medication | Strong | Note start dates of new prescriptions |
The threshold model: why “just avoid your triggers” fails
The old idea was binary: X causes your migraines, so avoid X. It rarely works. People cut out chocolate and still get attacks; they drink red wine and sometimes nothing happens.
The better model is a threshold, sometimes called trigger stacking:
- Everyone with migraine has an attack threshold.
- Each trigger adds some load toward it. One is often not enough; several close together can push you over.
- The threshold moves. It tends to be lower when you’re short on sleep, ill, stressed or around your period.
The alcohol data shows it well. In a Dutch study of 2,197 patients (Onderwater 2019), 35.6% named alcohol as a trigger and 77.8% of those named red wine, yet red wine consistently caused an attack in only 8.8%. The authors concluded alcohol probably works with a fluctuating threshold rather than on its own.
In practice, on a high-risk day (short night, falling pressure, day 1 of your period), don’t add easy triggers on top: eat on time, go easy on alcohol, keep caffeine steady.
Stress and lifestyle (1–6)
1. Acute stress
The most-reported trigger in nearly every survey. Diary studies show a messier picture: in Lipton 2014, the level of stress on a given day wasn’t generally linked to attacks. Test: rate stress 0–10 each evening so you can see both levels and drops.
2. Stress let-down (“weekend migraines”)
The same study found that a drop in stress from one evening to the next raised the odds of an attack over the following 6 to 18 hours (odds ratios 1.5–1.9). It explains the attack on the first day of a holiday. Test: look for attacks the day after a deadline or a hard week ends.
3. Anxiety and low mood
Both are more common in people with migraine, and the relationship runs both ways. Test: a weekly mood rating against monthly attack counts. Treating anxiety or depression can lower attack frequency for some people; that’s a conversation for your doctor.
4. Sudden hard exertion
Kelman found 22.1% reported exercise as a trigger, and in Hougaard’s provocation study, 4 of 12 people developed migraine after strenuous exercise. Regular, gradual exercise is generally protective. Test: note intense or unusual efforts and look at the next few hours.
5. Neck pain and posture
Reported by 38.4% in Kelman, but neck pain is also one of the most common premonitory symptoms, so it may be an early sign of an attack already underway (Marmura 2018). Test: note whether neck pain arrived before or alongside other early signs like yawning, tiredness or cravings.
6. Travel and time-zone change
Travel stacks several triggers: broken sleep, irregular meals, dehydration and stress. Test: tag travel days and watch the following one to three days.
Sleep (7–9)
7. Disturbed or poor-quality sleep
Patients report it often (49.8%), but the details matter. In a six-week study of 98 adults wearing sleep trackers (Bertisch 2020), short sleep (≤6.5 hours) wasn’t linked to next-day migraine, while low self-reported sleep efficiency (lots of time awake in bed) was linked to 39% higher odds of a headache, not the next day but the day after. Test: log how rested you felt and how often you woke, not only hours.
8. Sleeping late
Reported by 32.0% in Kelman. A lie-in often comes with a later first meal and a later first coffee, so it may partly be a meal or caffeine effect. Test: tag lie-ins and note when you ate and had caffeine.
9. Irregular schedule and shift work
Big swings in bedtime disrupt circadian rhythm, and a regular schedule is standard advice in headache clinics. Test: log your bedtime and look at the weeks with the biggest swings.
Hormonal (10–13)
10. Falling estrogen before a period
One of the best-supported triggers. In MacGregor 2006, women had more attacks during the late-luteal/early-follicular phase, when estrogen falls, and fewer while it was rising. “Menstrual migraine” is a recognised diagnosis. Test: log cycle day with every attack; a pattern often shows within two or three cycles.
11. Ovulation
Some women report a mid-cycle cluster, but MacGregor’s data didn’t support a post-ovulation estrogen-withdrawal effect. Test: look for attacks around mid-cycle before assuming it applies to you.
12. Perimenopause
Hormone levels become erratic in the years before menopause, and migraine often gets worse before it improves after menopause. Test: compare monthly attack counts across a year.
13. Hormonal contraception and HRT
Estrogen-containing contraception can make migraine better, worse, or leave it unchanged. For people with migraine with aura, combined estrogen contraception is usually avoided because of stroke risk, so it’s a prescriber’s decision, not a self-experiment. Test: compare attack frequency before and after any change.
Weather (14–17)
14. Falling barometric pressure
Marmura lists low pressure as “most likely” a genuine trigger, and small studies from Japan, such as Kimoto 2011, link attacks to pressure falls of a few hectopascals. The evidence is promising but comes from small groups, and individual sensitivity varies a lot. More detail is in our posts on the barometric pressure threshold and barometric pressure headaches. Test: compare your attacks with your local pressure history.
15. Heat and temperature swings
Kelman found 30.3% reported heat as a trigger. Test: log day-to-day temperature change, not just how hot it was.
16. High humidity
Often arrives with low pressure, so it’s hard to separate the two. Test: log it alongside pressure.
17. Dry, warm winds (Foehn, Chinook, Santa Ana)
People in these regions often report attacks during wind events, but the studies are few and mixed. Test: if you live in one of these regions, tag wind days separately.
Diet (18–23)
18. Alcohol
Reported by about a third of patients (37.8% in Kelman, 35.6% in Onderwater). Onset is fast: one-third within 3 hours, almost 90% within 10 hours. Yet red wine reliably triggered attacks in only 8.8% of people, a classic threshold pattern. Test: log drink type and amount, and check the following 10 hours.
19. Skipped meals and fasting
Reported by 57.3% and among the triggers Marmura lists as clinically verified. Working through lunch on a stressful day stacks two triggers at once. Test: log meal times and watch long gaps.
20. Dehydration
On the AMF top-10 list. It’s plausible and easy to fix, but there’s little controlled data. Test: rough daily water intake, especially on hot or active days.
21. Aged cheese and cured meat
The classic “tyramine” foods. Some people report them, but the evidence is thin. Test: name the specific food in your log and look for a repeatable pattern.
22. Chocolate
The most-blamed food, but a 2020 review of 25 studies (Nowaczewska 2020) found that none of the three controlled trials showed chocolate caused more attacks than a placebo. A craving for sweets is a common premonitory symptom, so the chocolate is often eaten because an attack has already started. Test: note whether the craving came first.
23. MSG, nitrates and aspartame
A systematic review (Obayashi 2016) found no consistent effect of MSG eaten with food; the positive results came from poorly blinded, high-dose drinks. Test: if you strongly suspect one, a planned elimination and reintroduction with your doctor’s input is the only way to know.
Sensory (24–27)
24. Bright or flickering light
Reported by 38.1%, but in Hougaard’s provocation study no one developed an attack after photo stimulation. Sensitivity to light is also a common early symptom. Test: note your light environment in the hour before, and whether it already felt too bright.
25. Strong smells
Reported by 43.7%: perfume, cleaning products, petrol. Sensitivity to smells is unusually specific to migraine. Test: name the smell and the time.
26. Smoke
Reported by 35.7% in Kelman. Test: tag exposure to cigarette smoke or wildfire smoke.
27. Loud noise
Concerts, building work, open-plan offices. It’s often part of general sensory overload. Test: tag noisy environments together with whatever else was going on.
Medication-related (28–30)
28. Caffeine withdrawal
Regular coffee drinkers can get an attack when they skip their usual dose. Caffeine helps some attacks at modest doses and triggers others when withdrawn. Test: log caffeine every day, not only on attack days.
29. Medication overuse headache (MOH)
Taking acute headache medication too often can itself cause more frequent headaches. ICHD-3 defines medication overuse as taking acute medication on 10 or more days a month (triptans, combination painkillers, opioids) or 15 or more days a month (simple painkillers like ibuprofen or paracetamol) for over three months, with headache on 15 or more days a month (ICHD-3). Test: count medication days per month. If you’re near those numbers, talk to your doctor before changing anything.
30. Vasodilating medicines
Nitroglycerin reliably provokes migraine in research settings (Marmura 2018), and other vasodilators can do the same. Test: if attacks increased after a new prescription, note the start date and raise it with your prescriber. Don’t stop a medicine on your own.
What the list can’t tell you
- Which triggers matter for you. In Peris 2017, 85% of personal trigger profiles were unique. Stress is #1 overall and might mean nothing for you.
- How your triggers combine. Most attacks come from a stack, and you only see stacks in your own history.
- Whether you’ve confused a symptom with a cause. Cravings, neck pain, yawning and light sensitivity can start hours before the pain.
Guessing from memory is also unreliable. Houle & Turner (2013) showed that day-to-day weather, stress and hormones vary so much that two otherwise comparable days, with and without a suspect trigger, rarely occur naturally. That’s why a structured log beats intuition.
How to find your triggers
The short version:
- Log every attack and a small set of daily factors for at least two to three months: sleep, stress, meals, alcohol and caffeine, cycle day and weather.
- Then compare the days before attacks with ordinary days. The Peris study used 90 days of diaries to build personal profiles.
The full method is in our step-by-step guide: how to find your migraine triggers with an app. It covers what to log, how to tell triggers from early symptoms, and how to read the results.
You don’t need an app. A paper migraine diary template, a spreadsheet or the Notes app all work, and consistency matters more than the tool. If you want something that does the comparison for you, see the comparison of migraine tracker apps. I built one of them, Migra, for iPhone:
- It reads live barometric pressure from the phone’s sensor.
- It can pull sleep and cycle data from Apple Health.
- Once you’ve logged five attacks, it shows which logged triggers come up most often.
Before you start, it’s worth taking the 5-question MIDAS test. It’s the disability score many neurologists use, and it gives you a baseline to compare against later.
The one-line takeaway
The top-30 list tells you where to look; your own log tells you what matters. Stress, hormones, missed meals, weather and sleep are the likeliest candidates. A few months of consistent logging will show which ones, and which combinations, apply to you.
Frequently asked
What is the most common migraine trigger?
▾
Stress. In Kelman's 2007 study of 1,207 migraine patients, 79.7% reported stress as a trigger. Next came hormonal changes in women (65.1%), not eating (57.3%), weather (53.2%) and sleep disturbance (49.8%). Other studies order them a little differently, but stress, hormones, missed meals, weather and sleep are the consistent top group.
How many triggers does a typical person with migraine have?
▾
Usually several. In a 90-day diary study of 326 patients (Peris 2017), the average person had about four factors linked to their attacks, and 85% of those personal profiles were unique. That's why a population list can only tell you where to look, not what your own triggers are.
What is trigger stacking?
▾
Trigger stacking is the idea that one trigger often isn't enough to cause an attack, but several close together can push you over your personal threshold. The threshold itself moves: it tends to be lower when you're short on sleep, stressed, or around your period. It explains why the same glass of wine causes an attack one week and not the next.
How long after a trigger does a migraine start?
▾
It depends on the trigger. In a large Dutch study, about 90% of alcohol-triggered attacks started within 10 hours. In a diary study, a drop in stress from one evening to the next raised the odds of an attack over the following 6 to 18 hours. Menstrual attacks cluster around the two days before and the first days of a period.
Are food triggers overrated?
▾
Mostly, yes. Chocolate is the most-blamed food, but a 2020 review found no controlled study where chocolate caused more attacks than a placebo. A systematic review of MSG found no consistent effect when it was eaten with food. Skipped meals and alcohol hold up much better than specific foods.
Can something feel like a trigger but actually be an early symptom?
▾
Yes. The premonitory (prodrome) phase can start hours before the pain and includes food cravings, neck stiffness, tiredness and sensitivity to light, sound or smells. So the chocolate you craved, or the light that suddenly felt too bright, may be an early sign of an attack that had already begun rather than its cause.
How do I find my own migraine triggers?
▾
Log every attack and a short set of daily factors (sleep, stress, meals, alcohol and caffeine, your cycle, weather) for at least two to three months, then compare days before attacks with days that weren't. Our step-by-step guide on finding your migraine triggers with an app explains what to log and how to read the results.
Sources
- Kelman, 2007 — The triggers or precipitants of the acute migraine attack (Cephalalgia, n=1,207)
- Marmura, 2018 — Triggers, Protectors, and Predictors in Episodic Migraine (Curr Pain Headache Rep)
- Pavlovic et al., 2014 — Trigger factors and premonitory features of migraine attacks: summary of studies (Headache)
- Peris et al., 2017 — Determining potential trigger factors in individual patients (Cephalalgia, n=326)
- Lipton et al., 2014 — Reduction in perceived stress as a migraine trigger: the 'let-down headache' hypothesis (Neurology)
- MacGregor et al., 2006 — Incidence of migraine relative to menstrual cycle phases of rising and falling estrogen (Neurology)
- Bertisch et al., 2020 — Nightly sleep duration, fragmentation, and quality and daily risk of migraine (Neurology)
- Onderwater et al., 2019 — Alcoholic beverages as trigger factor in patients with migraine (Eur J Neurol, n=2,197)
- Nowaczewska et al., 2020 — A review of the relationship between chocolate and migraines (Nutrients)
- Obayashi & Nagamura, 2016 — Does monosodium glutamate really cause headache? A systematic review (J Headache Pain)
- Hougaard et al., 2013 — Provocation of migraine with aura using natural trigger factors (Neurology)
- Houle & Turner, 2013 — Natural experimentation is a challenging method for identifying headache triggers (Headache)
- Kimoto et al., 2011 — Influence of barometric pressure in patients with migraine headache (Intern Med)
- ICHD-3 — 8.2 Medication-overuse headache
- American Migraine Foundation — Top 10 migraine triggers and how to deal with them
- NHS — Migraine: causes