Guide
How big a pressure drop is linked to migraine? What the studies actually report
Short answer
No study has found one pressure change that holds for everyone with migraine. The figures you see quoted come from small studies that measured different things. A Japanese study of 34 people found attacks clustered at 1003–1007 hPa, which is 6–10 hPa below the 1013 hPa standard; that is a level, not a drop over a set time. Another Japanese study linked more frequent migraine to a fall of more than 5 hPa from the day of the headache to the next day. A large app-based study found a significant decrease in the six hours before was associated with more headaches. Two older European studies found the opposite: fewer migraines at low pressure. The most useful threshold is the one that fits your own records.
The numbers, study by study
| Study | People | What was measured | Reported finding |
|---|---|---|---|
| Okuma et al., Japan, 2015 | 34 | Pressure recorded in a diary when an attack began, compared with 1013 hPa | Attacks most frequent at 1003–1007 hPa, 6–10 hPa below standard |
| Kimoto et al., Japan, 2011 | 28 | Daily pressure from two days before to two days after each headache | More migraine when pressure fell by more than 5 hPa from the headache day to the next day; less when it rose by more than 5 hPa over the following two days |
| Katsuki et al., Japan, 2023 | 4,375 app users | Hourly weather against 336,951 hourly headache records | Lower pressure and a significant decrease 6 hours before were associated with more headaches per hour |
| Cull, UK, and Osterman et al., Sweden, both 1981 (as summarised by Farah et al.) | 44 and 53 | Morning pressure and 24-hour changes | Fewer migraines when 6 a.m. pressure was below 1005 mb than at 1016–1025 mb; a rise of more than 15 mb in 24 hours went with fewer migraines |
| Hoffmann et al., Germany, 2015 | 100 | Pressure every 4 hours and its change over the preceding 24 hours | Significant weather sensitivity in a subgroup; no significant association when everyone was pooled |
| Zebenholzer et al., Austria, 2011 | 238 | 11 weather measures and their day-to-day changes over 90 days | No finding stayed significant after correcting for multiple testing |
The 2025 systematic review by Farah and colleagues, which pulled these and eight other studies together, summed it up as inconsistent: several studies linked pressure drops or rapid fluctuations with more frequent migraine, fewer found an effect on severity, and none found one on duration.
A level, a drop and a timing are three different things
The figure quoted most often online, “6 to 10 hPa”, comes from the Okuma study, and it is easy to misread. It describes how far below the standard 1013 hPa the pressure was when attacks began. It does not say pressure fell by 6 to 10 hPa within a day. On a day that started at 1020 hPa, reaching 1005 hPa is a 15 hPa fall; on a day that started at 1009 hPa, it is only 4.
The Kimoto figure is a change, but look at its timing: more migraine was reported on days before a further fall of more than 5 hPa. In weather terms, that fits headaches occurring while a low-pressure system is still approaching rather than after it arrives. The Katsuki study looked at a much shorter window, the six hours before each logged headache.
None of these can be turned into a rule such as “a 6 hPa drop means a migraine tomorrow”. They describe what was more common across a group, measured in one particular way.
How big is a 6 hPa change?
Weather services describe pressure tendency over three hours. In the UK Met Office's station-report terms, a change of 3.6 to 6.0 hPa in three hours is “quickly”, and more than 6.0 hPa is “very rapidly”. So 6 hPa in six hours, the definition of a major change in Barometric Pressure Tracker, averages 3 hPa every three hours: a clear, sustained change rather than an extreme one. In inches of mercury, 6 hPa is about 0.18 inHg.
How often such changes happen depends on where you live and the season. That is why the app's pattern card does not just count how many of your entries followed a big change; it also measures how often any hour at your locations falls within 24 hours of one. If big changes are common where you live, a lot of headaches will land near them by chance alone.
Why your own threshold beats a population average
The Hoffmann study is the best argument for looking at yourself rather than at averages. Across 100 people, pooled data showed no significant link, yet a subgroup was clearly sensitive. The authors also found that, even for sensitive patients, weather data could not tell in advance when the next attack would come, mainly because attacks were relatively rare compared with the number of weather changes.
In practice that means two things. First, a pressure threshold is a way of deciding which changes you want to notice, not a way of knowing what your head will do. Second, it is worth testing more than one threshold against your log. Some people's entries cluster after fast falls; others, if anything, after rises.
Do rising pressure changes count?
Possibly. The Okuma study looked at deviations both below and above the standard. The Vienna study found, before correction for multiple comparisons, that a ridge of high pressure went with more headache occurrence. The older Cull and Osterman results point the other way, with fewer migraines at low pressure. Barometric Pressure Tracker counts a change of 6 hPa or more within 6 hours in either direction as major, so a fast rise is included in your pattern as well as a fast fall.
Choosing what to be alerted about
If you use pressure alerts, start broad and tighten later. Alerts are part of Pro, and their sensitivity runs from major swings only, through a recommended medium setting for meaningful changes, to high for smaller changes, plus a custom threshold and time window of your own. A sensible approach:
- Begin with the medium setting and log every headache, including mild ones, for at least eight weeks.
- Look at your pattern card: are your entries more common after major changes than the normal rate where you live?
- If yes, consider a custom threshold close to the changes your entries followed. If no, a quieter setting will spare you alerts that do not tell you much.
Whatever you choose, an alert is information to plan with. Discuss any change to how you take medication with your doctor.