How to Tell If Your Migraines Are Hormonal: Count 3 Cycles

Nothing about a hormonal migraine feels different. There is no distinguishing symptom, no particular quality of pain, nothing you could notice mid-attack that would settle it. The question gets answered by arithmetic rather than sensation: whether your attacks land in a specific five-day window around your period, in at least two cycles out of three. Establishing that takes three months and about ten seconds a day.

Most articles on this answer a different question. They list symptoms said to accompany hormonal attacks and invite you to recognise yourself in the list, which is a test nobody fails. The definition used by the people who actually diagnose it contains no symptoms at all.

The window is five days, and there is no day 0

The International Headache Society sets the criteria, and they are public. Its classification of pure menstrual migraine without aura requires attacks occurring “on day 1 ± 2 (ie, days −2 to +3) of menstruation in at least two out of three menstrual cycles”.

Two notes attached to that sentence do most of the work.

The first day of bleeding is day 1. The day before it is day −1. ICHD-3 states plainly that there is no day 0, and that off-by-one is the single most common way people get this wrong. Written out, the window is: the two days before bleeding starts, the day it starts, and the two days after. Five days, not six.

If you count a day 0, every date after it shifts by one. An attack that genuinely fell on day 3 gets filed as day 4 and drops out of the window. Do that across three cycles and a real pattern dissolves into noise.

Two cycles out of three, not one convincing month

The two-of-three rule exists because a single month proves very little, and the arithmetic shows why.

Five days out of a 28-day cycle is about 18 percent of it. Suppose you have four attacks in a month and they fall entirely at random, with no hormonal component whatsoever. The chance that at least one of them lands inside the window anyway is better than even. Not a study finding — just what those numbers do.

That is the trap. One month where the attack arrived the day before your period feels like proof, and it is roughly a coin flip. Three cycles is the smallest count that separates a pattern from a coincidence, which is why the criteria ask for it.

What the count actually tells you

Two categories share the same window and the same two-of-three rule, and differ on one point.

Pure menstrual migraine means the attacks fall in that window and, in the classification’s words, “at no other times of the cycle.”

Menstrually-related migraine means the same window pattern holds “and additionally at other times of the cycle.”

If your count lands on the second, the cycle explains some of your attacks and leaves the rest unaccounted for. That is a useful result rather than a failed one, but it does mean the tracking job is not finished — the attacks outside the window still have something else going on, and finding it is the slower work covered in how to track migraine triggers without fooling yourself.

Both categories above are the versions without aura. The classification carries parallel entries for migraine with aura, on the same window and the same rule.

Withdrawal bleeds count, and most people assume they don’t

This is the detail that changes the answer for a lot of readers. For the purposes of the criteria, menstruation includes endometrial bleeding “resulting either from the normal menstrual cycle or from the withdrawal of exogenous progestogens, as in the use of combined oral contraceptives or cyclical hormone replacement therapy.”

The pill-break bleed is day 1. So is the bleed on cyclical HRT. Someone who assumed the whole question stopped applying to them when they started contraception has three cycles of countable data and has never counted it.

Whatever the count shows, it is information to bring to your GP or neurologist rather than a conclusion to act on alone, and that goes double if you use hormonal contraception, where the cycle and the treatment are entangled.

Doing it on paper, in ten seconds a day

You need exactly two marks. Everything else is arithmetic you do once, at the end.

1. On the first day of bleeding, write P.
2. On any day an attack starts, write M.
3. After three cycles, take each M and count its distance from the nearest P, remembering that the day before P is −1 and there is no zero. Days −2, −1, 1, 2 and 3 are inside.
4. Count how many of the three cycles contained at least one M inside the window.

Two of three, and you have the pattern. The monthly grid in the free printable migraine diary already has a box per day, so adding a P costs nothing.

One discipline matters more than the rest: record the attack when it starts, not when it ends and not the following morning. The onset date is the entire measurement here. A one-day error moves an attack across the edge of a five-day window, and by the evening you will genuinely misremember whether it began before or after you went to bed. Migraine Journal stamps the date at the moment you log, which removes the reconstruction step, but a pencil mark made at the time is worth more than a careful reconstruction made later by any method.

The error that erases a pattern that is really there

Irregular cycles break the count when you anchor on the wrong thing. If your cycle runs anywhere from 24 to 33 days, an expected date is a guess, and a window placed around a guess lands in the wrong week. Anchor on the day bleeding actually started, every cycle, even when that means the window falls somewhere you did not expect. There is a point past which that repair stops working: in perimenopause the cycle stops supplying a usable anchor at all, and the count has to move to a plain monthly rate instead.

The other one is filling the diary in at the end of the month. Retro-filling produces a tidy record of what you believe happened, and belief drifts toward the story you already suspect. If you think your migraines are hormonal, you will place ambiguous attacks nearer the window without noticing you did it. Marks made on the day cannot do that.

Three cycles of two marks a day gets you to a sentence worth saying out loud in an appointment: attacks inside the window in two of the last three cycles, plus four outside it. A neurologist can work with that, particularly alongside the four numbers they ask for anyway. If the count comes back positive, what helps once the pattern is confirmed is an appointment agenda rather than a treatment list. It also costs less effort than reading another symptom list that was never able to answer the question. The other tracking guides use the same marks-on-the-day method for different questions.