Why is my ADHD worse before my period?

What daily tracking found

This is the strongest study design in the batch, and worth understanding because it avoids a problem most of this literature has.

Zaritsky, Reed and Evans recruited 30 female participants already being treated with amphetamine salts for ADHD, and asked them to complete 35 daily online surveys tracking ADHD symptoms, mood and medication use through the menstrual cycle.

Because participants reported day by day rather than recalling afterwards, the design sidesteps the recall problem that affects retrospective surveys—where knowing you expect a bad week can shape how you remember it.

They found ADHD symptom severity was significantly associated with cycle phase: most severe during the menstruation phase, comparatively milder in the mid-follicular phase. That difference was positively correlated with subjective changes in negative mood. The authors framed the finding as something that could inform clinical and prescribing practice.

The limits: 30 participants is a pilot, everyone in it was already medicated, and everyone already had a diagnosis. It establishes that variation happens in this group, not how large it is generally or what it looks like unmedicated.

What women report, and where it differs

Self-report is far larger and agrees on the existence of a pattern, though not exactly on its timing.

In Osianlis and colleagues’ cross-sectional online survey of 600 female participants reporting an ADHD diagnosis, 88.6% of the premenopausal participants not taking hormonal therapy reported changes in their ADHD symptoms across the menstrual cycle. Most of those described a worsening during the luteal phase—the stretch after ovulation and before menstruation. Self-selected sample, recalled report, so it describes these participants rather than a rate.

That’s a slight mismatch with the daily tracking, which found the worst symptoms during menstruation itself and mildest in the mid-follicular phase. The two stretches are adjacent, and late luteal runs directly into menstruation, so this may be a difference in where people draw the boundary rather than a contradiction. It may also be the difference between what’s recalled and what’s recorded at the time.

Worth naming rather than smoothing over: two reasonable sources give slightly different answers to “which week is worst,” and your own tracking will tell you more about you than either.

Boyd and colleagues add a third data point. Their premenstrual comparison covers 600 of the 602 participants—376 reporting ADHD, 224 without—and found more severe premenstrual symptoms in the ADHD group (χ2 = 204.7, p < .001) alongside higher rates of menstrual irregularity (χ2 = 14.2, p < .001). Cross-sectional, self-reported diagnosis, recruited through support groups and social media, so this describes the sample rather than a population.

Premenstrual difficulty and PMDD

Premenstrual dysphoric disorder is a diagnosable condition involving severe mood symptoms in the premenstrual phase, and it appears more frequently in this population.

Groenman, Welling and Pieters compared 199 women aged 20 to 40 who weren’t using hormonal contraceptives: 89 with ADHD, 39 autistic, and 71 controls. Using the Premenstrual Symptoms Screening Tool, both the ADHD and autistic groups showed significantly elevated provisional PMDD rates compared with controls.

Their dimensional analysis found positive associations between premenstrual symptoms and neurodivergent traits—particularly inattention, hyperactivity-impulsivity, and social skills difficulties—with the impact of premenstrual problems similarly associated with those traits across diagnostic groups. One hypothesis didn’t survive: sensory hypersensitivity was not independently associated with premenstrual symptoms once ADHD and autism characteristics were controlled for.

The authors added a caveat worth repeating, because it applies to most studies on this page: prevalence estimates are only as reliable as the recruitment strategies behind them. A screening tool also produces provisional PMDD, not a diagnosis, which requires prospective symptom tracking across cycles.

If premenstrual mood symptoms are severe—and for some people they include hopelessness or thoughts of self-harm—that’s worth raising with a clinician rather than absorbing. PMDD is recognised and treatable, and the treatments differ from ADHD treatment.

It isn’t only an ADHD pattern

The Groenman comparison is the only study here that put ADHD and autistic women side by side, and its result is worth separating out: both groups screened positive for provisional PMDD at higher rates than controls, and there was no significant difference between the two neurodivergent groups.

So this isn’t a specifically-ADHD phenomenon as far as that study can tell. Autistic women in the same sample showed a comparable pattern.

That’s one study with 39 autistic participants, which is a small group on which to rest a conclusion. But it’s a caution against writing this up as an ADHD story when the evidence that exists doesn’t draw the line there.

The proposed mechanism

The usual explanation is that oestrogen supports dopamine signalling, ADHD involves dopamine regulation, and so the low-oestrogen stretch of the cycle worsens symptoms.

The 2025 review by Kooij and colleagues for the Eunethydis Special Interest Group on Female ADHD examines the interplay of oestrogen and progesterone with dopaminergic pathways and proposes that periods of lower oestrogen may affect cognition. A separate narrative review by Wynchank and colleagues covers menstrual-cycle hormonal fluctuations and cognitive functioning in ADHD specifically.

Both are narrative reviews rather than experiments, and several authors overlap between them and other papers cited here—this is a small research community, so agreement between these sources is less independent than it looks.

The mechanism is plausible and consistent with the observed pattern. It has not been demonstrated as a causal pathway from hormone level to ADHD symptom severity, and the reviews themselves call for the longitudinal studies that would test it.

What this doesn’t establish

Four things, so the page doesn’t leave a stronger impression than the evidence supports.

It doesn’t establish how large the effect is. The prospective study had 30 participants, all medicated.

It doesn’t establish that this happens to everyone, or how often. High proportions reported change in self-selected online samples of people already identified with ADHD; that is not a population rate.

It doesn’t establish a mechanism, only a proposed one.

And it doesn’t establish what to do about it. No dosing protocol, no tested intervention, no schedule anyone has trialled.

What you can do with it

Practical suggestions, not tested interventions.

Track it before you act on it. Two or three cycles noting the date, your symptoms, and what was hard tells you your own pattern, which the studies can’t. It also turns “I think this is a thing” into something a clinician can work with.

Plan around the pattern you find, not the one in the literature. If the studies disagree about which week is worst, yours is the one that matters. Where you have any control over deadlines, demanding meetings, or big decisions, that record is what to schedule around.

Separate the strands. Cycle-related mood change, ADHD symptoms, and PMDD are different things that can occur together. Noticing which of them is driving a given week is useful, and it is also hard, which is a reason to bring the record to someone rather than adjudicate it alone.

Take the bad week’s evidence less seriously. A recurring theme in what people describe is concluding, during the worst stretch, that they are fundamentally failing. If the pattern is real, that conclusion is arriving on schedule, which is a reason to distrust its timing.

Medication questions

The prospective study’s authors framed their finding as something that could inform prescribing practice, and it’s worth being precise about what that means: it means researchers think this is worth clinicians investigating, not that a protocol exists.

No dosing schedule across the cycle has been established. Reviews in this area note that pharmacological research tailored to hormonal fluctuation is lacking. So if your medication seems less effective in a particular week, that’s a real thing to raise with your prescriber, and not something to adjust on your own.

Bring the tracking. A record of dates and symptoms is the most useful thing you can put in front of someone, and it’s the part of this that only you can produce.