do suicides and homicides by women peak in the days before menstruation?
“We know that both suicides and homicides peak in this days and before the cycle.”
a guest, a psychiatrist and researcher in women's mental health said on Psychiatry & Psychotherapy Podcast, 2026-10-24, at 39:04. hear it in context →
the episode: Perimenopause and Mental Health: Depression, Anxiety, Brain Fog, Sleep, and Hormone Therapy (at 39:04). we link the show's own page; we do not re-host audio.
first, the part that holds
there is a real signal behind this. a 44-study systematic review found that non-fatal suicidal behaviour clusters in the low-oestrogen phases around menstruation (late luteal and early follicular). a 1980 study of 50 women charged with violent crimes found 44% offended in the days around menstruation.
what we found when we went and looked
we read the abstracts of the main reviews and autopsy studies and did not find that completed suicides and homicides by women are known to peak in the days before menstruation. the clearest finding, from a 2006 review of 44 studies, covers suicide attempts and self-harm, not deaths. the autopsy studies of completed suicide disagree with each other. one found deaths spread evenly across the cycle (50 suicides), one found too many in the post-menstrual proliferative phase (50 cases), and a 2023 meta-analysis of five studies put 45.2% in the secretory (luteal) phase, which is about the share you would expect from a phase that lasts around half the cycle. on homicide, the record is one 1980 study of 50 women charged with violent crimes of any kind. its window includes menstruation itself, and its abstract does not report how many of the offences were homicides.
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| completed suicides by women peak in the days before menstruation, and "we know" this | overstated | the autopsy studies of completed suicide disagree. one coroner series of 171 deaths, 50 of them suicides, found deaths "distributed evenly as one would expect on theoretical grounds". a 2025 histology study of 50 suicides found too many in the proliferative (post-menstrual) phases, with 36% early and 28% late proliferative (p<0.001). a 2023 meta-analysis of five studies put 45.2% (95% ci 36.7-53.7%) in the secretory phase, which by our own rough arithmetic is no more than the share of a phase lasting about half the cycle; its conclusion is only that the menstrual phase carries lower risk. the 2006 systematic review drew its conclusion for non-fatal behaviour and stated none for completed suicide. limitation: all of these are small, cross-sectional and abstract-level [1][3][4][5] |
| homicides by women peak in the days before menstruation | overstated | the only primary record we found is one study of 50 women charged with crimes of violence: "44% committed their offence during the paramenstruum (P less than 0.02)" and there were significantly fewer offences around and after ovulation. the paramenstruum covers the days around menstruation, menstruation included, not only the days before it. the abstract does not say how many of the offences were homicides, and the offences were "unrelated to symptoms of premenstrual tension". it is a single small sample from 1980 that has not been replicated in anything we found [2] |
| non-fatal suicidal behaviour (attempts, self-harm) is more common in the low-oestrogen phases around menstruation | confirmed | a systematic review of 44 studies: "studies with better methodology suggest that a positive relationship exists between aspects of the menstrual cycle and non-fatal suicidal behaviour". this behaviour was more frequent in phases when oestrogen is lowest (late luteal and follicular) and in people with premenstrual syndrome. this is the part of the claim the record supports, but it concerns attempts and not deaths [3] |
| the cycle phase of women who are homicide victims is skewed | false | in the same coroner series, 31 homicide victims were part of a distribution of deaths across cycle phases that was "distributed evenly". this is about victims, not perpetrators, so it does not test what the speaker described. we include it because "homicides peak" is often heard as a claim about women who die [4] |
both directions
for the speaker: the 2006 systematic review links suicide attempts and self-harm to the late luteal and early follicular phases and to premenstrual syndrome. the 1980 study found a significant cluster of violent offences around menstruation and a significant shortage around and after ovulation. both findings point toward the perimenstrual window the speaker described. the quote ("in this days and before the cycle") may also be meant to cover menstruation itself, and that is how the 1980 study defined its window.
against: for completed suicide, the autopsy studies disagree. one found no relationship with cycle phase, one found an excess in the post-menstrual proliferative phase, and the meta-analysis's secretory-phase figure is no larger than that phase's share of the cycle. the review that found a cycle effect limited its conclusion to non-fatal behaviour. for homicide, the evidence is one 50-woman sample of violent offences in general, with no homicide-specific figure in the abstract. its window includes menstruation, and its authors found the offences unrelated to premenstrual symptoms. \"we know\" claims more certainty than this record holds.
what this cannot say
pubmed and pmc article pages would not load in this session, so we read the abstracts through ncbi's e-utilities interface, not the full texts. we may therefore have missed study-level detail, such as how many of the 1980 offences were homicides or how the meta-analysis weighted its five studies. the autopsy studies date the cycle phase from endometrial histology, which is imprecise and cannot isolate the few days before menstruation. every study here is small (about 50 deaths or offenders) and cross-sectional. our point that 45.2% is about what a luteal phase lasting half the cycle would produce is our own arithmetic, assumes a typical cycle, and is not the meta-analysis authors' conclusion. we did not search non-english literature or every prospective ideation study. some prospective work on suicidal ideation does point to a perimenstrual peak, but ideation is not the same as death.
sources, each one fetched and read
- Alnashwan YA et al. Incidence and comparison of suicide in various phases of the menstrual cycle: a systematic review and meta-analysis. Acta Inform Med. 2023;31(1):76-83. doi:10.5455/aim.2023.31.76-83. PMID 37038486 What we read there: "Five articles were shortlisted... Incidence of suicide in the secretory phase was highest at 45.2% [95% CI, 0.367-0.537]." conclusion: "the menstrual phase has a lower risk of mortality due to suicide when compared to the other two phases." background: "there are no conclusive findings." Fetched 2026-11-03 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/37038486/
- d'Orbán PT, Dalton J. Violent crime and the menstrual cycle. Psychol Med. 1980;10(2):353-9. doi:10.1017/s0033291700044123. PMID 7189893 What we read there: "Of 50 women charged with crimes of violence, 44% committed their offence during the paramenstruum (P less than 0.02) and there was a significant lack of offences during the ovulatory and post-ovulatory phases... Offences were unrelated to symptoms of premenstrual tension." Fetched 2026-11-03 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/7189893/
- Saunders KE, Hawton K. Suicidal behaviour and the menstrual cycle. Psychol Med. 2006. doi:10.1017/S0033291706007392. PMID 16573848 What we read there: 44 studies (13 on suicide). "Studies with better methodology suggest that a positive relationship exists between aspects of the menstrual cycle and non-fatal suicidal behaviour", more frequent when oestrogen is lowest (late luteal and follicular phases) and in those with premenstrual syndrome. the conclusion addresses non-fatal behaviour only Fetched 2026-11-03 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/16573848/
- Vanezis P. Deaths in women of reproductive age and relationship with menstrual cycle phase. Forensic Sci Int. 1990. doi:10.1016/0379-0738(90)90284-6. PMID 2210551 What we read there: 171 autopsied women (46 natural, 44 accidental, 50 suicides, 31 homicides), cycle phase determined from endometrium. deaths "were distributed evenly as one would expect on theoretical grounds" Fetched 2026-11-03 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/2210551/
- Nath A, Jain P, Hazarika P. Cross-sectional study on the correlation between menstrual cycle phases and suicidal behavior: a histological analysis. Cureus. 2025. doi:10.7759/cureus.81346. PMID 40296969 What we read there: 50 women who died by suicide: "the early proliferative phase accounted for 18 cases (36%), followed by the late proliferative phase with 14 cases (28%)", with "an overrepresentation in the proliferative phases" (p<0.001). the authors note that sample size limits the detection of smaller associations Fetched 2026-11-03 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/40296969/
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam