A Nature Medicine study from Israel’s Weizmann Institute reconstructed blood-test trajectories for more than a million women in the United States and Israel. It reports that 52 of 90 routine measures, covering bone, liver, kidney, lipids and blood chemistry, change abruptly at the final menstrual period, and that hormonal changes start 12 to 15 years before it. Independent experts welcome the scale of the work but warn against reading it as a “cliff” of decline.
Menopause is one of the most universal events in human biology, yet one of the least studied. Most research has focused on hot flushes and fertility. A large new study published in Nature Medicine on 12 October argues that menopause is better understood as a sudden, whole-body transition, and that the hormonal run-up to it begins far earlier than clinicians usually assume.
The team, led by systems biologist Uri Alon of the Weizmann Institute of Science in Rehovot, Israel, with first author Glen Pridham, describes a “physiological cliff” at the final menstrual period. The paper went online on Monday; embargoed press materials were released at 20:30 IST.
The problem the study solves
The obvious way to study menopause would be to follow thousands of women for decades with repeated blood tests. That is slow and expensive, and it has been done only on modest scales. The big American cohort, SWAN, followed about 3,300 women. Large health databases contain vast numbers of blood tests, but they are cross-sectional snapshots of different women at different ages. Because women reach their final period at different ages, averaging by age blurs any sudden change at menopause into an apparently gentle slope.
The Weizmann team borrowed a solution from astronomy. Astronomers use deconvolution to sharpen images blurred by a telescope’s optics. The researchers treated variation in the age at menopause as a similar kind of blur. They then mathematically reconstructed how each lab measure would look if every woman’s data were aligned to her own final period.
The data
They applied the method to two very large datasets: – NHANES: the US National Health and Nutrition Examination Survey, 12 survey waves from 1999 to 2022. – Clalit: records from Israel’s largest health maintenance organisation, covering about half the country’s population, including 1.3 million women aged 30–70 between 2004 and 2024.
Together these held more than 300 million laboratory results from over a million women. The analysis covered 90 lab tests and 31 clinical outcomes, and men were used as a comparison group.
What they found
- A step at the final period. In 52 of 90 tests, there was a statistically significant jump at the final menstrual period, ranging from 0.1 to 1.4 standard deviations. The changes spanned sex hormones, bone markers, liver and kidney function, blood lipids and blood chemistry. Ferritin, the body’s iron store, jumped by 0.6 standard deviations.
- Not seen in men. Men showed no comparable jumps at the equivalent ages, which the authors read as evidence that the step is driven by menopause rather than ageing in general.
- Consistent across countries. Jump sizes in the US and Israeli data were strongly correlated (r = 0.89).
- An earlier start. Anti-Müllerian hormone, a marker of ovarian reserve, began an accelerated fall about 15 years before the final period. Follicle-stimulating hormone began rising faster about 14 years before, and oestradiol falling faster about 12 years before. The authors note that previous estimates put the start at about seven years before.
- Health outcomes. Osteoporosis and cardiovascular outcomes rose more steeply after menopause. Anaemia became less common, as expected when menstruation stops. Depression scores rose temporarily around the transition.
- Hormone therapy. Women using menopausal hormone therapy (MHT) had smaller jumps in many measures, especially bone, liver and kidney markers.
What it does not show
The authors themselves list important limitations. The data are retrospective and cross-sectional, and menopause status is self-reported. In NHANES, about 27 per cent of women had unknown menopause status and had to be imputed, while the Clalit data contain no menstrual information at all. The method depends on an assumed distribution of age at menopause, although sensitivity analyses suggested this had little effect.
Independent experts contacted by the Australian Science Media Centre were cautious. Professor Martha Hickey, Professor of Obstetrics and Gynaecology at the University of Melbourne, warned that calling the findings a “physiological cliff” was “potentially alarming for women”. She said the study does not show dramatic physical or mental decline, and that its design is low-level evidence. She also said the hormone-therapy finding is inconsistent with randomised trial evidence. Only about 9 per cent of women in NHANES used MHT, and those who do may be healthier in other ways, the so-called “healthy user” effect. The authors agree that randomised trials are needed before drawing conclusions about therapy.
Professor Andrew Shelling of the University of Auckland called the dataset unprecedented and the approach innovative, but described the findings as “hypothesis-generating rather than conclusive”. Associate Professor Mia Schaumberg of the University of the Sunshine Coast noted that the largest changes were in oestrogen-sensitive systems such as bone and liver.
Why it matters
If confirmed, the findings have practical implications. A sharp change around the final period argues for checking bone health, lipids and liver and kidney function around that time, rather than spreading attention evenly through midlife. A hormonal transition beginning in the late thirties or early forties suggests that perimenopausal symptoms may be recognised and managed earlier.
The Indian context
The study’s populations are American and Israeli, and its numbers should not be assumed to apply directly to Indian women, whose age at menopause, diet, prevalence of anaemia and access to hormone therapy differ. Indian population studies have documented a substantial share of women reaching natural menopause before 50. Two conditions highlighted by the study, osteoporosis and cardiovascular disease, are major and under-diagnosed burdens among older Indian women. Menopause care barely features in most primary-health programmes. A body-wide view of the transition strengthens the case for including midlife women’s health, including bone and cardiovascular screening, in routine care.
What comes next
The authors say prospective studies that follow the same women over time, and randomised trials of hormone therapy, are needed to confirm the step changes and test interventions. Replication in other populations, including South Asian cohorts, would show how general the pattern is.
– Rithvisha Kiran
Key facts
– Paper: Pridham G, Hayut Y, … Alon U. “Population-scale reconstruction of menopause dynamics reveals a physiological cliff at the final menstrual period.” Nature Medicine, published 12 Oct 2026. DOI: 10.1038/s41591-026-04614-0 (open access).
– Data: more than 300 million lab tests from over 1 million women (NHANES 1999–2022; Clalit 2004–2024); 90 lab tests; 31 outcomes.
– Main result: 52 of 90 tests jump at the final period (0.1–1.4 SD); no equivalent jumps in men; US–Israel agreement r = 0.89.
– Timing: AMH falls about 15 years, FSH rises about 14 years and oestradiol falls about 12 years before the final period.
– MHT: associated with smaller jumps; observational only; about 9% of NHANES women were users.
– Caveats: retrospective, cross-sectional and self-reported data; experts caution against the “cliff” framing.
