Menopause as a cliff: 52 of 90 blood tests jump at the final period
A Nature Medicine analysis of 300 million routine lab tests from US and Israeli records found that many measurements step abruptly at the final menstrual period instead of drifting, that men show no such step, and that women on hormone therapy showed smaller steps.
By Gale Staff · October 13, 2026 · Nature Medicine
Not reviewed by a clinician.

The short answer
In two large health datasets, 52 of 90 laboratory tests showed a significant step-like jump at the final menstrual period compared with men, with jump sizes ranging from 0.1 to 1.4 standard deviations across hormone, bone, liver, kidney and lipid tests. Sex hormones began changing more than a decade earlier, but the study is observational and cross-sectional, the timing of the final period was statistically reconstructed rather than measured in one of the two cohorts, and the smaller jumps seen in women using hormone therapy cannot be read as proof that the therapy caused them.
Who to see
Gynecologist or menopause specialist
Many lab results, from hormone and bone tests to liver, kidney and lipid tests, jumped at the final period; a menopause specialist can talk through your lab results.
Find care →Why a blood test at 53 can look so different from one at 47
Two routine blood panels, taken six years apart on either side of the last period, can differ in ways that neither the patient nor the chart has a tidy explanation for. Cholesterol is higher, a liver enzyme has crept up, a bone marker has moved, and the usual shorthand is that this is what getting older does. The question that follows the headline is a fair one: is menopause a slope that people slide down for years, or is something happening at a particular moment?
A paper in Nature Medicine, published online on 12 October 2026 by Glen Pridham, Uri Alon and colleagues, argues for the second picture. Its title says the final menstrual period (FMP) marks a physiological cliff, and the data it rests on are far larger than the small follow-up studies that shaped the earlier, gradual-transition view.
What the researchers did
The obstacle the authors name is simple. Almost no large medical dataset records the date of a woman's final period, so lab results are filed by age. But the final period arrives at different ages for different women, with a median of 50 and a middle range of 45 to 53 in the US survey data. Averaging by age therefore smears a sharp event across many years, much as a sharp astronomical image blurs when the telescope distorts it. The authors borrowed that astronomy idea: they used the known spread of final-period ages to mathematically undo the blur, producing the average trajectory of each test as if every result had been lined up on the final period.
They applied this to two independent sources. One is the US National Health and Nutrition Examination Survey, 12 two-year waves from 1999 to 2022, where women reported their menopause status and age. The other is the Israeli health insurer Clalit, covering about half the Israeli population, with 1.3 million eligible women aged 30 to 70 across 2004 to 2024. In total, the paper says, the analysis covers more than 300 million laboratory tests from more than 1 million women. For each test, the team set aside people taking medications or carrying diagnoses known to affect it, excluded pregnancy, and ran the same alignment on men as a comparison, assigning men an apparent menopause time of age 50.
What they found: steps, not slopes
Across 90 types of test, women showed a significant jump at the final period, compared with men, in 52. The jumps ranged from 0.1 to 1.4 standard deviations, measured against a reference group of non-pregnant women aged 25 to 45. Ferritin, the iron-storage marker, is the paper's worked example: a step of 0.6 standard deviations at the final period. Large steps, of 0.5 standard deviations or more, were found in the sex hormones estradiol, anti-mullerian hormone and follicle-stimulating hormone. The less expected result was how many other systems moved with them: liver tests, blood counts, blood chemistry, kidney function, bone markers, lipids and blood osmolality, the concentration of dissolved particles in blood. Men's results changed gradually with age instead.
The two cohorts agreed. Jump sizes in the US and Israeli data were strongly correlated (r = 0.89), and a repeat using only women with recorded menopause status and a different age range agreed even more closely (r above 0.95). Most of the jumps went in the direction associated with higher disease risk. The exceptions were iron, hemoglobin and red-cell tests, which moved toward less anemia, which the authors say is probably owing to the end of menstrual bleeding.
Sex hormones began shifting well before the cliff. Anti-mullerian hormone started an accelerated fall about 15 years before the final period, follicle-stimulating hormone an accelerated rise about 14 years before, and estradiol an accelerated fall about 12 years before. The authors place the estradiol and follicle-stimulating hormone changes roughly seven years earlier than a well-known prospective study had reported. Note the scope of that claim: the decade-long lead is shown for these three hormones, not for every test that later jumps.
The paper also looked at self-reported outcomes in the US survey. Osteoporosis prevalence rose faster after menopause than in men, and heart attack and stroke prevalence also climbed more steeply. Depression scores rose briefly: 46 of 236 women (19 percent) scored above 10 on a standard nine-item questionnaire in the year after the final period, versus 698 of 6,495 (11 percent) over the following 24 years.
Hormone therapy and the smaller jumps
Among US women who reported current menopausal hormone therapy, about 9 percent of the sample, 39 of 80 tests differed from matched women not using it, and the size of most jumps was reduced. The bigger a test's jump, the bigger the apparent correction (r = 0.76). A second approach in the Israeli data, following 9,103 women prescribed estrogen-containing therapy before, during and after use, pointed the same way with a moderate correlation (r = 0.63). One marker went the other direction: C-reactive protein, an inflammation marker, was higher with therapy.
That is an association in observational data, and the authors say so, writing that the effect of therapy needs testing in prospective randomized trials. Women who use hormone therapy can differ in access to care and health habits in ways matching only partly removes.
The caveat that matters most, and why it still matters
Nobody here was followed through menopause. Each woman contributes results at one or a few moments, and the step is inferred by combining thousands of such snapshots with an assumed distribution of final-period ages. The Israeli records do not capture menopausal status at all, so that cohort leans entirely on the assumption. The authors tested it with simulations in which the average final-period age shifted by four years and found the method largely unaffected, and they note that a prospective study with detailed menopause tracking would still be the stronger design.
The results are lab values, not diseases. A jump of a few tenths of a standard deviation in a liver enzyme or lipid fraction is a shift in a measurement and does not say how many women will develop a condition because of it. The practical weight of the paper is therefore in what it changes about the picture, a sharp systemic event around the final period and an earlier hormonal prelude, and in what it makes plausible to study next, rather than in any individual patient's result.
What this study can't tell you
- Whether any individual woman experiences a cliff: the method reconstructs average trajectories for groups lined up on the final period, and the authors state it does not predict the final-period date for any one person.
- Whether the step is truly sudden in each person's body: with only one or a few results per woman and no follow-up through the transition, an abrupt average step cannot be separated from a fast change whose timing varies between individuals.
- How accurate the reconstruction is in the Israeli cohort: menopausal status is not recorded there, so the final-period timeline depends on an assumed age distribution, tested by simulation and sensitivity checks but not by observing the real dates.
- Whether hormone therapy caused the smaller jumps: users were about 9 percent of the US sample and may differ from non-users in access to care and lifestyle, and the authors call for randomized trials; the abstract's own word is 'appear' attenuated.
- Whether the lab shifts lead to disease: the outcomes analysis uses self-reported conditions in the US survey, shows prevalence curves rising after the final period, and cannot show that the lab changes produced them.
- Whether the decade-long hormonal lead applies beyond anti-mullerian hormone, follicle-stimulating hormone and estradiol; the early dysregulation is shown for those three, not for the other tests that later step.
- What happens in groups with few records: the authors flag missing data on hormonal intrauterine devices and endometrial ablation, which can hide the true final period, and the US figures rest on self-reported menopause timing.
The Gale read
This is a paper whose strongest feature is also its limit. Deconvolving the spread of final-period ages is a clever way to get a menopause-anchored timeline out of data nobody collected for that purpose, and the agreement between a US survey and Israeli insurer records, together with the near absence of any step in men, makes a pure ageing explanation harder to hold. But the cliff is a statistical reconstruction of group averages, so what it supports is the claim that menopause is a systemic event visible in routine bloodwork, not a claim about the shape of any one woman's transition or about disease risk. The hormone therapy finding is the most tempting to over-read, and the authors do not over-read it: it is observational, in a self-selected minority, and flagged for trials. The most useful way to carry this paper is as evidence for the cliff as an average pattern and a hypothesis-generating signal for therapy, from a design that a prospective study could now test directly.
Common questions
does menopause suddenly change your blood test results
On average, yes for many tests. In a Nature Medicine analysis of more than 300 million lab tests from two cohorts, 52 of 90 tests showed a significant step-like jump at the final menstrual period compared with men, ranging from 0.1 to 1.4 standard deviations. The study reconstructs group averages, so it cannot say how any one woman's results will move, and the final-period timing was statistically inferred rather than directly recorded in the Israeli data.
when do hormone changes start before menopause
In this study, much earlier than the final period itself. Anti-mullerian hormone began an accelerated fall about 15 years before it, follicle-stimulating hormone an accelerated rise about 14 years before, and estradiol an accelerated fall about 12 years before. The authors place the estradiol and follicle-stimulating hormone changes roughly seven years earlier than a prior prospective study, and the finding is specific to these three hormones.
does hormone therapy prevent the changes in lab results at menopause
The study found that jumps looked smaller in women using hormone therapy, with 39 of 80 tests differing from matched non-users, but it is observational and cannot show that the therapy caused the difference. Users were about 9 percent of the US sample and may differ in access to care and lifestyle, and C-reactive protein, an inflammation marker, was higher with therapy. The authors call for randomized trials to test the effect.
is menopause a gradual change or an abrupt one
This analysis argues for abrupt in lab markers, with a step at the final menstrual period instead of a smooth slide, while men's results changed gradually with age. The step is an average pattern reconstructed from cross-sectional data, not an observation of individual women followed through the transition, so how sudden it feels or looks in one person is outside what the paper measured.
Sources
- 1.Pridham G, Hayut Y, Lavi-Shoseyov N, Neeman M, Hovav N, Toledano Y, Alon U. Population-scale reconstruction of menopause dynamics reveals a physiological cliff at the final menstrual period. Nature Medicine. 2026;32(10):3823-3831. Published online 12 October 2026. doi:10.1038/s41591-026-04614-0 link
- 2.Europe PMC open-access record (PMID 42864411, PMCID PMC13668668) of Pridham G, Hayut Y, Lavi-Shoseyov N, et al. Population-scale reconstruction of menopause dynamics reveals a physiological cliff at the final menstrual period. Nature Medicine. 2026 - the abstract and full text read for this story. link
2 sources, numbered by first appearance. General health information, not medical advice. AI-assisted editorial content — every citation independently verified. Editorial policy
More from Gale Staff
Research
Top-Third Vitamin D Came With 19 More Minutes of Sleep a NightResearch
A Stroke Drug Beat Placebo, 69.7% to 56.3%; An Earlier Trial Did Not
Published October 13, 2026 · Gale Staff