Menopause Showed Up as a Cliff in 300 Million Lab Tests. The Study Never Followed One Woman Through It.

Here is a quiet problem in menopause research: few big medical databases record the date of a woman's last period. Hospitals log her age, her cholesterol, her liver enzymes, her bone markers. The one date that anchors all of it goes missing, and the whole timeline blurs.

A team led by Glen Pridham at the Weizmann Institute of Science in Israel borrowed a fix from astronomy, where researchers sharpen a smeared telescope image by correcting for known distortion. Women reach their final menstrual period at different ages (the US survey data put the median at 50, with the middle half between 45 and 53), so lining everyone up by birthday smears the picture. Their algorithm reverses the smear. Published October 12 in Nature Medicine, the analysis covered more than 300 million lab tests from more than a million women, drawn from two sources. One was NHANES, a US health survey running from 1999 to 2022. The other was Clalit, an Israeli health insurer covering about half the country, which supplied 1.3 million women aged 30 to 70.

What jumped, and what did not

Sex hormones moved most, which surprises no one. Estradiol, AMH and FSH each jumped by half a standard deviation or more. The surprise was everything downstream: liver tests, kidney measures, blood fats, bone markers and the body's salt-and-water balance all stepped too. The authors call the pattern a step-like physiological cliff.

Most of the steps ran in the direction of higher disease risk. The exceptions were iron, hemoglobin and red blood cell tests, which improved, probably because monthly bleeding stops. The US and Israeli results agreed closely. Jump sizes in the two datasets correlated at r = 0.89, despite different populations and different ways of collecting the data.

The hormone clock starts earlier than the textbook window

AMH began dropping faster about 15 years before the final period. FSH started climbing about 14 years before, and estradiol started falling about 12 years before. For the last two, that is roughly seven years earlier than previous studies, including the SWAN cohort, had documented. Mia Schaumberg, a physiologist at the University of the Sunshine Coast in Australia who was not involved, said, as quoted by ScienceAlert, that the timeline suggests the body-wide changes could be predicted from hormone shifts far sooner than researchers assumed.

Sleep and mood: the part that fades

Not every effect stuck. In NHANES questionnaires, 46 of 236 women who reported their menopause timing, 19 percent, scored in the moderate-to-severe depression range in the year after their final period. Across the following 24 years the figure was 698 of 6,495, or 11 percent. Trouble falling asleep and insufficient sleep also rose briefly, then drifted back toward the slow pattern seen in men. Osteoporosis, heart attack and stroke went the other way, climbing faster after menopause and staying high.

Those depression numbers describe an average. The authors point out that a subset of women have depression that lasts after menopause, which a curve that fades back to baseline cannot show. The 236 is also a small slice of a very large study.

Is it really a cliff?

Martha Hickey, a professor of obstetrics and gynecology at the University of Melbourne who was not involved, said, as quoted by ScienceAlert, that the framing is potentially alarming for women and implies a dramatic physical and mental decline that the study does not demonstrate. The paper is plain that it describes group averages. Its algorithm does not predict any one woman's final period. It estimates what the average lab value looks like as if every woman had been lined up by her last period.

Andrew Shelling, a health scientist at the University of Auckland who was not involved, called the results interesting but not conclusive because the work is observational. The authors list their own limits: retrospective data, self-reported menopause status, and no way to follow the same women through the transition.

The hormone therapy comparison

About 9 percent of NHANES participants reported current menopausal hormone therapy. Their lab trajectories after menopause looked like a smoother continuation of their premenopausal ones, and 39 of 80 tests differed from matched women not on therapy, including bone, liver and kidney markers. One marker moved the other way. CRP, an inflammation marker, ran higher, which the authors link to how oral estrogen acts on the liver.

That comparison carries the weakest footing in the paper. Women who take hormone therapy may differ in access to care and health habits in ways records cannot capture, and the authors say randomized trials are needed to test the effect.

What the dataset offers is a sharper map of where to look. Checking whether any one woman's body follows the average needs the thing the databases lack, a recorded date and the same person measured on both sides of it. The authors say their method could be pointed next at puberty, pregnancy and disease onset.