What the research found

Researchers used a mathematical technique (adapted from astronomy) to align lab data by time-to-menopause rather than chronological age, solving a long-standing problem in menopause research: most medical records don't capture when a woman's final period actually occurred. They analysed over 300 million lab tests from two large datasets—US national health surveys spanning 1999–2022 and an Israeli health system covering 1.3 million women—and validated their approach against an existing long-term cohort study.

The results show that hormonal shifts begin substantially earlier than previously recognised. AMH (a marker of ovarian reserve) accelerates its decline roughly 15 years before the final period; FSH (the pituitary signal to the ovaries) rises about 14 years before; and estradiol falls about 12 years before. For a woman whose last period arrives around age 50, this places the first measurable hormone changes in her late 30s—roughly 7 years earlier than prior estimates suggested.

At the final menstrual period itself, 52 of 90 biomarkers showed an abrupt shift rather than a gradual trend: bone turnover markers, liver and kidney tests, blood lipids, blood counts, haemoglobin, and lead levels all jumped. Men showed no equivalent step-change at an equivalent age, suggesting this is specific to oestrogen loss. Clinically, osteoporosis incidence and cardiovascular event rates accelerated after menopause; depressive symptoms and sleep disturbance peaked in the immediate postmenopausal year before partially resolving.

Why it matters for you

If you're tracking biomarkers on MyKine, this reframes the timeline. Rather than waiting for obvious symptoms or a recognised diagnosis, the data suggests screening should begin in your late 30s or early 40s—particularly if you're female. A baseline lipid panel, liver panel, bone health assessment, and mood screening now give you a reference point to detect the magnitude of your individual shift when it comes, not just population averages.

For those using hormone therapy, the observational evidence suggests it may modulate the magnitude of the postmenopausal lab shift, particularly in bone, liver, and kidney markers. If you're already on or considering HRT for symptom management, you now have additional context for that conversation with your clinician: the therapy appears to preserve metabolic continuity across the transition, rather than simply masking acute symptoms. However, this doesn't establish causation or prove prevention of disease—only that the biomarker trajectory looks different.

If you track sleep quality, mood, or inflammatory markers (like CRP) in your app, expect potential volatility around the final menstrual period. Oral oestrogen appears to raise CRP; transdermal patches may not. Depressive symptoms are most common in the first year after menopause but don't persist uniformly—some users will see spontaneous improvement while others won't, so individual tracking matters more than population trends here.

Caveats

• Observational design for HRT outcomes: the hormone therapy analysis used matched comparisons, not randomisation. Women who access and use HRT may differ in unmeasured ways from those who don't (healthy-user bias).
• Limited HRT characterisation: the US data captured mostly oral pills; the Israeli data used two specific oral combinations. This tells us little about transdermal patches or other formulations.
• Population snapshot, not individual trajectories: the data combines women at different life stages rather than following single individuals across the transition. The "cliff" is a population average; individual experience will vary.
• No causation for chronic disease: observational alignment of biomarkers with later osteoporosis and cardiovascular outcomes doesn't prove HRT prevents these conditions. Bone findings align with existing randomised trial data; others do not.
• Subset effects on mood: the averaging masks meaningful individual variation. Some women remain depressed long-term; the postmenopausal year is not uniformly difficult.