Computational Reconstruction of 300 Million Lab Tests Shows Abrupt Cliff-Like Shifts in 52 of 90 Markers at Menopause

Uri Alon and colleagues applied a computational model to reconstruct trajectories around the final menstrual period using electronic health records spanning 1999-2024. Fifty-two markers spanning sex hormones, bone turnover, liver enzymes, renal function, and lipid profiles displayed abrupt inflections rather than linear decline. Anti-Müllerian hormone and estradiol began accelerating 15 and 12 years earlier, respectively, while follicle-stimulating hormone rose sharply around year minus 14. Menopausal hormone therapy correlated with attenuated post-transition slopes, particularly in bone and renal parameters, consistent with estrogen-receptor density in those tissues.

The pattern aligns with prior longitudinal cohorts such as the Study of Women’s Health Across the Nation, which documented accelerated bone loss and shifts in adipokines but lacked the scale to resolve the precise timing of the inflection. Estrogen-sensitive systems showed the largest deltas; systems with weaker receptor expression remained stable, supporting a causal rather than purely chronological mechanism. No comparable discontinuity appeared in age-matched male controls, reinforcing that the transition is reproductive rather than generic senescence.

Limitations include reliance on self-reported final menstrual period dates and absence of randomization for hormone therapy exposure. Next required studies are prospective cohorts that begin serial sampling 15 years before expected menopause and randomize early hormone therapy to test whether intercepting the pre-cliff acceleration reduces later fracture and cardiovascular incidence by measurable thresholds.