From menarche to menopause: the impact of reproductive factors on the metabolic profile of over 65,000 women

We explored the relation between age at menarche, parity and age at natural menopause with 249 metabolic traits, measured using nuclear magnetic resonance (NMR), in up to 65,487 UK Biobank women using multivariable regression (MV), Mendelian randomization (MR) and a male negative control (parity only). Older age of menarche was related to a less atherogenic metabolic profile in MV and MR, which was largely attenuated when accounting for adult body mass index. In MV, higher parity related to complex changes in lipoprotein-related traits; these were not observed in male negative controls and were imprecisely estimated in MR. In MV and MR, older age at natural menopause was related to lower concentrations of inflammation markers, but inconsistent results were observed for LDL-related traits due to chronological age-specific effects. Our findings support a role of reproductive traits on later life metabolic profile and provide insights into identifying novel markers for the prevention of adverse cardiometabolic outcomes in women. Summary box What is new?O_LIMarkers of womens reproductive health are associated with several common chronic conditions. Whilst some attempts have been made to explore the extent to which these associations are causal, metabolites could act as mediators of the relationship between reproductive markers and chronic diseases. C_LIO_LIOlder age of menarche was related to a less atherogenic metabolic profile in multivariable regression and Mendelian randomization, however, this was largely attenuated when accounting for adult body mass index. C_LIO_LIIn multivariable regression, higher parity related to complex changes in lipoprotein-related traits. Whilst these were not observed in male negative controls, suggesting a potential causal effect in females, they were not replicated in the Mendelian randomization, possibly due to imprecise estimates. C_LIO_LIOlder age at natural menopause was related to lower concentrations of inflammation markers in both multivariable regression and Mendelian randomization. Consistent results were observed for LDL-related traits when stratified by chronological age. C_LI ImplicationsO_LIGiven that the age at menarche results were largely attenuated to the null when accounting for adult BMI, it is likely that age at menarche itself may not causally relate to the metabolic profile. C_LIO_LIThese results, particularly for parity and age at menopause, could contribute to identifying novel markers for the prevention of adverse cardiometabolic outcomes in women and/or methods for accurate risk prediction. For example, consistent with other studies, higher parity was associated with unfavourable (e.g. higher number of particles and lipid content in VLDL and higher glycine) changes in the metabolic profile. Similarly, older age at menopause was related to higher lipid content in HDL particles and lower systemic inflammation, as proxied by GlycA. C_LI