Reassessing the association between age at menarche and cardiovascular disease: observational and Mendelian randomisation analyses

Abstract Background Observational studies have shown a non-linear U-shaped association between age at menarche and cardiovascular disease (CVD). Purpose The aim of this study was to reassess the observational association between age at menarche and CVD and investigate whether the causal relationship between age at menarche and CVD is non-linear. Methods We included women from the UK Biobank and EPIC-CVD study without a history of myocardial infarction or stroke at baseline. In observational analyses, we estimated study-specific hazard ratios (HRs) for myocardial infarction, ischaemic stroke, and haemorrhagic stroke using Cox-regression before meta-analysing them. In addition, we conducted non-linear Mendelian randomisation analyses based on fractional polynomials and performed linear Mendelian randomisation analyses using inverse-variance weighted regression. Results Our analyses included 283,210 women without a history of myocardial infarction or stroke at baseline, of which 8,468 had a myocardial infarction, 5,501 an ischaemic stroke, and 1,887 a haemorrhagic stroke. Mean age at menarche was 13.1 (SD 1.6) in the UK Biobank and 13.0 (SD 1.6) years in EPIC-CVD. The association between age at menarche and myocardial infarction and ischaemic stroke was U-shaped with higher risks in women aged 12 and ≥16 compared to those aged 13 years at menarche. We found no statistically significant association with haemorrhagic stroke. Our Mendelian randomisation analysis based on fractional polynomials suggested no non-linear relationship between genetically proxied age at menarche and any CVD endpoint, but each year higher genetically proxied age at menarche was statistically significantly related to a lower risk of myocardial infarction (HR 0.92 95% CI 0.86-0.99), but not to ischaemic (1.00 0.91-1.09) and haemorrhagic stroke (1.02 0.89-1.17). Conclusion In conclusion, although the observational association between age at menarche and myocardial infarction and ischaemic stroke was non-linear, our analysis revealed no indication of a non-linear relationship between higher genetically proxied age at menarche and risk of myocardial infarction, and there appeared to be no causal link with ischaemic and haemorrhagic stroke.