Identifying potential causal effects of age at menopause: a Mendelian randomization phenome-wide association study

Abstract

Age at natural menopause (ANM) is associated with a range of health-related traits, including bone health, female reproductive cancers, and cardiometabolic health. Our objective was to conduct a Mendelian randomization phenome-wide association study (MR-pheWAS) of ANM. We conducted a hypothesis-free analysis of the genetic risk score (GRS) for ANM with 18,961 health-related traits among 181,279 women in UK Biobank. We also stratified the GRS according to the involvement of SNPs in DNA damage response. We sought to replicate our findings in independent cohorts. We conducted a negative control MR-pheWAS among men. Among women, we identified potential effects of ANM on 221 traits (1.17% of all traits) at a false discovery rate (P value ≤ 5.83 × 10–4), and 91 (0.48%) potential effects when using Bonferroni threshold (P value ≤ 2.64 × 10–6). Our findings included 55 traits directly related to ANM (e.g. hormone replacement therapy, gynaecological conditions and menstrual conditions), and liver function, kidney function, lung function, blood-cell composition, breast cancer and bone and cardiometabolic health. Replication analyses confirmed that younger ANM was associated with HbA1c (adjusted mean difference 0.003 mmol/mol; 95% CI 0.001, 0.006 per year decrease in ANM), breast cancer (adjusted OR 0.96; 95% CI 0.95, 0.98), and bone-mineral density (adjusted mean difference − 0.05; 95% CI − 0.07, − 0.03 for lumbar spine). In men, 30 traits were associated with the GRS at a false discovery rate (P value ≤ 5.49 × 10–6), and 11 potential effects when using Bonferroni threshold (P value ≤ 2.75 × 10–6). In conclusion, our results suggest that younger ANM has potential causal effects on a range of health-related traits.