Fig. 2. Forest plot showing the effect of bioavailable testosterone on all binary outcomes for premenopausal women.

Main analysis MR: Instrumental variable regression using PGS to instrument bioavailable testosterone, adjusted for age at baseline assessment, time of blood collection, UK Biobank recruitment center, and 40 genetic principal components. Multivariable-adjusted: Regression without genetic instruments, adjusted for age at baseline assessment, time of blood collection, UK Biobank recruitment center, and 40 genetic principal components. Note: Employment outcomes were restricted to women younger than 60 years at the time of recruitment.