Background: Anastrozole, an aromatase inhibitor, is approved for breast cancer prevention in high-risk women. The long-term effects of aromatase inhibition, including its repurposing potential to other cancers, possible adverse effects, and treatment effect heterogeneity across patient subgroups, remain unclear. Methods: We used the rs727479 variant in CYP19A1 to mimic the effect of long-term pharmacological aromatase inhibition. To evaluate repurposing opportunities, genetic association data on five cancers (211,386 cases, 684,665 controls) were obtained from genome-wide association study consortia. Potential adverse effects were evaluated in a phenome-wide association study (PheWAS) of 449 health-related traits in 162,360 postmenopausal women in the UK Biobank. Effects were investigated across clinically relevant subgroups in the UK Biobank including those defined by body mass index (BMI). Results: Genetically-proxied aromatase inhibition was associated with reduced risk of ER+ breast cancer (OR:0.78, 95%CI:0.67-0.92) and decreased heel bone mineral density (-0.32SD change, 95%CI:-0.36,-0.28). When examining the repurposing potential of anastrozole to other cancers, we found that genetically-proxied aromatase inhibition reduced endometrial cancer risk (OR:0.34, 95%CI:0.26-0.44). In PheWAS, genetically-proxied aromatase inhibition was associated with 6 outcomes (PFDR