Exploring estrogen-related mechanisms in ovarian carcinogenesis: association between bone mineral density and ovarian cancer risk in a multivariable Mendelian randomization study

Abstract Background Estrogen may play a role in epithelial ovarian cancer (EOC) carcinogenesis, with effects varying by EOC histotype. Measuring women’s long-term exposure to estrogen is difficult, but bone mineral density (BMD) may be a reasonable proxy of longer-term exposure. We examined this relationship by assessing the association between genetic predisposition for higher BMD and risk of EOC by histotype. Methods We used Mendelian randomization (MR) to assess associations between genetic markers for femoral neck and lumbar spine BMD and each EOC histotype. We used multivariable MR (MVMR) to adjust for probable pleiotropic traits, including body mass index, height, menarcheal age, menopausal age, smoking, alcohol intake, and vitamin D. Results Univariable analyses suggested greater BMD was associated with increased risk of endometrioid EOC (per standard deviation increase; lumbar spine OR = 1.21; 95% CI 0.93,1.57, femoral neck: OR = 1.25; 0.99,1.57), but sensitivity analyses indicated that pleiotropy was likely. Adjustment using MVMR reduced the magnitude of estimates slightly (lumbar spine: OR = 1.13; 95% CI 1.00,1.28, femoral neck: OR = 1.18; 1.03,1.36). Results for lumbar spine BMD and high-grade serous EOC were also suggestive of an association (univariable MR: OR = 1.16; 95% CI 1.03,1.30; MVMR: OR = 1.06; 0.99,1.14). Conclusion Our study found associations between genetic predisposition to higher BMD, a proxy for long-term estrogen exposure, and risk of developing endometroid and high-grade serous EOC cancers. These findings add to existing evidence of the relationship between estrogen and increased risk of EOC for certain histotypes.