Reproductive and hormone‐related risk factors for epithelial ovarian cancer by histologic pathways, invasiveness and histologic subtypes: Results from the EPIC cohort

Whether risk factors for epithelial ovarian cancer (EOC) differ by subtype (i.e., dualistic pathway of carcinogenesis, histologic subtype) is not well understood; however, data to date suggest risk factor differences. We examined associations between reproductive and hormone-related risk factors for EOC by subtype in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort. Among 334,126 women with data on reproductive and hormone-related risk factors (follow-up: 1992-2010), 1,245 incident cases of EOC with known histology and invasiveness were identified. Data on tumor histology, grade, and invasiveness, were available from cancer registries and pathology record review. We observed significant heterogeneity by the dualistic model (i.e., type I [low grade serous or endometrioid, mucinous, clear cell, malignant Brenner] vs. type II [high grade serous or endometrioid]) for full-term pregnancy (phet = 0.02). Full-term pregnancy was more strongly inversely associated with type I than type II tumors (ever vs. never: type I: relative risk (RR) 0.47 [95% confidence interval (CI): 0.33-0.69]; type II, RR: 0.81 [0.61-1.06]). We observed no significant differences in risk in analyses by major histologic subtypes of invasive EOC (serous, mucinous, endometrioid, clear cell). None of the investigated factors were associated with borderline tumors. Established protective factors, including duration of oral contraceptive use and full term pregnancy, were consistently inversely associated with risk across histologic subtypes (e.g., ever full-term pregnancy: serous, RR: 0.73 [0.58-0.92]; mucinous, RR: 0.53 [0.30-0.95]; endometrioid, RR: 0.65 [0.40-1.06]; clear cell, RR: 0.34 [0.18-0.64]; phet = 0.16). These results suggest limited heterogeneity between reproductive and hormone-related risk factors and EOC subtypes.