RF-152 Agricultural exposures and risk of ovarian cancer in the AGRIculture and CANcer (AGRICAN) cohort

Abstract

<h3>Introduction</h3> Ovarian cancer is uncommon, with poor prognosis, and it ranks among the 1st causes of gynecological cancer death in females in developed countries. Few risk (age, genetic, hormonal) and protective (oral contraceptive) factors are established. The role of asbestos, estrogen menopausal therapy and tobacco smoking in ovarian cancers is supported by sufficient evidence in humans (CIRC). The influence of hormones and reproductive factors on ovarian cancer suggests that endocrine disrupters may impact risk. <h3>Objectives</h3> We assessed the risk of ovarian cancer among women farmers, and studied associations with livestock, crops and related tasks. <h3>Methods</h3> Our analysis included 83,048 women, enrolled in AGRICAN cohort in 2005–2007 in France, who completed an enrolment questionnaire on lifelong agricultural exposures and reproductive life. Ovarian cancer cases were identified by cross-linkage with cancer registries. <h3>Results</h3> Until 2015, 186 incident cases of ovarian cancer were identified. Late menopause increased the risk of ovarian cancer, while parity and oral contraceptive use were protective. Decreased risks of ovarian cancer were observed (1) in female farmers compared to non-farmers (HR=0.65[0.43–0.98]), but remained only for those who used oral contraceptive, (2) for some activities (wheat and corn, poultry and sheep/goat), probably due to residual confounding. However, increased risks were observed among (1) female pig-breeders (HR&gt;15 pigs = 1.93[1.08–3.44]), (2) female fruit-growers (HR HR=2.58[1.03–6.41]). Increased risks observed on crops were reinforced among women who did not use oral contraception and results in livestock were unchanged according to oral contraception use. <h3>Conclusion</h3> In the AGRICAN cohort established risk (late menopause) and protective factors (oral contraceptive and parity) were confirmed in female farmers. We found positive association for some crops and animal breeding but we also found decreased risks possibly due to residual confounding.