Article Text
Abstract
Context Exposure to inorganic arsenicals, including occupational use of pesticides, is carcinogenic to the lung (IARC group 1). However epidemiological data are scarce for agricultural exposures. This work assesses lung cancer (LC) risk, including duration-effect relationships, associated to arsenicals use in farming, by gender and histology.
Methods We linked data from two French projects: (1) the Agriculture and Cancer (AGRICAN) cohort, a large prospective cohort of farmers and people affiliated to the French agricultural insurance scheme and (2) the Pesticide Matrix (PESTIMAT), a crop-exposure matrix. Incident lung cancer cases were collected and their histological subtype ascertained from cancer registries, from enrolment (2005–2007) to December 31 st 2013. The enrolment questionnaire included items on smoking history, and the involvement in 18 different breedings/crops and specific tasks, including pesticide application, with years of beginning and end. We performed Cox models, with age as timescale, adjusted on gender, smoking, and two activities found to be protective in previous analyses – cattle breeding and corn growing. The reference group included farmers having never applied any pesticide on any crop. We assessed risks for each inorganic compound (lead, sodium, aluminum, copper and calcium arsenate) and for overall exposure.
Results Nearly 10% (n=14 359 people) of the population was potentially exposed to arsenicals, in vineyard growing before 2001, or in fruit-tree or potato growing before 1973. We observed 98 incident LC among exposed people. Only women exhibited a higher risk of LC (HR 3.14 95% CI(1.42–6.96) for exposure to any compound, n=7 exposed cases, all adenocarcinomas), but with no duration-effect relationship. Risks were significantly elevated for lead, copper and sodium arsenate.
Conclusion We found an increased risk of LC, especially adenocarcinomas, among women. At this stage, exposure assessment was broad: the use of an exposure index, based on probability, frequency and intensity of use, will help refine the analyses.