Introduction Epidemiologic studies have found lower risks of lung cancer (LC), partly due to a lower prevalence of smoking, and with increasing evidence of potential protection in some animal farmers. However, farmers are also exposed to some hazardous air contaminants (pesticides, diesel exhaust, mineral dust). In the prospective cohort AGRICAN, we assessed associations between several crop-related activities and tasks and LC risk, including adenocarcinomas (ADC), squamous cell carcinomas (SqCC) and small cell carcinomas (SmCC).

Methods AGRICAN includes French individuals affiliated to the agricultural health insurance scheme. Incident LC were identified by linkage with cancer registries from enrolment (2005–2007) to 2011. Data on crop exposure during lifetime (13 crops and from 2 to 5 specific tasks) were obtained from the enrolment questionnaire. Analyses were performed using a Cox model, with attained age as time scale, adjusted for gender, smoking history and exposure to potential protective activities – cattle and horses.

Results Among the 148,046 subjects, 487 incident LC were identified. Higher risks were observed in pea growers, especially in those who harvested, (harvesting: HR = 1.43[0.99–2.06], with a significant relationship with duration of exposure (p-trend for duration = 0.06)), for all subtypes except ADC. Increased risks were also observed among (i) farmers growing vegetables (HR = 1.26[0.93–1.72]), (ii) vine-growers (ADC: HR = 1.37[0.97–1.93]), especially in those performing re-entry tasks (ADC: HR 1.40 [0.96–2.04]) and harvesting (ADC: HR = 1.37[0.95–1.95]); (iii) pesticide users on beets (SqCC: HR = 1.88[1.13–3.11]) and (iv) farmers who treated rape seeds (SqCC: HR = 2.55[1.17–5.53]). No significant relationship was observed with duration for any of these activities and tasks.

Conclusions We found positive associations between LC risk and several crop-related tasks including pesticide exposure (application in fields or on seeds, re-entry tasks) or not like peas harvesting.