Introduction Workers in the greenspace industry are exposed to a range of occupational hazards including pesticides. Occupational exposure to pesticides and their health effects have been mainly studied among farmers while data on greenspace workers remain scarce. Exposures in greenspaces are not similar to those in farming: there are differences in applied substances, equipment, application scenario, general environment etc. Studying the impact of pesticides highly used in this specific population provides relevant data on some specific substances like total herbicides (glyphosate, paraquat, etc.).

Objectives To analyse the causes of death and the incidence of main cancers among greenspace workers.

Methods Within the AGRICAN cohort - that enrolled more than 181,000 workers affiliated to the health insurance for agriculture in 2005–2007 in 11 French areas – we defined a sub-cohort of 6,247 workers from the greenspace industry. We run survival analyses (Cox-proportional hazards models) on main causes of death and on cancer incidence from enrolment to the end of 2015. Comparisons with farmers and non-agricultural workers have been performed.

Results Overall mortality among greenspace workers was comparable to that of farmers and non-agricultural workers. However, greenspace workers’ overall cancer incidence (n=446) was higher than among farmers (HR=1.15 [1.04–1.27]). Compared to farmers, increased risks have been found in men for: skin melanoma (HR=2.15 [1.33–3.47]), prostate (HR=1.21 [1.02–1.44]), testicular (HR=3.98 [1.50–10.58]), and thyroid (HR=2.84 [1.60–6.41]) cancers; and in women for breast cancer (HR=1.71 [1.17–2.50]). Elevated risks were also found for cancers of the larynx and bladder and sarcomas. These associations have been found among pesticide applicators as well.

Conclusion The differences in cancer incidence between greenspace workers and farmers could suggest the impacts of occupational risks specific to this population. Additional research is underway to better characterize their exposures and will be used in further analyses.