Combining crop-exposure matrices and land use data to estimate indices of environmental and occupational exposure to pesticides.

backgroundExposure assessment represents a major challenge for studies on the relation between pesticides and health.

objectiveWe developed a method combining information from crop-exposure matrices (CEMs) and land use data, in order to compute indices of environmental and occupational pesticide exposure. We illustrate our approach using French data (1979-2010).

methodsWe used CEMs for five crops (straw cereals, grain corn, corn fodder, potatoes, vineyards) that describe use (annual probability, frequency, intensity) of pesticide subgroups, chemical families, and active substances by region and time since 1960. We combined these data with land use data from agricultural censuses (1979, 1988, 2000, 2010) to compute indices of environmental and occupational pesticide exposure in cantons (small French administrative units). Indices of environmental exposure were calculated based on the area of each crop in the cantons, while indices of occupational exposure depended on combinations of crops in each farm from the cantons. To illustrate our approach, we selected a pesticide group (herbicides), chemical family of herbicides (phenoxyacetic acids), and active substance from the phenoxyacetic acids chemical family (2,4-D).

resultsThe estimated proportion of the area of crops with CEMs and of farms sprayed with herbicides was close to 100% between 1979-2010, but the estimated average annual number of applications increased. There were decreasing time-trends for phenoxyacetic acids and 2,4-D over the same period for all indices of exposure. There was a high use of herbicides throughout France in 2010, except in the South coast. For phenoxyacetic acids and 2,4-D, the spatial distribution was heterogeneous for all indices of exposure, with the highest values in the Centre and North regions. IMPACT STATEMENT: Assessment of pesticide exposure is a key issue for epidemiological studies on their association with health outcomes. However, it presents some unique challenges, particularly for retrospective exposure and the investigation of chronic diseases. We present a method to compute indices of exposure by combining information from crop-exposure matrices for five crops and land use data. Specificities of environmental and occupational exposure are addressed using different methods. These methods are applied to pesticides used in agriculture in France for five crops (3 groups, 91 chemical families, 197 active substances) to produce indices at a small geographic scale from 1979 to 2010 for the whole metropolitan France. Besides using these indices in French epidemiological studies, our approach could be relevant for other countries.