Agricultural activities and the incidence of Parkinson's disease in the general French population.

Most studies on pesticides and Parkinson's disease (PD) focused on occupational exposure in farmers. Whether non-occupational exposure is associated with PD has been little explored. We investigated the association between agricultural characteristics and PD incidence in a French nationwide ecologic study. We hypothesized that persons living in regions with agricultural activities involving more intensive pesticide use would be at higher risk. We identified incident PD cases from French National Health Insurance databases (2010-2012). The proportion of land dedicated to 18 types of agricultural activities was defined at the canton of residence level. We examined the association between agricultural activities and PD age/sex-standardized incidence ratios using multivariable multilevel Poisson regression adjusted for smoking, deprivation index, density of neurologists, and rurality (proportion of agricultural land); we used a false discovery rate approach to correct for multiple comparisons and compute q-values. We also compared incidence in clusters of cantons with similar agricultural characteristics (k-means algorithm). We identified 69,010 incident PD cases. Rurality was associated with higher PD incidence (p < 0.001). Cantons with higher density of vineyards displayed the strongest association (RRtop/bottom quartile = 1.102, 95% CI = 1.049-1.158; q-trend = 0.040). This association was similar in men, women, and non-farmers, stronger in older than younger persons, and present in all French regions. Persons living in the cluster with greatest vineyards density had 8.5% (4.4-12.6%) higher PD incidence (p < 0.001). In France, vineyards rank among the crops that require most intense pesticide use. Regions with greater presence of vineyards are characterized by higher PD risk; non-professional pesticides exposure is a possible explanation.