BACKGROUND: Vitamin D is thought to contribute to brain health, but it is unclear whether low vitamin D levels are associated with increased incidence of Parkinson's disease (PD). Using ultraviolet B (UV-B) as a surrogate for vitamin D levels, we conducted a nationwide ecologic study in France in order to examine the association of UV-B with PD incidence. METHODS: We used French national drug claims databases to identify PD cases using a validated algorithm. UV-B data from the solar radiation database were derived from satellite images. We estimated PD incidence (2010-2012) at the canton level (small administrative French unit) and used multilevel Poisson regression to examine its association with UV-B (2005 annual average), after adjustment for age, sex, deprivation index, density of neurologists, smoking, proportion of agricultural land, and vitamin D supplementation. RESULTS: Analyses are based on 69,010 incident PD patients. The association between UV-B and PD incidence was quadratic (P<0.001) and modified by age (P<0.001). Below 70y, incidence was higher in the bottom quintile (relative risk, RR=1.18, 95% CI=1.08-1.29) compared with the middle UV-B quintile, and lower in the top quintile (RR=0.85 [0.77-0.94]). An opposite pattern was observed in older subjects (RR=0.92 [0.89-0.96]; RR=1.06 [1.02-1.11]). Analysis based on continuous UV-B yielded similar conclusions. CONCLUSIONS: In this nationwide study, there was an age-dependent quadratic association between UV-B and PD incidence. This study suggests that reasonable UV-B exposure is associated with lower PD risk in younger persons and that future studies should examine dose-response relations and take age into account.