Progress in pesticide exposure assessment : The case of Parkinson's Disease in the Netherlands

Pesticides are extensively used in agriculture and numerous chemical active ingredients and product formulations have been marketed since the 1950ies. Human populations can be exposed to pesticides via multiple routes, such as application of these substances in occupational settings, or via pesticides migrating into the (home) environment following agricultural pesticide applications in vicinity. Pesticide exposure has been associated with different health effects, such as Parkinson’s Disease (PD). Evidence for the role of specific active ingredients is still limited, however, as the wide range of pesticides used over time and space complicates accurate assessment of exposure. It is also not clear to what extent epidemiological findings obtained abroad are relevant for the Netherlands. The first aim of this thesis was to develop and apply improved methods for the assessment of occupational and environmental pesticide exposures in several retrospective epidemiological studies. The second aim was to investigate the association between pesticide exposures and PD in the Netherlands. The work described in this thesis has resulted in country-specific crop-exposure-matrices and a spatio-temporal model to assess environmental pesticide exposure at the residential address for the Netherlands. These methods address pesticide exposure at the level of active ingredients, while going back in time for several decades. Low specificity and high correlations between pesticide exposures remained an issue that cannot be easily solved however. In part this is inherent to pesticide applications, with several active ingredients being applied during the same time window and/or location. But also the exposure assessment methods have their limitations. High correlations limit the ability to distinguish effects of individual pesticides on health in epidemiological studies. Further improvements will require additional data collection, such as higher (spatial and temporal) resolution data on crop cultivation and pesticide use, for time periods where this is possible. Future studies should combine spatial models or exposure proxies with environmental and personal measurements, to determine validity of different measures, and to gain improved understanding of the relevance of different exposure pathways and routes. No significant association was found between occupational pesticide exposure and PD mortality in a Dutch cohort study. Environmental exposure to a number of specific (correlated) pesticides did appear to be associated with increased risk of PD in a Dutch case-control study. Bulb cultivation in vicinity of residences also appeared to be associated with higher PD risk. Risk estimates were generally higher for higher cumulative exposure and smaller distance between residences and crops, which suggests that the elevated risks observed are rather reflecting a true effect of agricultural exposures on PD, than a chance finding. These findings should be seen as hypothesis generating however, and should be further investigated and replicated. Future automated digital registration of pesticide applications by farmers and routine monitoring of pesticides in different environmental media, would be beneficial to future epidemiological studies. Given the potential adverse effects of pesticides on human health and ecosystems, the government should take a more active role in the construction of such detailed pesticide use databases, and provide access for scientific research.