Student Research: Jordan Firestone
, , 2002
Faculty Advisor: Harvey Checkoway
Pesticides and Parkinson's Disease, Risk Estimates from a Case-Control Study
Parkinson's Disease is a progressive neurodegenerative disorder of uncertain etiology. Pesticide exposures have been implicated as environmental risk factors, however previous studies have yielded inconsistent risk estimates.
From this large (cases=220; controls=352), population-based, case-control study of incident Parkinson's disease, risks were estimated from occupational and home-based pesticide exposures. Exposure data, including subjects' demographic characteristics; smoking status; occupational titles; exposures to specific pesticides and use categories; and histories of residential locale and well water consumption; were based on subjects' self-report in face-to-face interviews using a structured questionnaire. Conditional logistic regression controlled for the subjects' matching by age and gender, and for smoking, yielding adjusted odds ratios (ORadj) and 95% confidence intervals (CI).
The most significant findings related to occupational pesticide exposures, with increased risk in pesticide workers (ORadj=3.9, CI=1.0-14.5), orchard growers (ORadj=1.7, CI=0.6-5.1), and vegetable farmers (ORadj=1.5, CI=0.7-3.3), in a gradient that parallels the predicted exposure intensity. This supports the hypothesis that occupational pesticide exposure causes Parkinson's disease. However, given the known variability of exposures within occupational categories, these findings must be interpreted cautiously. Among occupational pesticide use categories, risk is slightly elevated for herbicides (ORadj=1.2; CI=0.4-3.8) and insecticides (ORadj=1.1; CI=0.5-2.4). Risks are even greater for specific pesticides of a priori interest, including paraquat (ORadj=1.9, CI=0.2-14.8), parathion (ORadj=9.8, 95% CI=1.0-90.5), diazinon (ORadj=1.3, CI0.4-4.4) and malathion (ORadj=1.3, CI=0.4-3.9), in a gradient paralleling their relative potency. While these risks may reflect chance associations, the gradient of risk, biological plausibility, and consistency with other reports suggest a causal relationship between specific pesticides and Parkinson's disease.
It is notable that all risk estimates from home-based exposures suggest protective effects, including herbicides (ORadj=0.6; CI=0.43=0.99) and insecticides (ORadj=0.7; CI=0.5-1.1). While this contrast with occupational exposures may reflect biased reporting, it is consistent with fundamental difference between occupational and home-based exposures. Surrogate indicators of home-based exposures demonstrate increased risk from ever living on a farm (ORadj=1.2; CI=0.8-1.8), residence on a farm for more than five years during childhood (ORadj= 1.6; CI=1.0-2.4), and well water consumption (ORadj=1.3; CI=0.9-2.0). Future work must focus on refining the measures of pesticide exposure to help reveal potentially meaningful associations.