Association of Pesticide Exposure With Neurologic Dysfunction and Disease

Freya Kamel; Jane A. Hoppin


Environ Health Perspect. 2004;112(9) 

In This Article

Pesticide Exposure

Pesticides are a broad range of substances most commonly used to control insects, weeds, and fungi (plant diseases). They are frequently classified by target organism or mode of use as insecticides, herbicides, fungicides, or fumigants. Insecticides are often subclassified by chemical type as organophosphates (OPs), organochlorines, carbamates, and pyrethroids. Individuals are frequently exposed to many different pesticides or mixtures of pesticides, either simultaneously or serially. These exposures are often highly correlated, particularly within functional or chemical groups, making it difficult to identify effects of particular agents.

Studies of pesticide neurotoxicity have typically evaluated either the long-term sequelae of pesticide poisoning or the effects of occupational exposure ( Table 1 ). Pesticide poisoning may go undiagnosed, especially among farmworkers with poor access to medical care (Moses et al. 1993) and particularly among women (London et al. 2002). Thus, workers who have never been diagnosed with pesticide poisoning may still have sustained high exposures or experienced pesticide-related illness; therefore using diagnosed poisoning as a criterion for inclusion in an exposed group or exclusion from a comparison group may incorrectly classify individuals.

Some studies of occupational pesticide exposure have classified as exposed all members of an occupational group—typically farmers or farmworkers—sometimes also considering job duration. The potential for misclassification with this approach is high. Farm owners who employ others to apply pesticides may have limited personal exposure to pesticides. Even among pesticide applicators, exposure can vary widely. For example, farmworkers with little access to information about safety practices or protective equipment (Gomes et al. 1999) may sustain far more exposure than well-trained and equipped commercial applicators (Maizlish et al. 1987). Further, farmworkers who do not apply pesticides as part of their job may still be exposed, and even family members with no direct occupational exposure may be exposed at home or elsewhere (Fenske 1997; Gladen et al. 1998), so neither of these may be an appropriate comparison group.

Factors such as application method, use of personal protective equipment, work practices related to hygiene, spills, and attitudes toward risk may all influence the degree of pesticide exposure and can be incorporated into exposure estimates (Alavanja et al. 2004; Buchanan et al. 2001; Dosemeci et al. 2002; Gomes et al. 1999; Hernandez-Valero et al. 2001; London and Myers 1998; Ohayo-Mitoko et al. 1999; Stewart et al. 2001). The relationship of these factors to exposure can be complex. For example, wearing gloves can increase exposure under some circumstances (Hines et al. 2001), perhaps because fabric (as opposed to chemically impervious) gloves can become impregnated with pesticide and serve as a reservoir of exposure. The same may be true of other types of protective clothing (Ohayo-Mitoko et al. 1999). In developing countries, use of closed pesticide mixing and loading systems may increase exposure when the equipment is used to speed up work and increase productivity rather than to protect workers (McConnell et al. 1992). Additional factors may be crucial for evaluating exposure in farmworkers, such as availability of washing and drinking water, interval between application of pesticides to a field and re-entry of workers, and housing conditions (Arcury and Quandt 1998; Gomes et al. 1999; Hernandez-Valero et al. 2001; Tielemans et al. 1999). Studies of neurotoxicity have used all these kinds of information to evaluate pesticide exposure (Gomes et al. 1999; Ohayo-Mitoko et al. 1999). The most sophisticated approaches were employed by London and Myers (1998), who used a crop- and job-specific job exposure matrix to evaluate exposure in a study of the neurotoxicity of chronic OP exposure among South African farmworkers, and by Buchanan et al. (2001), who developed an exposure algorithm to predict diazinon exposure for a study of chronic neurologic effects among sheep dippers in the United Kingdom.

Both historic and current exposures may be relevant to neurotoxicity and need to be characterized. Even among people who remain in the same occupation, current exposure may not reflect past exposure patterns because both available products and methods of use change over time. The need to evaluate past as well as current exposure has limited the utility of biomarkers; most modern pesticides are not persistent, so studies of chronic exposure rely primarily on questionnaire-based methods. Biomarkers are, however, useful in some situations. For example, organochlorines have a long half-life, so serum levels can be used as a marker of exposure to these pesticides. OP inhibition of erythrocyte acetylcholinesterase (AChE) can also be used as an exposure marker. The effect lasts 3-4 months, so AChE activity in whole blood or erythrocytes can be used to evaluate subchronic exposure, although interpretation can be complicated by acute exposure. Although the clinical utility of this biomarker in individuals may be limited by variability in baseline levels, in populations chronic OP exposure is associated with small but reliable decreases in erythrocyte AChE activity (Karr et al. 1992; Ohayo-Mitoko et al. 1997). OPs also inhibit plasma butylcholinesterase, but the effect lasts at most a few weeks and is therefore not useful for evaluating chronic exposure. Cholinesterase inhibition by carbamates lasts only minutes, so it is not a useful marker of chronic exposure to these pesticides.

Estimating lifetime pesticide exposure quantitatively is difficult because it is affected by many factors, including the multiple chemicals involved, uncertainty regarding the degree of exposure related to specific job tasks or other events, and contributions from multiple sources of exposure, including sources unrelated to occupation. Further, the biologically relevant exposure measure is not known: Peak or average exposure intensity might be more important than cumulative exposure. Thus, attempts to assess quantitative dose-response relationships may be problematic. The goal of exposure assessment in epidemiologic studies is not, however, to assign quantitative dose estimates but rather to rank individuals by relative exposure level. Assignment of either exposed or unexposed individuals to the wrong category can be a significant problem, as can combining individuals with low and high levels of exposure into one group. Random misclassification of exposure, unrelated to health outcome, will typically weaken studies by making associations more difficult to detect, although it will not undermine the validity of any association that is observed. As discussed above, assuming that all farmers or even all pesticide applicators are equally exposed is likely to entail significant misclassification, as is assuming that all farmworkers who are not applicators are not exposed. Further, studies that identify only a single highly exposed group for study cannot evaluate the neurotoxicity of moderate exposure, which may have great significance to public health. Methods described above can correctly categorize study participants with respect to their relative exposure levels, and using such methods to increase precision of exposure assessment may help minimize inconsistencies among studies.