No abstract

Agricultural workers are exposed to pesticide residues via dermal contact with foliage upon entry of treated fields. Restricted Entry Intervals (REIs) are established based on both toxicity and exposure. Key factors for estimation of potential worker exposures are dislodgeable foliar residues (DFRs) and the manner in which DFRs dissipate over time. DFR dissipation curves vary in form and shape, and are often biphasic, reflecting different rate processes, chemical-physical influences, and partitioning. Biphasic dissipation behavior for endosulfan [6,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-hexahydro-6,9-methano-2,4,3-benzodioxathiepin 3-oxide] previously reported for tomatoes and peppers has been confirmed here for endosulfan on melon, grape, and peach foliage. Use of biphasic kinetics results in more robust r2 values for the regression curves that describe foliar dissipation of endosulfan compared to use of simple first-order kinetics. For endosulfan, the use of biphasic kinetics to describe the overall dissipation accurately predicts daily DFR values. In contrast, first order kinetics may overestimate DFRs and, potentially, postapplication worker exposures during the critical period when entry of treated fields is most likely to occur.

Exposure estimates produced using predictive exposure assessment methods are associated with a number of uncertainties that relate to the inherent variability of the values for a given input parameter (e.g., body weight, ingestion rate, inhalation rate) and to unknowns concerning the representativeness of the assumptions and methods used. Despite recent or ongoing consensus-building efforts that have made significant strides forward in promoting consistency in methodologies and parameter default values, the potential variability in the output exposure estimates has not been adequately addressed from a quantitative aspect. This is exemplified by remaining tendencies within federal and state agencies to use worst-case approaches for exposure assessment. In this study, range-sensitivity and Monte Carlo analyses were performed on several different exposure scenarios in order to illustrate the impact of the variability in input parameters on the total variability of the exposure output. The results of this study indicate that the variability associated with the example scenarios range up to more than four orders of magnitude when just some of the parameters are allowed to vary. Comparison of exposure estimates obtained using Monte Carlo simulations (in which selected parameters were allowed to vary over their observed ranges) to exposure estimates obtained using standard parameter default assumptions demonstrate that a default value approach can produce an exposure estimate that exceeds the 95th percentile exposure in an exposed population.

There are a number of sources of variability in food consumption patterns and residue levels of a particular chemical (e.g., pesticide, food additive) in commodities that lead to an expected high level of variability in dietary exposures across a population. This paper focuses on examples of consumption pattern survey data for specific commodities, namely that for wine and grape juice, and demonstrates how such data might be analyzed in preparation for performing stochastic analyses of dietary exposure. Data from the NIAAA/NHIS wine consumption survey were subset for gender and age group and, with matched body weight data from the survey database, were used to define empirically-based percentile estimates for wine intake (microliter wine/kg body weight) for the strata of interest. The data for these two subpopulations were analyzed to estimate 14-day consumption distributional statistics and distributions for only those days on which wine was consumed. Data subsets for all wine-consuming adults and wine-consuming females ages 18 through 45, were determined to fit a lognormal distribution (R2 = 0.99 for both datasets). Market share data were incorporated into estimation of chronic exposures to hypothetical chemical residues in imported table wine. As a separate example, treatment of grape juice consumption data for females, ages 18-40, as a simple lognormal distribution resulted in a significant underestimation of intake, and thus exposure, because the actual distribution is a mixture (i.e., multiple subpopulations of grape juice consumers exist in the parent distribution). Thus, deriving dietary intake statistics from food consumption survey data requires careful analysis of the underlying empirical distributions.

Alternative methods of human exposure assessment that reduce and/or allow quantification of the uncertainties associated with exposure estimates are surveyed and illustrated. These alternative approaches include (1) use of more appropriate exposure parameter default values rather than values that result in extreme exposure estimates; (2) incorporation of time-activity data to better define appropriate exposure duration values; (3) the use of reasonable exposure scenarios rather than the traditional Maximally Exposed Individual (MEI) approach; (4) the use of stochastic approaches such as Monte Carlo-based and information analysis-based methods; (5) use of bivariate analysis to identify the extent to which interdependencies between different exposure parameters affect the distribution of exposure estimates; (6) use of less-than-lifetime exposure and risk assessment; and (7) incorporation of physiological considerations relevant to absorbed dose estimation, including route-specific impacts, use of improved absorption factors, and application of pharmacokinetic models. Other ways to improve the exposure assessment process, including assuring statistical equivalency in comparing different exposure estimates and incorporation of sensitive subpopulation considerations are also discussed, as are key research needs.