The variation of pesticide residues in fruits and vegetables and the associated assessment of risk

Abstract

High levels of triazophos residues detected in carrots during routine monitoring led to the discovery of a wide variability between levels in individual roots. Conventional point estimates of consumer exposure were carried out. Due to the assumptions used, these calculations were likely to give rise to gross overestimates. In 1997, data were obtained for individual apples, pears, peaches, nectarines, oranges, bananas, and tomatoes that showed similar levels of variability in a range of organophosphate and carbamate residues. Point estimate models that had previously been used for intake estimates for carrots were not appropriate since it was necessary to take account of not only the variation of residue levels from crop item to crop item but also the variation in eating patterns in individual consumers. Probabilistic modeling was identified as a suitable way to produce multifactorial submodels and address some of the problems of combining distributions of consumption and residues. Consumption data from 1675 toddlers were linked with residue distributions from individual crop items not only to allow combinations of fruit consumed but also to allow for the variability in residue levels that occur between individual crop items. The model was also capable of taking account of the percentages of crops that did not contain any detectable residues; this information was available from initial screens of bulked samples and percentage of crop not treated in the case of carrots. The outputs from the models were given as percentages of consumers that could exceed a toxicological end point; this could be the acute reference dose or a factor of the no-observable-adverse-effect level. Modeling in this way was considered to give a realistic view of the likely short-term exposure and the output was used as an aid to decision making in terms of necessary regulatory action.

Background: As a result of high levels of triazophos detected in carrots during routine monitoring, studies were carried out to determine the variability of organophosphate residue levels in individual roots. Results obtained indicated that the highest residue levels could be 25 times the mean level in bulked samples (which were used in routine monitoring). Since sufficient levels of organophosphate compounds can give rise to toxicological effects after a single exposure, it was considered necessary to carry out assessments of short-term or acute consumer risk. At that time, models available worldwide were designed only to carry out point estimates of long-term exposure. From consumption data, it was possible to derive the levels of carrot consumption during a single day and calculations were carried out assuming all carrots contained the highest levels of residues found in trials. This led to a gross overestimate of likely exposure but was considered to give to intakes that eroded margins of safety; these were not a cause for extreme regulatory action. Further studies were carried out on other crops that may be eaten whole, at one sitting, and without processing to consider whether the large variability of organophosphate residues was a phenomenon that was common to other fruits and vegetables.