Assessment of uncertainty in a probabilistic model of consumer exposure to pesticide residues in food

The assessment of consumer exposure to pesticides is an important part of pesticide regulation. Probabilistic modelling allows analysis of uncertainty and variability in risk assessments. The output of any assessment will be influenced by the characteristics and uncertainty of the inputs, model structure and assumptions. While the use of probabilistic models is well established in the United States, in Europe problems of low acceptance, sparse data and lack of guidelines are slowing the development. The analyses in the current paper focused on the dietary pathway and the exposure of UK toddlers. Three single food, single pesticide case studies were used to parameterize a simple probabilistic model built in Crystal Ball™. Data on dietary consumption patterns were extracted from National Diet and Nutrition Surveys, and levels of pesticide active ingredients in foods were collected from Pesticide Residues Committee monitoring. The effect of uncertainty on the exposure estimate was analysed using scenarios, reflecting different assumptions related to sources of uncertainty. The most influential uncertainty issue was the distribution type used to represent input variables. Other sources that most affected model output were non-detects, unit-to-unit variability and processing. Specifying correlation between variables was found to have little effect on exposure estimates. The findings have important implications for how probabilistic modelling should be conducted, communicated and used by policy and decision makers as part of consumer risk assessment of pesticides.     

农药残留急性膳食风险评估研究进展

农药残留急性膳食风险评估直到最近才引起大家的关注。目前,JMPR研究国际范围农药急性膳食风险评估;美国、英国、荷兰、澳大利亚和新西兰也开始进行国家农药急性膳食风险评估。农药残留急性膳食风险是急性或短期接触毒性与农药残留急性膳食摄入量的函数。急性膳食风险评估包括设定急性毒性参考剂量、急性膳食摄入量评定和急性膳食风险描述。急性毒性参考剂量是根据现有的认知水平,在24h或少于24h的期间,人体摄入的食品或水中某物质对消费者不产生可察觉的健康危险的量,单位为毫克/公斤体重(mg/kgbw)。急性膳食接触量评估常用的方法有:定点或确定性方法和概率模型法。在确定性方法中选取食物的大部分人群消耗量和高残留量来计算膳食摄入量,为了解决混合样品中食品个体之间的残留差异,在计算中引入了变异因子。概率模型法是将食品中农药残留分布曲线与膳食摄入食品分布曲线进行整合得到农药膳食摄入量的分布曲线。目前最常用的是蒙特卡洛分析法。使用该方法需要收集个体在每天或每顿饭的食品膳食消耗数据和田间残留试验、市场残留检测数据以及关于加工、削皮、清洗、烹饪等行为对农药残留影响的研究数据。我国应该尽快建立健全膳食结构和农产品性状数据库,建立健全市场中农产品的农药残留数据库,并在高毒和中等毒性农药登记前,进行急性膳食风险评估,提高农药膳食摄入的安全性。首先应采用JMPR的确定性方法,然后再开发适合我国使用的概率模型法。     

Effects of refining predicted chronic dietary intakes of pesticide residues: a case study using glyphosate

A chronic consumer risk assessment based on a worst-case scenario, conducted as part of the European Union review leading to Annex I inclusion for glyphosate, was evaluated and refined. An extensive database of information on the effects of processing on the levels of glyphosate residues in food is available. This database together with refined consumption data from the UK's surveys of adults and toddlers and extensive monitoring data of glyphosate residues in mainly cereal products conducted in the UK were combined to examine the potential overestimates of dietary intakes that are predicted using the current regulatory methodology developed by the Food and Agricultural Organization/World Health Organization and applied as part of the European Union regulatory process. Analysis focussed on the chronic exposure from treated cereals, the crop group contributing significantly to the dietary intake of glyphosate residues. A steep reduction of predicted intake was seen when progressively realistic measures of residues were incorporated into the models, giving a strong indication of the conservative nature of current regulatory procedures. Calculations using even the most unrefined methodology gave rise to intakes of up to 11% of the acceptable daily intake; this was reduced to 0.6% of the acceptable daily intake when justifiable refinements based on extensive monitoring data collected in the UK were made. Consumption data for processed foods abstracted from the UK Food Standard Agency's database were used to refine further the predicted dietary intakes as a result of residue reductions or concentration from processing. The current regulatory model used in the UK generally only has the potential to use a single value for consumption of a particular food. The Pesticides Safety Directorate model consistently predicted the highest intakes with the exception of intakes by adults using the supervised trials median residue and median monitoring data. This suggests that conservatism in the regulatory model exists particularly where specific processing factors cannot be applied to individual fractions of the diet.     

Variability assessment of groundwater exposure to pesticides and its consideration in life-cycle assessment

Pesticide leaching from agricultural fields to groundwater is an environmentally relevant and highly variable process. In the present paper, leaching scenarios typical in European agriculture are defined. These scenarios consider important sources of pesticide leaching variability, namely site factors, farming practice, and substance properties. The logic-tree method was used to structure these scenarios. For each scenario, leached fractions of pesticide applied in agriculture were calculated with data and models used in the registration process of the European Union (EU). Contributions of all parameters to variability were calculated for 11 pesticides. Substance properties (Koc and DT50,soil) contributed the mostto variability, followed by site, weather, season of application, crop, and macropore flow. The results of the variability assessment may be directly applied in policy making or they may be used in the environmental assessment of pesticides, e.g. with the life-cycle assessment (LCA) method. Several approaches are suggested for howthe variability assessment presented in this paper may be incorporated in LCA. The application of these approaches is illustrated by a case study on atrazine.     

Establishing quality management systems for European food composition databases

As part of the activities of EuroFIR (European Food Information Resource Network), the scientific validation and exploitation of food composition data will be assured by a quality framework. This paper outlines the processes and procedures implemented during the initial development of this quality framework for national food composition data compilers and for laboratories generating new data. EuroFIR’s strategic quality objectives are based upon requirements of national food composition data compilers, laboratories, and users/stakeholders. To implement thorough quality principles, a range of standard procedures and tools are being developed. Elements of a quality management system include criteria for quality control and quality assurance of the compilation process and the computerised systems in which data are hosted, as well as data quality assessment models. The results and findings obtained so far for the proposed quality framework indicate that existing guidelines and standards for laboratories should be updated to meet EuroFIR requirements. The generic compilation process has been described, including the identification of critical stages of the data compilation process and standard operating procedures are being prepared. A data quality assessment system has been developed for application to all nutrients and has been tested by national compilers within the EuroFIR network. The development of the EuroFIR quality policy through open and constructive discussions among food composition data compilers, laboratories, and users, strengthens the proposed quality standards and procedures and is fundamental to improving quality of data exchanged across Europe and beyond.     

The US Food Quality Protection Act--policy implications of variability and consumer risk.

The passage of the Food Quality Protection Act (FQPA) in August of 1996 increased the role of risk assessment in the decision-making process of the US Environmental Protection Agency (EPA). Although the law and guidance issued by the EPA provide for more sophisticated risk assessments, the databases for many chemicals may not be robust enough for such data-sensitive analyses. FQPA mandated a major change in how the EPA evaluates the safety of pesticides. This change was immediate, without provision for a phase-in period. Consequently, the EPA is still in the process of learning how to evaluate pesticides under the new paradigm. The EPA's task was further compounded by the lack of scientifically tested methodologies for evaluating aggregate and cumulative risk, as required under the new law. Clearly, the EPA is still in a state of transition between evaluating aggregate and cumulative risks to pesticides and evaluating them one chemical and one exposure route at a time. In all likelihood, the transition period will continue as the discipline of risk assessment develops the mental constructs and computational methodologies to fulfil the requirements of the law. In this interim period, therefore, policies are needed so the regulators and the regulated industry know what is currently acceptable, and how the EPA's thinking is evolving.     

Challenges in Risk Assessment: Quantitative Risk Assessment

The process of risk analysis consists out of three components, risk assessment, risk management and risk communication. These components are internationally well spread by Codex Alimentarius Commission as being the basis for setting science based standards, criteria on food safety hazards, e.g. setting maximum limits of mycotoxins in foodstuffs. However, the technical component risk assessment is hard to elaborate and to understand. Key in a risk assessment is the translation of biological or chemical pathways into a mathematical framework. Within the International Training Program ‘ITP food safety, quality assurance and risk analysis’ of Ghent University, department of Food Safety and Food Quality, we developed for low and middle income countries and emerging countries a training module on risk assessment. In where (semi-) quantitative probabilistic risk assessment calculations or qualitative risk rankings are trained for both microbial and chemical food safety hazards along the agro-food chain. This presentation will explain these methodologies demonstrated with examples from former ITP trainees.     

Estimates of dietary exposure to bisphenol A (BPA) from light metal packaging using food consumption and packaging usage data: a refined deterministic approach and a fully probabilistic (FACET) approach

The FACET tool is a probabilistic model to estimate exposure to chemicals in foodstuffs, originating from flavours, additives and food contact materials. This paper demonstrates the use of the FACET tool to estimate exposure to BPA (bisphenol A) from light metal packaging. For exposure to migrants from food packaging, FACET uses industry-supplied data on the occurrence of substances in the packaging, their concentrations and construction of the packaging, which were combined with data from a market research organisation and food consumption data supplied by national database managers. To illustrate the principles, UK packaging data were used together with consumption data from the UK National Diet and Nutrition Survey (NDNS) dietary survey for 19–64 year olds for a refined deterministic verification. The UK data were chosen mainly because the consumption surveys are detailed, data for UK packaging at a detailed level were available and, arguably, the UK population is composed of high consumers of packaged foodstuffs. Exposures were run for each food category that could give rise to BPA from light metal packaging. Consumer loyalty to a particular type of packaging, commonly referred to as packaging loyalty, was set. The BPA extraction levels used for the 15 types of coating chemistries that could release BPA were in the range of 0.00005–0.012 mg dm^–2. The estimates of exposure to BPA using FACET for the total diet were 0.0098 (mean) and 0.0466 (97.5th percentile) mg/person/day, corresponding to 0.00013 (mean) and 0.00059 (97.5th percentile) mg kg^–1 body weight day^–1 for consumers of foods packed in light metal packaging. This is well below the current EFSA (and other recognised bodies) TDI of 0.05 mg kg^–1 body weight day^–1. These probabilistic estimates were compared with estimates using a refined deterministic approach drawing on the same input data. The results from FACET for the mean, 95th and 97.5th percentile exposures to BPA lay between the lowest and the highest estimates from the refined deterministic calculations. Since this should be the case, for a fully probabilistic compared with a deterministic approach, it is concluded that the FACET tool has been verified in this example. A recent EFSA draft opinion on exposure to BPA from different sources showed that canned foods were a major contributor and compared results from various models, including those from FACET. The results from FACET were overall conservative.     

Assessing exposure to transformation products of soil-applied organic contaminants in surface water: comparison of model predictions and field data

Transformation products (TPs) of chemicals released to soil, for example, pesticides, are regularly detected in surface and groundwater with some TPs even dominating observed pesticide levels. Given the large number of TPs potentially formed in the environment, straightforward prioritization methods based on available data and simple, evaluative models are required to identify TPs with a high aquatic exposure potential. While different such methods exist, none of them has so far been systematically evaluated against field data. Using a dynamic multimedia, multispecies model for TP prioritization, we compared the predicted relative surface water exposure potential of pesticides and their TPs with experimental data for 16 pesticides and 46 TPs measured in a small river draining a Swiss agricultural catchment. Twenty TPs were determined quantitatively using solid-phase extraction liquid chromatography mass spectrometry (SPE-LC-MS/MS), whereas the remaining 26 TPs could only be detected qualitatively because of the lack of analytical reference standards. Accordingly, the two sets of TPs were used for quantitative and qualitative model evaluation, respectively. Quantitative comparison of predicted with measured surface water exposure ratios for 20 pairs of TPs and parent pesticides indicated agreement within a factor of 10, except for chloridazon-desphenyl and chloridazon-methyl-desphenyl. The latter two TPs were found to be present in elevated concentrations during baseflow conditions and in groundwater samples across Switzerland, pointing toward high concentrations in exfiltrating groundwater. A simple leaching relationship was shown to qualitatively agree with the observed baseflow concentrations and to thus be useful in identifying TPs for which the simple prioritization model might underestimate actual surface water concentrations. Application of the model to the 26 qualitatively analyzed TPs showed that most of those TPs categorized as exhibiting a high aquatic exposure potential could be confirmed to be present in the majority of water samples investigated. On the basis of these results, we propose a generally applicable, model-based approach to identify those TPs of soil-applied organic contaminants that exhibit a high aquatic exposure potential to prioritize them for higher-tier, experimental investigations.     

农药残留膳食暴露评估模型研究进展

农药残留膳食暴露风险评估为化学品风险管理、食品安全与人类健康领域的重点关注问题，有效的定量风险评估模型是开展农药残留膳食暴露评价的基础工具。本文结合国内外农药残留膳食暴露风险评估研究进展，旨在归纳农药残留膳食暴露风险评估模型，包括确定性评估模型、概率性评估模型和累积性评估模型；根据模型演变、适用情形与评估需求，比对分析模型间优缺点，探讨模型不确定度、评估软件、数据库应用等关键因子。同时立足我国农药残留膳食风险评估现状，展望模型相关参数整合的紧迫性与重要性，为完善农药残留膳食暴露风险评估体系建设提供科学参考。     

Household hazardous waste data for the UK by direct sampling

The amount of household hazardous waste (HHW) disposed of in the United Kingdom (UK) requires assessment. This paper describes a direct analysis study carried out in three areas in southeast England involving over 500 households. Each participating householder was provided with a special bin in which to place items corresponding to a list of HHW. The amount of waste collected was split into nine broad categories: batteries, home maintenance (DIY), vehicle upkeep, pesticides, pet care, pharmaceuticals, photographic chemicals, household cleaners, and printer cartridges. Over 1 T of waste was collected from the sample households over a 32-week period, which would correspond to an estimated 51,000 T if extrapolated to the UK population for the same period or over 7,000 T per month. Details of likely disposal routes adopted by householders were also sought, demonstrating the different pathways selected for different waste categories. Co-disposal with residual household waste dominated for waste batteries and veterinary medicines, hence avoiding classification as hazardous waste under new UK waste regulations. The information can be used to set a baseline for the management of HHW and provides information for an environmental risk assessment of the disposal of such wastes to landfill.     

Joint FAO/WHO Geneva consultation acute dietary intake methodology

Significant developments have been made at the international level in the methodology for conducting dietary risk assessments. The existing chronic exposure assessment methodology has been updated so that it takes account of the level of residues to which consumers are most likely to be exposed. In addition, short-term exposure assessment methodology has also been developed. This uses portion size data and takes account of the variability in residue levels between individual units where this is appropriate to the way the commodity is consumed. Other refinements to the assessment can also be made. Although the short-term methodology has been used successfully by a number of regulatory authorities, there is a need for data on portion sizes and typical unit weights before it can be fully implemented internationally.     

A risk assessment approach applied to the growth of Erwinia carotovora in vegetable juice for variable temperature conditions

Risk assessment for food spoilage relies on probabilistic models of microbial growth to predict the likelihood that microbial populations will exceed predefined spoilage levels. To assist in the design and management of industrial food quality systems, predictive microbiological models have to incorporate major risk factors such as the variability in the microbial strain, environment and initial contamination levels. In addition, the application of results measured under laboratory conditions to the less controlled environment of an industrial process usually also involves uncertainty. Extra information regarding this uncertainty must be factored into industrial microbial risk assessment. In this paper, based on our previous analysis of the growth of Erwinia carotovora we show how different factors contribute to the risk of microbial spoilage of vegetable juice and we demonstrate an effective way of including these factors into risk assessment models. The association of risk components with different unavoidable and manageable factors is also valuable for the development of optimal strategies for reducing microbial risk.     

Application of predictive modelling techniques in industry: from food design up to risk assessment

In this communication, examples of applications of predictive microbiology in industrial contexts (i.e. Nestlé and Unilever) are presented which cover a range of applications in food safety from formulation and process design to consumer safety risk assessment. A tailor-made, private expert system, developed to support safe product/process design assessment is introduced as an example of how predictive models can be deployed for use by non-experts. Its use in conjunction with other tools and software available in the public domain is discussed. Specific applications of predictive microbiology techniques are presented relating to investigations of either growth or limits to growth with respect to product formulation or process conditions. An example of a probabilistic exposure assessment model for chilled food application is provided and its potential added value as a food safety management tool in an industrial context is weighed against its disadvantages. The role of predictive microbiology in the suite of tools available to food industry and some of its advantages and constraints are discussed.     

Statistical analysis of chemical transformation kinetics using Markov-Chain Monte Carlo methods

For the risk assessment of chemicals intentionally released into the environment, as, e.g., pesticides, it is indispensable to investigate their environmental fate. Main characteristics in this context are transformation rates and partitioning behavior. In most cases the relevant parameters are not directly measurable but are determined indirectly from experimentally determined concentrations in various environmental compartments. Usually this is done by fitting mathematical models, which are usually nonlinear, to the observed data and such deriving estimates of the parameter values. Statistical analysis is then used to judge the uncertainty of the estimates. Of particular interest in this context is the question whether degradation rates are significantly different from zero. Standard procedure is to use nonlinear least-squares methods to fit the models and to estimate the standard errors of the estimated parameters from Fisher's Information matrix and estimated level of measurement noise. This, however, frequently leads to counterintuitive results as the estimated probability distributions of the parameters based on local linearization of the optimized models are often too wide or at least differ significantly in shape from the real distribution. In this paper we identify the shortcoming of this procedure and propose a statistically valid approach based on Markov-Chain Monte Carlo sampling that is appropriate to determine the real probability distribution of model parameters. The effectiveness of this method is demonstrated on three data sets. Although it is generally applicable to different problems where model parameters are to be inferred, in the present case for simplicity we restrict the discussion to the evaluation of metabolic degradation of chemicals in soil. It is shown that the method is successfully applicable to problems of different complexity. We applied it to kinetic data from compounds with one and five metabolites. Additionally, using simulated data, it is shown that the MCMC method estimates the real probability distributions of parameters well and much better than the standard optimization approach.     

Bayesian analysis of a reduced-form air quality model

Numerical air quality models are being used for assessing emission control strategies for improving ambient pollution levels across the globe. This paper applies probabilistic modeling to evaluate the effectiveness of emission reduction scenarios aimed at lowering ground-level ozone concentrations. A Bayesian hierarchical model is used to combine air quality model output and monitoring data in order to characterize the impact of emissions reductions while accounting for different degrees of uncertainty in the modeled emissions inputs. The probabilistic model predictions are weighted based on population density in order to better quantify the societal benefits/disbenefits of four hypothetical emission reduction scenarios in which domain-wide NO(x) emissions from various sectors are reduced individually and then simultaneously. Cross validation analysis shows the statistical model performs well compared to observed ozone levels. Accounting for the variability and uncertainty in the emissions and atmospheric systems being modeled is shown to impact how emission reduction scenarios would be ranked, compared to standard methodology.     

Predicting the frequency of water quality standard violations: a probabilistic approach for TMDL development

To address the impaired condition of the water bodies listed under Section 303(d) of the Clean Water Act, over 40 000 total maximum daily loads (TMDLs) for pollutants must be developed during the next 10-15 years. Most of these will be based on the results of water quality simulation models. However, the failure of most models to incorporate residual variability and parameter uncertainty in their predictions makes them unsuitable for TMDL development. The percentile-based standards increasingly used by the EPA and the requirement for a margin of safety in TMDLs necessitate that model predictions include quantitative information on uncertainty. We describe a probabilistic approach to model-based TMDL assessment that addresses this issue and is suitable for use with any type of mathematical model. To demonstrate our approach, we employ a eutrophication model for the Neuse River Estuary, North Carolina, and evaluate compliance with the state chlorophyll a standard. Any observed variability in chlorophyll athatis notexplained bythe model is explicitly incorporated via a residual error term. This probabilistic term captures the effects of any processes that are not considered in the model and allows for direct assessment of the frequency of standard violations. Additionally, by estimating and propagating the effects of parameter uncertainty on model predictions, we are able to provide an explicit basis for choosing a TMDL that includes a margin of safety. We conclude by discussing the potential for models currently supported by the EPA to be adapted to provide the type of probabilistic information that is necessary to support TMDL decisions.     

From dishwasher to tap? Xenobiotic substances benzotriazole and tolyltriazole in the environment

There is increasing evidence that the use of chemicals frequently results in widespread environmental contamination with little understanding of the toxicological implications. Benzotriazoles are used in, among other applications, dishwashing formulations for home use, and are a class of chemicals recently reported to be present in European waters. This study demonstrates their presence in UK wastewaters, rivers, and drinking water. It also estimates that their use as silver polishing agents in dishwasher tablets and powders may account for a significant proportion of inputs to wastewaters. The lack of a complete set of good quality (eco)toxicological data on possible chronic effects of these high use chemicals should caution against using them in a manner which may have contributed to such widespread environmental contamination.     

The use and limitations of current 'standard' toxicological data packages in the setting of acute reference doses.

To perform risk assessments for short term dietary intakes of pesticides it is necessary to determine an acceptable level of exposure from the available toxicity studies. This type of assessment is a relatively new development in regulatory toxicology and often the studies currently available are not optimal for such uses. The strengths and weaknesses of currently performed study types are presented. Consideration of the entire toxicity data base is essential.     

Swift probabilistic risk assessment model of chemical hazards in food by tiered approach: A case study of pesticide residues in packaged red pepper powder

Development of a swift probabilistic risk assessment (sPRA) model is needed to obtain information on exposure to food hazards before performing complex full-scale risk assessment. A sPRA model was developed using a tiered approach, which uses qualitative information, quantitative analysis of focused hazards, and a Monte Carlo simulation. A case study was conducted on residual pesticides in red pepper powder using sPRA model. Of a wide range of pesticides, chlorpyrifos and cypermethrin were selected for sPRA modeling. As a result of the sPRA model simulation, the estimated dietary exposure to pesticide residue from the consumption of red pepper powder was below 0.04% of the acceptable daily intake (ADI) for the 95^th percentile exposure. Risk of dietary exposure to hazards was quantitatively estimated by the sPRA model developed, and the risk of residual pesticide in red pepper powder in Korea was found to be very low.