Probabilistic risk assessment of dietary exposure to single and multiple pesticide residues or contaminants: Summary of the work performed within the SAFE FOODS project

This introduction to the journal’s supplement on probabilistic risk assessment of single and multiple exposure to pesticide residues or contaminants summarizes the objectives and results of the work performed in work package 3 of the EU-funded project SAFE FOODS. Within this work package, we developed an electronic platform of food consumption and chemical concentration databases harmonised at raw agricultural commodity level. In this platform the databases are connected to probabilistic software to allow probabilistic modelling of dietary exposure in a standardised way. The usefulness of this platform is demonstrated in two papers, which describe the exposure to pesticides and glycoalkaloids in several European countries. Furthermore, an integrated probabilistic risk assessment (IPRA) model was developed: a new tool to integrate exposure and effect modelling, including uncertainty analyses. The use of this model was shown in a paper on the cumulative exposure to anti-androgen pesticides. Combined with a health impact prioritization system, developed within this work package to compare heath risks between chemicals, the IPRA tool can also be used to compare health risks between multiple chemicals in complex risk assessment situation such as risk–benefit and risk trade-off analyses. Both the electronic platform of databases as the IPRA model may proof to be powerful tools to tackle the challenges risk managers are or will be faced with in the future.     

Using exposure bands for rapid decision making in the RISK21 tiered exposure assessment

The ILSI Health and Environmental Sciences Institute (HESI) Risk Assessment in the Twenty-first Century (RISK21) project was initiated to address and catalyze improvements in human health risk assessment. RISK21 is a problem formulation-based conceptual roadmap and risk matrix visualization tool, facilitating transparent evaluation of both hazard and exposure components. The RISK21 roadmap is exposure-driven, that is, exposure is used as the second step (after problem formulation) to define and focus the assessment. This paper describes the exposure tiers of the RISK21 matrix and the approaches to adapt readily available information to more quickly inform exposure at a screening level. In particular, exposure look-up tables were developed from available exposure tools (European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC) Targeted Risk Assessment (TRA) for worker exposure, ECETOC TRA, European Solvents Industry Group (ESIG) Generic Exposure Scenario (GES) Risk and Exposure Tool (EGRET) for consumer exposure, and USEtox^® for indirect exposure to humans via the environment) and were tested in a hypothetical mosquito bed netting case study. A detailed WHO risk assessment for a similar mosquito net use served as a benchmark for the performance of the RISK21 approach. The case study demonstrated that the screening methodologies provided suitable conservative exposure estimates for risk assessment. The results of this effort showed that the RISK21 approach is useful for defining future assessment efforts, focusing assessment activities and visualizing results.     

Risk assessment of engineered nanomaterials: a review of available data and approaches from a regulatory perspective

Abstract

It has been largely recognised that substantial limitations and uncertainties make the conventional risk assessment (RA) of chemicals unfeasible to apply to engineered nanomaterials (ENMs) today, which leaves the regulators with little support in the near term. The aim of this paper is to discuss the state of the art in the area of the RA of nanomaterials, focusing on the available data and approaches. There is a paucity of reliable information in the online safety databases and the literature is dominated by (eco)toxicity studies, while the nano-exposure research lags behind. Most of the reviewed nano-RA approaches are designed to serve as preliminary risk screening and/or research prioritisation tools and are not intended to support regulatory decision making. In this context, we recommend to further study the possibilities to apply complementary/alternative tools for near-term RA of ENMs in order to facilitate their timely regulation, using the data that are currently available in the literature.     

Use of read-across and tiered exposure assessment in risk assessment under REACH – A case study on a phase-in substance

REACH requests the exploration of alternative strategies for hazard identification before resorting to (in vivo) testing. Here, we combined read-across as non-testing strategy with a tiered exposure assessment for the risk characterisation of 1-methoxypropan-2-ol (PGME) as a representative for phase-in substances to be registered under REACH. Read-across from the selected source substances provided data which were comparable with experimental data available for target substance PGME, resulting in a realistic starting point for both qualitative and quantitative risk assessment. Greater variability was observed in the exposure estimates from a first Tier model (ECETOC TRA) or less conservative further Tier models (Stoffenmanager; RISKOFDERM), when these results were compared with results from a data-rich approach using measured data.     

Exposure Endpoint Selection in Acute Dietary Risk Assessment

Current USEPA Office of Pesticide Program approaches to acute dietary risk assessment do not adequately address uncertainty in the distributional analysis of exposure. This is especially true with respect to regulatory decision points (bright lines) located at the far extreme of the cumulative output distribution. Use of the 99.9th centile as a risk assessment endpoint necessitates confidence in food consumption and residue input distributions that cannot be demonstrated with currently available data and analysis approaches. Even for a pesticide with a rich residue database, data limitations are sufficient to skew results to significantly overestimate exposure. This is compounded when extremes in food consumption are used that go beyond the stated error bounds for the database used. Risk management decision making need not consider endpoints at extremes of exposure output distributions in order for mitigation to be protective of sensitive populations. In fact, such decision making is better informed by utilizing more statistically reliable endpoint selection in the risk assessment process. The richest data content in cumulative exposure distributions occurs in regions well removed from output tails, where the pattern of exposure distribution is driven by the effects of residue concentrations. In contrast, the extreme upper tail of the exposure distribution is data poor and is characterized by high uncertainty reflecting extremes in food consumption patterns. At present, risk managers are better served with exposure endpoints removed from the highly uncertain tails of exposure distributions such as the 99.9th centile bright line. The selection of more appropriate risk management decision points should consider the nature of the distribution, the severity of the effect being assessed, and robustness of the data available for assessing acute dietary risk.     

Application of a quantitative weight of evidence approach for ranking and prioritising occupational exposure scenarios for titanium dioxide and carbon nanomaterials

Substantial limitations and uncertainties hinder the exposure assessment of engineered nanomaterials (ENMs). The present deficit of reliable measurements and models will inevitably lead in the near term to qualitative and uncertain exposure estimations, which may fail to support adequate risk assessment and management. Therefore it is necessary to complement the current toolset with user-friendly methods for near-term nanosafety evaluation. This paper proposes an approach for relative exposure screening of ENMs. For the first time, an exposure model explicitly implements quantitative weight of evidence (WoE) methods and utilises expert judgement for filling data gaps in the available evidence-base. Application of the framework is illustrated for screening of exposure scenarios for nanoscale titanium dioxide, carbon nanotubes and fullerenes, but it is applicable to other nanomaterials as well. The results show that the WoE-based model overestimates exposure for scenarios where expert judgement was substantially used to fill data gaps, which suggests its conservative nature. In order to test how variations in input data influence the obtained results, probabilistic Monte Carlo sensitivity analysis was applied to demonstrate that the model performs in stable manner.     

A proposed framework for assessing risk from less-than-lifetime exposures to carcinogens

Quantitative methods for estimation of cancer risk have been developed for daily, lifetime human exposures. There are a variety of studies or methodologies available to address less-than-lifetime exposures. However, a common framework for evaluating risk from less-than-lifetime exposures (including short-term and/or intermittent exposures) does not exist, which could result in inconsistencies in risk assessment practice. To address this risk assessment need, a committee of the International Life Sciences Institute (ILSI) Health and Environmental Sciences Institute conducted a multisector workshop in late 2009 to discuss available literature, different methodologies, and a proposed framework. The proposed framework provides a decision tree and guidance for cancer risk assessments for less-than-lifetime exposures based on current knowledge of mode of action and dose-response. Available data from rodent studies and epidemiological studies involving less-than-lifetime exposures are considered, in addition to statistical approaches described in the literature for evaluating the impact of changing the dose rate and exposure duration for exposure to carcinogens. The decision tree also provides for scenarios in which an assumption of potential carcinogenicity is appropriate (e.g., based on structural alerts or genotoxicity data), but bioassay or other data are lacking from which a chemical-specific cancer potency can be determined. This paper presents an overview of the rationale for the workshop, reviews historical background, describes the proposed framework for assessing less-than-lifetime exposures to potential human carcinogens, and suggests next steps.     

Framework for use of toxicity screening tools in context-based decision-making.

One of the principal applications of toxicology data is to inform risk assessments and support risk management decisions that are protective of human health. Ideally, a risk assessor would have available all of the relevant information on (a) the toxicity profile of the agent of interest; (b) its interactions with living systems; and (c) the known or projected exposure scenarios: to whom, how much, by which route(s), and how often. In practice, however, complete information is seldom available. Nonetheless, decisions still must be made. Screening-level assays and tools can provide support for many aspects of the risk assessment process, as long as the limitations of the tools are understood and to the extent that the added uncertainty the tools introduce into the process can be characterized and managed. Use of these tools for decision-making may be an end in itself for risk assessment and decision-making or a preliminary step to more extensive data collection and evaluation before assessments are undertaken or completed and risk management decisions made. This paper describes a framework for the application of screening tools for human health decision-making, although with some modest modification, it could be made applicable to environmental settings as well. The framework consists of problem formulation, development of a screening strategy based on an assessment of critical data needs, and a data analysis phase that employs weight-of-evidence criteria and uncertainty analyses, and leads to context-based decisions. Criteria for determining the appropriate screening tool(s) have been identified. The choice and use of the tool(s) will depend on the question and the level of uncertainty that may be appropriate for the context in which the decision is being made. The framework is iterative, in that users may refine the question(s) as they proceed. Several case studies illustrate how the framework may be used effectively to address specific questions for any endpoint of toxicity.     

Risk assessment of local dermal effects and skin sensitisation under the EU Chemicals Regulation REACH: a proposal for a qualitative, exposure scenario specific, approach

Within the framework of REACH, an assessment regarding local dermal effects and skin sensitisation should be performed for substances. Quantitative hazard information for these effects is often not available. Furthermore, it is difficult to relate the way in which animals are exposed in dermal toxicity studies directly to dermal exposure in practice. In the absence of quantitative information, a qualitative assessment for dermal effects is the most reasonable option. The qualitative approach as proposed in the REACH guidance recommends only general risk management measures (RMM) for three categories with a low, moderate and high identified hazard, without specifying which RMM are needed for a specific exposure scenario. We propose to differentiate frequency of exposure based on differences in activities and to compare measured and estimated local skin exposure levels with rules of thumb for evaluation of control of risks per hazard category. For workers, specific RMM regimes are assigned to each combination of hazard category and process category (PROC). For consumers, a strategy in which RMM are arranged from product-integrated measures to the use of personal protective equipment (PPE) is presented. Our approach may be transferred into automated assessment tools like Chesar and CEFIC GES.     

Cumulative risk assessment of the exposure to pyrethroids through fruits consumption in China – Based on a 3-year investigation

In the present study, the long-term and short-term cumulative risks of pyrethroids exposured for the Chinese general population and children through fruits consumption were evaluated. A total of 1450 fruit samples and seven pyrethroids were included based on the pesticide residues monitoring programme of China from 2013 to 2015. The exposure was estimated using both deterministic approach and semi-probabilistic model for comparison. The hazard index approach was used to assess cumulative risk. 26% of samples contained pyrethroid residues with concentrations ranged from 0.0050 mg/kg to 1.2 mg/kg, of which 30% simultaneously with 2–4 mixture residues. Results demonstrated that the cumulative health risks were extremely low for both general population and children (1–6 years old) of China in the long term. Acute risk estimations calculated by deterministic method were several or many times overestimated than the results based on semi-probabilistic method. Acute cumulative exposure of children to pyrethroid compounds in 0.76% samples were exceeded 1 in worst case scenario. More detailed assessments with adequate data in the future use probabilistic method is expected to reduce the uncertainties of cumulative dietary exposure.     

Exposure Assessment of Multiple Mycotoxins and Cumulative Health Risk Assessment: A Biomonitoring-Based Study in the Yangtze River Delta,China

The extensive exposure to multiple mycotoxins has been demonstrated in many countries; however, realistic assessments of the risks related to cumulative exposure are limited. This biomonitoring study was conducted to investigate exposure to 23 mycotoxins/metabolites and their determinants in 227 adults (aged 20–88 years) in the Yangtze River Delta, China. Eight mycotoxins were detected in 110 urine samples, and multiple mycotoxins co-occurred in 51/227 (22.47%) of urine samples, with deoxynivalenol (DON), fumonisin B1 (FB1), and zearalenone (ZEN) being the most frequently occurring. For single mycotoxin risk assessment, FB1, ZEN, aflatoxin B1 (AFB1), and ochratoxin A (OTA) all showed potential adverse effects. However, for the 12 samples containing DON and ZEN, in which none had a hazard risk, the combination of both mycotoxins in two samples was considered to pose potential endocrine disrupting risks to humans by hazard index (HI) method. The combined margin of exposure (MOE_T) for AFB1 and FB1 could constitute a potential health concern, and AFB1 was the main contributor. Our approach provides a blueprint for evaluating the cumulative risks related to different types of mycotoxins and opens a new horizon for the accurate interpretation of epidemiological health outcomes related to multi-mycotoxin exposure.     

The applicability of in vitro-derived data in hazard identification and characterisation of chemicals

Toxicological hazard and risk assessments for chemicals presently are mainly based on highly standardised protocols for animal experimentation and exposure assessment. In this paper the possibilities are being discussed of developing systems in which the systemic (acute and chronic) toxicity of chemicals can be quantified, without the heavy reliance on animal experiments. On the basis of a chemical's structure, in vitro data on its toxicity, and biokinetic modelling a decision/flow scheme is presented. Key elements are the evaluation of chemical functionalities representing structural alerts for toxic actions, the construction of biokinetic models on the basis of non-animal data (e.g. tissue–blood partition coefficients (PCs), in vitro biotransformation parameters), tests or batteries of tests for determining basal cytotoxicity and more specific tests for evaluating tissue- or organ toxicity. It is concluded that such a flow chart is a useful tool for different steps in the toxicological hazard and risk assessment, especially for those forms of toxicity for which validated in vitro and other non-animal tests have already been developed.     

Occupational exposure to formaldehyde: genotoxic risk evaluation by comet assay and micronucleus test using human peripheral lymphocytes

Formaldehyde (FA) is a world high-production compound with numerous applications ranging from production of resins to medicines. Due to its sensitizing properties, irritating effects and potential cancer hazard FA is of great environmental health concern. Numerous studies in humans and experimental animals demonstrated that inhaled FA produced toxicity, genotoxicity, and cancer at distal sites. IARC, based on sufficient data, reclassified FA as a human carcinogen. The highest level of human exposure to this aldehyde occurs in occupational settings, namely, in pathology and anatomy laboratories, where FA is commonly used as a fixative and tissue preservative. Several studies consistently showed that the levels of airborne FA in anatomy laboratories exceeded recommended exposure criteria. In order to assess the genotoxic effects of chronic occupational exposure to FA, a group of pathology/anatomy workers was assessed using a micronucleus (MN) test and comet assay. The level of exposure to FA was also determined and the time-weighted average (TWA) of exposure was calculated for each subject. The TWA mean value for FA exposed workers was 0.43 ± 0.06 ppm, exceeding national and international recommended limit levels of 0.3 ppm. Both MN frequency and comet assay parameters were significantly higher in exposed subjects. Data obtained confirm a correlation between genetic damage and occupational exposure to FA. These data, along with recent implications of human carcinogenicity, point out the need for close monitoring of occupational exposure to FA. Implementation of security and hygiene measures as well as good practices campaigns may be crucial to decrease risk.     

Screening of chemicals for human bioaccumulative potential with a physiologically based toxicokinetic model

Human bioaccumulative potential is an important element in the risk assessment of chemicals. Due to the high number of synthetic chemicals, there exists the need to develop prioritisation strategies. The purpose of this study was to develop a predictive tool for human bioaccumulation risk assessment that incorporates not only the chemical properties of the compounds, but also the processes that tend to decrease the concentration of the compound such as metabolisation. We used a generic physiologically based toxicokinetic model that based on in vitro human liver metabolism data, minimal renal excretion and a constant exposure was able to assess the bioaccumulative potential of a chemical. The approach has been analysed using literature data on well-known bioaccumulative compounds and liver metabolism data from the ECVAM database and a subset of the ToxCast phase I chemical library—in total 94 compounds covering pharmaceuticals, plant protection products and industrial chemicals. Our results provide further evidence that partitioning properties do not allow for a reliable screening criteria for human chemical hazard. Our model, based on a 100% intestinal absorption assumption, suggests that metabolic clearance, plasma protein-binding properties and renal excretion are the main factors in determining whether bioaccumulation will occur and its amount. It is essential that in vitro metabolic clearance tests with metabolic competent cell lines as well as plasma protein-binding assays be performed for suspected bioaccumulative compounds.     

Application of Bayesian networks for hazard ranking of nanomaterials to support human health risk assessment

In this study, a Bayesian Network (BN) was developed for the prediction of the hazard potential and biological effects with the focus on metal- and metal-oxide nanomaterials to support human health risk assessment. The developed BN captures the (inter) relationships between the exposure route, the nanomaterials physicochemical properties and the ultimate biological effects in a holistic manner and was based on international expert consultation and the scientific literature (e.g., in vitro/in vivo data). The BN was validated with independent data extracted from published studies and the accuracy of the prediction of the nanomaterials hazard potential was 72% and for the biological effect 71%, respectively. The application of the BN is shown with scenario studies for TiO₂, SiO₂, Ag, CeO₂, ZnO nanomaterials. It is demonstrated that the BN may be used by different stakeholders at several stages in the risk assessment to predict certain properties of a nanomaterials of which little information is available or to prioritize nanomaterials for further screening.     

Micronucleus frequencies in lymphocytes and reticulocytes in a pesticide-exposed population in Portugal

A wide range of chemical products known to be acutely toxic is currently used in the agricultural sector, including numerous pesticides with different compositions. Nevertheless, the effects in human health as result of chronic exposure to low levels are not yet completely understood. The methodology for determination of micronuclei (MN) in lymphocytes (CBMN) is well established, and accumulating data demonstrated a correlation to enhanced risk of cancer development. However, analysis of MN in reticulocytes (MN-RET) in humans is a recent tool on human biomonitoring. The aim of this study was to examine the influence of pesticide exposure on MN-RET and CBMN frequencies. In total, 177 individuals were studied (93 controls and 84 exposed). All individuals included in the exposed group were exposed regularly to various chemicals. Both MN-RET and CBMN were significantly higher in the exposed subjects compared to controls. The CBMN frequencies were quantitatively higher in females than males, especially within the exposed group. Smoking habits exerted no marked influence on the frequency of the biomarkers studied. A significant and positive correlation was found between both indicators. Within the exposed group, data showed that there was a significant correlation between MN-RET and recent exposure (exposure in the previous 10 d) that is not found when considering CBMN. It is conceivable that due to the short life span of reticulocytes, MN-RET were found to be more reliable to characterize recent genetic damage as opposed to CBMN.     

Risk assessment of substances that are both genotoxic and carcinogenic report of an International Conference organized by EFSA and WHO with support of ILSI Europe.

The European Food Safety Authority (EFSA) and the World Health Organization (WHO), with the support of the International Life Sciences Institute, European Branch (ILSI Europe), organized an international conference on 16-18 November 2005 to discuss how regulatory and advisory bodies evaluate the potential risks of the presence in food of substances that are both genotoxic and carcinogenic. The objectives of the conference were to discuss the possible approaches for risk assessment of such substances, how the approaches may be interpreted and whether they meet the needs of risk managers. ALARA (as low as reasonably achievable) provides advice based solely on hazard identification and does not take into account either potency or human exposure. The use of quantitative low-dose extrapolation of dose-response data from an animal bioassay raises numerous scientific uncertainties related to the selection of mathematical models and extrapolation down to levels of human exposure. There was consensus that the margin of exposure (MOE) was the preferred approach because it is based on the available animal dose-response data, without extrapolation, and on human exposures. The MOE can be used for prioritisation of risk management actions but the conference recognised that it is difficult to interpret it in terms of health risk.     

Assessing the safety of cosmetic chemicals: Consideration of a flux decision tree to predict dermally delivered systemic dose for comparison with oral TTC (Threshold of Toxicological Concern)

Threshold of Toxicological Concern (TTC) aids assessment of human health risks from exposure to low levels of chemicals when toxicity data are limited. The objective here was to explore the potential refinement of exposure for applying the oral TTC to chemicals found in cosmetic products, for which there are limited dermal absorption data. A decision tree was constructed to estimate the dermally absorbed amount of chemical, based on typical skin exposure scenarios. Dermal absorption was calculated using an established predictive algorithm to derive the maximum skin flux adjusted to the actual ‘dose’ applied. The predicted systemic availability (assuming no local metabolism), can then be ranked against the oral TTC for the relevant structural class. The predictive approach has been evaluated by deriving the experimental/prediction ratio for systemic availability for 22 cosmetic chemical exposure scenarios. These emphasise that estimation of skin penetration may be challenging for penetration enhancing formulations, short application times with incomplete rinse-off, or significant metabolism. While there were a few exceptions, the experiment-to-prediction ratios mostly fell within a factor of 10 of the ideal value of 1. It can be concluded therefore, that the approach is fit-for-purpose when used as a screening and prioritisation tool.     

Mathematical models in quantitative assessment of carcinogenic risk

Regulatory guidelines have specified default assumptions for the calculation of upper bounds on potential carcinogenic risk. These default "science policy" assumptions are used in the absence of more appropriate biological information, but can, in theory, be overridden by compound-specific data. As a screening tool, the standard LMS procedure provides a useful yardstick for regulators to compare hazard potential across compounds and across exposure scenarios. This standard methodology does not estimate actual risk, however. When specific compounds are being considered for detailed evaluation, methodology for more accurately estimating the true risk is often available. Pharmacokinetic models can be used to more accurately define the high-dose to low-dose extrapolation as well as to provide insight into across route and across species extrapolations. "More central" estimates of the upper bound on risk can be derived, or decision analysis techniques can be used to define the probability distribution of risk estimates, as the assumptions are changed. These data give the risk manager more information as to the relevance of the upper bound, and should be presented in any risk estimate to provide a richer demonstration as to the true risk.     

Long-term exposures to low doses of titanium dioxide nanoparticles induce cell transformation,but not genotoxic damage in BEAS-2B cells

Abstract

There is a great interest in a better knowledge of the health effects caused by nanomaterials exposures and, in particular to those induced by titanium dioxide nanoparticles (nano-TiO₂) due to its high use and increasing presence in the environment. To add new information on its potential genotoxic/carcinogenic risk, we have carried out experiments using chronic exposures (up to 4 weeks), low doses, and the BEAS-2B cell line that, as a human bronchial epithelium cells, can be considered a good cell target. Cell uptake has been assessed by transmission electron microscopy (TEM) and flow cytometry (FC); genotoxicity was evaluated using the comet and the micronucleus (MN) assays; and cell-transforming ability was evaluated using the soft-agar assay to detect anchorage-independent cell growth. Results show an important cell uptake at all the tested doses and sampling times used (except for 1?µg/mL and 24-h exposure). Nevertheless, no genotoxic effects were observed in the comet and in the MN assays. This lack of genotoxic effect agrees with the FC results showing no induction of intracellular reactive oxygen species (ROS), the data from the comet assay with formamidopyrimidine DNA glycosylase (FPG) enzyme showing no induction of oxidized bases, and the lack of induction of expression of heme-oxygenase (HO-1) gene both at the RNA and protein level. On the contrary, significant increases in the number of clones growing in an anchorage-independent way were observed. This study would indicate a potential carcinogenic risk associated to nano-TiO₂ exposure, not mediated by a genotoxic mechanism.