Acropolis logo

Session 3: Future perspective on aggregate exposure and toxicological testing

The cumulative exposure model produced by ACROPOLIS is now a workable concept in line with EFSA guidance. Aggregated exposure assessment, as defined in the ACROPOLIS project, refers to exposure assessment via different routes e.g. skin, inhalation and food. Aggregated exposure assessment is still in its infancy and todays discussions are mainly related to improving non-dietary exposure. However, the legislation also requires aggregated exposure assessment to be implemented in future risk assessment and risk management. In this session two speakers looked at the future perspective on aggregate exposure in risk assessment and how ACROPOLIS can contribute to that future needs. The third speaker elaborated on new toxicological testing in relation to confirm the assumption of dose-addition made in current EFSA methodology and also to narrow the number of pesticides in a common assessment group.

Richard Glass (FERA-UK) presented the achievements of the ACROPOLIS project with respect to aggregated exposure model for pesticide risk assessment. Aggregate exposure is an estimate of the exposure of a defined population to a given compound by all relevant routes and from all relevant sources. This includes exposure via the diet, via the skin (dermal) and via inhalation. Different scenarios can be defined for occupational exposure of operators and workers, incidental exposure of bystanders and residents (often living adjacent to farm land), and use of amateur or consumer products in the home and garden.

Several collaborative projects have contributed to the improvement of data collection and concepts for aggregated exposure assessment. Results of these projects are becoming available now. In addition to the ACROPOLIS project, Richard Glass highlighted the FP7 Browse project on improving models for non-dietary exposure to pesticide of bystanders, residents, operators and worker. For aggregated exposure assessment, several aspects need to be considered amongst which are mentioned the toxicological properties of the pesticides e.g. acute or chronic toxicology, and use patterns of pesticides in agriculture.

Richard Glass presented several case studies. The first case study illustrates a combination of dietary exposure and occupational exposure for UK operators to the triazole group of compounds. The ACROPOLIS IT tool combines the non-dietary and dietary exposure input together and the output of this case study was presented in understandable output graphics and diagrams. A second case study illustrates the combination of exposure of a pesticide used in agriculture and which pesticide was also present in consumer products. In this case study the output for non-dietary exposure is generated using the ConsEXPO model and the outputs was linked to the dietary exposure for a defined population group. The aggregate ACROPOLIS model allows for flexibility in the input and the output selections.

Although the first experience with the aggregated exposure models seems promising, a number of critical data gaps were identified for both exposure data and use patterns. Most critical data gaps were identified and the need for these to be addressed in the future. An EFSA funded project aiming to collect farm survey data for cumulative risk assessment for plant protection product CFT/EFSA/PPR/2010/04) was highlighted as a good example to make this possible.

Acro12 Acro13

Sebastian Denys and Valerie Pernelet (ANSES –France) presented an example of how exposure via food and exposure via environmental contamination can be combined in an aggregated exposure assessment. They showed a case study using the chemical bisphenol A, which occurs as a contaminant in food and the environment. The presentation outlined the current single chemical and single source approach as being the state of the art in today’s risk assessment approaches. However, bisphenol A is present in different concentrations in several food items, in air dust and air. Exposure assessment via different routes requires additional information regarding absorption factor after inhalation and bioavailability factors after oral ingestion. Finally, results of the exposure assessments needs to be summed and expressed on an internal dose dimension in the human body. PB-PK models are necessary to elaborate on potential metabolism and distribution of bisphenol A within the human body before the chemical poses effect on the critical target organ.

ANSES worked on a conceptual overview of all potential sources and routes relevant for exposure to bisphenol A. In a second step, ANSES identified relevant input data either coming from existing database such as exposure data from the French Total Diet Study or from literature research. The final bisphenol A hazard characterization was based on a peer review of all available data including a bioavailability factor for bisphenol A after oral ingestion and a skin absorption factor.

The case study addressed the exposure of bisphenol A to pregnant women. The bioavailability of bisphenol A was assumed as high as 3% after oral ingestion for food and an adsorption factor of 100% for exposure via inhalation. The results are expressed as a probabilistic distribution showing a major contribution to the total internal exposure dose from food (84%), air (12%) and dust (4%).

ANSES concluded that aggregated exposure is a challenge for all of us and not only for pesticide risk assessment. The aggregated exposure assessment can be improved in the future by including more uncertainty analyses or links with bio-monitoring data and/or PB-PK modeling. Furthermore, a link with other population groups and a standardized approach to link aggregated models with exposure data in France or various member states in a standardized manner is highly recommended.

Angelo Moretto (University of Milan) completed the section with a lecture addressing the new toxicological in-vitro testing performed in the ACROPOLIS project. The aim of the work is to define which pesticides should be included in the common assessment group of triazoles and group of pesticides affecting neurotransmission. A second aim was to test the dose-addition assumption for mixtures of pesticide exposure. Furthermore, the work on toxicology testing in ACROPOLIS elaborated on the usefulness of PB-PK models for estimating internal exposure: this was applied to both extrapolation from in vitro concentrations to in vivo doses (“reverse dosimetry”), and to estimate exposure once a consumer is exposed to two pesticides simultaneously.

Angelo Moretto highlighted the results of the in-vitro study. A nice dose-response relationship was seen for the tested pesticides triadimefon and flusilazole. Apart from the pesticides tested, a mixture of both pesticides was added to the test system. A stronger effect was observed compared to the separate effect of the single chemicals, but the magnitude of the effect of the mixture was according to expectation based on the dose-additivity assumption made in the EFSA guidance.

In vitro test are also very helpful to narrow down the number of pesticides assumed to belong to a common assessment group. Furthermore, in vitro testing helps to reduce the number of animal testing. However, in-vitro test results are not always quantitatively the same as in vivo results. Compounds are activated or metabolized via different pathways or with different efficiency. Therefore, in-vitro to in vivo extrapolation is not always straightforward and in-vitro validation with in-vivo experiments are necessary to validate in-vitro experiments. Within the ACROPOLIS project, the University of Milan has performed such a validation study. Angelo Moretto highlighted a few results and he recommends gaining more experience with finding the correct scaling factor between animal in-vitro testing and the real human situation.

Within the ACROPOLIS project, a PB-PK model was developed addressing exposure to two pesticides simultaneously. The PB-PK model also allows for entering information of exposure via different routes. The PB-PK model is internet compatible and in future developments the PB-PK model and the exposure modelling presented by Polly Boon might be linked directly. The PB-PK model was very helpful to link in vitro data with the in vivo situation and for the interpretation of the toxicological findings resulting from a field study, which was set up for validation reasons. In this field study relevant exposure routes were measured such as exposure via skin and via a duplicate diet study as well as the concentration of the pesticide in urine. The measured and predicted aggregated exposure results are not ready now, but will be published in a scientific paper after the project has ended.

Acro15 Acro18