Acropolis logo


Cumulative dietary exposure modelling

Based on a case study into the cumulative dietary exposure to triazole residues performed in the ACROPOLIS project, the following recommendations were formulated:

  1. The use of the summary information on processing factors as provided by the German database developed by the Federal Institute for Risk Assessment (BfR) was very useful, but the link with concentration databases containing residue data of multiple substances and raw agricultural commodities needs to be optimised for use on a routine basis.
  2. Assessing the exposure according to the pessimistic basic model run, as described in the EFSA Guidance on probabilistic modelling, was very laborious. Within the ACROPOLIS project an automated procedure was developed to overcome the major limitations. However further optimization of data compatibility is still recommended.
  3. More experience is needed with the EFSA Guidance on probabilistic modelling. Some kind of intermediate ‘realistic’ scenario may be needed that combines the settings of the optimistic and pessimistic model run in such a way that it results in more realistic exposure estimates. These estimates should not be over-conservative as now seems the case with the EFSA Guidance basic pessimistic model run. Based on our experience, optimal settings for the probabilistic assessment can be proposed to EFSA and the European Commission.
  4. The results of probabilistic modelling should be compared with a safe intake level. Currently such a safe intake level is based on a deterministic approach. It is recommended that a safe level of intake is also set by a probabilistic methodology. To establish these numbers risk managers are recommended to take into account the experience of the US Environmental Protection Agency (EPA) to globally harmonize the way of interpreting exposure results. Furthermore, it is recommended to explore the Margin of Exposure approach to establish safe intake levels as set out in the EFSA guidance on the use of probabilistic modeling.

Aggregate exposure modelling

Within the ACROPOLIS project, an aggregated exposure model was developed. With this model, the non-dietary exposure to pesticide residues can be estimated. This model will contribute to a better understanding of the issues related to aggregate exposure and will facilitate an open and transparent discussion on the required level of protection regarding cumulative and aggregated exposure.

The recommendations regarding aggregate exposure are

  1. More experience is needed to understand the aggregated exposure in full detail. We therefore recommend to test the developed aggregate exposure model in more detail than done in the project using more aggregate exposure scenarios.
  2. The EFSA Guidance on non-dietary exposure to pesticides (EFSA calculator), the BROWSE project and other European initiatives are working on standardizing and improving non-dietary exposure assessments. The experience of aggregate exposure modelling gained in the ACROPOLLS project might help to discuss future directions.
  3. We explored successfully a potential infrastructural link between the ACROPOLIS model on the one side and the BROWSE model and EFSA calculator on the other side. This exploration should be continued for as far as feasible in future EFSA or national authorities work on aggregate exposure assessment.
  4. The ACROPOLIS aggregate exposure model was tested for the exposure to pesticide residues. The model can however easily be adapted to estimate exposures to other groups of compounds providing the exposure to these compounds can be weighted relative to a reference compound. We recommend gaining more experience in this area in line with current and future EFSA opinion(s) on non-dietary and aggregate exposure assessment.

New toxicological testing approaches

  1. In vitro test systems can be very helpful in setting criteria for in- or exclusion of pesticides or chemicals in a common assessment group. Dose addition assumptions can be tested. More chemicals with a assumed common effect might be tested. It is recommended to develop in vitro test systems for the new common assessment groups set by EFSA.
  2. In vitro test systems are also useful to test the assumption of dose addition which means that the dose of chemicals being equally toxic can be summed up. When chemicals are not equal toxic the relative potency factor has to be considered. The ACROPOLIS project delivered two examples, which can be seen as proof-of-principle. One example is addressing dose-addition for triazoles and one for neuro-toxic compounds. More experience using these test systems with other chemicals is recommended.
  3. PB-PK modelling is a powerful tool to extrapolate external exposure and in vitro-test results to internal exposure doses at the relevant target organ. Furthermore, PB-PK models have been developed to predict the internal dose effect of being exposed to two pesticides sharing the same effect. Since there will be more future attention for mixtures and extrapolation from external to internal exposure, further exploration and developments of PB-PK modelling is highly recommended.

Model developments, from a proof of principle to routine production

  1. The new ACROPOLIS cumulative exposure model has been tested by various member states using their national triazole monitoring data. The EFSA cumulative risk assessment methodology presently results in common assessment groups with a far higher number of pesticides belonging to such a group compared to the group of triazole pesticides. It is recommended to develop the model also for higher numbers of chemicals that can be addressed simultaneously.
  2. Performance of the ACROPOLIS cumulative exposure model might be influenced by the amount of monitoring data used in the simulations. A country-by-country approach has been tested and worked according to expectation. When the data of 6 countries were pooled the model runs worked as well. A full pan-European approach using all European monitoring data of pesticides belonging to a common assessment group might require more computer capacity or a redesign of certain modules in the software. If future approaches are in this direction, more performance tests are recommended. These tests should be in line with stakeholder needs to routinely apply models and upgrades of the software.
  3. The ACROPOLIS project has harmonised the input data flow and modelling efforts in Europe according to current EFSA standards as well as newly developed approaches. These standards may change in the nearby future and consequently the ACROPOLIS models should be updated.
  4. Some stakeholders might use the ACROPOLIS model on a routine basis in the future. Pesticide industry is by European legislation responsible to check the safety of new pesticides to be authorised on the market. This includes a safety assessment of cumulative exposure. Pesticide industry needs to optimise their data collection and data flow to be optimally connected to the ACROPOLIS model.
  5. A good exploitation plan of the ACROPOLIS cumulative exposure model, including a hardware platform able to fulfil all user requirements, has been developed as part of the project: the model is unique in the world and the need to perform probabilistic exposure assessments is increasing. National food authorities, EFSA and / or DG Sanco should feel responsibility to support further exploitation.


  1. The results of probabilistic exposure assessment are more transparent compared to results from deterministic approaches; by their nature they might also be more complicated to communicate. Experience from the USA might be helpful to Europe. The development of a communication strategy including a stakeholder training programme is recommended. Such a training programme should focus on the impact of tool usage on decision making and interpretation of the findings from the cumulative and aggregated assessments.
  2. Different stakeholders have different opinions on the acceptability of probabilistic modelling in risk management decisions. In comparison to other stakeholders industry stakeholders expressed a higher perceived complexity of the model and tools, a lower self-reported understanding of cumulative and aggregate exposure assessment, and a lower conviction that cumulative and aggregate exposure are issues of concern. Regulators and food authorities were considered as the stakeholder groups for whom the outcomes of the ACROPOLIS project will be most useful and acceptable. NGO representatives reported a very critical attitude towards cumulative and aggregate exposure assessment, which is mainly related to their low perceived necessity of using pesticides and highly critical attitude towards the possible outcomes of the assessment tool. The ACROPOLIS model was very well received during the first and the second ACROPOLIS stakeholder meetings. However, for future dynamics and implementation, we recommend a targeted communication plan taking into account the concerns of different stakeholder groups.