Acropolis logo

Session 2: Addressing cumulative risk assessment and the experience of different user groups

This session reviewed the new ACROPOLIS IT tool addressing cumulative exposure assessment via food. The two main users of the software for setting Maximum Residue Limits (MRL) and to establish safe use of new pesticides are pesticide industry and the regulators, including EFSA peer reviewing draft reports of the member states. National Food Authorities and NGOs have been trained or have been offered training in an early stage of the ACROPOLIS project and further training tools will become available to them after the project ends. The speakers in this session evaluated the experiences and lessons learnt by the different user groups and their view on the way forward.

Introducing the session, Polly Boon (RIVM) provided an overview of the ACROPOLIS IT tool and its achievements on cumulative exposure modeling. The presentation addressed the key elements of the Monte Carlo Risk Assessment (MCRA) IT tool. This tool follows the EFSA guidance on probabilistic modeling including an optimistic and pessimistic model run. The presentation also addressed the goal of WP2 of ACROPOLIS to generate an e-platform of input data for the stakeholders involved in pesticide risk assessment in Europe, and to use the developed tool to estimate the cumulative exposure to a group of triazole pesticides according to the requirements of the EFSA guidance.

The IT tool was tested and proved to be user-friendly and easy to operate, and able to assess acute and chronic cumulative exposure. The optimistic model was easy to apply, while the performance of the pessimistic model run was more laborious because a lot of additional data is required to fulfill all the details in the EFSA guidance. During the project this has been a struggle to follow the pessimistic requirements, but based on the experience a workable approach was achieved. Certain assumptions made in the EFSA guidance turned out to result in unrealistic results, like the assumption of including MRLs for animal products resulting in a more than 50% contribution of these commodities to the overall exposure.

Based on the experience it was recommended to consider an intermediate “realistic” scenario combining the optimistic and pessimistic model run. It was also recommended to gain more experience with the EFSA guidance and the ACROPOLIS IT tool with larger Cumulative Assessment Groups.

Acro7 Acro8

Monika Bross (ECPA/BASF) reported about the experience of the industry user group . Her presentation was made on behalf of seven leading EU and USA pesticide manufacturers. The user group of pesticide industry started to get experience with the ACROPOLIS IT tool from December 2012 onwards. Since then the seven companies have organized meeting on two occasions and had bi-monthly teleconference for exchanging and agreeing on testing procedure using the ACROPOLIS IT tool. Furthermore, the industry user group closely cooperated with the ECPA Expert group on toxicology and the two working groups worked on common view of developing a higher tier dietary exposure assessment procedure for cumulative risk assessment.

The main focus of the user group tests was on acute exposure for single compound (triazole fungicide) using European consumption data and the user group became familiar with the probabilistic tool MCRA software 8 (also referred to as IT tool of ACROPOLIS). They were able to enter their own company specific compound information and compared results generated with deterministic and probabilistic approaches. During the evaluation, the calculation procedures of MCRA and DEEM were directly compared by using US consumption data. The industry user group was very pleased to see how the ACROPOLIS IT tool can be used with the data from two different continents, because it helps to harmonize procedures in pesticide risk assessment. The industry user group studied the impact that the different input parameters such as processing and variability factors had on the results of probabilistic modeling.

Although the pesticide industry companies were able to upload their data to the ACROPOLIS IT tool, a more automated procedure is recommended.

Monika Bross shared the view of pesticide industry companies on the way forward with the audience. Today there is a large volume amount of consumption and monitoring data available in the ACROPOLIS IT tool, but still data from key-countries such as Germany can make the tool even more valuable in the near future. A plea was held to regularly upload all EU monitoring data collected in EU/national programs in order to maintain the usefulness of the IT tool in the nearby future. While the ACROPOLIS IT tool is easy to use, probabilistic assessment itself is, however, still a very complex process. A more straightforward and general accepted ‘cook book’ complementary to the current EFSA guidance is highly recommended. This ‘cook book’ should explain calculations steps in easy-understandable language to non-experts. The help offered by the experts of the ACROPOLIS project was always very timely and appreciated by the industry companies. The EFSA on-going work establishing new common assessment groups needs to be reflected in the future ACROPOLIS IT tool.

In conclusion, industry companies felt that probabilistic modeling is necessary for cumulative risk assessment and that the ACROPLIS project has brought the relevant IT tools and data accessible to industry. Monika Bross recommends discussing the pending issues between DG Sanco and all relevant stakeholders, before deciding to implement probabilistic modeling in a regulatory framework. That discussion should also address intentions to work towards one agreed procedure over Europe and inclusion of realistic assumption related to the input variables. Furthermore, the European Commission should strive for a set of basic criteria for risk managers to decide on the required level of protection. The USA experience herewith, might be helpful to guide Europe through the still open discussion in how to implement probabilistic modeling in future risk management practice in Europe.

Paul Hamey ( HSE-UK) reported on the experience collected by the regulators user group. The regulators were trained in May 2013 following an invitation of the European Commission. Paul Hamey started with a reviewed of the history, the current state of the art and future direction in probabilistic risk assessment and risk management needs from the perspective of European regulators.

Prior to the ACROPOLIS project experiences were already gained both in the UK and in the USA on probabilistic modeling. Probabilistic assessments were explored in several European funded projects such as the Monte Carlo project, SAFE Foods and the current ACROPOLIS project. In the early part of the century the EU regulatory framework changed significantly with the launch of EFSA , acceptance of MRL regulation 396/2005, and new PPP regulation 1107/2009 including articles describing the future regulatory obligation to address cumulative and aggregated exposure assessment. EFSA organized a colloquium on cumulative assessments, which was followed by publications of several EFSA opinions (criteria for cumulative assessment, triazole study, guidance on exposure assessments needed for cumulative dietary assessments, identification of CAG). EFSA is in the lead to set and to finalize the cumulative risk assessment methodology.

The ACROPOLIS project is well received by the regulatory process because it addresses complexities of factoring multiple sources of exposure from a wide range of related compounds. Despite all the complex issues, the ACROPOLIS IT tool is now user-friendly and it is possible to follow the EFSA guidance on the use of probabilistic modeling for dietary intake to pesticide residues using the ACROPOLIS IT tool. It enables implementation of cumulative dietary assessment, following the approaches recommended by EFSA into the risk management deliberations in Brussels. Furthermore, ACROPOLIS also provides an approach to consider non-dietary pathways in addition to the cumulative exposure assessment. Representatives of nearly all EU member states were trained and the general feeling is that ACROPOLIS is a useful tool for the assessment of cumulative exposures to multiple compounds via food in the coming year(s). However, still a number of issues related to the acceptance and final use in risk assessment needs to be considered by the regulators and the Standing Committee on the Food Chain and Animal Health of DG SANCO. These issues are, among other things, related to uncertainty in the data and input variable and the required level of protection to be set in the European Union.

For aggregated exposure, the methodology is still in its infancy. It is therefore recommended to gain more experience and to include robust pesticide usage data and behavior patterns of operator, worker, bystander and resident into the aggregated exposure models. In a follow-up of the ACROPOLIS project, the stakeholders and DG SANCO should secure the achievements of the project for better risk assessment and risk management of real risks in Europe, while avoiding additional complexity, precaution and cost with little benefit. In addition we should strive for acceptance of probabilistic modeling within the European Commission decision making process.

Acro9 Acro10

To complete this user group review, Jorgen Schlundt (Director DTU Denmark ) presented the perspectives of the food authorities on the ACROPOLIS achievements. This presentation first reminded the organization of risk assessment, risk management and risk communication including national and international food safety institutes and responsibilities. Specificities of chemical and microbiological risk assessment were highlighted. Where the microbiological risk assessment had a more refined modeling approach compared to chemical risk assessment ten years ago, the ACROPOLIS project now shows that chemical risk assessment has potentially improved.

Jurgen Schlundt emphasized that human risk assessment of combined exposure to multiple chemicals poses several challenges including the complexity of the terminology and problem formulation. Furthermore, the diversity of chemical entities and the toxicological profiles and exposure pattern share complicated issues to be included into today’s risk assessment. The probabilistic modeling approach set by ACROPOLIS can be seen as a good example not only for pesticides but also for all types of chemical risk assessment. Particular focus was drawn on the need and relevance of an international perspective on this issue of cumulative exposure assessment. Therefore, we should look at other national experience outside Europe and we also should see the project in perspective of WTO standards (SPS agreement) or CODEX developments. Today, a big step forwards is set with the ACROPOLIS IT tool presented, but the tool and issues should also remain open for continuous improvement. The current tool is certainly not the final stage of developments.

Jurgen Schlundt made a plea not to forget the importance of risk communication. Often risk managers and scientist believe that probabilistic exposure assessment is too complicated for consumers, but this might be a misunderstanding. Consumers will be able to understand results of probabilistic modeling, but it is necessary to include consumers and their perception of risk into the risk management discussions and decisions.

The session was finalized with a lively debate. The audience and the speakers discussed issues about how to improve the use of the ACROPOLIS model for risk management practice and how to improve some practicalities in current use of the ACROPOLIS IT tool. Suggestions were given on extrapolating the ACROPOLIS model experience to other international organizations including FAO/WHO and how to get other international organization connected to a harmonized approach (e.g. via collaborating WHO centers and/or US EPA involvement).

Acro11 Acro12