The use of chemical occurrence data at European vs. national level in dietary exposure assessments: a methodological study.

A typical EFSA approach to assess dietary exposure is to combine data from national consumption surveys with chemical occurrence data that have been pooled across the EU Member States (pooled approach). This approach was compared to the case where occurrence data were stratified by country and used for food categories where national data were abundant (semi-pooled approach), using cadmium as a case study. Some differences in estimated dietary exposure were observed between the pooled and semi-pooled approach. They were explained by differences, between the national and the European occurrence data, with respect to (1) contamination values and (2) sample proportions of food items classified in the food categories the assessment was based on. The latter aspect highlighted the sensitivity of the approach of directly aggregating monitoring data into food categories. Both the pooled and semi-pooled approach tended to be conservative relative to approaches used at national level. This appears to be attributed to differences in the way the available occurrence data is aggregated. Refinement of the studied methodologies would include a better separation of the food items with high concentration from those with low concentration.