Update on acrylamide levels in food from monitoring years 2007 to 2010
Recommendations on the monitoring of acrylamide levels in food by the European Commission (EC) were extended in 2010 (EC, 2010). The new Recommendation includes a revised categorisation of food products for monitoring purposes. This report describes the results of European acrylamide monitoring during the period from 2007 to 2010 using the revised product categorisation with 10 main food categories and an additional 26 sub-categories. Twenty-five European countries submitted a total of 13 162 acrylamide results for the four-year period including 2 200 results for the year 2010. During the monitoring period, time trends in acrylamide levels for the food categories were estimated. In 2010, middle bound mean acrylamide values ranged from 31 μg/kg for ‘other processed cereal based foods for infants and young children’ to 1 350 μg/kg for ‘coffee substitutes’. The highest 95th percentile value of 8 044 μg/kg was reported for ‘instant coffee’. The trend analysis showed only few changes in acrylamide levels from 2007 to 2010. At main food category level, a ‘common European trend’ was a decrease in acrylamide levels for ‘processed cereal based foods for infants and young children’ and an increase for ‘coffee and coffee substitutes’. As a ‘common European trend’ at sub-category level, acrylamide levels of ‘biscuits and rusks for infants and young children’ and ‘non-potato savoury snacks’ showed a decrease and an increase was seen for ‘crisp bread’. A marginal decrease was observed for the sub-category ‘other processed cereal based foods for infants and young children’ and a marginal increase was observed for ‘French fries from fresh potatoes’ as well as for ‘instant coffee’. Although only applicable from 2011, acrylamide levels were compared with indicative values recommended by the EC. Indicative values were exceeded in the case of 3-20 % of samples in different food categories based on 2010 monitoring data. An extended time period and detailed descriptions of sample sources would be needed for a more accurate trend evaluation.