Time Trends of Acrylamide Exposure in Europe: Combined Analysis of Published Reports and Current HBM4EU Studies

Abstract

More than 20 years ago, acrylamide was added to the list of potential carcinogens found in many common dietary products and tobacco smoke. Consequently, human biomonitoring studies investigating exposure to acrylamide in the form of adducts in blood and metabolites in urine have been performed to obtain data on the actual burden in different populations of the world and in Europe. Recognizing the related health risk, the European Commission responded with measures to curb the acrylamide content in food products. In 2017, a trans-European human biomonitoring project (HBM4EU) was started with the aim to investigate exposure to several chemicals, including acrylamide. Here we set out to provide a combined analysis of previous and current European acrylamide biomonitoring study results by harmonizing and integrating different data sources, including HBM4EU aligned studies, with the aim to resolve overall and current time trends of acrylamide exposure in Europe. Data from 10 European countries were included in the analysis, comprising more than 5500 individual samples (3214 children and teenagers, 2293 adults). We utilized linear models as well as a non-linear fit and breakpoint analysis to investigate trends in temporal acrylamide exposure as well as descriptive statistics and statistical tests to validate findings. Our results indicate an overall increase in acrylamide exposure between the years 2001 and 2017. Studies with samples collected after 2018 focusing on adults do not indicate increasing exposure but show declining values. Regional differences appear to affect absolute values, but not the overall time-trend of exposure. As benchmark levels for acrylamide content in food have been adopted in Europe in 2018, our results may imply the effects of these measures, but only indicated for adults, as corresponding data are still missing for children.

Keywords: HBM; acrylamide; exposure level; glycidamide; human biomonitoring; time-trend.

Figure A1

Plots of AAMA vs. GAMA (in µg/L), in 9 different European populations ((A–I), individual data) as indicated. Linear fit in red, gray = 95%. AAMA to GAMA correlation is linear within the range of observations and the slope is depending on region/country, but not age, as indicated by data from France based on kids and adults (H,I). Higher GAMA-to-AAMA level ratio is reported from Italy (EPIUD, (A)) and Porugal (INSEF, (B)), low ratio from France (H,I).

Figure 1

Flowchart of study design.

Figure 2

Harmonized overview of mean reported acrylamide biomonitoring levels based on previously published literature and HBM4EU Aligned Studies (2000–2021). Mean/estimated mean (+standard error) for AAMA (µg/L, (A)) and GAMA (µg/L, (B)), color/pattern coded for applied harmonization calculation method. Standard error of the mean (SEM) is shown for visual comparison of variance taking into account of differences in study population size. Study index labels (LH-FR6) are shown according to Table 1.

Figure 3

Linear model of corresponding acrylamide biomarkers in blood and urine matrix. Regression of AAMA vs. AAVal ((A), AAMA (µg/L) = 1.64 × AAVal, pmol/g Hb, p < 0.0001) and GAMA and GAVal ((B), GAMA (µg/L) = 0.35 × GaVal, pmol/g Hb, p < 0.0001) in individuals (data kindly provided by T. Schettgen, M.B. Boettcher, J. Angerer). Red = regression line, blue = data points, grey = 95% confidence interval.

Figure 4

Linear model (regression) of harmonized means of acrylamide biomarkers AAMA and GAMA (non-smokers) as reported by European published papers and HBM4EU Aligned Studies (AAMA 30 studies, GAMA 29 studies). Regional distribution is indicated by colors and relative population size is shown by the size of the circles indicating single surveys. (A): AAMA in µg/L, (B): GAMA in µg/L. Gray: 95%-confidence interval. Regression line in blue.

Figure 5

Locally fitted scatter-plot fit (LOESS) for AAMA time-trend data (harmonized, non-smokers). Study means represented by country color-coded dots with size dependent on count number and 95%-confidence interval (gray). LOESS fit line in blue. The red line shows the trend of a multi-segmented regression analysis (significant break-point detected in 2017, p < 0.001). Dashed line indicating the time-point of Commission Regulation 2017/2158 put in force (2018).

Figure 6

Linear model (regression) of harmonized means of acrylamide biomarkers AAMA and GAMA in smokers as reported by European published papers and HBM4EU Aligned Studies. Study means represented by country based color-coded dots with size dependent on the corresponding sample number. Population size is represented by the size of the circles indicating single surveys. (A): AAMA in µg/L, (B): GAMA in µg/L. Gray: 95%-confidence interval. Regression line is in blue.