Exposure of the US Population to Acrylamide in the National Health and Nutrition Examination Survey 2003-2004

BACKGROUND: The lifelong exposure of the population to acrylamide has raised concerns about the possible health effects of the chemical. Data on the extent of exposure to acrylamide and its primary metabolite, glycidamide, arc needed to aid in the assessment of potential health effects. OBJECTIVES: The aim of this study was to assess human exposure to acrylamide and glycidamide in the general U.S. population through the measurement of hemoglobin adducts of acrylamide (HbAA) and glycidamide (HbGA). METHODS: HbAA and HbGA were measured in 7,166 subjects front the National Health and Nutrition Examination Survey. Stratified HbAA and HbGA data were reported by sex, age groups, race/ethnicity (Mexican American, non-Hispanic black, non-Hispanic white), and smoking status based on serum cotinine levels. Covariate-adjusted geometric means for each demographic group were calculated using multiple regression analysis. RESULTS: HbAA and HbGA levels ranged from 3 to 910 and from 4 to 756 pmol/g hemoglobin, respectively, with smokers having the highest levels overall. Tobacco smoke exposure in nonsmokers had a small but significant effect on HbAA and HbGA levels. Adjusted geometric mean levels for children 3-11 years of age were higher than for adults >= 60 years of age [mean (95% confidence interval): HbAA, 54.5 (49.1-51-5) and HbGA, 73.9 (71.3-76.6) vs. HbAA, 46.2 (44.3-48.2) and HbGA, 41.8 (38.7-45.2)]. Levels were highest in Mexican Americans [HbAA: 54.8 (51.9-57.8), HbGA: 57.9 (53.7-62.5)], whereas non-Hispanic blacks had the lowest HbGA levels [43.5 (41.1-45.9)]. CONCLUSIONS: U.S. population levels of acrylamide and glycidamide adducts are described. The high variability among individuals but modest differences between population subgroups suggest that sex, age, and race/ethnicity do not strongly affect acrylamide exposure. Adduct concentration data can be used to estimate relative exposure and to validate intake estimates.