Validation of a database on acrylamide for use in epidemiological studies

Abstract

Acrylamide, a probable human carcinogen, was detected in various heat-treated foods such as French fries and potato crisps. Recently, positive associations have been found between dietary acrylamide intakes, as estimated with a food frequency questionnaire using an acrylamide database, and cancer risk in some epidemiological studies. As acrylamide levels vary considerably within the same type of foods, a validation study was performed to investigate whether use of an acrylamide food database containing calculated mean acrylamide content, based on extensive sampling and chemical analysis of Dutch foods (several samples per food), can classify subjects with respect to true acrylamide intake. We used the data from a 24-h duplicate diet study. The acrylamide content of 39 Dutch 24-h duplicate diets collected in 2004 was estimated using the mean acrylamide levels of foods available from the database and the menu list, on which the participants of the duplicate diet study had listed the amounts of individual foods and drinks in household units. Next, the acrylamide content of the total duplicate diets was analytically measured and correlated to the estimated acrylamide contents. The Spearman's correlation coefficient between chemically determined acrylamide content and the calculated acrylamide content of the duplicate diets was 0.82 (P<0.001). This study indicates that it is possible to classify subjects with respect to acrylamide intake if mean instead of actual content of each food is applied. The database can therefore be applied in epidemiological studies on acrylamide intake and cancer risk, such as the Netherlands Cohort Study on Diet and Cancer.