Questionnaire-based self-reported nutrition habits associate with serum metabolism as revealed by quantitative targeted metabolomics

Nutrition plays an important role in human metabolism and health. However, it is unclear in how far self-reported nutrition intake reflects de facto differences in body metabolite composition. To investigate this question on an epidemiological scale we conducted a metabolomics study analyzing the association of self-reported nutrition habits with 363 metabolites quantified in blood serum of 284 male participants of the KORA population study, aged between 55 and 79 years. Using data from an 18-item food frequency questionnaire, the consumption of 18 different food groups as well as four derived nutrition indices summarizing these food groups by their nutrient content were analyzed for association with the measured metabolites. The self-reported nutrition intake index "polyunsaturated fatty acids" associates with a decrease in saturation of the fatty acid chains of glycero-phosphatidylcholines analyzed in serum samples. Using a principal component analysis dietary patterns highly associating with serum metabolite concentrations could be identified. The first principal component, which was interpreted as a healthy nutrition lifestyle, associates with a decrease in the degree of saturation of the fatty acid moieties of different glycero-phosphatidylcholines. In summary, this analysis shows that on a population level metabolomics provides the possibility to link self-reported nutrition habits to changes in human metabolic profiles and that the associating metabolites reflect the self-reported nutritional intake. Moreover, we could show that the strength of association increases when composed nutrition indices are used. Metabolomics may, thus, facilitate evaluating questionnaires and improving future questionnaire-based epidemiological studies on human health.