Biomarkers of Red Meat Intake: Are We There Yet? Applications of Untargeted Metabolomics in the Discovery and Validation of Biomarkers

Research output: Book/ReportPh.D. thesisResearch

Assessment of dietary intake is of central importance in nutrition research. Current tools to estimate the human diet are, however, regarded as flawed and subjective as they rely on the will and memory of the study participants to report the foods they consume. Imprecise assessment of dietary intake may hinder the associations between specific foods and their effects on human health, thus complicating public recommendations. Meat stands as a good example. While protein-rich diets have been proposed as a potential strategy to combat the global obesity epidemic, the health effects of meat consumption are controversial. Red and processed meats, in particular, have been associated with an increased risk of colorectal cancer and diabetes. The nature of this association and whether or not it is causal is often debated. Such debates highlight the need to assure a high accuracy of dietary assessment tools. Objective biomarkers measured in biofluids hold great promise in complementing traditional methods and aid their accuracy. Several metabolomics studies have already pointed at candidate biomarkers for intake of meat, such as, for example, carnosine and carnitine for meat, in general, or anserine and 3-methylhistidine for chicken. This thesis starts off providing a comprehensive overview of existing biomarkers of meat intake while also indicating additional work needed to determine their usefulness as dietary tools (paper I). For red meat, a shortage of biomarkers to specifically reflect its intake was observed. In this thesis I therefore sought to unravel biomarkers for meats of decreased redness (red, pink and white) by applying untargeted metabolomics, in an attempt to document biomarkers of red meat intake (paper II). Confirmation in independent studies, validation according to established criteria, and evaluation of their ability to predict intake are essential steps needed to bring real progress into the field of biomarkers of intake. I thereby confirmed and partially validated the newly discovered biomarkers in an independent, less controlled, intervention study. Although no specific red meat biomarkers were observed, combined markers were shown to be superior to single markers in predicting intake of red meat (paper II). Biomarkers are routinely discovered by comparing a food of interest to a matched-control, in this case, one type of meat to a meat-free control. In real life subjects consume different subtypes of meats simultaneously; biomarkers should, therefore, be able to disentangle mixed exposures and thereby indicate, in a qualitative manner, firstly, if any meat was consumed, and, secondly, what type of meat it was. In the absence of specific red meat markers, the discrimination of red from white meats remains a challenge. By combining partially validated biomarkers of white and general meat intake, I aimed to separate i) omnivorous and vegetarian subjects, and ii) consumers of red and white meats, in a free-living population. Both separations were established with very high sensitivity (AUROC > 0.9) using single and combined markers (paper III). The ideal biomarkers should also indicate quantitatively the amount of meat consumed. In the present thesis, it is emphasized that not all biomarkers have both qualitative and quantitative characteristics; some single meat biomarkers are useful in estimating quantity once the source of meat is defined, while others only work qualitatively (paper III). This observation may help in a broader perspective to explain the low concordance between biomarkers and dietary records observed in many biomarker studies. In this thesis I conclude that even in the absence of red meat specific biomarkers, it is possible to estimate red meat intake by applying a step-wise objective assessment strategy. Moreover, 24h food diaries used to determine consumption of various kinds of meat, have high quality with only a few percent misclassifications, especially for general and white meats. Improvements in the food grouping of dietary software are pivotal in understanding the effect of single foods on human health.
Original languageEnglish
Place of PublicationCopenhagen
PublisherDepartment of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen
Number of pages81
ISBN (Print)978-87-7209-276-8
Publication statusPublished - 2019

    Research areas

  • The Faculty of Science - Validation of biomarkers of intake of food, Red meat intake, Untargeted metabolomics

ID: 223566349