Assessment of flavonoid and fatty acid intake by chemical analysis of biomarkers and the duplicate diets

Dietary intake is important to investigate the relationship between diet and the occurrence of disease. However, it is difficult to assess the intake of nutrients such as flavonoids, minor fatty acids and plant sterols because the data on these nutrients in food composition tables are insufficient or because the bioavailability of these nutrients differs between foods. The results of studies investigating the relationship of these nutrients to disease are inconsistent, perhaps because of errors in the methods used to assess nutrient intake. The aim of this thesis was to evaluate physically and chemically based methods of measuring the intake of flavonoids, fatty acids, sterols and energy.

We found differences of up to 80 mg per day in the intake of flavonols and flavones in subjects eating a variety of diets. The ratio of the within- to the between-subject variation in the intake of flavonols and flavones was lower than one, indicating that it is possible to study the relationship between flavonoid intake and disease. The food frequency questionnaire used in this study was suitable for classifying subjects by their flavonol intakes.

We also found that the bioavailability of the flavonol quercetin differs between the major dietary sources. The bioavailability of quercetin from red wine was 75% of that from onions and from tea 50% of that from onions. Therefore, flavonols from red wine can probably not explain the lower incidence of coronary heart disease in France compared to other western countries. Concentrations of quercetin in plasma can be used as biomarkers to distinguish between subjects with a low and with a high flavonol intake. This is possible because of a relatively small variation of plasma quercetin and a linear relationship between quercetin intake to its concentrations in plasma and excretions in urine.

Chemical analysis of food composites is a suitable method to assess the intake of a large number of fatty acids and sterols for subsamples of populations. We found large differences in the amount of these nutrients in the diets of middle-aged men living in 16 cohorts in seven countries. Three-day records are not suitable to measure individual energy intake, but they can be used to classify subjects by their energy intakes.

In conclusion, the most feasible method to assess flavonol intake in epidemiological studies appears to be a food frequency questionnaire specially devised to assess flavonol intake and validated in a sub-population by biomarkers of flavonol intake. In addition, chemical analysis of food composites is a good tool to assess specific fatty acids and sterols in the diet of a population. Finally, when using 3-day food records, even from a motivated, lean, well-educated population, underestimation of intakes has to be taken into account.