Dynamic modelling of time series data in nutritional metabonomics — A powerful complement to randomized clinical trials in functional food studies

Abstract

Functional foods are foods or dietary ingredients that provide a health benefit beyond basic nutrition. A new legislation, known as the Nutrition and Health Claims Regulation, defines the legal framework for such claims within the European Union. Any claim about the nutritional or physiological effects of a product must be scientifically demonstrated. In this study, we have focused on the exploration of metabonomics as a complementary profiling technology to establish monitoring/data analysis procedures of randomized nutritional trials. More specifically, a combined intake of soybean and grapefruit in a human intervention study was analyzed with respect to both pharmacological and physiological effects. Resulting multivariate models showed a diet-induced decrease of lactate, cholesterols and triglycerides. The most drastically elevated metabolite, myo-inositol, was found to accompany a marked reduction of triglyceride levels. Suggestively, this is due to the biotransformation of myo-inositol to phosphatidylinositol, which results in a decrease of available precursors to form triglycerides. Strong inter-subject variation was present that required special attention. Dynamic modelling of collected time series data that provided the opportunity to identify slow, medium or fast responders as well as groups of subjects showing different response profiles, was also highlighted in the study. The applied strategy of time series data has proven to be a powerful complement to randomized nutritional studies adopting a clinical trial design.