We are encountering a significant progress in nutritional knowledge, relevant to the practice of all aspects of human nutrition, medicine and public health. So, it is conceivable that we may view the role of nutrition differently in the future. The diet may not only provide an adequate amount of nutrients to meet the metabolic requirements, but could also contribute to improving human health status. As a consequence, extracts of plants or single compounds thereof which are believed to benefit human health need to be identified and developed for the food market to complement a balanced diet. The assessment of risk and benefit of constituents of a diet or plants will be a challenge for scientists working in this area. The number of compounds to be tested is enormous. Their impact on human health is supposed to be through prevention. Their effect on the human system may be modest, yet still significant when consumed over an entire lifespan. Trials employing traditional clinical endpoints, for many reasons, appear not to be feasible to investigate the relevance of these compounds on human health. Rather biomarkers, which are 'surrogate endpoints' for clinical events, may be used in the field of nutritional science. These biomarkers have to be highly sensitive and specific, non- or minimally invasive and inexpensive and they need to be validated and standardized. The biomarker concept may prove essential in nutritional science to demonstrate the effect of diet constituents on human health.