Human intervention studies have provided clear evidence that dietary polyphenols (eg, flavonoids--eg, flavonols--and isoflavones) are at least partly absorbed and that they have the potential to exert biological effects. Biological activity of polyphenols is often assessed by using cultured cells as tissue models; in almost all such studies, cells are treated with aglycones or polyphenol-rich extracts (derived from plants and foods), and data are reported at concentrations that elicited a response. There are 2 inherent flaws in such an approach. First, plasma and tissues are not exposed in vivo to polyphenols in these forms. Several human studies have identified the nature of polyphenol conjugates in vivo and have shown that dietary polyphenols undergo extensive modification during first-pass metabolism so that the forms reaching the blood and tissues are, in general, neither aglycones (except for green tea catechins) nor the same as the dietary source. Polyphenols are present as conjugates of glucuronate or sulfate, with or without methylation of the catechol functional group. As a consequence, the polyphenol conjugates are likely to possess different biological properties and distribution patterns within tissues and cells than do polyphenol aglycones. Although deconjugation can potentially occur in vivo to produce aglycone, it occurs only at certain sites. Second, the polyphenol concentrations tested should be of the same order as the maximum plasma concentrations attained after a polyphenol-rich meal, which are in the range of 0.1-10 micromol/L. For correct interpretation of results, future efforts to define biological activities of polyphenols must make use of the available data concerning bioavailability and metabolism in humans.