Large inter-individual variation in isoflavone plasma concentration limits use of isoflavone intake data for risk assessement

Velpen, V. van der; Hollman, P.C.H.; Nielen, M. van; Schouten, E.G.; Mensink, M.R.; Veer, P. van 't; Geelen, A.


Background/objectives: Isoflavones are present in soy foods and soy-based supplements. Despite low plasma isoflavone concentrations in the general Western population, concentrations in supplement users exceed those suggested to be beneficial for health in Asian populations, raising concerns for adverse effects. To aid risk assessment, quantification of the relation between isoflavone intake and plasma concentrations is essential. Subjects/methods: Plasma samples were collected from postmenopausal women in three placebo-controlled crossover studies with 8-week periods for supplements (two studies, ~100¿mg isoflavones/day, n=88) or 4-week periods for soy foods (one study, ~48¿mg isoflavones/day, n=15). Plasma isoflavone concentrations (daidzein, equol, genistein and glycitein) were quantified using high-performance liquid chromatography and electrochemical detection. The association between plasma concentrations and isoflavone intake, equol producer status, intake–producer interaction and background dietary intake was assessed based on the assumption of a log-linear relation. Results: Median plasma total isoflavone concentrations after the soy food and supplement interventions were respectively 2.16 and 3.47¿µmol/l for equol producers and 1.30 and 2.39¿µmol/l for non-producers. Regression analysis showed that doubling isoflavone intake increased plasma concentrations by 55–62% (±s.e. 1–2%, R2>0.87) for daidzein, genistein, equol (only for producers) and total isoflavones; for glycitein the association was weaker (15±1%, R2=0.48). Adjustments for energy, carbohydrate and fat intake did not affect these estimates. Inter-individual variation, estimated based on repeated measures in one of the studies, was 30–96%. Conclusions: Although the relation between isoflavone intake and plasma concentrations was adequately quantified, the use of isoflavone intake data for risk assessment needs caution due to large inter-individual variation in plasma concentrations.