A national sampling plan was developed to select the most widely used isoflavone-containing foods in the United States. Foods were selected based on their retail volume and sampled in five geographical areas representing seven metropolitan areas. Isoflavones were analyzed from composite samples, raw and cooked, and reported by brand. Quality control measures were evaluated throughout the study. Isoflavone levels ranged from 1 microg/g in soy sauces to 540 microg/g in tempeh. Soymilk and tofu represented the major portion of soy foods evaluated. These data will appear in the electronic version of USDA Handbook No. 8 of Food Composition Data in 1999.

Development of a database of the soy isoflavone content of foods requires accurate and precise evaluation of different food matrixes. To evaluate accuracy, we estimated recoveries of both internal and external standards in 5 different soyfoods weekly. Standards were evaluated daily for system quality assurance. To evaluate sample precision, we analyzed soybeans and soymilk bimonthly for within-day precision and over 4 d for day-to-day precision. CVs should be < or = 8%. We validated our methods for single and multiple recovery concentrations by using our new internal standard, 2,4,4'-trihydroxydeoxybenzoin, and the external standards daidzein, genistein, and genistin. Concentrations of 12 isoflavone isomers, 3 aglycones (daidzein, genistein, and glycitein), and 9 glucosides (daidzin, genistin, glycitin, acetyldaidzin, acetylgenistin, acetylglycitin, malonyldaidzin, malonylgenistin, and malonylglycitin) were measured in a variety of soybeans and soyfoods. The extraction methods used depended on soyfood type. The HPLC conditions for soy isoflavone analysis were improved, leading to good separation with a short analysis time (60 min/sample). A data bank of concentration and distribution of isoflavones in different soybean products was assembled. A wide range of isoflavone concentrations, from < 50 microg/g to > 20,000 microg/g, was found in different soy products. The glucoside forms are almost twice the molecular weight of the aglycones; reported isoflavone concentrations should be normalized to the aglycone mass (or an isoflavonoid equivalent) rather than a simple sum of all isomers.

Six of the major soy-based infant formulas marketed in the United States were assayed for their isoflavone levels. Samples were taken from the east coast, midwestern, and west coast regions of the United States. Isoflavone levels were variable across brands probably due to different amounts of soy isolate used in product formulation. Total isoflavones ranged from 214 to 285 μg/g of dry formula or ≈25−30 μg/mL of reconstituted formula. Keywords: Phytoestrogens; genistein; glycitein; daidzein