Validation of a soy food frequency questionnaire with plasma concentrations of isoflavones in US adults.

OBJECTIVE: To validate assessment of soy intake using food frequency questionnaires (FFQs) compared with plasma isoflavone (genistein and daidzein) concentrations.
DESIGN: Cross-sectional analysis of soy isoflavone intake and plasma analysis of isoflavones.
SUBJECTS: 77 men and women, age range 20 to 40 years, recruited from the Seattle metropolitan area. MAIN OUTCOME MEASURES: Isoflavone intake was determined from responses to a 40-item soy FFQ and from tofu and soymilk intake assessed as part of a comprehensive FFQ used for the Women's Health Initiative (WHI FFQ). Isoflavone concentrations in fasting blood samples were determined by liquid chromatography-mass spectrometry. STATISTICAL ANALYSES: Correlation coefficients were calculated for: a) isoflavone intake assessed by the soy FFQ and the WHI FFQ, b) intake assessed by the soy FFQ and plasma isoflavone concentrations, and c) intake assessed by the WHI FFQ and plasma isoflavone concentrations.
RESULTS: Isoflavone intake was highly correlated between the soy FFQ and the WHI FFQ (r = 0.84). Genistein and daidzein intakes determined by the soy FFQ were significantly correlated with plasma concentrations (r = 0.53 and 0.45, respectively). Isoflavone intake assessed from the WHI FFQ was also correlated with plasma concentration (r = 0.46 and 0.45). Soymilk and tofu were the two major contributors to isoflavone intake (38.6%).
CONCLUSIONS: A soy-specific, 40-item FFQ assessed isoflavone intake with good validity. Isoflavone intake assessed by the WHI FFQ (tofu and soymilk) had lower correlations with plasma concentrations compared with the soy FFQ. Nonetheless, assessment of the two foods is a reasonably good marker for soy food consumption in this sample.