Soy food frequency questionnaire does not correlate with baseline isoflavone levels in patients with bladder cancer

Background: The isoflavone genistein, a natural soy product with receptor tyrosine kinase-inhibiting activity, as well as phytoestrogenic and other potential anticarcinogenic effects, is being studied as an anticancer agent. Since isoflavones are commonly consumed in food products containing soy proteins, a method to control for baseline isoflavone consumption is needed. Methods: HPLC was used to evaluate baseline plasma and urine concentrations of isoflavone in fifty-four participants with bladder cancer enrolled on a phase II chemoprevention study of G-2535. The soy food frequency questionnaire was used to assess participant's baseline soy intake. The association between baseline isoflavone concentrations and intakes for genistein and daidzein was assessed by the Spearman's rank correlation coefficient. Results: The majority of participants had no detectable genistein or daidzein in plasma at baseline. The median and range of values were 0 (0-1480) nmol/L for genistein, and 0 (0-1260) nmol/L for daidzein. In urine, the median and range of values were 91.0 (0-9030) nmol/L for genistein and 623 (0-100,000) nmol/L for daidzein. The median and range of weekly estimated genistein intake was 0 (0-236) mg/wk; the median and range of weekly estimated daidzein intake was 0 (0-114) mg/wk. There was no relationship to soy intake as measured by the food frequency questionnaire and baseline isoflavone levels in plasma or urine and the Spearman's rank correlation coefficients were not significant. Conclusion: The soy food frequency questionnaire did not correlate with plasma or urine concentrations of either isoflavone. Impact: Alternative methods for controlling for soy consumption, including measuring plasma and urine concentrations, in isoflavone chemoprevention trials should be considered.