Validation of an antioxidant nutrient questionnaire in whites and African Americans.

High antioxidant intakes are inversely related to risk for many diseases. However, there is no comprehensive instrument that captures consumption of antioxidant nutrients from both foods and dietary supplements. This report examines the validity of a newly developed questionnaire assessing self-reported dietary and supplemental intakes of antioxidant nutrients (carotenoids, vitamin C, and vitamin E). Between March and December 2005, participants (n=164), 20 to 45 years old, completed the new 92-item antioxidant nutrient questionnaire, a demographic/health questionnaire, four 24-hour dietary recalls, a dietary supplement inventory, and provided semi-fasting blood samples that were analyzed for plasma antioxidant levels. Data analyses included descriptive statistics, correlation coefficients, and linear regression. Mean age of participants was 31.9 years, 51% were African American, and 52% were female. Median antioxidant intakes from the questionnaire and mean of the four recalls were generally comparable. Adjusted Pearson's correlations of questionnaire- and recall-derived intakes ranged from r=0.06 to r=0.56; correlations for the questionnaire and biomarkers ranged from r=0.10 to r=0.33. Agreement rates for classification of intakes from the questionnaire and recalls into the same/adjacent quartiles were 65% to 89%; misclassification to the opposite quartile was rare (range=0 to 12%). For most nutrients, there were linear trends of increasing plasma concentrations with higher questionnaire-derived intakes (P<0.01). Correlations of supplement use between the questionnaire and a supplement inventory were r=0.33 to r=0.84. The new antioxidant nutrient questionnaire demonstrated good validity for collecting self-reported antioxidant nutrient intakes from foods and supplements in both whites and African Americans. The study also underscores the importance of examining the performance characteristics of dietary assessment instruments separately in different population subgroups.