Validation of a food frequency questionnaire for use in a Scottish population: correlation of antioxidant vitamin intakes with biochemical measures

Background: The food frequency questionnaire (FFQ) has become established as the primary method for assessing dietary intake in epidemiological studies of diet and disease; attention must now be given to developing and evaluating FFQs that extend the range of populations studied. Aim: To assess the validity of antioxidant vitamin intake estimates derived from a self-administered food frequency questionnaire which was developed for use in Scottish populations. Method: Intake estimates of vitamins A, C, E and beta-carotene were compared with corresponding biochemical measurements in 273 Scottish men and women aged 39-45 years. Results: The mean ratios of energy intake to calculated basal metabolic rate were 1.40 for men and 1.46 for women suggesting that the intake values derived from the FFQ were not biased towards under-reporting. Significant positive correlation coefficients ranging from 0.14 (for vitamin A) to 0.42 (for vitamin C) were observed between total intake (including diet and supplements) and plasma values. Adjustment for other known determinants of diet and plasma values (i.e. energy intake, body mass index, gender, smoking and plasma lipid levels) improved the correlations for vitamin E and p-carotene. Excluding supplement users generally reduced the correlations suggesting that failure to include supplemental sources may result in misclassification of antioxidant vitamin intake. The food frequency questionnaire assigned 68-89% of subjects correctly into the upper or lower plus adjacent tertiles of plasma vitamin concentration. Conclusions: These results demonstrate that the food frequency questionnaire is a valid measure of antioxidant vitamin intake.