Carotenoid intakes, assessed by dietary questionnaire, are associated withplasma carotenoid concentrations in an elderly population

High intakes of fruits and vegetables and of carotenoids are associated wit h a lower risk for a variety of chronic diseases. It is therefore important to test the validity of dietary questionnaires that assess these intakes. We compared intakes of five carotenoids, as calculated from responses to th e Willett 126-item food-frequency questionnaire, with corresponding biochem ical measures. Subjects included 346 women and 201 men, aged 67-93 y, in th e Framingham Heart Study. Unadjusted correlations were higher among women t han men as follows: alpha-carotene 0.33 and 0.18, beta-carotene, 0.36 and 0 .25; beta-cryptoxanthin, 0.44 and 0.32; lycopene, 0.35 and 0.21; and lutein + zeaxanthin, 0.27 and 0.10, respectively. Adjustment for age, energy inta ke, body mass index (BMI, kg/m(2)), plasma cholesterol concentrations and s moking reduced the gender differences, respectively, to the following: alph a-carotene 0.30 and 0.28; beta-carotene, 0.34 and 0.31; beta-cryptoxanthin, 0.45 and 0.36; lycopene, 0.38 and 0.31; and lutein + zeaxanthin, 0.24 and 0.14. Plots of adjusted mean plasma carotenoid concentration by quintile of respective carotenoid intake show apparent greater responsiveness among wo men, compared with men, to dietary intake of alpha- and beta-carotene and b eta-cryptoxanthin, but similar blood-diet relationships for lycopene and lu tein + zeaxanthin. Reported daily intake of fruits and vegetables correlate d most strongly with plasma beta-cryptoxanthin and beta-carotene among wome n and with plasma alpha- and beta-carotene among men. With the exception of lutein + zeaxanthin, this dietary questionnaire does provide reasonable ra nkings of carotenoid status among elderly subjects, with the strongest corr elations for beta-cryptoxanthin, Appropriate adjustment of confounders is n ecessary to clarify these associations among men.