Background: In order to assess intake of antioxidants and retinol, we evaluated the validity and reproducibility of a semiquantitative food frequency questionnaire.
Methods: Data were collected from 110 women, aged 15-54 years, who resided in the southern part of Mexico City. Study participants were asked to complete the food frequency questionnaire before and after completing four series of 4 days of 24-h diet recalls. We obtained the 24-h recalls at 3-month intervals designed to account for seasonal variability. In addition, we measured serum blood levels of retinol and various carotenoids at 3 and 9 months after baseline.
Results: Intraclass correlation coefficients for intake of energy-adjusted daily nutrients, assessed by the questionnaires 1 year apart, ranged from 0.23 for lutein/zeaxanthin to 0.50 for total carotenes. Correlation coefficients between energy-adjusted nutrient intakes were measured by means of the sixteen 24-h dietary recalls, and the second questionnaire ranged from 0.09 for lutein/zeaxanthin to 0.45 for total carotenes. These correlations were higher after adjusting for day-to-day variations in diets ranging from 0.11 for lutein/zeaxanthin to 0.49 for vitamin C. Correlation coefficients between intake and biochemical measures were 0.32 for alpha-carotene, 0.34 for beta-carotene, and 0.40 for total carotene after adjusting for age, body mass index, cholesterol and triglyceride plasma levels, smoking status, and caloric intake.
Conclusions: These data support the ability of our food frequency questionnaire to provide valid information on the relative intake of these nutrients and to confirm its usefulness in epidemiological studies of diet-disease relations. However, its application outside Mexico City or in different age or gender populations will require additional modifications and validation efforts.