Dietary antioxidants such as carotenoids and vitamin E have been hypothesized to reduce the risk of some chronic diseases such as atherosclerosis, cancer, and cataracts. They could also act in modulating the effects of some environmental toxins mediated by means of free radicals. It is, therefore, important to develop a valid method for assessing the intake of antioxidants to estimate their potential beneficial effect on health.

Food-frequency questionnaires (FFQ) have been widely used in studies on the relation of food consumption to the risk of chronic diseases, because such questionnaires allow for the evaluation of individual dietary intake at low cost, including a large population. Food-frequency questionnaires are designed to measure average long-term diet rather than current short-term dietary intakes. The information collected with FFQ allows individuals to be ranked on the basis of levels of past nutrient intake. However, due to the large variability of dietary patterns and food availability among populations, the food items included in the FFQ need to be carefully selected to reflect the usual food consumption of the population under study.

In most studies on the validity of dietary assessments, the reference measurements, or “gold standards,” have been based on more detailed dietary assessment methods, such as food records or series of 24-h recalls. Because different methods of dietary assessments may have correlated measurement errors, an alternative validation strategy is to use biochemical measurements such as plasma concentrations of nutrients. The primary advantage of this form of validation is that it is based on independent measurements; therefore, the sources of error in the two methods should be uncorrelated. Plasma carotenoids and vitamin E levels vary with intake and consequently may be used as indicators of intake.

In this article, we present the results of a study whose purpose was to assess the validity and reproducibility of a semi-quantitative food-frequency questionnaire designed to assess dietary intake of carotenoids, tocopherol, retinol, and vitamin C among female residents of Mexico City.

The FFQ was developed using the methodology proposed by Willett et al. We identified food items by tabulating the results of a dietary survey carried out in 1983 by the National Institute of Nutrition in Mexico City from a random sample of 240 low- to medium-income families in Mexico City. Data collected in this survey included both 24-h recalls and home visits to weigh and measure the food items actually consumed. Using stepwise linear regression, we identified the foods that were the best predictors of each nutrient of interest. We then compiled all the foods identified using this methodology. In a second step, we identified important foods that may not have been reported by the population who participated in the survey administered by the National Institute of Nutrition. For this purpose, a group of Mexican dietitians and nutritionists were invited to participate, and were asked to identify relevant food items that may not have been identified before with the stepwise linear regression approach. A final format that included 85 food items was then pilot-tested using a convenient sample of women from medium- to low-socioeconomic status living in Mexico City. The final instrument was composed of a matrix listing 116 food items and 10 frequencies of consumption. The consumption frequencies were as follows: 6 or more times per day; 4–5 times per day; 2–3 times per day; once a day; 5–6 times per week; 2–4 times per week; once a week; 2–3 times per month; once a month or less, and never.

For each food item, we compiled the nutrient content per average unit (specified serving size: slice, glass, or natural unit). The nutritional composition of each food included in the questionnaire was obtained from the U.S. Department of Agriculture (USDA) food composition tables and, when necessary, complemented by the nutrient database developed by the National Institute of Nutrition in Mexico. We estimated daily dietary nutrient intakes for study participants by calculating a nutrient score for each food, using the food nutrient content taken from the food table adjusted to the specified portion size of the questionnaire. To estimate individual intakes, we then multiplied this score by the weight corresponding to the reported frequency of use. The weights used were the following: 6 for reported frequencies of six or more times per day; 4.5 for four to five times per day; 2.5 for two to three times per day; 1 for once a day; 0.8 for five to six times per week; 0.43 for two to four times per week; 0.08 for two to three times per month, and 0.016 for once a month or less. We then added the product of frequencies multiplied by the nutrient scores, producing scores for overall food items, in order to obtain a total nutrient score.

The number of servings of fruits and vegetables per day was computed as the sum of the frequency of the 18 fruits and 19 vegetables specified in the questionnaire (the fruits and vegetables included are as follows: bananas; oranges; orange juice; cantaloupe; apples; watermelons; pineapples; papaya; pears; mangoes; mandarin oranges; strawberries; peaches; grapes; prickly pears; plums; mammey apples, and zapote fruit, while vegetables and legumes included tomato paste; raw tomatoes; beans; potatoes; carrots; lettuce; corn; cactus leaves (paddles); spinach; zucchini; avocados; zucchini (squash) blossoms; cauliflower; string beans; peas; green or yellow lima beans; raw hot peppers; canned hot peppers, and dry hot peppers) Vitamin A and E intakes were assessed using information from various food groups of the questionnaire and included the consumption of oils, dairy products, main dishes (beef, liver, fish, chicken), eggs, breads, and cereals. Portion size was estimated by means of “natural” units (for example, one orange) or standard quantities (for example, a glass of milk) wherever possible. No additional question on portion size was added to the questionnaire. To account for the seasonal availability of some foods, we included a section in the questionnaire concerning the frequency of consumption of these foods during the season (for example, for mangoes and mammey apples). This was also taken into consideration in the estimation of nutrient densities with the introduction of a weight factor.

The data from the sixteen 24-h dietary recalls were coded in a specifically elaborated data entry program. We calculated the specific nutrient content of the reported portion size and estimated a nutrient intake for each day.

We randomly selected an age-stratified sample of 154 women, aged 15–54 years, who resided in the southern part of Mexico City. Among these, 110 women agreed to participate in the study and provided blood samples. All participants signed an informed consent. Over a 1-year period, we obtained four 24-h recalls every 3 months (corresponding to 16 days of diet recalls) in batches of 4 days, representing all seasons in Mexico. Interviewers visited each participating woman at her home and recorded the type and quantity of food (prepared using local cooking utensils) she had consumed during the previous 24 h. We arranged the visits of each subject on different days of the week (including Saturdays and Sundays) to take into account differences in day-to-day dietary habits. At baseline, a trained dietitian administered the FFQ. At 3 and 9 months after the beginning of the study, blood samples were collected and after 1 year we administered a second FFQ.