Skip to main content
ARS Home » Research » Publications at this Location » Publication #65146

Title: NUTRIENT DENSITY ESTIMATES FROM AN AVERAGE OF FOOD FREQUENCY AND FOOD RECORDS CORRELATE WELL WITH SERUM CONCENTRATIONS OF VITAMINS E AND THE CAROTENOIDS IN FREE-LIVING ADULTS

Author
item Burri, Betty
item DIXON, ZISCA
item Neidlinger, Terry

Submitted to: International Journal of Food Sciences and Nutrition
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/1/1996
Publication Date: N/A
Citation: N/A

Interpretive Summary: Many reports have suggested that people who eat low amounts of vitamins A, E, and beta-carotene are at high risk for cancers, heart disease, and degenerative diseases. Unfortunately, finding out what people eat is very difficult. In fact, all the common tests for estimating what people eat have serious defects, and are especially poor at estimating vitamins A and E. We hypothesized that the average of two common tests (food frequency and food records) for estimating what people eat might give better results for vitamins A and E than either test alone. Our results from this study suggest that this is true, at least for vitamin E. Therefore, we recommend that scientists who are attempting to study the influence of diet on disease should consider collecting and combining information for both of these tests.

Technical Abstract: Dietary intakes are usually estimated by either a food frequency test, or by food records. We hypothesized that dietary intake estimates for fat soluble vitamins might be more accurate if information from both food frequency tests and food records were used. We estimated dietary intakes in 10 healthy adults by 4 food frequency questionnaires and ten-3 day food records collected over a year. Serum antioxidant nutrient concentrations (vitamins A, E, and the carotenoids) were measured by HPLC throughout the year. Only carotenoid intake appeared to change over the year. Estimates of nutrient densities correlated more often than estimates of nutrient intakes to serum nutrient concentrations. Nutrient density estimates derived from the average of food frequency and food record estimates were significantly correlated with serum nutrient concentrations more often then dietary estimates derived from either food frequency or food records alone. We suggest that nutrient density estimates derived from a combination of food frequency and food records may be useful for free-living studies.