Comparison of a 21-item food questionnaire with a 7-day dietary registration andbiomarkers of fat intake in a Swedish cohort of 60-year-old adults.

• AbstractPurpose. To evaluate a 21-item food frequency questionnaire (FFQ) in men participating in alarge cohort of Swedish 60-year-old adults (60YO).Methods. A self-reported FFQ (including 21 qualitative and semi-quantitative questions) was completed by >2000 men as part of a detailed baseline examination. A subsample of 301 men was included in a subsequent study in which detailed dietary habits were determined by a 7-day food record. Spearman rank correlations (r) and weighted Kappa (Kw) statistics were used to compare food intake categories in FFQ and 7-day food record. Furthermore, fatty acid (FA) composition in serum cholesteryl esters, assessed concurrent with the FFQ completion, was used to evaluate intakes of specific fat-rich foods from the FFQ.Results. We found good agreement between FFQ and food records for the reported intake ofalcohol (r=0.72, Kw=0.51), margarine (r=0.60, Kw=0.33), and fruit (r=0.49, Kw=0.31), reasonably good agreement for total fish (r=0.25, Kw=0.23), and egg (r=0.35, Kw=0.28), but poor agreement for other food groups such as bread, cheese, vegetables and cookies. In addition, serum proportions of long-chain n-3 polyunsaturated FAs and pentadecanoic acid were significantly higher in men with greater intakes of fish and cheese, respectively.Conclusion. In this evaluation of a short FFQ against 7-d food records and serum biomarkers of fat intake we found that the FFQ reasonably well reflected the intake of certain food groups (e.g. alcohol, fish, and margarine), but performed less accurate for other food groups. Firm and overallconclusions on validity are confined by the time-lag between the test and the reference method.

Nyckelord

validation

food frequency questionnaire

biomarkers of fat intake

Kostvetenskap

Food, Nutrition and Dietetics

Publikations- och innehållstyp

vet (ämneskategori)

ovr (ämneskategori)