A valid food frequency questionnaire for measuring dietary fish intake

Abstract

There is considerable interest in the potentially protective effects of high fish consumption on many chronic diseases. Many epidemiological studies use food frequency questionnaires (FFQ) to quantify usual dietary fish intake, so it is important to validate this assessment against objective markers. The objective of this study was to determine the relationship between plasma percentage fatty acids and dietary fish intake as assessed by a FFQ. A semiquantitative FFQ was completed by 174 adults from the community (aged 26-49 years) who also had venous blood analysed for plasma percentage fatty acids. Following linear regression modelling, total non-fried fish intake was a significant predictor of n-3 (regression coefficient, B = 0.94; 95% CI = 0.60-1.28), docosahexaenoic acid (DHA; B = 0.73; 95% CI = 0.47-0.99) and the ratio of n-6: n-3 fatty acids (B = -1.0; 95% CI = - 1.35- -0.65). Steamed, grilled or baked fish was a small but significant predictor of eicosapentaenoic acid (EPA) levels (B = 0.13; 95% CI = 0.05-0.21) while total fish intake was a predictor of n-6 fatty acids (B = -0.88; 95% CI = -1.41- -0.36). This semiquantitative FFQ could be useful for ranking subjects according to their likely plasma n-3, DHA, and n-6 fatty acid intake and the ratio of n-6: n-3 fatty acids, when the available resources may simply not permit biological markers to be used.