Validation of an abbreviated FFQ for assessing DHA and EPA intake in pregnant women.

Abstract

Background: Docosahexaenoic acid (DHA) is an omega-3 fatty acid found in variable amounts in the diet and is often accompanied by another omega-3 fatty acid, eicosapentaenoic acid (EPA) when the food chosen is seafood. While both nutrients support optimal health, DHA is of particular interest for pregnant women because of it's hypothesized effect on infant visual and cognitive development. Results from randomized clinical trials (RCTs) of DHA supplementation in pregnant women, however, are mixed, and because dietary DHA intake is variable. A valid baseline measure of intake is imperative if trials are to be properly compared. The National Cancer Institute Diet History Questionnaire-II (DHQ-II) can estimate DHA and EPA dietary intake but it requires answers to 150 questions. An abbreviated 7-question food frequency questionnaire (FFQ) measures DHA and EPA intake more efficiently, however, neither the DHQ-II nor the abbreviated FFQ have been validated for DHA and EPA intake of pregnant women. Our goal was to determine the validity of each method for assessing DHA and EPA intake by comparing reported dietary intakes to red blood cell (RBC) phospholipid (PL) DHA and EPA, a standard measure of nutrient status. Methods: 309 pregnant women between 12 and 20 weeks gestational age who were enrolled in an on-going RCT completed both the DHQ-II and abbreviated FFQ as well as a venipuncture blood draw to assess RBC PL DHA and EPA. Pearson correlations between measures were determined. We additionally examined other potential covariables the might influence baseline DHA and EPA status. Results: The abbreviated FFQ was moderately correlated with both DHA and EPA status (r=0.49 and r=0.56, respectively) while the DHQ-II was poorly correlated with these two nutrients (r=0.24 and 0.30, respectively). Parity and paternal education (collinear with income, non-Black race) were also significantly correlated with DHA and EPA status. Together with dieraty intake estimated by the abbreviated FFQ these three variables accounted for about 37% of the variance in DHA status and 44% of the variance in EPA status (r2=0.365 and r2=0.44, respectively). Conclusion: The abbreviated 7-question FFQ better represents DHA and EPA intake of pregnant women compared to the longer DHQ-II. Importantly this finding supports use of the abbreviated measure to increase efficiency in clinical trials and reduce participant burden. It is also important to consider variables other than dietary intake when predicting DHA and EPA status.