Validation of fatty acid intakes estimated by a food frequency questionnaire using erythrocyte fatty acid profiling in the Montreal Heart Institute Biobank.

Abstract

To improve the prevention, treatment and risk prediction of cardiovascular diseases, genetic markers and gene-diet interactions are currently being investigated. The Montreal Heart Institute (MHI) Biobank is suitable for such studies because of its large sample size (currently, n = 17 000), the availability of biospecimens, and the collection of data on dietary intakes of saturated (SFAs) and n-3 and n-6 polyunsaturated (PUFAs) fatty acids estimated from a 14-item food frequency questionnaire (FFQ). We tested the validity of the FFQ by correlating dietary intakes of these fatty acids with their red blood cell (RBC) content in MHI Biobank participants. Seventy-five men and 75 women were selected from the Biobank. We successfully obtained RBC fatty acids for 142 subjects using gas chromatography coupled to mass spectrometry. Spearman correlation coefficients were used to test whether SFA scores and daily intakes (g day(-1)) of n-3 and n-6 PUFAs correlate with their RBC content. Based on covariate-adjusted analyses, intakes of n-3 PUFAs from vegetable sources were significantly correlated with RBC α-linolenic acid levels (ρ = 0.23, P = 0.007), whereas n-3 PUFA intakes from marine sources correlated significantly with RBC eicosapentaenoic acid (ρ = 0.29, P = 0.0008) and docosahexaenoic acid (ρ = 0.41, P = 9.2 × 10(-7)) levels. Intakes of n-6 PUFAs from vegetable sources correlated with RBC linoleic acid (ρ = 0.18, P = 0.04). SFA scores were not correlated with RBC total SFAs. The MHI Biobank 14-item FFQ can appropriately estimate daily intakes of n-3 PUFAs from vegetable and marine sources, as well as vegetable n-6 PUFAs, which enables the possibility of using these data in future studies.