Association of three different dietary patterns with the metabolomic profiles of children: the BiomarKid project

Abstract

Abstract Background and aims Metabolomics offers a promising approach to objectively assess food intake and nutrition-related health risks. Three different dietary patterns (DPs) associated with cardiometabolic health were previously identified at 2 years of age in the EU Childhood Obesity Project and were used as reference patterns for longitudinal scoring throughout childhood: “Core foods” DP (CORE), rich in vegetables, fruits, fish, meat, and olive oil; “Animal protein sources” DP (PROT), high in white meat, processed meat, flavoured milk, snacks, fish, and vegetables; and “Poor-quality fats and sugars” DP (F&S), high in soft cheese, butter, added sugars, and juices. This study aimed to identify metabolomic signatures of these DPs. Methods Targeted liquid chromatography-tandem mass spectrometry quantified 236 metabolites in blood samples collected at ages 5.5 and 8 years from 421 children, with 156 children contributing samples at both timepoints. Associations between DPs and metabolite levels were assessed using adjusted linear mixed models. Metabolite set enrichment analysis (MSEA) was conducted across 23 distinct sets of related metabolites. Results Forty-eight metabolites associated with DPs, although they did not remain significant after false discovery rate correction. Five showed associations with multiple DPs. For example, LPC a C22:6 showed associations with opposite effect directions for CORE and F&S (β = 0.108, p = 0.038, q = 0.539; β=− 0.140, p = 0.025, q = 0.484, respectively). The ratio PC aa C40:5/PC aa C40:6 was negatively associated with CORE (β=− 0.207, p = 3.81 × 10 − 5 , q = 0.016) and positively with F&S (β = 0.136, p = 0.020, q = 0.463). CORE and F&S showed opposite directions in seven of the eight metabolite groups significant in the MSEA. A group of phosphatidylcholines containing polyunsaturated fatty acids (including DHA) was positively associated with PROT (NES: 1.837, p = 0.001, q = 0.017). Conclusions Distinct metabolite groups were associated with adherence to specific dietary patterns in children. Consistent findings across metabolite classes and dietary components suggest biologically plausible metabolomic signatures that require confirmation in independent cohorts.