Identification of Poly-Metabolite Scores for Diets High in Ultra-Processed Food in an Observational Study with Validation in a Randomized Controlled Crossover-Feeding Trial

Background: Ultra-processed food (UPF) accounts for a majority of calories consumed in the United States, but the impact on human health remains unclear. We aimed to identify poly-metabolite scores in blood and urine that are predictive of UPF intake. Methods and findings: IDATA participants (n=718), aged 50-74 years, with serially collected blood and urine and up to 6 24-hour dietary recalls (ASA-24s), collected over 12-months, were included in metabolomics analysis, which used ultra-high performance liquid chromatography with tandem mass spectrometry to measure >1000 serum and urine metabolites. Average daily UPF intake was estimated as percentage energy according to the Nova system. Partial Spearman correlations and Least Absolute Shrinkage and Selection Operator (LASSO) regression were used to estimate UPF-metabolite correlations and build poly-metabolite scores of UPF intake, respectively. Scores were tested in a previously conducted randomized, controlled, crossover-feeding trial of 20 domiciled participants who consumed ad libitum diets that were 80% and 0% energy from UPF for 2 weeks each. IDATA participants were 51% female, and 97% completed ³4 ASA-24s. Mean intake was 50% energy from UPF. UPF intake was correlated with 187 (of 952) serum and 284 (of 1111) 24-hour urine metabolites (FDR-corrected P-value ≤ 0.01), including lipid (n=53 serum, n=21 24-hour urine), amino acid (n=33, 59), carbohydrate (n=3, 8), xenobiotic (n=33, 69), cofactor and vitamin (n= 9, 11), peptide (n=7, 6), and nucleotide (n=6, 8) metabolites. Using LASSO regression, 28 serum and 33 24-hour urine metabolites were selected as predictors of UPF intake; biospecimen-specific scores were calculated as a linear combination of selected metabolites. Overlapping metabolites included (S)C(S)S-S-methyl cysteine sulfoxide (rs= -0.19, -0.23), N2-N5-diacetylornithine (rs= -0.26, -0.27), pentoic acid (rs= -0.28, -0.31), and N6-carboxymethyllysine (rs=0.15, 0.21). Within the cross-over feeding trial, the poly-metabolite scores differed, within individual, between UPF diet phases (P-value for paired t-test <0.001). Conclusions: Poly-metabolite scores, developed in IDATA participants with varying diets, are predictive of UPF intake and could advance epidemiological research on UPF and health.