BACKGROUNDHuman milk offers significant health benefits for infants; however, when not feasible, infant formula serves as an alternative. The higher protein content in infant formula is thought to contribute to the distinct metabolic profiles observed in formula-fed infants compared with those fed human milk.OBJECTIVESThis study investigates the impact of formula protein quantity and whey protein types on the serum, urine, and fecal metabolomes of infants.METHODSA secondary analysis was performed on a random subset of 200 well-characterized per-protocol infants who completed a prospective, randomized, double-blind controlled trial. Infants were randomly assigned to 1 of 3 groups: standard formula, protein-reduced formula with α-lactalbumin-enriched whey, or protein-reduced formula with casein glycomacropeptide-reduced whey, along with an observational reference group of exclusively breastfed infants. Serum, urine, and fecal metabolites were quantified using 1H nuclear magnetic resonance spectroscopy at baseline (1-2 mo), 4, and 6 mo of age. Dietary intake was assessed monthly ≤6 mo of age.RESULTSFormula protein content and type of whey protein used significantly influenced the amino acid profile and associated catabolic markers in serum and urine but had minimal impact on the fecal metabolome. Reduced protein formulas yielded metabolome profiles closer to those of breastfed infants compared with standard formula. Despite these improvements, infants fed human milk still demonstrated enhanced branched-chain amino acid (BCAA) oxidation and a greater capacity to eliminate catabolic waste products from BCAA metabolism over infants consuming protein-reduced formulas.CONCLUSIONSComprehensive metabolomics profiling of serum, urine, and feces captures molecular-level changes and informs potential strategies for formula optimization. Both the quantity and source of protein significantly influenced the metabolic profiles of formula-fed infants. However, modifications in protein alone cannot fully resolve the metabolic differences between formula-fed and breastfed infants, highlighting the complexity of mimicking the human milk feeding-associated metabolic profile. This study was registered at clinicaltrials.gov as NCT02410057.