Minimal hepatic encephalopathy (MHE) reduces quality of life of cirrhotic patients, predicts overt hepatic encephalopathy, and is associated with poor prognosis. We hypothesized that MHE arises once metabolic alterations derived from the liver reach a particular threshold. Our aim was to assess whether metabolic profiling of serum samples by high-field (1)H-nuclear magnetic resonance spectroscopy ((1)H NMR) and subsequent multivariate analyses would be useful to characterize metabolic perturbations associated with MHE and to identify potential metabolic biomarkers. Metabolic serum profiles from controls (n = 69) and cirrhotic patients without MHE (n = 62) and with MHE (n = 39) were acquired using high field NMR. Supervised modeling of the data provided perfect discrimination between healthy controls and cirrhotic patients and allowed the generation of a predictive model displaying strong discrimination between patients with and without MHE (R(2)Y = 0.68, Q(2)Y = 0.63). MHE patients displayed increased serum concentrations of glucose, lactate, methionine, TMAO, and glycerol, as well as decreased levels of choline, branch amino acids, alanine, glycine, acetoacetate, NAC, and lipid moieties. Serum metabonomics by (1)H NMR offers a useful approach for characterizing underlying metabolic differences between patients with and without MHE. This procedure shows great potential as a diagnostic tool of MHE as it objectively reflects measurable biochemical differences between the patient groups and may facilitate monitoring of both disease progression and effects of therapeutic treatments.