Metabolic phenotypes in sepsis: A conceptual framework for phenotype-informed nutrition

Sepsis is a biologically heterogeneous syndrome characterized by profound alterations in immune function and intermediary metabolism that challenge standardized nutritional management. Current clinical nutrition guidelines recommend early enteral nutrition after hemodynamic stabilization, progressive caloric advancement, and avoidance of early full feeding, but they do not incorporate metabolic phenotyping. Emerging evidence from immunometabolism, metabolomics, and systems biology shows recurring, partially overlapping metabolic patterns in sepsis, including hyperglycolytic-hyperlactatemic, mitochondrial-bioenergetic dysfunction, hyperinflammatory Warburg-like, immunoparalytic-tolerant, lipidomic-acylcarnitine, gut-microbiota, and hypercatabolic-muscle wasting domains. These patterns share mechanisms and may coexist or evolve over time; their biological basis should be distinguished from their still unvalidated use as mutually exclusive bedside diagnostic classes. This narrative review examines these biological domains and their nutritional relevance while explicitly separating established, phase-based critical-care nutrition from phenotype-informed treatment hypotheses. In particular, sepsis-associated mitochondrial abnormalities may be tissue- and cell-specific and do not imply generalized suppression of whole-body oxygen consumption. Likewise, hyperglycemia, hyperlactatemia, or increased carbon dioxide production are not validated markers of overfeeding. Residual endogenous substrate production persists during acute critical illness but cannot currently be quantified as total energy flux at the bedside, limiting precise assessment of the combined endogenous and exogenous substrate load. No prospective randomized trial has yet tested nutrition prescribed according to the metabolic domains proposed here. The seven-domain framework therefore organizes recurrent components of sepsis biology; what remains to be validated is its operational use for patient classification and for predicting differential response to nutritional interventions.