Early enteral nutrition and dynamic metabolic-inflammatory trajectories in sepsis: a retrospective cohort study

Background Early enteral nutrition (EEN) is an important part of sepsis management, but its physiological effects are not fully understood. This study examined whether EEN influences the time course of metabolic and inflammatory biomarkers in patients with sepsis. Methods We performed a retrospective cohort study of 3,354 adult ICU patients with sepsis admitted to West China Hospital from 2011 to 2025. Group-based trajectory modeling was used to characterize longitudinal patterns of albumin, lactate, and procalcitonin. Associations between EEN and trajectory membership were assessed using multinomial logistic regression. The relationship between trajectory groups and 28-day mortality was further evaluated. Results Distinct trajectory groups were identified for each biomarker, reflecting heterogeneous nutritional, metabolic, and inflammatory responses. EEN was associated with more favorable albumin trajectories and lower odds of belonging to elevated lactate and procalcitonin patterns. In the multivariate joint trajectory model integrating all three biomarkers, three classes emerged: a stable–low inflammation pattern (67.5%), an intermediate–transient pattern (21.7%), and a high-risk inflammatory surge pattern (10.8%). EEN was independently associated with reduced likelihood of assignment to the intermediate (OR 0.66; 95% CI, 0.52–0.84) and high-risk (OR 0.57; 95% CI, 0.38–0.88) classes. The high-risk trajectory group showed significantly increased 28-day mortality. Conclusion Initiation of EEN was linked to a higher probability of remaining in low-risk albumin–lactate–PCT trajectories and a lower probability of entering the high-risk inflammatory surge pattern. This pattern-level shift may partly explain the observed reduction in short-term mortality associated with EEN.