Directional immuno–metabolic signatures following severe trauma are associated with organ-specific clinical endpoints in a prospective cohort

Background Severe trauma elicits evolving immune-metabolic responses that may differ across clinical endpoints. We characterized time-resolved innate and adaptive immune patterns associated with higher or lower endpoint odds during the first 96 h after injury. Methods Of 62 patients prospectively enrolled in the observational PRIME study (ISS >15), 52 met the prespecified analytical completeness criterion. Clinical and immune biomarkers were assessed at emergency room admission (ER), 24, 48, and 96 h. Time-point-specific quartile deviations were evaluated using logistic regression against microbiological culture positivity during hospitalization, creatinine-defined renal impairment, Berlin-defined ARDS of any severity, hepatic enzyme elevation, and four hemodynamic endpoints. Complementary sensitivity analyses and bootstrap resampling assessed the exploratory associations and their internal stability. Results Microbiological culture positivity was associated with lower admission base excess and temperature, reduced CD4/CD8 ratios and CD4 + T-cell proportions at 24–48 h, and lower monocyte measures with higher IL-8 at 96 h. The 96-h monocyte CBC/IL-8 combination showed an apparent within-cohort AUC of 0.78. Creatinine-defined renal impairment was associated with lower admission pH and base excess and a higher-IL-8/lower-CD4/CD8 combination (apparent AUC 0.88), alongside later myeloid and lymphocyte-related alterations. Berlin-defined ARDS showed predominantly low-odds associations, including higher early pH and phagocytic capacity; its 96-h combination of lower phagocyte CD64 expression and lower lymphocyte percentage yielded an apparent AUC of 0.80. Hepatic enzyme elevation showed low-odds associations involving higher 96-h monocyte HLA-DR expression and lower CD64 + phagocyte counts (apparent AUC 0.78). Hemodynamic endpoints showed distinct, endpoint-specific high-odds configurations. Bootstrap selection reproducibility was limited for the retained multivariable combinations. Conclusion The evaluated clinical endpoints were associated with distinct directional immune-metabolic configurations rather than a uniform inflammatory response. Their hypothesis-generating contribution lies in the endpoint-specific organization of established biological components across serial sampling points. These exploratory candidate signatures require independent validation before their potential clinical relevance can be established.