Text

Objectives and questions: Trace elements modulate secondary injury cascades following traumatic spinal cord injury (TSCI). Selenium is a central regulator of redox homeostasis and immune signaling; however, its translational behavior between animal models and humans remains poorly defined. We aimed to model cross-species serum selenium dynamics after TSCI and compare them with spinal column injuries without neurological deficits.

Material and methods: Serum selenium concentrations were analyzed in a rat thoracic compression–contusion spinal cord injury model (n=120; 80 SCI, 40 controls; all female) and a prospective human cohort (n=59; 49 TSCI, 10 controls; 17 female; median age 50 years, IQR 32–64). Sampling followed a harmonized cross-species protocol at 1, 3, 7, 14, and 28 days post injury (dpi). Values were z-standardized and compared to controls. Linear regression modeled temporal trajectories per species and jointly. Cross-species transferability was evaluated using RMSE, MAE, and R2.

Results: Both species showed significant selenium depression at 1 dpi compared to controls (animal: p < 0.001; human: p = 0.016). Temporal modeling identified a significant time effect without improvement from injury subclassification. The joint model demonstrated a species-dependent intercept shift but conserved temporal structure. Cross-prediction revealed asymmetric transfer: the animal model moderately predicted human data (R2 = 0.24, RMSE=0.90), whereas the human model poorly predicted animal dynamics (R2 = 0.04, RMSE=1.02). Residual diagnostics confirmed stable model assumptions.

Discussion and conclusions: Selenium exhibits conserved early injury dynamics across species, supporting translational relevance of experimental SCI models and underscoring its link to trauma immunology. However, asymmetric prediction indicates additional human-specific variability. Cross-species modeling provides a quantitative framework for evaluating biomarker translation in TSCI.