Multidimensional immunometabolic changes and energy expenditure during intensive neurorestorative rehabilitation in chronic motor-complete thoracolumbar spinal cord injury: a quasi-experimental study.

STUDY DESIGN: Single-group prospective quasi-experimental pre-post study. OBJECTIVES: To evaluate metabolic, structural, immunological, and functional changes in individuals with motor-complete chronic thoracolumbar spinal cord injury (SCI) undergoing intensive neurorestorative rehabilitation. SETTING: International Center for Neurological Restoration (CIREN), Havana, Cuba. METHODS: Forty-one males (20-39 years) with AIS A/B thoracolumbar SCI completed a 12-week rehabilitation program (≈ 33.5 h/week). Dietary intake, estimated resting metabolic rate (RMR), estimated total and rehabilitation-related energy expenditure (time-motion methodology), anthropometry-based muscle mass, amino acid profile, lymphocyte subset percentages, and functional indices (Barthel and Katz) were assessed at baseline and post-intervention. Longitudinal changes were analyzed using covariate-adjusted linear mixed-effects models based on a Directed Acyclic Graph framework. Principal Component Analysis (PCA) was used to explore multidimensional adaptation. RESULTS: Estimated rehabilitation-related energy expenditure (β = 181.293, p < 0.001) and estimated RMR (β = 50.78, p = 0.042) increased, whereas dietary energy intake relative to estimated muscle mass decreased (β = -19.637, p = 0.001), despite a stable total energy intake. The anthropometric indicators of muscle mass, branched-chain amino acids, phenylalanine, and lymphocyte subset percentages increased significantly. Functional independence improved (Barthel β = 10.456, p < 0.001; Katz β = 9.305, p = 0.001). The PCA identified four interrelated dimensions that explained 47.1% of the variance. CONCLUSIONS: Intensive neurorestorative rehabilitation is associated with concurrent changes in multiple physiological domains in patients with chronic motor-complete SCI. Given the quasi-experimental design and reliance on estimated metabolic measures, the findings should be interpreted as hypothesis-generating.