Longitudinal cohort-study of consecutive 742 patients admitted to the acute Spine Unit of the Vancouver General Hospital from 2004-2016 becoming enrolled in the Rick Hansen Spinal Cord Injury Registry up to 7 years after SCI. Investigation of the association between BMI and mortality in categorical manner using WHO criteria was complemented by a data driven approach using unbiased recursive partitioning (Classification and Regression Tree). Cox regression modes were used to determine mortality hazard risk ratios.

After exclusion of incomplete datasets 643 patients were studied. 541 patients survived and 102 (18.5%) patients died. Mortality was the lowest in obese (BMI > 30 kg/m²), followed by overweight (BMI 25- 29.9 kg/m²), increased further with normal weight (BMI 18.5-24.9 kg/m²), and was the highest in underweight patients (BMI < 18.5 kg/m²) (Cox-regression, Pearson Chi² 10.852). Unbiased recursive partitioning confirmed, that BMI was negatively associated with long-term mortality and identified three distinct mortality-risk cohorts i) BMI: > 30 kg/m², ii) 30 – 17.5 kg/m² and < 17.5 kg/m². The predictive model corrected for i) demographic factors (age, gender) and ii) SCI characteristics (level and completeness of injury (AIS ASIA, impairment scale)) or severity of associated poytrauma (ISS).

Obesity is a protective factor associated with lower mortality in SCI patients (hazard ratio 0.33, 95% CI: 0.77 – 0.33), concordant with an “obesity paradox”. Moreover, being underweight is a highly significant risk factor for death up to 7 years after SCI (hazard ratio 2.5, 95% CI: 5.1 – 1.2). The results suggest unknown neuro-metabolic pathways crucial for patient survival.