Brain protein biomarker clearance to blood in traumatic brain injury (TBI) is not fully understood. The aim of this study was to analyze the effect that a disrupted blood-brain barrier (BBB) had on biomarker clearance. Seventeen severe TBI patients admitted to Karolinska University Hospital were prospectively included. Cerebrospinal fluid (CSF) and blood concentrations of S100 calcium binding protein B (S100B) and neuron-specific enolase (NSE) were analyzed every 6-12 h for ∼1 week. Blood and CSF albumin were analyzed every 12-24 h, and BBB integrity was assessed using the CSF:blood albumin quotient (QA). We found that time-dependent changes in the CSF and blood levels of the two biomarkers were similar, but that the correlation between the biomarkers and QA was lower for NSE (ρ = 0.444) than for S100B (ρ = 0.668). Because data were longitudinal, we also conducted cross correlation analyses, which indicated a directional flow and lag-time of biomarkers from CSF to blood. For S100B, this lag-time could be ascribed to BBB integrity, whereas for NSE it could not. Upon inferential modelling, using generalized least square estimation (S100B) or linear mixed models (NSE), QA (p = 0.045), time from trauma (p < 0.001), time from trauma2 (p = 0.023), and CSF biomarker levels (p = 0.008) were independent predictors of S100B in blood. In contrast, for NSE, only time from trauma was significant (p < 0.001). These findings are novel and important, but must be carefully interpreted because of different characteristics between the two proteins. Nonetheless, we present the first data that indicate that S100B and NSE are cleared differently from the central nervous system, and that both the disrupted BBB and additional alternative pathways, such as the recently described glymphatic system, may play a role. This is of importance both for clinicians aiming to utilize these biomarkers and for the pathophysiological understanding of brain protein clearance, but warrants further examination.
Keywords: BBB; NSE; S100B; TBI; albumin quotient.