Comparison between capillary, venous and arterial levels of protein S100B in patients with severe brain pathology.

BACKGROUND Protein S100B is soon in clinical use as a sensitive marker after mild traumatic head injury in adults. Initial studies of S100B in pediatric head injury have shown promising results. Venous sampling can be challenging in children and capillary samples are often a preferred option. The aim of the study was to investigate the relation between capillary, venous and arterial measurements of protein S100B, primarily by determining whether capillary S100B differ from venous and if capillary S100B can predict venous S100B levels, and secondarily, if arterial S100B samples can substitute venous samples in severely brain-injured patients. METHODS Venous, arterial and capillary blood samples for S100B were collected simultaneously once a day for a maximum of 6 days. Patients were ≥18 years old and admitted to neurointensive care due to severe brain pathology. RESULTS Capillary S100B samples were on average 0.08 μg/L higher than venous S100B samples. Prediction of venous concentration from capillary samples yielded a prediction error of 0.07 μg/L. The mean difference between venous and arterial samples was 0.01 μg/L. The mean prediction error was 0.03 μg/L. CONCLUSIONS Capillary and venous serum S100B are not interchangeable, and should be considered as two separate, although related, variables. Arterial measurements of S100B can successfully predict the corresponding venous concentration.