부산대학교 도서관

Purpose: Traumatic brain injury causes morbidity and mortality worldwide. S100B is the most documented emergency brain biomarker and its urine-assay might be advantageous because of easier sampling. The primary aim was to evaluate urine S100B's ability to rule out intracranial hemorrhage. Secondary aims included S100B temporal pattern for 48 h post-trauma and chemical properties of urine that affect urine S100B. Methods: Patients with head trauma were sampled for serum and urine S100B. Patients who were admitted for intracranial hemorrhage were sampled for 48 h to assess S100B-level, renal function, urine-pH, etc. Results: The negative predictive value of serum S100B was 97.0% [95% confidence interval (CI) 89.5–99.2%] and that of urine S100B was 89.1% (95% CI 85.5–91.9%). The specificity of serum S100B was 34.4% (95% CI 27.7–41.6%) and that of urine was 67.1% (95% CI 59.4–74.1%). Urine-pH correlated strongly with urine S100B during the first 6-h post-trauma. Trend-analysis of receiver operator characteristics of S100B in serum, urine the arithmetic difference between serum and urine S100B showed the largest area under the curve for arithmetic difference, which had a negative predictive value of 93.1% (95% CI 89.1–95.8%) and a specificity of 71.8% (95% CI 64.4–78.4%). Conclusion: This study cannot support ruling out intracranial hemorrhage with urine S100B. Urine-pH might affect urine S100B and merits further studies. Serum and urine S100B have poor concordance and interchangeability. The arithmetic difference had a slightly better area under the curve and can be worth exploring in certain subgroups.