459

Abstract

Introduction: The presence of brain injury in patients with sepsis is associated with higher mortality. Currently, diagnosis of brain injury in the pediatric intensive care unit (PICU) is based on clinical examination, electroencephalography (EEG), computed tomography (CT), and/or magnetic resonance imaging (MRI); however, these tests are performed only if there is a high index of clinical suspicion. Furthermore, many of these patients are sedated, thereby limiting the clinical evaluation for brain injury. Therefore, validation of biomarkers that will enable detection of brain injury in patients with sepsis is needed. Two potential candidates include NSE, a neuronal glycolytic enzyme, and S100ß, a calcium-binding protein secreted by astrocytes. Both are established biomarkers of cerebral injury in traumatic brain injury and stroke. Methods: This IRB-approved study included children who were admitted to the PICU from December 2009 to December 2010 with the initial diagnosis of systemic inflammatory response syndrome or sepsis. In addition to the collection of demographic and clinical data, serum NSE and S100ß were measured in these patients by enzyme-linked immunosorbent assay. Based on clinical suspicion, EEG, CT, or MRI tests were performed by the clinical team if deemed necessary. Based on published reports, NSE levels ≥ 11.4 ng/mL and S100ß levels ≥ 0.6 ng/mL were considered to be diagnostic for brain injury. The Mann-Whitney test was used for statistical analysis. Results: The mean age of this cohort (n=24) was 5.7 years, and 54% were males. A majority of patients were diagnosed with septic shock (n=16 [67%]) or multi-organ failure (MOF, n=17 [71%]), and most required mechanical ventilation (n=22 [92%]). Two patients required extracorporeal membrane oxygenation (ECMO), and mortality for the cohort was 4% (n=1). Six patients (25%) had abnormal NSE or S100ß levels -- 5 patients had elevated NSE levels (mean 25.8 ± 4.82 ng/mL) and 1 patient had elevated S100ß (0.69 ng/mL). Only 3 out of these 6 patients (50%) demonstrated abnormal EEG or MRI findings (pertussis/ECMO, respiratory failure/ECMO, and seizure disorder). The diagnoses for the other 3 children were line sepsis, peritonitis, and cardiac arrest. MOF was present in all 6 patients. Interestingly, these 6 patients had lower C-reactive protein (CRP) levels than those with normal NSE levels (CRP 6.7 ± 2.4 vs 9.9 ± 2.2 mg/dL, p=0.41). Consistent with existing literature, patients with MOF and/or septic shock had higher serum inflammatory markers, such as CRP and ferritin, compared to patients without MOF or septic shock (CRP 10.8 ± 2.0 vs 2.9 ± 0.8 mg/dL, p=0.08 and ferritin 774.2 ± 241.7 vs 440.0 ± 120.0 ug/L, p=0.78). Conclusions: In this heterogeneous group of pediatric patients with sepsis, use of biomarkers identified additional patients with brain injury that were undetected on clinical examination. Understanding the time course of these biomarkers may provide insight into the prevalence and pathophysiology of brain injury in sepsis. Further verification of the relationship between systemic inflammation and central nervous system inflammation in a larger cohort of pediatric septic shock and MOF is required.