Neuron-specific enolase as a marker of fatal outcome in patients with severe sepsis or septic shock.

Abstract

Research Fellow Clinic of AnesthesiologyAssociate Professor Department of Internal MedicineAssociate Professor Clinic of AnesthesiologyProfessor and Chair Clinic of AnesthesiologyProfessorClinic of AnesthesiologyProfessorClinic of AnesthesiologyUniversity of HeidelbergHubert_Bardenheuer@med.uni-heidelberg.deTo the Editor:—The mortality rate of septic patients increases progressively with increasing severity of encephalopathy caused by sepsis. 1Several mechanisms have been suggested as the cause of septic encephalopathy, including direct central nervous system infection, endotoxin and cytokine effects on the brain, inadequate cerebral perfusion, metabolic derangements, or complications of medical therapy. 2,3Neuropathologic examinations of 12 septic patients who died during protracted, severe septic encephalopathy showed microabscesses in eight cases and vascular lesions in approximately half. 4We assessed the ability of the neuron-specific enolase (NSE), the neuronal isomer of the glycolytic enzyme 2-phospho-D-glycerate hydrolase, and the central nervous system–specific isoforms of S100 to predict mortality in patients with severe sepsis or septic shock. Both NSE and S100 have been shown to be specific parameters for the assessment of cerebral damage caused by hypoxia–ischemia. 5In addition, NSE has been found to be increased in nonsurvivors in a baboon model of sepsis. 6Twenty-nine consecutive patients of the surgical intensive care unit were enrolled in the study within the first 24 h after onset of severe sepsis or septic shock according to the criteria of the American College of Chest Physicians/Society of Critical Care consensus conference. 7At enrollment, we measured NSE and S100 in the serum of these patients using immunoluminometric assays (Byk-Sangtec, Dietzenbach, Germany). In addition, we determined interleukin 6 values by enzyme-linked immunosorbent assay (Bender MedSystems, Vienna, Austria). All patients were mechanically ventilated and were cared for by the intensive care unit staff. The severity of the patients’ illness was estimated using the APACHE II score. Patients with chronic altered mental status, acute primary central nervous system disorders (e.g. , meningitis or cerebrovascular accident), previous cardiac arrest, acute metabolic disorders, and acute primary liver disease were excluded from the study.In the present study, APACHE II score, mean arterial pressure, heart rate, arterial oxygen partial pressure/fraction of inspired oxygen ratio, and serum interleukin-6 concentration did not distinguish between survivors and nonsurvivors (table 1). Although S100 was increased in septic patients compared with the normal range (< 0.12 μg/l), no significant difference was observed between survivors and nonsurvivors. In contrast, serum concentrations of NSE were significantly higher in nonsurvivors than in survivors. The receiver operating characteristic curve in prediction of mortality for NSE is shown in figure 1. Based on 14 μg/l as the best cutoff point, NSE predicted mortality in patients with severe sepsis and septic shock with a sensitivity and specificity of 85% and 94%, respectively.Increased serum NSE may serve as a useful prognostic marker of outcome in severe sepsis and septic shock. However, it must be kept in mind that NSE can also be elevated in the case of small-cell lung cancer or benign pulmonary disease. 8Because pulmonary dysfunction as estimated by the arterial oxygen partial pressure/fraction of inspired oxygen ratio was not significantly worse in nonsurvivors and did not correlate with NSE values, it is unlikely that the difference in NSE in the present study reflects pulmonary dysfunction in septic patients. False-positive increases in NSE can also be caused by hemolysis, because erythrocytes contain NSE. 9The diverging result of NSE and S100 measurements was an unexpected finding, because both markers have been shown to be associated with ischemic cerebral disease. This could possibly be explained by different release kinetics of these two proteins 10or the finding that NSE more sensitively reflects small cerebral infarcts or transient ischemic attacks. 5Furthermore, NSE has a long half-life compared with S100. In contrast to S100, which is present in high concentrations in glial cells and Schwann cells, 11NSE originates predominantly from neurons and neuroendocrine cells. 12Therefore, in sepsis, it is possible that S100 just reflects a glial inflammatory reaction, whereas NSE may serve as a marker of neuronal damage. Further studies involving a greater number of patients are necessary to evaluate NSE as a parameter of outcome in sepsis.