Significance and Diagnostic Accuracy of Early S100B Serum Concentration after Aneurysmal Subarachnoid Hemorrhage

Abstract

Background: Early brain injuries (EBI) are one of the most important causes of morbidity and mortality after subarachnoid hemorrhage. At admission, a third of patients are unconscious (spontaneously or sedated) and EBI consequences are not evaluable. To date, it is unclear who will still be comatose (with severe EBI) and who will recover (with less severe EBI) once the aneurysm is treated and sedation withdrawn. The objective of the present study was to determine the diagnostic accuracy of S100B levels at hospital admission to identify patients with severe neurological consequences of EBI.

Methods: Patients were consecutively included in this prospective blinded observational study. A motor component of the Glasgow coma score under 6 on day 3 was used to define patients with severe neurological consequences of EBI.

Results: A total of 81 patients were included: 25 patients were unconscious at admission, 68 were treated by coiling. On day 3, 12 patients had severe consequences of EBI. A maximal S100B value between admission and day 1 had an area under the receiver operating characteristic curve (AUC) of 86.7% to predict severe EBI consequences. In patients with impaired consciousness at admission, the AUC was 88.2%.

Conclusion: Early S100B seems to have a good diagnostic value to predict severe EBI. Before claiming the usefulness of S100B as a surrogate marker of EBI severity to start earlier multimodal monitoring, these results must be confirmed in an independent validation cohort.

Keywords: S100B protein; early brain injury; patient outcome assessment; subarachnoid hemorrhage.

Figure 1

Flow chart. EBI: Early brain injury, GCS: Glasgow coma scale, M-GCS: motor score of GCS, SAH: subarachnoid hemorrhage. Patients’ clinical evolution over the first 3 days and their affiliation to the EBI-mild, EBI-moderate, and EBI-severe groups is presented. Among the three patients who worsened within 3 days, two had hydrocephalus requiring external ventricular drainage (EVD). The third was intubated for respiratory disorder, her M-GCS at day 3 was 5 without sedation and her modified Rankin Scale (mRS) at intensive care unit discharge was 3 (i.e., moderate disability). None had re-bleeding or aneurysm treatment complications.

Figure 2

Receiver–operator characteristic (ROC) curves of early brain injury (EBI) biomarkers. (a) ROC analyses of the total population (n = 81). (b) ROC analyses of patients with a motor component of the Glasgow coma score < 6 at admission (n = 25). Several S100B thresholds are plotted on both panels: (black square) the best threshold; (gray line) the gray zone of diagnostic uncertainty with (black dots) as its lower limit (i.e., to exclude severe EBI with almost certainty (Se ≥ 90%)) and upper limit (i.e., to predict severe EBI with almost certainty (Sp ≥ 90%)).

Figure 3

Algorithm chart for early brain injury severity prognostication. Early brain injury (EBI) severity prognostication and the proposed resulting management strategies are presented based on early S100B serum concentration. Low S100B: concentrations below the lower threshold of the gray zone rejecting severe EBI; high S100B: concentrations above the upper threshold of the gray zone predicting severe EBI. An S100B concentration within the gray zone has an uncertain diagnostic value but suggests significant brain injury. EEG: electroencephalography; ICH: intracerebral hematoma; TCD: transcranial doppler.