Introduction: Most patients with cardiac arrest (CA) admitted to hospitals after successful cardiopulmonary resuscitation (CPR) are discharged with various degree of neurological deficits. To determine predictor of neurological outcome early and accurately, and to determine cutoff values, serum levels of protein S-100B and neuron-specific enolase (NSE) within 24h after CA were assessed.
Methods and results: A multicenter prospective observational study was conducted between May 2007 and April 2008 at three medical institutions in Japan on 107 consecutive non-traumatic CA patients with return of spontaneous circulation after CPR. Based on "best-ever achieved" Glasgow-Pittsburgh cerebral performance categories (CPC) score within 6 months after CA, patients were classified into a "poor neurological outcome" group (CPC3 to CPC5) (n=67) and "favorable neurological outcome" group (CPC1 and CPC2) (n=13). Blood was sampled on admission, at 6 and 24h after CA. Serum S-100B and NSE in "poor outcome" group were higher than those in "favorable outcome" group (P<0.01). On ROC analysis, area under the curve of S-100B was 0.85, 0.94 and 1.0, respectively. These were greater than those of NSE at all sampling points. The "100%-specific" cutoff values of S-100B predictive of poor neurological outcome were 1.41, 0.21, and 0.05ng/mL, respectively. These values corresponded to sensitivities of 20.9%, 62.8%, and 100%, respectively, each of which was higher than those of NSE.
Conclusions: S-100B is more reliable as an early predictor of poor neurological outcome within 24h after CA than NSE and can be applied clinically.