BACKGROUND: We aimed to investigate the diagnostic performance of S100 as an outcome predictor after out-of-hospital cardiac arrest (OHCA) and the potential influence of two target temperatures (33 °C and 36 °C) on serum levels of S100.

METHODS: This is a substudy of the Target Temperature Management after Out-of-Hospital Cardiac Arrest (TTM) trial. Serum levels of S100 were measured a posteriori in a core laboratory in samples collected at 24, 48, and 72 h after OHCA. Outcome at 6 months was assessed using the Cerebral Performance Categories Scale (CPC 1-2 = good outcome, CPC 3-5 = poor outcome).

RESULTS: We included 687 patients from 29 sites in Europe. Median S100 values were higher in patients with a poor outcome at 24, 48, and 72 h: 0.19 (IQR 0.10-0.49) versus 0.08 (IQR 0.06-0.11) μg/ml, 0.16 (IQR 0.10-0.44) versus 0.07 (IQR 0.06-0.11) μg/L, and 0.13 (IQR 0.08-0.26) versus 0.06 (IQR 0.05-0.09) μg/L (p < 0.001), respectively. The ability to predict outcome was best at 24 h with an AUC of 0.80 (95% CI 0.77-0.83). S100 values were higher at 24 and 72 h in the 33 °C group than in the 36 °C group (0.12 [0.07-0.22] versus 0.10 [0.07-0.21] μg/L and 0.09 [0.06-0.17] versus 0.08 [0.05-0.10], respectively) (p < 0.02). In multivariable analyses including baseline variables and the allocated target temperature, the addition of S100 improved the AUC from 0.80 to 0.84 (95% CI 0.81-0.87) (p < 0.001), but S100 was not an independent outcome predictor. Adding S100 to the same model including neuron-specific enolase (NSE) did not further improve the AUC.

CONCLUSIONS: The allocated target temperature did not affect S100 to a clinically relevant degree. High S100 values are predictive of poor outcome but do not add value to present prognostication models with or without NSE. S100 measured at 24 h and afterward is of limited value in clinical outcome prediction after OHCA.

TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT01020916 . Registered on 25 November 2009.