Table 1.
Demographic data for all patients surviving first 3 days after admission investigated in the HAnnover COoling REgistry (HACORE).
Characteristics are also shown divided in survivors and non survivors within day 3–30 and good (CPC 1 and 2) to poor (CPC 3,4 and 5) outcome.
Table 2.
Univariate and multivariate cox regressions analysis for mortality at 30-day follow-up in the HAnnover COoling Registry (HACORE).
Table 3.
Univariate and multivariate regressions analysis for poor (CPC ≥3) outcome in the HAnnover COoling Registry (HACORE).
Table 4.
Distribution of neuromarkers according to the cause of death.
Fig 1.
ROC curves for determining cut-off values of NSE and S-100b for mortality and neurological outcome: ROC analysis for NSE (A and C) and S-100b (B and D) are presented for mortality (A and B), and neurological outcome (C and D), respectively, with corresponding areas under the curve (AUC) and confidence-intervals (CI). Cut-off values (co) identified by highest Youden-index (Yi) are shown above downwards pointing arrowheads (▼).
Table 5.
Sensitivity, specificity, positive predictive value, negative predictive value and accuracy for predicting poor neurological outcome (CPC≥3) according to determined cut off values for NSE and S-100b.
Table 6.
False positive rate, sensitivity, specificity and corresponding NSE (a) and S-100b (b) for prediction of poor neurological outcome (CPC≥3).
Fig 2.
Distribution of CPC-classes according to interquartiles of NSE and S-100b and relation to cut-off values: Proportions of CPC classes according to the interquartile distributions of NSE (A) and S-100b (C), respectively, and scatter plots of survivors with good (CPC 1 and 2) and poor neurological outcome (divided in CPC 3 and 4 and CPC 5 separated by *anoxic and #other causes of death) for NSE (B) and S-100b (D) with cut-off values for mortality (dashed line) and poor neurological outcome (dotted line) and poor neurological outcome of survivors (dotted and dashed line).