Routine blood markers from different biological pathways improve early risk stratification in cardiac arrest patients: Results from the prospective, observational COMMUNICATE study

Introduction: Prognostication of cardiac arrest patients admitted to the intensive care unit (ICU) may influence treatment decision, but remains challenging. We evaluated the incremental usefulness of routine blood markers from different biological pathways for predicting fatal outcome and neurological deficits in cardiac arrest patients. Methods: We prospectively included consecutive, adult cardiac arrest patients upon ICU admission. We recorded initial clinical parameters and measured blood markers of cardiac injury/stress (troponin, BNP, CK), inflammation/infection (WBC, CRP, procalcitonin) and shock (lactate, creatinine, urea). The primary and secondary endpoints were all-cause in-hospital mortality and bad neurological outcome defined by the Cerebral Performance Category (CPC) score. Results: Mortality in the 321 included patients was 49% (n= 156). Procalcitonin (adjusted odds ratio 1.84, 95% CI 1.34 to 2.53, p < 0.001; AUC 0.73) and lactate (adjusted odds ratio 7.29, 95% CI 3.05 to 17.42, p < 0.001; AUC 0.70) were identified as independent prognostic factors for mortality and significantly improved discrimination of a parsimonious clinical model including resuscitation measures (no-flow time, shockable rhythm) and initial vital signs (Glasgow coma scale, respiratory rate) from an AUC of 0.79 to 0.84 (p < 0.001). Cardiac markers did not further improve the model. Results for neurological outcome were similar with model improvements by procalcitonin and lactate from AUC 0.83 to 0.87 (p= 0.004). Conclusion: Assessment of routine markers of inflammation/infection and shock provide significant improvements for prognostication of cardiac arrest patients, while cardiac markers did not further improve statistical models. Combination of blood markers and clinical parameters may help to improve initial management decisions in this vulnerable patient population.