Pediatric in-intensive-care-unit cardiac arrest: incidence, survival, and predictive factors.

Objective: To describe the incidence, survival, and neurologic outcome of in-intensive-care-unit (ICU) cardiac arrest and to identify factors predictive of survival to hospital discharge. Methods: We performed a retrospective cohort study. Eligible patients were <18 yrs of age and experienced a cardiac arrest during their admission to a multidisciplinary pediatric intensive care unit in the 5.5-yr period ending June 2002. Cardiac arrest was defined as the administration of chest compressions or defibrillation for a nonperfusing cardiac rhythm. Mortality and the Paediatric Cerebral Performance Score were measured and presented according to the Utstein style. Factors predictive of survival to hospital discharge were identified by univariate analysis and independent predictors were identified by multivariate analysis. Main Measurements and Results: Ninety-one children had cardiac arrest, yielding an incidence of 0.94 cardiac arrests per 100 admissions. Resuscitation was successful in 75 (82%) children, 61 (67%) survived 24 hrs, 25 (27%) children survived to ICU discharge and 23 (25%) to hospital discharge. At hospital discharge, the median Pediatric Cerebral Performance Category score was 2 (range, 1-3) and the median Pediatric Overall Performance Category score was 3 (range, 1-4). No child was assessed as normal on both scores. The independent positive predictors of hospital mortality were the presence of renal failure before cardiac arrest (odds ratio [OR], 6.1; 95% confidence interval [Cl], 1.8-31), being on epinephrine infusion at time of cardiac arrest (OR, 9.5; 95% Cl, 1.5-62), and the administration of one or more calcium boluses during resuscitation (OR, 5.4; 95% Cl, 1.1-25). The use of extracorporeal membrane oxygenation (ECMO) within 24 hrs after cardiac arrest was associated with reduced hospital mortality (OR, 0.18; 95% Cl, 0.04-0.76). Conclusions: In-ICU cardiac arrest is associated with high in-hospital mortality and subsequent morbidity in survivors. Prearrest renal dysfunction and epinephrine infusion were associated with increased in-hospital mortality. The use of post-arrest ECMO within 24 hrs was associated with reduced mortality. Rigorous prospective evaluation of the role of ECMO following cardiac arrest is needed.