Characteristics and Timing of Mortality in Children Dying With Infections in North American PICUs*

OBJECTIVES: To investigate the characteristics and timing of death of children with severe infections who die during PICU admission. DESIGN: We analyzed demographics, timing of death, diagnoses, and common procedures in a large cohort obtained from the Virtual Pediatrics Systems database, focusing on early deaths (< 1 d). SETTING: Clinical records were prospectively collected in 130 PICUs across North America. PATIENTS: Children admitted between January 2009 and December 2014 with at least one infection-related diagnosis at time of death. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Analysis included data from 106,464 children admitted to PICUs. The 4,240 children (4%) who died were older than PICU survivors. The median (interquartile range) duration in PICU prior to death was 7.1 days (2.1-21.3 d), with 635 children (15%) dying early (< 1 d of PICU admission). Children who died early were older, more likely to have septic shock, and more likely to have received cardiopulmonary resuscitation than those who died later. Withdrawal of care was less likely in early deaths compared with later deaths. After adjusting for age, sex, sepsis severity, procedures (including cardiopulmonary resuscitation and heart, lung, and renal support), and number of admissions contributed per PICU, it was found that children admitted from the emergency department, inpatient floors, or referring hospitals had significantly greater risk of early death compared with children admitted from the operating room. CONCLUSIONS: A substantial proportion of children admitted to PICU with severe infections die early and differ from those dying later in diagnoses, procedures, and admitting location. The emergency department is a key source of critically ill patients. Understanding characteristics of early deaths may yield recruitment considerations for clinical trials enrolling children at high risk of early death.