Timing of antibiotics, volume, and vasoactive infusions in children with sepsis admitted to intensive care

INTRODUCTION Early administration of antibiotics for sepsis, and of fluid boluses and vasoactive agents for septic shock, is recommended. Evidence for this in children is limited. METHODS The Alberta Sepsis Network prospectively enrolled eligible children admitted to the Pediatric Intensive Care Unit (PICU) with sepsis from 04/2012-10/2014. Demographics, severity of illness, and outcomes variables were prospectively entered into the ASN database after deferred consent. Timing of interventions were determined by retrospective chart review using a study manual and case-report-form. We aimed to determine the association of intervention timing and outcome in children with sepsis. Univariate (t-test and Fisher's Exact) and multiple linear regression statistics evaluated predictors of outcomes of PICU length of stay (LOS) and ventilation days. RESULTS Seventy-nine children, age median 60 (IQR 22-133) months, 40 (51%) female, 39 (49%) with severe underlying co-morbidity, 44 (56%) with septic shock, and median PRISM-III 10.5 [IQR 6.0-17.0] were enrolled. Most patients presented in an ED: 36 (46%) at an outlying hospital ED, and 21 (27%) at the Children's Hospital ED. Most infections were pneumonia with/without empyema (42, 53%), meningitis (11, 14%), or bacteremia (10, 13%). The time from presentation to acceptable antibiotic administration was a median of 115.0 [IQR 59.0-323.0] minutes; 20 (25%) of patients received their antibiotics in the first hour from presentation. Independent predictors of PICU LOS were PRISM-III, and severe underlying co-morbidity, but not time to antibiotics. In the septic shock subgroup, the volume of fluid boluses given in the first 2 hours was independently associated with longer PICU LOS (effect size 0.22 days; 95% CI 0.5, 0.38; per ml/kg). Independent predictors of ventilator days were PRISM-III score and severe underlying co-morbidity. In the septic shock subgroup, volume of fluid boluses in the first 2 hours was independently associated with more ventilator days (effect size 0.09 days; 95% CI 0.02, 0.15; per ml/kg). CONCLUSION Higher volume of early fluid boluses in children with sepsis and septic shock was independently associated with longer PICU LOS and ventilator days. More study on the benefits and harms of fluid bolus therapy in children are needed.