Inherent Risk Factors for Nosocomial Infection in the Long Stay Critically Ill Child Without Known Baseline Immunocompromise: A Post -Hoc Analysis of the CRISIS Trial.

BACKGROUND Nosocomial infection remains an important health problem in long stay (> 3 days) pediatric intensive care unit (PICU) patients. Admission risk factors related to the development of nosocomial infection in long stay immune competent patients in particular are not known. METHODS Post-hoc analysis of the previously published Critical Illness Stress induced Immune Suppression (CRISIS) prevention trial database, to identify baseline risk factors for nosocomial infection. Because there was no difference between treatment arms of that study in nosocomial infection in the population without known baseline immunocompromise, both arms were combined and the cohort that developed nosocomial infection was compared with the cohort that did not. RESULTS There were 254 long stay PICU patients without known baseline immunocompromise. Ninety (35%) developed nosocomial infection, and 164 (65%) did not. Admission characteristics associated with increased nosocomial infection risk were increased age, higher PRISM III score, the diagnoses of trauma or cardiac arrest, and lymphopenia (p < 0.05). The presence of sepsis or infection at admission was associated with reduced risk of developing nosocomial infection (p < 0.05). In multivariable analysis, only increasing age, cardiac arrest, and existing lymphopenia remained significant admission risk factors (p < 0.05); whereas trauma tended to be related to nosocomial infection development (p = 0.07). CONCLUSIONS These data suggest that increasing age, cardiac arrest, and lymphopenia predispose long stay PICU patients without known baseline immunocompromise to nosocomial infection. These findings may inform pre-hoc stratification randomization strategies for prospective studies designed to prevent nosocomial infection in this population.