Incidence and risk factors for health care-associated pneumonia in a pediatric intensive care unit

Objectives: To determine the incidence and risk factors for health care-associated pneumonia in a pediatric intensive care unit. Design: Prospective cohort study. Setting: Pediatric intensive care unit with 16 medical and surgical beds in a tertiary teaching hospital in Recife, northeast Brazil. Patients: Patients aged <18 yrs were consecutively enrolled between January 2005 and June 2006 into a cohort set to investigate health care-associated infections. Newborns and patients admitted for surveillance and those staying for <24 hrs were excluded. Patients were followed up daily throughout the stay and until 48 hrs after discharge from the unit. Interventions: None. Measurements and Main Results: This report focuses on health care-associated pneumonia, defined as pneumonia that occurs >48 hrs after admission but that was not incubating at the time of admission, as the primary outcome. Intrinsic and extrinsic variables were prospectively recorded into a standardized form. Statistical analyses, including multivariable logistic regression, were performed in Stata version 9.1. There were 765 eligible admissions. Health care-associated pneumonia occurred in 51 (6.7%) patients with an incidence density of 13.1 episodes/1,000 patient-days. There were 366 (47.8%) patients on mechanical ventilation, of whom 39 (10.7%) presented with ventilator-associated pneumonia with an incidence density of 27.1/1,000 days on ventilation. Longer stay on ventilation (odds ratio [OR], 1.04; 95% confidence interval [CI], 1.01-1.08), use of gastric tube (OR, 2.88; 95% CI, 1.41-5.87), and of sedatives/analgesics (OR, 2.45; 95% CI, 1.27-4.72) were identified as independent risk factors for healthcare-associated pneumonia. Conclusion: Identification of independent predictors of health care-associated pneumonia may inform preventive measures. Strategies to optimize use of sedatives/analgesics, reduce the use of gastric tubes, and reduce the time on ventilation should be considered for inclusion in future intervention studies. (Crit Care Med 2011; 39:1968-1973)