Risk factors for nosocomial sepsis in newborn intensive and intermediate care units

A multicentre prospective study was performed to estimate the incidence of hospital infections and to identify the most relevant risk factors for sepsis in a large and unselected population of high-risk newborns. The study involved 49 neonatal intensive care units and 17 neonatal intermediate care units in Italy. Newborns were followed up from admittance to the units until discharge. Data on demographics and clinical characteristics, exposure to the principal invasive procedures, and onset of infectious complications were prospectively collected. Only infections developing after 48 h from admittance to the unit were recorded. A multiple logistic regression was performed to identify which factors were independently associated with sepsis. Among the 8263 newborns included in the analysis, the incidence of infected newborns was 14.4 per 100 newborns and 0.9/100 days of stay. The incidence of infections was 19.1/100 newborns and 1.2/100 days of stay. Sepsis represented 15.4% of all infections (incidence 2.9/100 newborns and 0.2/100 days of stay). The following factors were independently associated with sepsis: umbilical catheterization, both through the vein and the artery for more than 5 days; mechanical ventilation for more than 5 days; necrotizing enterocolitis; birth weight equal to or less than 2500 g; nasogastric tube; total parenteral nutrition; and transfer from other hospitals. Umbilical catheters accounted for the highest proportion of sepsis (62%), followed by arterial catheters (31%), nasopharyngeal cannulae (26%), tracheal cannulae (20%), and nasal cannulae (20%). The population attributable risk for the other procedures was less than 10%.

This study demonstrates that in a large and unselected newborn population, several host factors and invasive procedures are independently associated with an increased risk of sepsis. After adjustment for clinical severity, intravascular catheterization and assisted ventilation were found to be responsible for a considerable proportion of observed sepsis. They should therefore be considered as priorities for interventions, aimed both at reducing unnecessary use and promoting more strict compliance with aseptic practices.