Central Line-Associated Bloodstream Infection in Neonatal Intensive Care Units

OBJECTIVE. Describe the epidemiology of central line-associated bloodstream infections (CLABSIs) in neonatal intensive care units (NICUs) participating in a standardized and mandatory CLABSI surveillance program. DESIGN. Retrospective cohort. SETTING. We included patients admitted (April 2007-March 2011) to 7 level II/III NICUs who developed a CLABSI (as defined by the National Healthcare Safety Network). METHODS. CLABSIs/1,000 central line-days and device utilization ratio were calculated; chi(2) test, Student t test, Kruskal-Wallis, and Poisson regression were used. RESULTS. Overall, 191 patients had 202 CLABSI episodes for a pooled mean rate of 4.0 CLABSIs/1,000 central line-days and a device utilization ratio of 0.20. Annual pooled mean CLABSI rates increased from 3.6 in 2007-2008 to 5.1 CLABSIs/1,000 central line-days in 2010-2011 (P = .01). The all-cause 30-day case fatality proportion was 8.9% (n = 17) and occurred a median of 8 days after CLABSI. Coagulase-negative Staphylococcus was identified in 112 (50.5%) cases. Staphylococcus aureus was identified in 22 cases, and 3 (13.6%) were resistant to methicillin. An underlying intra-abdominal pathology was found in 20% (40/202) of CLABSI cases, 50% of which were reported in the last year of study. When adjusted for mean birth weight, annual CLABSI incidence rates were independently associated with the proportion of intra-abdominal pathology (P = .007) and the proportion of pulmonary pathology (P = .016) reported. CONCLUSION. The increase in CLABSI rates in Quebec NICUs seems to be associated with an increased proportion of cases with underlying intra-abdominal and pulmonary pathologies, which needs further investigation.