Abstract for SHEA 18th Annual Scientific Meeting

Trends in Central Line-associated Bloodstream Infections in Intensive Care Units-United States, 1997–2007

Deron C. Burton1, Jonathan R. Edwards1, Teresa C. Horan1, Scott K. Fridkin1
Author's affiliations: 1Centers for Disease Control and Prevention, Atlanta, GA

Background. Previously reported data from CDC's National Nosocomial Infections Surveillance (NNIS) System demonstrated declines in central line-associated bloodstream infection (CLABSI) rates for several types of intensive care units (ICUs) from 1990–1999. In recent years, CLABSI prevention efforts in U.S. ICUs intensified, and CDC transitioned from NNIS to the National Healthcare Safety Network (NHSN).

Objective. We sought to determine the presence and magnitude of trends in CLABSI rates in major ICU types over the most recent 10 years of surveillance under NNIS and NHSN.

Methods. Data were reported through NNIS from 1997–2004 and through NHSN from 2006-2007. In both systems during the months and ICUs selected for surveillance, infection control professionals used standard methods and definitions to identify and report all CLABSIs and central line days. Data from all facilities were aggregated to calculate pooled mean annual CLABSI rates per 1,000 central-line days, by ICU type. Trends in CLABSI rates were assessed by comparing annual mean rates and by using Poisson regression of ICU-specific rates; slopes for linear trends were expressed as annual rate ratios. Log linear models were used to estimate 2005 rates for display purposes. Observed trends were confirmed by limiting analysis to facilities and units reporting at least minimum data in every year of the study period.

Results. Overall, 37,655 CLABSIs were reported from 603 facilities representing 1,766 ICUs. We focused our analysis on the following predominant ICU types: combined medical/surgical (656 units, 206 of which had a major affiliation with a medical school teaching program), medical (239 units), surgical (237 units), coronary (185 units), cardiothoracic (143 units), and pediatric (137 units). Overall linear declines in annual CLABSI rates were observed for five ICU types: medical (slope 0.93, i.e., 7% rate reduction per year; 95% confidence interval [CI] 0.92–0.93); medical/surgical with major teaching affiliation (slope 0.92; CI 0.91–0.93); medical/surgical without major teaching affiliation (slope 0.93; CI 0.93–0.94); surgical (slope 0.95; CI 0.94–0.95); and pediatric (slope 0.95; CI 0.94–0.96) (Figure). No significant trends were observed for coronary or cardiothoracic ICUs. Significant reductions in CLABSI rates were observed separately within NNIS years and NHSN years for each ICU type with an overall decline, except surgical (no decline in NHSN).

Conclusion. Over the past decade, CLABSI rates declined significantly in most, but not all, major ICU types in facilities reporting to CDC. Further study is needed to assess the contribution of specific prevention efforts and participation in national surveillance to observed declines.

Trends in Central Line-associated Bloodstream Infections in Intensive Care Units

The chart above shows trends in the incidence rates of central line-associated bloodstream infections (or CLABSIs) reported to CDC from 1997 to 2007 for five major intensive care unit (ICU) types. Values shown for 2005 are estimated, because no surveillance data are available from the National Healthcare Safety Network (NHSN) for that pilot year.

The y-axis represents the pooled mean annual CLABSI incidence rate per 1,000 central line days and ranges from 0 to 9 in increments of 1. The x-axis runs from the year 1997 to the year 2007 in 1-year increments.

There are five separate lines plotted on the figure, one for each ICU type:

  • For pediatric ICUs, CLABSI incidence in 1997 was 7.8 per 1,000 central line days, declining to 3.7 per 1,000 central line days in 2007
  • For medical ICUs, CLABSI incidence in 1997 was 6.3 per 1,000 central line days, declining to 2.8 per 1,000 central line days in 2007
  • For surgical ICUs, CLABSI incidence in 1997 was 5.4 per 1,000 central line days, declining to 3.1 per 1,000 central line days in 2007
  • For medical/surgical ICUs in facilities with a major teaching affiliation with a medical school, CLABSI incidence in 1997 was 6.0 per 1,000 central line days, declining to 2.6 per 1,000 central line days in 2007
  • For medical/surgical ICUs in facilities without a major teaching affiliation with a medical school, CLABSI incidence in 1997 was 4.1 per 1,000 central line days, declining to 1.9 per 1,000 central line days in 2007

     

Date last modified: April 9, 2008
Content source: 
Division of Healthcare Quality Promotion (DHQP)
National Center for Preparedness, Detection, and Control of Infectious Diseases