Multi-facility, Hospital-Wide Surveillance of Line-associated Bacteremias: Looking Beyond the ICU

Background: Central-line associated bloodstream infections (CLABSI) are a significant cause of morbidity and mortality. Traditionally, surveillance and interventions have targeted intensive care units (ICU). Peripherally inserted central catheters (PICC) are often placed by interventional radiology (IR) or intravenous therapy teams (IVT) and are used with increasing frequency to facilitate early discharge.

Objective: Describe the incidence and epidemiology of CLABSIs; compare and contrast PICC and other central venous catheter (CVC) associated CLABSIs within and outside the setting of ICUs.

Methods: Hospital-wide surveillance of adult CLABSIs using NHSN criteria was conducted 10/1/2007-9/30/2009 across a 3 hospital system located in the Chicago metropolitan area. Automated device days were extracted from the electronic medical record. Charlson scores were calculated based on medical conditions present on admission. Medians with interquartile ranges (IQR) are reported while Wilcoxon two-sample and Fisher's exact tests are used for significance testing.

Results: There were 52 CLABSIs; 2 lacked sufficient data regarding line insertion and were excluded. There were over 33,000 line days for a combined incidence of 1.56 CLABSIs per 1,000 lines days. Among the 50 cases with sufficient data, infections occurred in 27 patients with a PICC only, 16 with non-PICC catheters only and 7 with both. 70% were attributed to non-ICU settings. 10% of patients died during their hospitalization. The average age at admission was 63. See Table 1. Over 2/3 of line days occurred outside of the ICU. Organism distribution was not significant with the exception that all 8 episodes of fungemia occurred in patients with PICCs. Infection rates did not differ between patients with PICCs (1.76) and other CVCs (1.14) (rate-ratio = 1.54, p=0.15). Similarly, differences in infection rates attributed to the ICU (1.35) and non-ICU (1.57) setting lacked statistical significance (rate-ratio = 0.86, p=0.63). 

and Conclusions: Central venous catheters are used with increasing frequency beyond critical care areas. PICCs are comparable to other tunneled and non-tunneled catheters with respect to rate of infection. Neither surveillance nor interventions to prevent CLABSIs should be restricted to the ICU.

Table 1: Summary statistics. Patients with PICCs (either alone or in combination with another line) are compared to patients without PICCs and CLABSIs attributed to the ICU are contrasted with CLABSIs occurring outside the ICU.