Skin: the first battlefield.

C atheter-related bloodstream infections (CRBSI) are a common and costly complication in hospitalized patients, especially the critically ill (1). CRBSI affects over 200,000 patients per year in the United States, and is associated with an increased risk of dying (though whether this association is causal is controversial) (2). The complexity of the pathogenesis of CRBSI is illustrated by the 50-fold range in the risk of bloodstream infection per 1000 catheter days for different types of vascular access devices (Table 1) (1). A central question important to the pathogenesis and prevention of vascular catheter infections is where do the microorganisms on vascular catheters originate? The elegant investigation of Jeske et al. (3) reported in this issue of Anesthesia & Analgesia provides important insight into this question by demonstrating that in intensive care unit (ICU) patients, 71% of bacteria found on the tip of short-term central venous catheters genetically match bacteria found on the insertion device. Thus, these microorganisms probably come from the patient’s own skin. Additional evidence supports the proposed mechanism whereby skin organisms colonize vascular catheters during the insertion process. A study of pulmonary artery catheters inserted just prior to cardiac surgery and cultured intraoperatively 90 min after placement found that 5 of 30 catheters inserted percutaneously had coagulase negative staphylococci isolated from their tip, whereas only 1 of 30 catheters inserted through an introducer was colonized (4,5). Molecular typing by pulsed field gel electrophoresis confirmed that 20% to 40% of the bacteria found on the catheter tips originated from the skin (5). Taken together, these data support the concept that the physical penetration of the skin barrier during catheter insertion is a mechanism that can result in early bacterial colonization of the catheter. The findings by Jeske et al. (3) detailed in this issue, and by Elliott et al. (4) and Livesley et al. (5) previously, that the catheter tip can be colonized at the time of catheter insertion have several significant implications. First, it has previously been shown for ICU patients with central venous catheters that the initial site of catheter colonization is divided about evenly 1/3 subcutaneous segment, 1/3 catheter tip, and 1/3 catheter lumen (6). This has been interpreted as indicating that 33% of the time microorganisms originate from the skin and extend down the external catheter surface (subcutaneous segment 1st week), one third originate from hematogenous seeding (catheter tip 2nd week), and another third arise from catheter breaks (catheter lumen 3rd week on). These findings were obtained by culturing the catheter tip segment, catheter lumen segment, and catheter lumens separately. It was then assumed that if both the catheter subcutaneous segment and catheter lumens were culture negative and the catheter tip segment was positive that the only way that this could happen was by hematogenous seeding. That conclusion is clearly incorrect, as more current data suggest that the majority of the early catheter colonization (first 2 weeks) in ICU patients come from skin organisms. The clinical implication of these findings is that interventions aimed at minimizing the risk of catheter colonization from skin microorganisms should be quite successful at preventing early (first 2 weeks) catheter-related infections in ICU patients. This has clearly been shown to be the case in studies examining what drug to use for skin preparation at the time of catheter insertion (7,8), a study of the use of maximal sterile barriers for catheter insertion (9), studies of educational interventions aimed at improving sterile technique (10–12), and studies of catheters with antiinfective coatings (13–15). In ICU patients, Maki et al. (7) randomized skin preparation solutions for 668 catheters comparing 2% chlorhexidine, 10% povidoneiodine, and 70% alcohol. Chlorhexidine was associated with the lowest incidence of CRBSI (2.3 per 100 catheters, P 0.02), whereas alcohol and povidoneiodine were associated with 7.1 and 9.3 infections per 100 catheters. A recent meta-analysis analyzed 8 studies with a total of 4,143 catheters (8) which identified a risk reduction for CRBSI 0.49, 95% confidence interval [0.28–0.88]. Chlorhexidine gluconate decreased Accepted for publication June 6, 2003. Address correspondence to Richard C. Prielipp, MD, FACCM, Department of Anesthesiology, Wake Forest University, WinstonSalem, NC 27157. Address e-mail to prielipp@wfubmc.edu.