Vascular catheters inserted in the trenches versus guideline documents: can the discrepancies be resolved?

Abstract

Infection control personnel familiar with the studies underlying recent guideline recommendations for the prevention of intravascular catheter–related infections1 could conclude that it would be easy to minimize the risk of such infections. Randomized, controlled trials have demonstrated that prepping the skin with chlorhexidine antiseptic can reduce the risk of catheter-related bloodstream infection (BSI) to less than 1%,2 using maximal sterile barriers also can reduce the risk of catheter-related BSI to less than 1%,3 choosing a central venous catheter (CVC) with anti-infective properties can reduce the risk of catheter-related BSI to 1% or less,4-6 and, finally, education can reduce the risk of catheter-related BSI associated with physicians-in-training.7,8 Yet, despite these impressive results in short prospective studies, in practice it is uncommon to find endemic catheter-related BSI rates less than 1%. In this issue of Infection Control and Hospital Epidemiology, there are two large prospective observational studies that address the in-hospital standard of practice related to CVCs.9,10 Alonso-Echanove et al. performed a study sponsored by the Centers for Disease Control and Prevention (CDC) of 4,535 patients with CVCs in 8 intensive care units (ICUs) in 6 different National Nosocomial Infections Surveillance (NNIS) System hospitals in the United States between 1997 and 1999 that examined 60 potential risk factors for catheter-related BSI.9 Braun et al. performed a study sponsored by the Society for Healthcare Epidemiology of America (SHEA), the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), and the CDC regarding the standard of practice associated with 3,320 CVC insertions in ICU patients in 55 hospitals (41 in the United States and 14 from other countries) between 1998 and 1999.10 The large size of both of these studies gave them substantial power to examine questions previously not answerable at single institutions. Several important observations were made. Braun et al. found that 91% of the CVCs inserted were nontunneled, 85% had multiple lumens, and 22% were impregnated with antimicrobials.10 Forty-four percent of the CVCs were inserted in the internal jugular vein, followed by 32% in the subclavian location. Fifty-eight percent of the time, a large drape was used at the time of CVC insertion. Although the clinicians inserting the CVCs were experienced (median of 30 CVC insertions in the past 6 months), 12% of the time insertions were attempted at two or more sites and 20% of the time two or more attempts were required at the final insertion site. Overall, few CVCs (< 20%) were managed by an intravenous therapy team as opposed to ICU nurses, despite the fact that intravenous teams have been associated with lower catheter-related BSI rates. More than 50% of the time nurse staffing involved float nurses and more than 30% of the time nurse staffing involved agency staff, although float nurses made up only 8% of the 15 nurse-to-patient–days ratio that was the average during the study. Both a lower nurse-to-patient ratio and the use of float nurses have been associated with higher catheter-related BSI rates in ICU patients.11-13 The catheter-related BSI rates for the participating institutions were not provided in this publication, but it is hoped that an additional publication will use multivariate analysis to determine which of the practices they examined have the greatest independent effect on the risk of catheter-related BSI. Alonso-Echanove et al. found that the average duration of CVC placement was 6.6 days; CVCs with multiple lumens made up 88% of the CVCs; CVCs were most commonly inserted in the internal jugular vein (44%), followed by the subclavian vein (37%); 21% of the CVCs were impreg-