Sources: Braun BI, et al. Preventing central venous catheter-associated primary bloodstream infections: Characteristics of practices among hospitals participating in the evaluation of processes and indicators in infection control (EPIC) study. Infect Control Hosp Epidemiol. 2003;24:926-935; Climo M, et al. Prevalence of the use of central venous access devices within and outside of the intensive care unit: Results of a survey among hospitals in the prevention epicenter program of the Centers for Disease Control and Prevention. Infect Control Hosp Epidemiol. 2003;24:942-945; Kim SH, et al. Outcomes of Hickman catheter salvage in neutropenic cancer patients with Staphylococcus aureus bacteremia. Infect Control Hosp Epidemiol. 2003;24:897-904; Alonso-Echanove J, et al. Effect of nurse staffing and antimicrobial-impregnated central venous catheters on the risk of bloodstream infections in the intensive care unit. Infect Control Hosp Epidemiol. 2003;24:916-925.
Many questions continue to circle around the use of central venous access devices (CVC). What are the demographics of their use and the practices of their insertion? How dangerous are they? What location in the hospital is central venous catheter use most prevalent? What are ways to limit infection? Are there special considerations to treat infections, including bloodstream infections, resulting from the use of CVCs?
The December issue of Infection Control and Hospital Epidemiology has 4 articles addressing these issues, which are discussed below.
Barbara Braun, who works for the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), assembled epidemiologists from the study group known as EPIC (Evaluation of Processes and Indicators in Infection Control). They in turn contacted those hospitals that had members in the Society for Healthcare Epidemiology of America. The goal of this study was to uncover the methodology of CVC insertion and practices to limit primary bloodstream infections related to CVCs. Novel methods were used to capture the exact characteristics of the insertion of the CVC in 54 hospitals that completed the CVC survey; 41 were from the United States.
There were 3320 CVC insertions available for study, an average of 58 per hospital enrolled. There was a wide variation in the characteristics of insertion. Most insertions (91%) were nontunneled devices. About 20% of patients had a follow-up CVC inserted in the original site of insertion. Most CVC were placed by physicians, but their experience with insertion varied widely (ie, the number of years the clinician had inserted CVCs ranged from 0 to 39). Mask and gown barriers were used for most insertions. Up to 25% of CVCs were impregnated with an antibiotic or antiseptic. The mean time required for insertion was 10 minutes. Of note, in 8.2% of insertions, physicians had difficulties, mostly related to problems inserting the CVC at multiple sites. In almost 4% of insertions, a supply item was not available.
Hospital policies were also quite variable in dealing with prevention of BSI. Only 49% of hospital committees met 7 or more times per year. Before the study, less than a quarter of hospitals were collecting BSI data. Just more than half of the hospitals convened an immediate investigation if there was a substantial excess of bacteremias. More than 80% of hospitals had no IV team. Use of needleless systems for insertion occurred in about half of the episodes.
One interesting measure of hospital BSI surveillance was the ratio of blood cultures performed per 100 patient-days; the average in this study was 10. Hospitals in the study had an average of 2.2 individuals used for infection surveillance, and these workers spent an average of 13 hours in ICU surveillance. Almost all the hospitals had an epidemiologist.
To answer the question regarding the hospital location of patients with CVCs, Michael Climo in Richmond, Va, and a group of prominent hospital epidemiologists performed a 1-day prevalence study at 6 medical centers. The centers participated in the Prevention Epicenter Program of the CDC. During the day’s study, medical personnel visually examined all the patients hospitalized in their respective academic centers. Patients in emergency rooms, outpatient areas, and psychiatry, obstetric, and ophthalmology wards were excluded from the study.
Four classes of catheter were studied: tunneled CVCs; nontunneled CVCs; peripherally inserted central catheters (PICCS); and totally implantable devices that were Portacths by name (Deltec, Inc., St. Paul, Minn). Subclavian, jugular, femoral, or PICCs were all considered for the study. Chi square tests were applied for statistical significance.
Nearly 2500 patients were enrolled, and 29% had CVCs. Rate of use averaged 55.4% in ICU patients and 24.4% in non-ICU patients, but the absolute number of patients outside the ICU was more than twice that for those patients in the ICU.
The most common access sites were subclavian (55%), jugular (22%), and femoral (6%), with the jugular and femoral sites being more frequently used inside the ICU than outside (P < .001 for both). The sites in more than 80% of CVCs in ward patients were either subclavian or PICCs. The most common type of catheter was nontunneled CVC (46%) and then tunneled CVC (23%). In the ICU, as expected, most catheters (74%) were nontunneled CVCs.
Separately, Alonso-Echanove, with colleagues from the CDC, performed a study called DISC (Detailed ICU Surveillance Component), a prospective, observational, multicenter cohort study of bloodstream infecions (BSI) associated with CVCs in 8 ICUs.
A total of 4535 patients who used 8593 CVCs were studied. A total of 293 organisms were isolated: Gram-positive cocci accounted for 73% (coagulase-negative staphylococci, 44.5%), Gram-negative bacteria for 17%, and fungi for 10%.
There were 28 variables analyzed as risk factors. Those variables that were significantly associated with BSI were TPN with a nonimpregnated CVC, no antibiotics for 48 hours post-insertion, unarrousable patients, care by a float nurse, and a patient age of 45-55 years. Only 7.4% of CVCs were PICC lines, but those had a 74% lower risk of BSI. All the impregnated CVCs used chlorhexidine and silver sulfadiazine.
In a more limited study Kim and colleagues from Seoul National University sought to determine the rate of salvage of Hickman catheters associated with Staphylococcus aureus bacteremia in neutropenic cancer patients. The study covered 1998-2002 and involved 32 episodes in 29 patients. Vancomycin was not routinely used for empiric therapy. Empiric antibiotic therapy of the staphylococcemic episode was considered appropriate or inappropriate.
Overall mortality was 69%, and that due to S aureus was 38%. MRSA infections had a higher mortality rate (P = .04). Salvage was attempted in 24 of the patients (75%). Only 12 patients received appropriate empiric antistaphylococcal therapy, but the difference in salvage between appropriate (58%) and inappropriate (42%) empiric treatments was not significant.
Of 7 cases with persistent S aureus bacteremia in the face of specific antistaphylococcal therapy, 4 Hickman catheters were not removed. Three of the 4 died with staphylococcal bacteremia.
For patients in whom salvage was attempted, the overall success rate was 50% (12 of 24), more often in those patients with subsequent negative blood cultures (11 of 17). There was a trend toward a worse outcome in those patients with extraluminal infection and negative subsequent blood cultures.
Comment by Joseph F. John, Jr., MD
There are several cogent findings from these studies. First, use of CVCs does differ with regard to hospital location. Hospital policies for CVC use may be quite variable. Infections feature Gram-positive cocci much more than other bacteria and fungi, and coagulase-negative staphylococci are the current bane. A major risk factor for infections remains the concomitant use of total parenteral nutrition. Antibiotic impregnated CVCs reduce the risk of infection, and their use seems to be increasing. Finally, there may be some room for salvage when S aureus is the bloodstream pathogen, but when there is persistent staphylococcal bacteremia, the CVC should be removed.
Climo’s multicenter group found that central catheters are very commonly used outside the ICU, suggesting that new guidelines for surveillance outside the ICU are needed.
The study by Braun and associates is in fact the most provocative since it highlights the lack of assumption of leadership by some hospitals with regard to policies for these important devices. There is a disconnect between the findings that almost all (96.4%) hospital committees discussed the issue of bloodstream infection, but only about half initiated an investigation if there was an increased frequency. Only 19% of the hospitals had IV teams to manage the CVCs in the study ICU, and only 38% used needleless systems for insertion. This variation in the control and monitoring of CVC, thus, is a national issue that hospitals desperately need to address. Their attention should be directed to structural factors like catheter characteristics and hospital demographics; to process-of-care factors like the skill of operators and monitoring of outcomes; and to patient factors like severity of illness and urgency of insertion.
Only catheters coated with chlorhexidine and silver sulfadiazine were examined in the Alonzo-Echanove study, but the number studied was substantial—1775 of the 8593 CVCs studied. The impregnated CVCs had a 43% lower rate of infection than nonimpregnated CVCs. The reduction in infection was most pronounced in patients on TPN, so perhaps there is at work a catheter/microbial interaction like those influencing biofilm deposition. It was encouraging that the effect of the impregnated CVC was not dependent on the duration of use. Clearly, we are going to hear more about antibiotic-impregnated catheters and the role they may be playing in high-risk patients.
The further finding that PICC lines reduced the risk of infection by 74% when compared to nonimpregnated CVC is a very important aspect of this large, multicenter study. A broader use of PICCs may require additional resources for placement, but the extra time and price may be worth it.
All of us clinicians are increasingly faced with those patients with S aureus bacteremia associated with an indwelling CVC. The work of Kim et al from Seoul in neutropenic cancer patients with S aureus bacteremia may not be totally analogous to those patients in Western hospitals, but I suspect their outcomes are similar to studies here. The rate of salvage—50%—is about what a lot of ID practitioners would have guessed, but now here are the data.
Taken together, these 4 studies emphasize the importance CVCs have assumed in modern medicine. They are relatively safe devices, but their very vascular access make them more potentially dangerous than other medical devices. We can generalize from these studies to make the following recommendations:
- CVC should be placed by experienced operators and monitored rigorously by experienced nurses;
- Hospitals should regularly collect and review data on bloodstream infections, particularly those associated with CVC;
- PICC lines are preferable to jugular and subclavian lines;
- Attempts should be made to reduce staphylococcal colonization of CVC; and
- Antibiotic-impregnated catheters should be used for patients at high risk of infection.