FIGURE 32.1 Exit-site infection around an indwelling double-lumen left subclavian central venous catheter that clinically manifested with pain, tenderness, erythema, and swelling.
FIGURE 32.1 Exit-site infection around an indwelling double-lumen left subclavian central venous catheter that clinically manifested with pain, tenderness, erythema, and swelling.
Source of Pathogens
The four potential sources of pathogens are the patient’s skin around the site of catheter insertion, a contaminated catheter hub, hematogeneous seeding from a distant site of infection, and infected infusate. The source of infecting organisms is not determined in about 25% of patients with CRBSI. Not only do both of the first two sources listed above originate from the skin—of the patient and health care providers—but they collectively are responsible for the vast majority of catheter-associated infections. The patient’s skin around the catheter insertion site is the most common source of organisms that colonize central venous catheters with a short-term (mean duration, less than 7 to 10 days) indwelling time (16). As shown in Figure 32.2, the concentration of bacteria on the skin differs between various body sites, with the highest concentration generally present in the femoral area, which is greater than the jugular area, which itself is greater than the subclavian area. Factors that favor higher bacterial concentration on the skin where catheters are inserted include soiling by bacteria-containing bodily fluids and secretions, presence of hair follicles, high temperature, and moist environment. After colonizing the external surface of the catheter, skin-derived flora migrate along the subcutaneous segment into the distal intravascular segment to potentially result in bloodstream infection. In this clinical scenario of infection associated with long-term vascular catheters that are subjected to more extensive manipulation at the catheter hub, the catheter hub becomes contaminated by organisms that originate from the hands of medical personnel, then migrate along the internal surface of the catheter into the intravascular segment before causing bloodstream infection (17). The difference in the pathogenesis of infections associated with short-term versus long-term catheters helps explain why a surface-modified vascular catheter with antimicrobial activity only along the external surface is likely to protect against infection associated with short-term but not long-term catheters.
Milieu of Pathogens
Like other medical devices, infection of vascular catheters centers around the universal formation of a layer of biofilm surrounding the indwelling catheter (Fig. 32.3). The biofilm is composed of both bacterial products (fibroglycocalyx in the case of coagulase-negative staphylococci) and host factors—platelets and tissue ligands such as fibronectin, fibrinogen, and fibrin that variably adhere to well-described receptors on the surface of certain organisms, including staphylococcal and Candida organisms (18). Not only does the biofilm act as a protective barrier for embedded organisms against host immune defenses, including phagocytosis and opsonization (19,20), but it also can impair the activity (21) and, possibly, the penetration (22) of antibiotics against the slowly growing sessile organisms that inhabit the biofilm. This unique biofilm environment may explain why surface-modified vascular catheters containing antimicrobial agents that retain activity within the biofilm, and leach from the catheter surface to produce a zone of inhibition against deeply embedded organisms within the biofilm, tend to be clinically protective (Fig. 32.4).
Since bloodstream infection is the most common serious complication of indwelling vascular catheters, early and accurate diagnosis of this infectious complication is essential. According to the Centers for Disease Control and Prevention (CDC) (23), CRBSI is defined as the isolation of the same organism (i.e., the same species with identical antimicrobial susceptibility) from the colonized catheter and from peripheral blood in a patient with clinical manifestations of sepsis and no other apparent source of bloodstream infection. Until recently, catheter colonization was almost always defined as the growth in cultures from either the tip or subcutaneous segment of the catheter of greater than or equal to 15 colony-forming units (CFU)/mL by the semi-quantitative roll-plate method (24), or greater than or equal to 1,000 CFU/mL by the quantitative sonication method (25).
This standard manner of diagnosing CRBSI, however, is retrospective, as it requires removal and culture of the vascular catheter. Regrettably, only 15% to 25% of central venous catheters removed because of suspected catheter-related infection yield growth from cultures of the catheter tips (26). This explains the escalating interest in assessing and implementing procedures that are intended to prospectively diagnose catheter-related infection without removal of the catheter (27). The potential roles of two such microbiologic methods that could indicate whether the catheter is the source of bloodstream infection have been recently assessed. Both methods require concurrent collection of peripheral and central (i.e., through the lumen of the catheter) blood cultures. The first qualitative method—differential time to positivity (DTP)—which relies on the understanding that the culture of a blood sample that contains higher bacterial concentration would become positive, as detected by production of carbon dioxide by multiplying organisms, at least 2 hours before this would occur in a culture from peripheral blood (4). In the setting of a CRBSI, in which the catheter itself was the source of infection, the supposition is that the bacterial load of the infected catheter is higher than that seen in peripheral blood. The other quantitative method—paired quantitative blood cultures (PQBC)—is based on the anticipation that the number of CFU retrieved from a central blood culture would be greater than or equal to fivefold higher than that grown from cultured peripheral blood (28). Although a meta-analysis of 51 studies of both short- and long-term catheters published from 1996 to 2004 demonstrated that the PQBC method is the more accurate method in diagnosing intravascular device–related bloodstream infection (29), the PQBC method is less accurate than the DTP method for diagnosing bloodstream infection associated with short-term catheters (4,28). Furthermore, the PQBC method is more laborious and less implemented in hospital microbiology laboratories than the DTP method. In addition to considering the sensitivity, specificity, and the positive and negative predictive values of different diagnostic methods that do not require catheter removal, other factors—such as availability, ease of performance, cost, and clinical scenarios of individual patients—often affect the frequency of implementing various diagnostic methods in different medical facilities.
Infections associated with vascular catheters can also present as an exit-site infection (see Fig. 32.1), which manifest as erythema, tenderness, swelling, and drainage. Since inflammatory skin changes can be detected in only about one-fourth of patients with bloodstream infection associated with central venous catheters, the absence of exit-site infection does not negate the existence of CRBSI. Patients with a tunneled vascular catheter can also develop tunnel infection which, like bloodstream infection but unlike exit-site infection, usually requires the removal of the infected catheter to establish cure.
As CRBSI is a clinical diagnosis, it is not typically used for surveillance. It is often difficult to establish that the bloodstream infection is a CRBSI due to patient needs (catheter not always removed), limited laboratory technology (PQBC or DTP), or procedural noncompliance (inaccurate labeling). The CDC’s National Healthcare Safety Network (NHSN) defines the term central line–associated bloodstream infection (CLABSI) as a primary bloodstream infection in a patient who had a central catheter within 48 hours before the development of the bloodstream infection, after exclusion of other foci (30).
Bloodstream infection is the most common serious complication of indwelling vascular catheters. Although catheter colonization is a prelude to catheter-associated infection, most colonized catheters do not become clinically infected (18). Therefore, a significant reduction in the rate of catheter colonization does not, in and of itself, constitute proof of clinical efficacy. The ultimate proof of clinical efficacy is a significant reduction in the rate of CRBSI in a sufficiently powered, prospective, randomized clinical trial. As a corollary, if an inadequately powered clinical trial that fails to demonstrate a significant reduction in the rate of CRBSI despite a significantly lower rate of catheter colonization in the experimental versus control group, it is implied that the experimental strategy is either not clinically protective or needs to be examined in a larger clinical trial. In that regard, a properly conducted meta-analysis that adjusts well to confounding variables may help address the benefit of a potentially preventive approach. A critical analysis of the peer-reviewed literature allows the categorization of potentially preventive measures into three groups: (i) approaches that do not significantly reduce catheter colonization or CRBSI, (ii) approaches that significantly reduce catheter colonization but not CRBSI, and (iii) approaches that significantly reduce CRBSI. Although the most desirable impact of potentially preventive measures is a reduction in mortality associated with CRBSI, it would be impractical to conduct a several thousand patient clinical trials that would be sufficiently powered to assess this outcome, which has a relatively low incidence—equivalent to the incidence of CRBSI times the risk of dying from CRBSI.
Approaches that do not Significantly Reduce Catheter Colonization or CRBSI
This represents the most investigated approach in this category and focuses on modification of the catheter surface with different silver-containing moieties. Not only did in vitro studies yield conflicting findings with regards to efficacy—since some showed reduced bacterial adherence to the surfaces of polyurethane silver-coated catheters (31) and others indicated that the use of silicone silver-coated catheters is ineffective (32)—but the results of animal models also yielded inconclusive results (33). Although one prospective randomized clinical trial reported that silver-coated central venous catheters were protective (34), subsequent prospective randomized clinical trials found no evidence of clinical efficacy (35,36). The most recent assessment showed that short-term, central venous catheters impregnated with silver ions bonded to an inert ceramic zeolite reduce neither catheter colonization nor CRBSI (37). Not only is the silver application to the surface of short-term catheters mostly ineffective, but its incorporation onto the surface of long-term catheters can negatively impact the incidence of infection and cause adverse events. For instance, a prospective randomized clinical trial of tunneled long-term (mean dwell time, 92 days) hemodialysis catheters demonstrated a statistically insignificant trend for higher rates of catheter colonization (2.8 vs. 1.3 cases per 1,000 catheter-days) and catheter-related infection (1.8 vs. 1.1 cases per 1,000 catheter-days) in patients with silver-coated versus silver-uncoated catheters, respectively (38). In addition to being clinically ineffective, the silver-coated hemodialysis catheters were removed in 2 of 47 (4%) patients because of the chronic development of hyperpigmented skin lesions at the site of catheter insertion, thereby contributing to the decision to abandon the clinical use of that particular silver-coated catheter (38). Several factors are responsible for the poor clinical efficacy of silver-coated catheters (39):
- Since incorporated silver molecules do not effectively leach off the surface of most coated catheters, they do not produce effective zones of inhibition around the catheter surface that would ensure access of the coating agents to biofilm-embedded organisms.
- Silver tends to bind to host proteins, thereby resulting in lower concentration of free active silver molecules.
- The antimicrobial activity of silver can be impaired in the presence of bodily fluids.
Catheters Coated with Benzalkonium Chloride
In vitro studies showed that heparin-coated catheters possess some antimicrobial activity, possibly attributable to the weak antiseptic benzalkonium chloride, which is applied to the catheter surface primarily for its surfactant activity to allow bonding with heparin (40). However, small prospective randomized clinical trials failed to show a decrease in the rate of CRBSI associated with the benzalkonium chloride–coated catheters (41,42).
Approaches that Significantly Reduce Catheter Colonization But Not CRBSI
Dipping Catheters in Antibiotic Solutions
This bedside approach relies on dipping positively charged surfactant (usually tridodecyl methyl ammonium chloride, TDMAC)-pretreated catheters in a solution of negatively charged antibiotics such as cephalosporins and glycopeptides just prior to catheter insertion. Although a prospective randomized clinical trial showed that short-term central venous and arterial catheters that were pretreated with TDMAC and dipped at the bedside in cefazolin were sevenfold less likely to be colonized than undipped catheters, there was no demonstrated impact on the occurrence of CRBSI (43). A not-so-well-designed prospective randomized clinical trial also reported that immersion of short-term central venous catheters in vancomycin just prior to insertion was associated with a 22% reduction in the rate of catheter colonization (defined in that study as any level of bacterial growth by roll-plate culture of the catheter tip) as compared with nonimmersed catheters (44). The drawbacks of dipping catheters in vancomycin are the absence of impact on the incidence of CRBSI and the occurrence of Candida overgrowth.
Catheters Coated with Silver–Platinum–Carbon
A prospective randomized clinical trial showed that short-term central venous catheters coated with silver–platinum–carbon (so-called Oligon) were significantly less likely to become colonized than conventional uncoated catheters (18.6% vs. 29.6%, p = 0.003) (45). However, there was no significant reduction in the rate of bloodstream infection associated with coated versus uncoated catheters (3.3% vs. 4.3%).
Approaches that Significantly Reduce CRBSI
Quite understandably, the most recent CDC guidelines included new recommendations for using all of the following clinically protective measures, including cutaneous antisepsis with chlorhexidine (category IA), maximal sterile barriers (category IA), and catheters coated with the combinations of chlorhexidine plus silver sulfadiazine, or minocycline plus rifampin (category IB) (23).
Cutaneous Antisepsis with Chlorhexidine
The objective of antiseptic cleansing of the skin is to achieve a major reduction—preferably a greater than or equal to 3 log reduction—in bacterial concentration on the skin surface at the catheter insertion site. Prompted by an expanding body of supporting evidence, almost a dozen clinical guidelines issued by various scientific and regulatory organizations, both individually and in collaboration, have recommended the use of chlorhexidine-containing preparations rather than povidone–iodine or alcohol for cleansing the skin around the vascular catheter insertion site (23,46).
A prospective randomized clinical trial showed that vascular catheters inserted when using 2% aqueous chlorhexidine were significantly less likely to become colonized than catheters placed by using either 10% povidone–iodine or 70% alcohol (5). More important, CRBSI was fourfold to fivefold less likely in the former group than in the latter two groups (0.55% vs. 2.6% and 2.3%, respectively). In a meta-analysis of eight prospective randomized clinical trials that included a total of 4,143 vascular catheters, the relative risk of CRBSI was twice as high among patients who receive povidone–iodine versus chlorhexidine (47). The superiority of chlorhexidine over iodophor and alcohol could be predicted by comparing their characteristics, as shown in Table 32.1. Unlike iodophor and alcohol, chlorhexidine provides a residual and persistent antimicrobial activity that is not impaired by exposure to organic matter such as blood, does not irritate the skin, and has minimal absorption through the skin. Although 2% chlorhexidine compounds appear to be optimal in terms of both efficacy and safety, only recently has aqueous chlorhexidine become available in the United States, where the most frequently used form of chlorhexidine is in combination with an alcohol, usually isopropyl alcohol. As Table 32.2 delineates, the combination of chlorhexidine and alcohol has many more favorable properties than the combination of iodophor and alcohol. These recognizable differences have contributed to the escalating application of antiseptic solution that contains the combination of chlorhexidine and alcohol on the skin surrounding the insertion site of vascular catheters.
|TABLE 32.1 Comparison of Individual Antiseptic Agents|
|TABLE 32.2 Comparison of Solutions that Contain Combinations of Antimicrobial Agents|
Maximal Sterile Barriers
In contrast to traditional sterile precautions that include the use of gloves and a small drape, maximal sterile precautions comprise the use of gloves, a large drape, a cap, a mask, and a gown. When compared in a prospective randomized fashion, the use of maximal sterile barriers versus traditional sterile precautions when inserting long-term (mean duration of placement, 70 days), noncuffed silicone vascular catheters was associated with a significantly lower incidence of catheter colonization (0.3 vs. 1 per 1,000 catheter-days; p = 0.007) and CRBSI (0.08 vs. 0.5 per 1,000 catheter-days, p = 0.02) (48,49). Although this protective measure is intended to be used for insertion of all central venous catheters, it is currently used less frequently outside the ICUs and specialty care areas in which reside patients with a high risk of infection, including those with bone marrow transplant or leukemia.
Catheters Coated with Chlorhexidine and Silver Sulfadiazine
There exist two catheters coated with the combination of chlorhexidine and silver sulfadiazine. The first-generation, and most studied, catheter (50–59) has antimicrobial agent incorporated only along the external surface of the catheter. The second-generation catheter differs in two ways from the first-generation catheter: Both antimicrobial agents are incorporated onto the external and internal surfaces, and it contains three times the amount of chlorhexidine (60).
The largest prospective randomized clinical trial of the first-generation devices coated with chlorhexidine/silver sulfadiazine in 403 short-term (mean duration of placement, 6 days), polyurethane central venous catheters demonstrated a significant reduction in the rate of catheter colonization (13.5% vs. 24.1%; p = 0.005) and CRBSI (1.0% vs. 4.6%; p = 0.03) as compared with uncoated catheters (53). Although most other clinical trials (50,54–59) showed that chlorhexidine/silver sulfadiazine–coated catheters were significantly less likely to be colonized than uncoated catheters, they could not demonstrate a significant reduction in the rate of CRBSI; these studies were not sufficiently powered to detect significant differences in the rates of CRBSI. Additionally, however, a meta-analysis of 12 clinical trials showed that these antimicrobial-coated catheters resulted in a significant reduction in the rates of both catheter colonization (odds ratio = 0.44; p < 0.001) and CRBSI (odds ratio = 0.56; p = 0.005) (61).
Since this first-generation chlorhexidine/silver sulfadiazine–coated catheter provided only short-lived (about 1 week) antimicrobial activity, and only along the external surface of the catheter (56), it was not likely to protect against infection of long-term catheters that frequently become contaminated with bacteria migrating from the contaminated hub along the internal surface of the catheter. Not unexpectedly, a large (680 catheters) prospective randomized clinical trial showed that placement of chlorhexidine/silver sulfadiazine–coated central venous catheters for a mean of 20 days in patients with hematologic malignancy did not reduce the rate of CRBSI as compared with uncoated catheters (5% vs. 4.4%) (62).
The second-generation chlorhexidine/silver sulfadiazine–coated catheters have a longer durability of antimicrobial activity than the first-generation catheters (63). A recent report of a large (842 catheters) prospective randomized clinical trial demonstrated that second-generation chlorhexidine/silver sulfadiazine–coated polyurethane short-term central venous catheters are less likely to become colonized than uncoated catheters (9% vs. 16%, p < 0.01) but had a statistically insignificant trend for a lower rate of CRBSIs (0.3% vs. 0.8%) (60). Since the incidence of CRBSI in the uncoated catheter group was lower than usual, this study may not have had sufficient power to assess the desired outcome. Because both the first-generation and second-generation chlorhexidine/silver sulfadiazine–coated catheters generally reduce catheter colonization to a similar degree, it is reasonable to regard these two catheters as being equally protective against infection.
Catheters Coated with Minocycline and Rifampin
This unique combination of antibiotics was selected for the following reasons:
- Both agents are active against the vast majority of staphylococcal isolates, including MRSA and MRSE (64).
- The combination of agents provides broad-spectrum antimicrobial activity against most pathogens that can cause CRBSI, thereby reducing the likelihood of developing superinfection by gram-negative bacteria or Candida spp. (65,66).
- Since minocycline and rifampin have different mechanisms of activity, with minocycline retarding protein synthesis and rifampin inhibiting DNA-dependent RNA polymerase, it is unlikely that a bacterial strain will become concomitantly resistant to both agents.
Unlike many antibiotics—including vancomycin, ciprofloxacin, and the aminoglycosides—that are much less active against biofilm bacteria than planktonic bacteria, rifampin (67) and minocycline (68) are particularly active against biofilm-embedded bacteria.
The clinical efficacy of this catheter surface modification was initially confirmed in a prospective randomized clinical trial that showed that polyurethane short-term (mean duration of placement, 6 days) central venous catheters coated with minocycline and rifampin were significantly less likely than uncoated catheters to become colonized (8% vs. 26%; p < 0.001) and cause bloodstream infection (0% vs. 5%; p < 0.01) (69). In a large prospective randomized clinical trial of 738 catheters, polyurethane short-term (mean duration of placement, 8 days) central venous catheters coated with minocycline and rifampin were more protective than the first-generation chlorhexidine/silver sulfadiazine–coated catheters, with a 3-fold lower rate of catheter colonization (7.9% vs. 22.8%; p < 0.001) and a 12-fold lower rate of CRBSI (0.3% vs. 3.4%; p < 0.002) (70).
|TABLE 32.3 Comparison of Antibiotic Dipped Versus Antimicrobial Coating of Vascular Catheters|