There has been a correlation of catheter site and risk of infection in the past; however, one study has shown that there is no difference in the risk of CRBSI when comparing the femoral and internal jugular sites (30) and that concept has been enforced in the CDC guidelines (31). Avoiding the use of the femoral is for reasons other than infections such as thrombosis (31). In general, the preferential site for nontunneled CVCs is the subclavian vein (31). In terms of catheter site dressings, the CDC recommends using a sterile gauze or sterile transparent dressings to cover the catheter which are to be changed every 2 or 7 days, respectively (31). With the exception of dialysis catheters, the use of antibiotic creams or ointments is not advised as they may promote antimicrobial resistance and fungal growth. In the setting of increasing central line–associated bloodstream infection (CLABSI) rates, the use of chlorhexidine (CHX)-impregnated dressings is acceptable for short-term catheters provided the patient is older than 2 months (31).
Clinical manifestations of CRBSIs can be divided into two categories: local and systemic. Local manifestations include erythema, edema, tenderness, and purulent discharge. These signs and symptoms are neither sensitive nor specific, and cannot be relied on to identify catheter colonization or CVC-related BSI. On the one hand, they could be completely absent, especially in immunocompromised and neutropenic patients. On the other hand, peripherally inserted central catheters (PICCs) (inserted in the basilic or cephalic veins) are associated with a 26% rate of sterile local exit site inflammation secondary to irritation of small veins (i.e., cephalic vein) by insertion of a large catheter (32); to this must be added the finding that coagulase-negative staphylococci, the most frequent pathogen involved, incites little local or systemic inflammation (33).
The CDC suggests the following definitions:
- Exit-site infection: Purulent drainage from the catheter exit site, or erythema, tenderness, and swelling within 2 cm of the catheter exit site, and colonization of the catheter, if removed, with or without concomitant BSI.
- Port-pocket infection: Erythema, tenderness, induration, and/or necrosis of the skin or subcutaneous tissues either over or around the reservoir of a totally implanted IVD, and colonization of the device if removed, with or without concomitant BSI.
- Tunnel infection: Erythema, tenderness, and induration of the tissues above the catheter and more than 2 cm from the exit site, along the subcutaneous tract of a tunneled catheter and colonization of the catheter if removed, with or without concomitant BSI.
The systemic features of CRBSIs are generally indistinguishable from those of secondary BSIs arising from other foci of infection, and include fever and chills, which may be accompanied by hypotension, hyperventilation, altered mental status, and nonspecific gastrointestinal manifestations such as nausea, vomiting, abdominal pain, and diarrhea. Deep-seated infections such as endocarditis, osteomyelitis, retinitis, and organ abscess may complicate CRBSIs caused by some virulent organisms such as S. aureus, P. aeruginosa, and Candida albicans.
A clinical diagnosis of CRBSI is frequently inaccurate. At this juncture, it is worth noting the difference of CRBSI and CLABSI, as these terms are often used interchangeably but incorrectly. CLABSI is used by CDC’s National Healthcare Safety Network (NHSN) to describe any primary BSI in a patient who had a central line within 48 hours of the development of the BSI, and is not related to an infection at another site, such as pyelonephritis or pneumonia. However, other sources of infection, such as mucositis, that are less identifiable, may have led to the BSI, so the diagnosis of CLABSI is nonspecific and overestimates the true incidence of CRBSI. On the other hand, CRBSI is a subcategory of CLABSI for which specific laboratory testing has been done and identifies the catheter as the source of BSI.
Removal of the CVC has been mandatory to prove the CRBSI. Microbiologic methods requiring removal of the CVC were studied with the semi-quantitative roll-plate catheter cultures, developed by Maki et al. (34) in 1977, and considered the gold standard. However, the majority of the catheters were removed unnecessarily, exposing the patient to the complications related to reinsertion of a new CVC and adding to the cost of care. To prevent that, techniques allowing accurate diagnosis without removing the line have been elaborated; these are reviewed (Table 88.1.2).
Catheter-Sparing Diagnostic Methods
Simultaneous Quantitative Blood Cultures
This method consists of obtaining paired quantitative blood cultures (QBCs) simultaneously from the CVC and a peripheral vein. The target is to have both samples drawn less than 10 minutes apart with the same volume of blood. The hypothesis is that the higher load of organisms on the internal lumen of the CVC signifying CRBSI would translate into a colony count from the CVC greater by many folds than the peripheral stick. A CVC/peripheral ratio of CFU/mL of 3:1 has been chosen by the Infectious Diseases Society of America (IDSA) to represent true infection, meaning that the colony count of microbes grown from blood obtained through the catheter hub must be at least threefold greater than the colony count from blood obtained from a peripheral vein (35). However, there are variable reports among the literature which recommend up to a 10:1 ratio (5). A meta-analysis found that QBC is the most accurate, with a pooled sensitivity of 75% to 93% and specificity of 97% to 100% (36). That same study recommended not culturing all catheter tips, but rather culturing only if CRBSI is suspected clinically. This method of QBCs is limited by the fact that it is expensive and labor intensive, in addition to the difficulty in obtaining samples through the catheter in some cases (37). The diagnosis of CRBSI should be strongly considered, though not definitively established, when QBCs are drawn from two lumens of the CVC and the colony count for blood from one lumen is at least threefold greater than that from the other lumen.
|TABLE 88.1.2 Sensitivity and Specificity of Tests Used in the Diagnosis of CRBSIa|
Differential Time to Positivity
The differential time to positivity (DTP) of qualitative paired CVC and peripheral blood culture has been a more practical test for centers that lack the logistics for QBCs, especially with the introduction of automated radiometric blood culture systems that record the time at which a culture turns positive. The hypothesis suggests that time to positivity of a culture is closely related to the inoculum size of micro-organisms. The technique involves measuring the difference between the time required for culture positivity in simultaneously drawn samples of catheter blood and peripheral blood. In a single-center trial evaluating for CRBSI, a DTP of 120 minutes was associated with 81% sensitivity and 92% specificity for short-term catheters and 93% sensitivity and 75% specificity for long-term catheters (38). A meta-analytic study showed that the DTP of 120 minutes predicts CRBSI, with a pooled sensitivity and specificity of 89% and 87% for short-term catheters and 90% and 72% for long-term catheters, respectively (36). This technique also demands a simultaneous blood draw (within 10 minutes) from the line and the peripheral vein with the same amount of blood. One limitation of this study is that its sensitivity could be compromised when antibiotics are given intraluminally at the time of drawing the blood cultures through the catheters (38).
Acridine Orange Cytospin Leukocyte Technique
This test involves 1 mL of ethylenediaminetetraacetic acid (EDTA) blood aspirated through the CVC. The sample is added to 10% formalin saline solution for 2 minutes; the sample is then centrifuged, the supernatant decanted, and the cellular deposit homogenized and cytocentrifuged. A monolayer is stained with 1 in 10,000 acridine orange and viewed under ultraviolet light; a positive test is indicated by the presence of any bacteria (39). This method is expensive but takes only 30 minutes, with a sensitivity of 87% and specificity of 94% (40). This technique has been tested only by a small group of investigators and is not easy to perform correctly in order to reproduce the Kite method (41). One trial showed that this technique anticipated CRBSI earlier than routine measures (42); it is not recommended by the current guidelines of IDSA.
Fluorescence In Situ Hybridization on Peptide Nucleic Acid Probes
Fluorescence in situ hybridization (FISH) using peptide nucleic acid (PNA) probes is a novel technique in detecting several organisms (43). PNA probes are basically similar in structure to DNA or RNA, but have an uncharged backbone which accounts for its superior stability and improved hybridization when compared to DNA or RNA (43). These characteristics of PNA probes improve binding to certain molecules such as rRNA which makes FISH PNA a superior diagnostic test (43). Interestingly, a study comparing acridine orange cytospin leukocyte technique, FISH PNA and DTP found similar results in terms of sensitivity, specificity, positive predictive value, and negative predictive value (91%, 100%, 100%, and 97%, respectively) (44). In another study, even though FISH PNA was a successful diagnostic tool in patients who experience a BSI, positive results from random CVC sampling did not predict clinical progression to CRBSI as this phenomenon was most likely due to CVC colonization (45).
Diagnostic Methods Requiring Catheter Removal
Semi-Quantitative Roll-Plate Catheter Culture
This method was described by Maki et al. (46) in 1977 and remains the international reference diagnostic method. It consists of rolling a 3- to 5-cm section of the distal tip of the CVC at least four times back and forth over an agar plate surface and incubating overnight. A cutoff of ≥15 CFU defines catheter colonization; if at the same time, a peripheral culture grows the same organism, then a CRBSI is diagnosed. However, this method does not sample the internal lumen of a CVC that is the source of the infection in long-term catheters. Nevertheless, pooled sensitivity and specificity in 14 trials involving short-term catheters were 84% and 85%, respectively (36); this number decreased to 45% and 75%, respectively, with long-term CVCs (i.e., those with more than 30 days of dwell time) (10,47).
Quantitative Catheter Cultures
This type of culture involves flushing or sonicating a catheter segment in broth with the target of retrieving organisms from both surfaces of the line. A threshold of ≥100 CFU (48) correlated best with colonization, although older work (49) used a 1,000 CFU cutoff. CRBSI would be defined by the cutoff of 100 CFU accompanied by a high clinical suspicion and absence of evidence of other sites of infection. As would be expected, the sonication method had a higher sensitivity than the roll-plate method for long-term CVCs (10); however, both sonication and vortexing had the same sensitivity and specificity of the roll-plate method for short-term CVCs (50). Meta-analysis revealed a pooled sensitivity and specificity of 82% and 89% for short-term catheters and 83% and 97% for long-term catheters, respectively (36).
It should go without saying—but obviously does not—that CVCs should only be used when medically necessary, and should be removed as soon as possible to prevent potential complications. In a large study that included 1,981 ICU months of data, collective antiseptic measures consisting of hand-washing, maximal sterile barriers during insertion, cutaneous antisepsis with CHX, avoidance of femoral site, and removal of CVCs determined to be unnecessary were associated with a significant decrease in CRBSI rate—from 7.7 per 1,000 catheter days to 1.4 per 1,000 catheter days (p < 0.001) over 18 months of follow-up (51). In 1992, Cobb et al. (13), in an attempt to reduce catheter-related infection, conducted a controlled study whereby CVCs or pulmonary artery catheters were changed or exchanged over guidewire every 3 days; the former procedure actually resulted in an increase in the risk of mechanical complications, whereas the latter technique increased the risk of BSI. Table 88.1.3 provides a listing of preventive strategies to decrease the risk of CVC colonization. We review below the novel strategies implemented by the Healthcare Infection Control Practices Advisory Committee (HICPAC) and other professional organizations, including the IDSA, Society for Healthcare Epidemiology of America (SHEA), and American Society of Critical Care Anesthesiologists (ASCCA) aiming at controlling all factors that could lead to colonization of the CVC, and hence decreasing the rate of CRBSI.
|TABLE 88.1.3 Preventive Measures to Decrease the Risk of Colonization of Central Venous Catheters|
The HICPAC/CDC guidelines recommend with level 1A evidence—data derived from multiple randomized clinical trials proving general agreement on its effectiveness—the usage of 2% CHX-based preparation (52). Maki et al. (53) prospectively randomized 68 ICU patients to 10% povidone–iodine, 70% alcohol, or 2% aqueous CHX to disinfect the site before insertion of CVCs and for site care every other day thereafter, and demonstrated that 2% aqueous CHX preparation tended to decrease the rate of CRBSI substantially; using lower concentrations of CHX decreased the effectiveness of this method. Tincture of chlorhexidine gluconate 0.5% is no more effective in preventing CRBSI or CVC colonization than 10% povidone–iodine, as demonstrated by a prospective, randomized study in adults (54). A meta-analysis of eight randomized trials found an overall reduction of 49% in catheter-associated BSIs when a disinfectant containing CHX was used (55). A French trial randomly assigned 1,181 ICU patients to CHX-based preparation and 1,168 to povidone–iodine. The CHX preparation was associated with lower incidence of CRBSI, 0.28 versus 1.77 per 1,000 catheter days (95% confidence interval [CI] 0.05 to 0.41, p = 0.0002) (56). Finally, the use of a dilute CHX solution for daily baths has been shown to decrease CRBSIs—among other infections—in a variety of settings (57–63).
Maximal Sterile Barrier
This involves wearing a sterile gown, gloves, and a cap, and using a large drape similar to those used in the operating room during the insertion of catheters as opposed to the regular precautions consisting of sterile gloves and a small drape only. The HICPAC/CDC guidelines recommend this technique while inserting CVCs, PICC lines, and pulmonary artery catheters (52) (category 1A) based on a number of studies (64–66). A prospective study conducted by Raad et al. (64) with long-term, nontunneled silicone CVCs and PICC lines in a cancer patient population demonstrated not only a reduction of CRBSIs (p = 0.03), but also that this practice was cost effective. Mermel et al. (65), in another prospective study with pulmonary artery catheters, found that less stringent barrier precautions were associated with a significantly increased risk of catheter-related infection (relative risk = 2.1, p = 0.03). Of note is that this technique failed to reduce the colonization of CRBSIs associated with arterial catheters (66). It has been shown that dedicated physician education courses can improve compliance with maximal sterile barrier and decrease the incidence of CRBSI (67).
Antimicrobial Catheter-Lock Solutions
Antimicrobial catheter lock involves flushing the catheter lumen and then filling it with 2 to 3 mL of a combination of an anticoagulant plus an antimicrobial agent. The dwell (lock) time varies between clinicians, but 20 to 24 hours is the most preferred. However, this might not be possible if the catheter has to be used (68). This intervention has often been used in long-term CVCs that remain in place longer than 30 days. Henrickson et al. (69) showed that a combination of vancomycin and heparin, with or without ciprofloxacin, was equivalent, but each was superior to heparin alone. Of six studies, four revealed a significant reduction in CRBSI with the above lock solution (70–72), and two demonstrated no benefit (73,74). However, vancomycin–heparin lock solutions may promote the risk of vancomycin resistance and the risk of superinfection with GNB and Candida is present since the vancomycin spectrum is limited to gram-positive bacteria. A meta-analysis concluded that the use of a vancomycin lock solution in high-risk patient populations being treated with long-term central IVDs may reduce the risk of BSI with a risk ratio of 0.34 (95% CI, 0.12 to 0.98; p = 0.04) (75).
Minocycline and EDTA (M-EDTA), another lock solution, was reported in a prospective randomized trial to significantly reduce the risk of catheter colonization and infection when compared with heparin in long-term hemodialysis CVCs (76). This solution was superior in an in vitro biofilm model and in an animal model to vancomycin–heparin lock solution (76–78). A clinical study of pediatric cancer populations showed that M-EDTA significantly reduces the risk of catheter infection and colonization when compared to heparin (79).
In a prospective nonrandomized study of tunneled CVCs in a pediatric cancer population, ethanol as a lock solution reduced the risk of relapse of CRBSI and was well tolerated (80). However, symptoms of fatigue, nausea, dizziness, and headache were reported. The study involved filling the catheter lumen with 2.3 mL of a 74% ethanol solution for 20 to 24 hours. The solution was then flushed through to prevent clotting inside the catheter. Each port was alternately blocked for 3 days, allowing the unblocked port to be used. In a study by Raad et al. (81), M-EDTA in 25% ethanol was found to be highly effective in eradicating organisms embedded in biofilm, even after a short exposure of 15 to 60 minutes. Hence, the addition of a low concentration of ethanol (25%) to M-EDTA could expedite its activity and decrease the necessary dwell time. A prolonged dwell time of more than 8 hours is often required for nonalcohol-based antibiotic lock solutions, which makes their use limited, particularly in critically ill patients or patients requiring TPN.
A meta-analytic study from 2014 (82), which included 23 studies and 2,896 patients, showed that antimicrobial lock solutions led to a 69% reduction in CLABSI compared with heparin, without significantly causing catheter failure due to noninfectious complications. However, one must keep in mind that all of the trials were done in special population patients, such as hemodialysis and oncology patients, patient receiving TPN, and so forth. As there is some concern for the emergence of bacterial resistance associated with utilization of a sole antibiotic agent and not in combination, some have recommended caution when it comes to the widespread indiscriminate prophylactic use of antibiotic-based catheter-lock solutions (82).
Given the fact that heparin, an effective antithrombotic agent, was shown in a study by Shanks et al. (83) to promote the kinetics of biofilm formation by S. aureus by enhancing cell–cell interaction and the concern of antibiotic resistance, the future trend will be to move away from heparin and/or antibiotic-based lock solutions. Based on that, several studies have tried to compare heparin with other agents. One study has shown superiority of a combination of ethanol and sodium citrate over heparin in terms of CRBSI prevention and catheter survival (84). Another comparative study has shown superiority of minocycline–EDTA, taurolidine–polyvinylpyrolidine, or ethanol over other current antibiotic-based lock solutions (85). Recently, a novel antimicrobial solution containing 15 μg of nitroglycerin, 4% of citrate, and 22% of ethanol has shown to have in vitro activity against all biofilm-producing organisms in as a little at 2 hours (86). However, and while this particular field is very promising, there has yet to be large clinical trials confirming the in vitro activity of these novel nonantibiotic, nonheparin-based lock solutions (85).
Antimicrobial Impregnation of Catheters
This technique consists of the impregnation of the external and/or internal surface of the catheter with antiseptic or antibiotics; the slow release of antimicrobials would prevent initial bacterial adherence and biofilm formation, with virtually undetectable serum levels. The HICPAC/CDC, with a category 1B, recommends the use of the coated CVCs described herein. The first-generation catheters were impregnated on the external surface with CHX and silver sulfadiazine (CHX/SSD) (Arrow Gard and Arrow Gard Plus, Arrow International, Inc.). That technique lowered the rate of CRBSI from 7.6 cases per 1,000 catheter days to 1.6 cases per 1,000 catheter days (p = 0.03), with a decrease in the rate of colonization (relative risk, 0.56 [95% CI, 0.36 to 0.89]; p = 0.005) (9); the estimated cost savings per CVC insertion was $196 (87). However, three subsequent studies failed to show that difference (88–90). This was explained by the fact that short-term catheter infection is due to external colonization, whereas long-term CRBSIs due to internal colonization are not prevented by external coating. Moreover, Mermel (6) showed that these catheters do not protect if the CVC dwell time is more than 3 weeks, secondary to wearing off of the antimicrobial activity. The second-generation CHX/SSD-coated catheters were impregnated on both surfaces. In a multicenter, randomized double-blind prospective study from 14 French ICUs, second-generation catheters failed to decrease the rate of CRBSI (91,92) when compared to noncoated catheters, although they significantly decreased the rate of colonization (11/1,000 catheter days to 3.6/1,000 catheter days, p = 0.01).
In 1997, Raad and colleagues (92) developed a catheter impregnated on both surfaces with minocycline–rifampin (M/R). In a prospective randomized, double-blind trial, M/R CVCs showed more efficacy when compared to noncoated catheters. Another prospective trial comparing M/R catheters with first-generation CHX/SSD-impregnated catheters found that the former were three times less likely to be colonized (p < 0.001), and CRBSI was 12-fold more likely to occur in the CHX/SSD catheters (p < 0.002) (93). The use of antibiotic-impregnated CVCs in medical and surgical units was associated with a significant decrease in nosocomial BSIs, including vancomycin-resistant enterococci (VRE) bacteremia, catheter-related infections, and length of hospital and ICU stay (94). Furthermore, the M/R-coated catheters saved $9,600 per each CRBSI and $81 per each catheter placed when compared to first-generation CHX/SSD (95).
The concern for the emergence of antibiotic-resistant organisms was raised with the catheters coated with M/R. Four prospective studies evaluated the skin at the catheter insertion site before and after the insertion of antibiotic-coated catheters and failed to detect any emergence of resistance (92,93,96,97). A retrospective review of the M/R-coated CVC experience in bone marrow transplant patients also detected no emergence of resistance of staphylococci to either component (98). In a series of prospective randomized studies, the M/R-coated CVCs were shown to bring the risk of CRBSI to a level ≤ 0.3 per 1,000 catheter days in nontunneled, noncuffed CVCs (92,93,97), lower than the 1.4 per 1,000 catheter days achieved with multiple other aseptic measures applied collectively (such as the maximal sterile barrier, CHX cutaneous antisepsis, and hand hygiene).
Two studies by Raad et al. and Jamal and colleagues (99,100) found that minocycline–rifampin impregnated CVCs coated internally and externally with CHX, termed CHX–M/R catheters, were superior to CHX–/SSD (chlorhexidine–silver sulfadiazine) or M/R-coated catheters in preventing biofilm formation and catheter colonization particularly when it comes to Pseudomonas and Candida. A novel approach using gendine (CHX and gentian violet)-coated CVCs showed promising results in terms of prevention of biofilm formation with no acute systemic exposure of CHX or gentian violet (101); however, clinical trials involving the use of gendine-coated catheters have yet to be initiated.
When all are said and done, CHX/SSD and M/R catheters both reduce CRBSI when compared to noncoated ones. The HICPAC/CDC recommends the use of either if the medical center continues to have higher than national average CRBSI rates despite successful implementation of provider education, maximal sterile barrier and use of CHX preparation with alcohol for skin antisepsis.
Other catheters incorporate silver, platinum, and carbon (SPC) into the polyurethane, allowing topical silver ion release (Vantex CVC with oligan, Edwards Life Sciences, Irvine, CA). One prospective randomized study compared these catheters to the M/R-coated type; the latter was more efficacious in reducing, to a significant degree, CVC colonization with gram-positive and gram-negative bacteria (p = 0.039); however, the CRBSI rates were low and similar between the two groups (102). In another prospective, randomized, controlled, open-label, multicenter clinical trial, the SPC CVCs failed to show any benefit in reducing CRBSI or colonization (103). A meta-analysis study failed to prove any association between reduced rates of colonization or CRBSI and the use of silver-impregnated catheters (104).
The management of CRBSIs involves confirming the source and cause of infection, determining the choice of antimicrobials, determining the duration of therapy, and deciding whether to remove the invasive device. Confirmation of the infection is dependent on the diagnostic measures outlined above. The duration of therapy depends on whether the infection is complicated (i.e., by a septic phlebitis or endocarditis) or uncomplicated.
Coagulase-negative staphylococci are the primary organisms involved in CRBSIs because they are the most common skin organisms; however, and for the same reason, they are the most frequent blood contaminants. One study indicated that QBC collected through CVC, with a cutoff point of 15 CFU/mL, could be a useful laboratory criterion, together with positive clinical findings, for differentiating true bacteremia from false-positive contaminated blood cultures, with a sensitivity of 96%, specificity of 94%, positive predictive value of 86%, and negative predictive value of 98% (105); the IDSA guidelines recommend removing the CVC and treating for 5 to 7 days. Otherwise, if the CVC is to be retained, duration of treatment should be 10 to 14 days, and antibiotic lock therapy should be considered (106). Leaving the CVC in place carries a risk of recurrence of 20% (107). Finally, in the absence of endovascular or orthopedic hardware, and with catheter removal, the patient can be monitored off antibiotics while new blood cultures are drawn to confirm the resolution of bacteremia.
Lock solutions used included vancomycin plus heparin. The limited activity of vancomycin against Staphylococcus embedded in biofilms (73,75,108) led investigators to consider other alternatives; minocycline and EDTA, ethanol, or the triple combination (109,110) was used as an alternative. While systemically, vancomycin has been the most frequently used glycopeptide, dalbavancin, a new, long-acting glycopeptide that is dosed weekly, was noted to be superior to vancomycin for adult patients with CRBSIs caused by coagulase-negative Staphylococcus and S. aureus, including methicillin-resistant S. aureus (MRSA) in a phase 2, open-label, randomized, multicenter study; the side effect profile was comparable (111). Linezolid and daptomycin were also used successfully (112,113).
S. aureus CRBSI is associated with high rates of deep-seated infection such as osteomyelitis, septic phlebitis, and endocarditis (114). In addition, Fowler et al. (114) showed that patients whose IVD was not removed were 6.5 times more likely to relapse or die of their infection than were those whose device was removed. IDSA guidelines recommend removing the CVC, as this results in a more rapid response and lower relapse rate but, at the same time, gives the option of keeping it and initiating systemic and lock solutions in the rare and extreme cases of lack of other vascular access, bleeding diathesis, and quality-of-life issues intervene (106). Capdevila et al. (115) used the antibiotic lock technique in addition to standard parenteral therapy for patients with a hemodialysis catheter–related infection. All 40 CRBSIs—including all 12 cases reported to involve S. aureus—were cured and the catheter salvaged. The lock solutions most frequently used in vivo and in vitro are vancomycin plus heparin, or minocycline plus EDTA (71,110). However, the former combination—with or without ceftazidime, depending on the organism—was associated with a 60% failure rate in hemodialysis MRSA catheter infections (116). Another study showed that even though systemic antibiotic therapy was not successful in eradicating most CRBSIs without catheter removal, attempted CVC salvage appeared to have not increased the complication rate even in the setting of S. aureus (117). Low-concentration ethanol (25%) is another very appealing component for use in combination lock solutions; Raad and colleagues (81) found that the combination of minocycline–EDTA in 25% ethanol was highly efficacious in eradicating S. aureus in biofilm within 60 minutes of dwell time.
For methicillin-sensitive S. aureus, nafcillin or first-generation cephalosporins are the first-line agents (100). Vancomycin, linezolid, daptomycin, and dalbavancin (111–113) are all appropriate options for MRSA. Duration of therapy usually consists of 10 to 14 days of intravenous therapy if the CVC is removed, with no deep-seated infection present (106). If fever or bacteremia persists for more than 72 hours after catheter removal, transesophageal echocardiography should be performed to rule out IE, with the intravenous therapy duration expanded to at least 4 weeks (106,118). This is especially important as the frequency of IE in S. aureus bacteremia is 25% to 32% (119).
Enterococcus is the third most common pathogen seen in CRBSI, accounting for about 10% of nosocomial BSIs (120). The IDSA recommends catheter removal and treatment with systemic antibiotics, beginning with ampicillin as the first-line agent of choice; the organism can be treated with vancomycin if ampicillin-resistant. However, 60% of Enterococcus faecium and 2% of Enterococcus faecalis are now resistant to vancomycin (120); in such cases, linezolid and daptomycin are the agents of choice.
Antibiotics are recommended for 7 to 14 days in the setting of catheter removal, as well as long-term catheter salvage with systemic antibiotics and lock therapy. A transesophageal echocardiogram (TEE) should be pursued to evaluate for IE if the patient has prolonged bacteremia or fever more than 72 hours after the initiation of appropriate antimicrobial therapy, evidence of septic emboli, a new murmur, or embolic phenomena, all of which are also, in the cases of long-term catheters, indications of salvage therapy failure and the need for removal.
There are data regarding combination therapy for enterococcal CRBSI, namely a cell wall–active antimicrobial and an aminoglycoside. Several retrospective cohort studies found no statistically significant difference in outcomes when treated with combination therapy versus monotherapy (121–123).
GNB bacteremia is rarely due to a CVC; rather, it generally arises from a visceral source of infection such as the genitourinary, pulmonary, or gastrointestinal tracts. However, CRBSIs caused by such organisms as K. pneumoniae, Enterobacter spp., P. aeruginosa spp., Acinetobacter spp., and Stenotrophomonas maltophilia have been reported (124,125). Elting and Bodey (124) reported a 15-year experience of 149 episodes of septicemia caused by Xanthamonas maltophilia and Pseudomonas spp. in cancer patients where the CVC was the most common source. Hanna et al. (125) demonstrated that catheter removal within 72 hours of the onset of the catheter-related GNB was the only independent protective factor against the relapse of infection (OR, 0.13; 95% CI, 0.02 to 0.75; p = 0.02). IDSA guidelines (107) recommend removing nontunneled CVCs and treating for 10 to 14 days with systemic antibiotics. Patients at high risk for colonization or infection with multidrug-resistant gram-negative pathogen (i.e., critically ill, neutropenia) should be covered with either two antibiotics of different classes that provide gram-negative activity or a carbapenem as initial therapy. It is considered appropriate to attempt to salvage the CVC in certain situations (see above) using systemic and lock solution therapies. However, lock therapy for GNB CRBSIs is anecdotal; successful cases were salvaged using gentamicin, amikacin, or ceftazidime (106,118).
Five large prospective studies proved that catheter retention was associated with increased mortality and an increase in the mean duration of candidemia in cases of Candida CRBSI (126–130). Hung et al. (128) investigated the predisposing factors and prognostic determinants of candidemia in a Taiwan hospital, and concluded that higher severity scores, nonremoval of the catheter, persistent candidemia, and lack of antifungal therapy adversely affect the outcome. Raad and colleagues (131), in a retrospective study of 404 patients with candidemia and an indwelling CVC, using a multivariate analysis, demonstrated that catheter removal 72 hours or sooner after onset of candidemia improved the response to antifungal therapy exclusively in patients with catheter-related candidemia (p = 0.04). IDSA guidelines recommend removing the CVC and treating for 14 days after the last positive blood culture in uncomplicated cases; endophthalmitis merits 6 weeks of therapy (106). Further studies are needed to define the role of antifungal lock solution in these cases. Fluconazole and caspofungin were equivalent to amphotericin B in candidemia, but with a better safety profile (130,131); therefore, fluconazole or caspofungin should be considered in documented cases of catheter-related candidemia. If the rates of fluconazole-resistant Candida glabrata and Candida krusei in the hospital are high, an echinocandin (caspofungin, micafungin, or anidulafungin) would be the best alternative to amphotericin B.
PERIPHERALLY INSERTED CENTRAL CATHETERS
The use of PICCs is very common in cancer patients, patients receiving TPN, and long-term i.v. antibiotics. Their complication rates, including infection and thrombosis, are low in the outpatient setting (0.4 per 1,000 catheter days) (132,133). There has been speculation that perhaps PICCs are less prone to infection compared to CVCs in the critically ill; this has been proved to be untrue.
In a prospective study in which 115 patients had 251 PICCs placed, Safdar and Maki (127) showed that PICCs used in ICU patients are associated with a rate of CRBSI similar to CVCs placed in internal jugular or subclavian veins (2.1 vs. 2 to 5 per 1,000 catheter days). Nolan et al. (134) performed a retrospective cohort study of 200 PICCs and 200 CVCs placed in the medical ICU adults at Mayo Rochester between 2012 and 2013. Overall, thrombotic and infectious complications were rare following PICC and CVC insertion, with no significant difference in complication rates observed. Finally, the meta-analytic study of Chopra and colleagues (135) identified 23 studies that met eligibility criteria for comparing infection risk of PICC versus CVC. Thirteen studies reported CLABSI rates; PICC-related CLABSI occurred as frequently as CLABSI in CVC. The subcategory where PICC showed a significant infection reduction is in outpatients, not in critically ill, hospitalized patients.
- CVCs are as much a part of modern ICU practice as are mechanical ventilators and antibiotics.
- When CVCs are placed with the appropriate technique, accessed, and cared for, it is possible to use these devices while approximating a zero incidence of infection.
- PICCs are not lower risk for infectious complication when used in critically ill, hospitalized patients.
- As is typically true in the practice of critical care medicine, it is in the details that the battle is won or lost.