Prevention of Bloodstream Infections in Patients Undergoing Hemodialysis

Abstract

Bloodstream infections are an important cause of hospitalizations, morbidity, and mortality in patients receiving hemodialysis. Eliminating bloodstream infections in the hemodialysis setting has been the focus of the Centers for Disease Control and Prevention (CDC) Making Dialysis Safer for Patients Coalition and, more recently, the CDC’s partnership with the American Society of Nephrology’s Nephrologists Transforming Dialysis Safety Initiative. The majority of vascular access–associated bloodstream infections occur in patients dialyzing with central vein catheters. The CDC’s core interventions for bloodstream infection prevention are the gold standard for catheter care in the hemodialysis setting and have been proven to be effective in reducing catheter-associated bloodstream infection. However, in the United States hemodialysis catheter–associated bloodstream infections continue to occur at unacceptable rates, possibly because of lapses in adherence to strict aseptic technique, or additional factors not addressed by the CDC’s core interventions. There is a clear need for novel prophylactic therapies. This review highlights the recent advances and includes a discussion about the potential limitations and adverse effects associated with each option.

Introduction

In the 2018 US Renal Data Report, 80% of patients initiated hemodialysis with a catheter and 21% are still in use 1 year after hemodialysis initiation (1). One of the major complications of hemodialysis catheter use is bloodstream infection, which is associated with an increased risk of systemic infectious complications, hospitalizations, and death (2–6). Hemodialysis catheter use is associated with an eight-fold higher rate of vascular access-related bloodstream infections when compared with an arteriovenous fistula (2). The definition of catheter-associated bloodstream infection used in various clinical trials lacks consistency (7–12). In 2018, a multidisciplinary group of experts formed the Kidney Health Initiative’s Catheter End Points Workgroup to establish a standardized definition of catheter-associated bloodstream infection in patients on hemodialysis (12). The workgroup proposed the following criteria for the diagnosing hemodialysis catheter–associated bloodstream infection: (1) clinical suspicion of infection (fever/temperature >37.5°C or rigors or altered mental status or new predialysis hypotension [systolic BP <90 mm Hg]), (2) confirmation of bacteremia (blood cultures growing the same organism from the hemodialysis catheter and a peripheral vein or dialysis bloodline), and (3) exclusion of any alternate sources of infection.

The accurate reporting of bloodstream infection rates in the outpatient hemodialysis setting may be jeopardized by methodological challenges. Patients who become symptomatic with fever or chills oftentimes present directly to the hospital, and not to the hemodialysis facility. Furthermore, on rare occasion, antibiotics are administered without sending blood cultures in patients who become symptomatic when in the hemodialysis unit. In the National Healthcare Safety Network reporting system, bloodstream infection events are determined solely by a positive blood culture result and does not include intravenous antibiotic starts (2,13). A determination of the completeness of capture of blood cultures for all events in which obtaining blood cultures may have been indicated is missing in this data. One proposal to mitigate inaccuracies in bloodstream infection reporting is to provide data about positive blood culture results as a percentage of all intravenous antibiotic starts in the hemodialysis facility. However, this assumes that intravenous antibiotics are ordered solely for suspected cases of bacteremia. In addition, this proposal does not account for cases of bloodstream infection in which blood cultures were obtained in the hospital and the intravenous antibiotic course was completed during the hospitalization.

In the most recent National Healthcare Safety Network Dialysis Event Surveillance report for 2014, the mean rate of all bloodstream infections in catheter-dependent patients was 2.16 per 100 patient-months and the mean rate for access-related bloodstream infection was 1.83 per 100 patient-months (combined permanent and temporary catheters) (2). This was a marked improvement compared with the previous report for 2006 when the rate of all bloodstream infections in catheter-dependent patients was 4.2 (permanent catheters) and 27.1 (temporary catheters) per 100 patient-months, and access-related bloodstream infection rates were 3.1 (permanent catheters) and 17.8 (temporary catheters) per 100 patient-months (13). Similarly, from 2003 to 2013 the hospitalization rate for vascular access infections has declined in both incident and prevalent patients on dialysis (30% and 46% reduction, respectively) (14). However, there was a simultaneous rise in the rate of hospitalization for bacteremia/sepsis in both incident (38% increase) and prevalent (40% increase) patients (14). One explanation may be a change in the frequency of coding from the diagnosis “vascular-access infection” to an increase in use of the code “bacteremia.” Further investigation into these apparent paradoxical findings would be informative. In the recent US Renal Data Systems annual report for 2018, septicemia accounted for 8% of all deaths (1). Sepsis is a potentially preventable cause of mortality in hemodialysis population, and a significant percentage is vascular access related. Interventions to reduce infections in the hemodialysis setting include protocols for catheter reduction, new tools to improve compliance with the existing Centers for Disease Control and Prevention’s core interventions for catheter care, improving patient and staff education, and the development of novel devices for preventing catheter colonization (15).

Interventions for Catheter Reduction

There are several interventions that can decrease the incidence and prevalence of catheters in patients on hemodialysis. There is evidence that early referral of patients to nephrologists, multidisciplinary teams, and vascular access coordinators who provide education to patients; implantation of early-stick grafts; and urgent-start peritoneal dialysis can play a role in decreasing catheter use.

Early referral of patients with CKD to the nephrologist results in dialysis modality choice discussions and, thereafter, leads to timely referral for permanent vascular access placement, to a transplant center for enlistment or planning for peritoneal catheter insertion. In a small study of 135 patients, early (>4 months before dialysis initiation) referral to a nephrologist was associated with a greater likelihood of starting hemodialysis using a permanent vascular access (48% early referral versus 4% late referral) (16). Similarly, in study of a large cohort of 2398 incident patients on hemodialysis, late referral to a nephrologist (<90 days of hemodialysis initiation) was associated with a 42% higher risk of initiating hemodialysis with a catheter compared with those seen by a nephrologist earlier in the course of their kidney disease (odds ratio [OR], 1.42; 95% confidence interval [95% CI], 1.17 to 1.71) (17).

There is modest evidence that when multidisciplinary teams (consisting of physicians, nurses, social workers, and dietitians) provide education to patients, either prehemodialysis or at initiation of hemodialysis, the prevalence of catheters decreases (18,19). In the Treatment Options Program, implemented by a large dialysis organization, patients were educated on modality choice and vascular access prehemodialysis. Patients undergoing hemodialysis enrolled in this program (n=2800) had a lower likelihood of initiating hemodialysis with a catheter compared with matched control patients (n=2800) (63.2% versus 75.8%; P<0.001) (18). Similar findings were reported in the Incident Management of Patients Actions Centered on Treatment program, also implemented by a large dialysis organization. When patients received education at the initiation of hemodialysis, using a multidisciplinary team, the proportion of patients with an arteriovenous fistula or graft versus a catheter was significantly higher for patients in the intervention group at 6 months (0.60 versus 0.52; P<0.001) and 1 year (0.63 versus 0.48; P<0.001) (19). Contrary to the positive results of these studies, a quality improvement initiative using a multidisciplinary vascular access team to educate patients initiating hemodialysis in Canadian units did not show a decrease in catheter-free arteriovenous fistula use at 1 year (20).

The utilization of a vascular access coordinator can significantly decrease hemodialysis initiation with a catheter. A quality improvement project in Australia reported that prehemodialysis patient education and coordination by a vascular access coordinator resulted in a significant decrease in catheter-initiated hemodialysis (from 39% to 25%) (P=0.007) and a reduction in the total number of catheter days (2833 versus 4685 days) (21). Similar findings were reported from a program in the United States, where the implementation of a comprehensive access program led by a vascular access coordinator resulted in a significant reduction in prevalent catheter use at >90 days after hemodialysis initiation (from 11% to 6%; P<0.001) (22). The use of a vascular access coordinator in the hemodialysis unit can also improve outcomes after an episode of catheter-associated bloodstream infection. A quality improvement project in the Bronx and Connecticut (n=223 episodes of catheter-associated bacteremia) reported a significant reduction in recurrent bacteremia in 3 months (from 18% to 6%; P<0.02) and death due to sepsis (6% versus 0%; P=0.05) in hemodialysis units with a vascular access coordinator (23).

In patients in need of urgent-start dialysis, the creation of an early-cannulation arteriovenous graft and placement of a peritoneal dialysis catheter for immediate use are additional options available to reduce catheters. In a randomized, controlled trial, patients in need of an urgent vascular access for hemodialysis were randomized to receive an early-cannulation arteriovenous graft (n=60) or a tunneled catheter (n=61) (24). At 6 months of follow-up, bloodstream infections developed more frequently in catheter group (16.4%) compared with the arteriovenous graft group (3.3%; P=0.02). In a retrospective study from China, 96 patients starting dialysis with a peritoneal dialysis catheter were compared with 82 patients starting with a hemodialysis catheter (25). A significantly higher incidence of bloodstream infection occurred in the hemodialysis catheter group at 30 days (11% hemodialysis catheter versus 0 peritoneal catheter; P=0.003). Although both studies comprised a single center, with small numbers of patients, they suggest that avoiding catheters can reduce bloodstream infections in patients with ESKD on dialysis.

Challenges to Adherence with the Centers for Disease Control and Prevention’s Core Interventions

Implementation of the Centers for Disease Control and Prevention’s core interventions for hemodialysis catheter care has been associated with a 20%–50% reduction in bloodstream infection rates and hospitalizations due to sepsis (Figure 1) (15,26–28). Several objective factors may potentially interfere with optimal adherence with the Centers for Disease Control and Prevention’s core interventions in outpatient hemodialysis units in the United States. Until studies are completed that quantify and account for missing blood culture results, conclusions about the effect of specific factors remain speculative. Possible factors present in the outpatient hemodialysis setting that may play a role include the high patient-to-staff ratio, the proportion of patients dependent upon dialysis catheters, and a high staff turnover. The rate of catheter use is high among new hemodialysis starts. The newly started patients are typically placed on the available late dialysis shifts at the three medical centers where the authors work. To the extent that this holds true across the United States, the result would be a disproportionately high number of catheter-dependent patients on the late shifts. Additionally, in about half of the states, only nurses but not patient care technicians are permitted to connect and disconnect catheter patients (29). The high burden of catheter-dependent patients on those late shifts is compounded by the restriction of catheter care to nurses and presents a huge burden on nursing staff. It is not unusual for one nurse to be responsible for five or more catheter-dependent patients. A high dialysis patient-to-nurse ratio has been associated with more frequent access infections (30). It is possible that this high workload may cause lapses in adhering to the Centers for Disease Control and Prevention guidelines, thereby contributing to an increased risk of catheter-associated bloodstream infections. In contrast, in Canada, Europe, and Japan, hemodialysis care is provided exclusively by nurses, and the ratio of patients to nurses is only 3–4:1, as compared with 8–12:1 in the United States.

Figure 1.

Potential strategies for the prevention of bloodstream infections in patients undergoing dialysis. Educating patients and hemodialysis staff about the risks of long-term catheter use and about optimal catheter care are key components to reduce bloodstream infections. Attention to proper catheter exit-site care and hub disinfection using recommended antiseptic agents, and the use of recommended topical ointments during exit site dressing changes are important core interventions. Novel therapies, including chlorhexidine containing hub devices, and chlorhexidine dressings may further reduce bloodstream infections in catheter-dependent hemodialysis patients. In select patients, novel non-antibiotic locking solutions and Staphylococcus decolonization protocols may also be considered. CDC, Centers for Disease Control and Prevention.

A recent observational study reported the counterintuitive finding that the frequency of catheter-associated bloodstream infection in a given hemodialysis unit was inversely related to the proportion of patients using a catheter at that unit (31). In other words, the greater the proportion of catheter-dependent patients, the lower the frequency of catheter-associated bloodstream infection. One possible explanation for this surprising finding may be that in hemodialysis units where a greater proportion (>20%) of patients use a catheter, there may be a higher percentage of healthier patients, who are therefore less prone to infection. One can speculate that these healthier patients may be suitable for an arteriovenous fistula or graft, and might not have a catheter if they were dialyzing in units with lower catheter percentages (<10%). Catheter-dependent patients in facilities with a lower catheter percentages (<10%) may have more comorbidities and unsuitable for other forms of vascular accesses, and therefore be at higher risk for infection. Consistent with this hypothesis is the finding that catheter-associated bloodstream infection rates are particularly low in Alberta, Canada (0.19 per 1000 days or 0.57 per 100 patient-months), despite the fact that in Canada, 45% of patients on hemodialysis are catheter-dependent (32,33). An alternative explanation is that in high-catheter-percentage units, experienced staff, who are frequently performing catheter care may better adhere to core infection prevention techniques.

It is possible that lapses in adherence with the Centers for Disease Control and Prevention core interventions is a factor contributing to the relatively high catheter-associated bloodstream infection rate in the United States. There is clearly a need for studies on the use of catheter care audit checklists to assess whether they improve adherence with the Centers for Disease Control and Prevention guidelines. In addition, in an effort to reduce catheter-associated bloodstream infection, novel interventions have been developed to reduce exit-site and catheter lumen bacterial colonization. These measures are described in following sections.

Patient Self-Care and Showering

Frequent and consistent patient education about the associated infectious risks of prolonged hemodialysis catheter use and on self-care of the catheter are important for engaging patients to be an active participant in improving safety (34). Historically, showering was discouraged in catheter-dependent patients on hemodialysis; however, showering is important from a quality-of-life perspective. The Centers for Disease Control and Prevention now recommends showering be permitted, using exit-site and hub protection with an impermeable cover to reduce the likelihood of introducing organisms into the hemodialysis catheter. Sheet and pouch products have been designed to prevent water contamination of the hemodialysis catheter during showering, but may be cost-prohibitive. There is insufficient evidence to determine whether showering with a protocol that includes hemodialysis catheter hub protection or after-shower exit-site care decreases catheter-associated bloodstream infection rates (35,36). In a small study, comparable bloodstream infection rates were reported in patients on hemodialysis trained in after-shower exit-site care and a control group for which standard catheter care was performed by hemodialysis nurses (35). Similarly, another small trial compared showering without an exit-site dressing (with catheter hub protection) to a no-showering policy. There was no difference in bloodstream infection rates between the groups, suggesting that showering without an exit-site dressing may be safe (36). Further investigation with a large, randomized clinical trial is warranted.

Interventions Targeting the Catheter Exit Site

Ointments

Topical antimicrobial ointments that are applied to the catheter exit site are recommended at time of catheter insertion and at each hemodialysis session (Table 1) (37–42). The Centers for Disease Control and Prevention recommends using polysporin triple antibiotic ointment (bacitracin/gramicidin/polymyxin B) or povidone iodine ointment, which have been shown to be associated with a 75%–93% reduction in catheter-associated bloodstream infection (15,37,40). Polysporin ointment was also associated with a significant reduction in mortality, and long-term follow-up over 6 years was not associated with a change in microbiologic isolates over time (40,41). Unfortunately, gramicidin is not available in the United States, although polysporin ointments containing bacitracin/zinc/polymyxin B are in clinical use, but have not been rigorously studied for catheter-associated bloodstream infection prophylaxis. Mupirocin has also been associated with an 85% reduction in the bloodstream infection rate; however, there are reports of developing resistant microbes with the long-term routine use of mupirocin (38,39). Medicinal honey has been shown to be similar in efficacy to mupirocin in a small study (42). The advantage of medicinal honey is that it has a low likelihood for selecting resistant strains and is effective against antibiotic-resistant microorganisms. Well designed, adequately powered studies are needed before medicinal honey can be recommended for catheter-associated bloodstream infection prophylaxis. When using any ointment, checking catheter compatibility is mandatory, and a chart is available on the Centers for Disease Control and Prevention’s website (15).

Table 1.

Topical ointments and dressings used for hemodialysis catheter exit-site application

Study	N	Methods	Standard Definition of Bloodstream Infection Used	Definition Used	Ointment	Control	Outcome	Rate Ratio (95% CI)	P Value	Concerns
Levin et al. (37)	129	Randomized, double-blind, placebo-controlled trial	No	Positive peripheral blood culture without another demonstrable focus	Povidone-iodine	No ointment	Bloodstream infection	0.07 (0.06 to 0.24)	<0.01	Nontunneled catheters included
Sesso et al. (38)	136	Randomized, prospective trial	No	(1) One or more peripheral blood cultures with Staphylococcus aureus; (2) fever >37.8°C, with or without other signs of infection	Mupirocin	Povidone-iodine	Staphylococcus aureus bloodstream infection	0.14 (0.03 to 0.63)	<0.001	Potential for resistance with long-term use
Johnson et al. (39)	50	Prospective, randomized, controlled, open-label trial	No	(1) A single positive culture together with a positive culture from the catheter tip or exit site with same organism; (2) two or more positive cultures with no evidence of other infection source	Mupirocin	No ointment	Bloodstream infection	0.15 (0.03 to 0.80)	<0.01	Potential for resistance with long-term use
Lok et al. (40)	169	Double-blind, placebo-controlled, randomized trial	Yes	Health Canada guidelines (7)	Bacitracin-gramicidin-polymyxin B (polysporin triple ointment)	Placebo	a) Bloodstream infection b) Death	a) 0.25 (0.19 to 0.34) b) 0.22 (0.07 to 0.74)	a) <0.001 b) 0.004	Gramicidin containing polysporin triple ointment is not available in United States
Battistella et al. (41)	1478	Quality improvement report	Yes	Health Canada guidelines (7)	Bacitracin-gramicidin-polymyxin B (polysporin triple ointment)	Historical controls	Bloodstream infection	<1.0/1000 catheter-d	—	Gramicidin containing polysporin triple ointment is not available in United States
Johnson et al. (42)	101	Prospective, open-label, randomized, controlled trial	Yes	Health Canada guidelines (7)	Antibacterial honey	Mupirocin	Bloodstream infection	0.94 (0.27 to 3.24)	0.92	Underpowered to prove equivalence
Camins et al. (46)	121	Prospective, nonblinded, crossover intervention trial	Yes	CDC/NNIS (9)	Chlorhexidine-impregnated sponge	Transparent dressing	Bloodstream infection	1.22 (0.75 to 1.97)	0.46	Lack of efficacy
Apata et al. (47)	N/A	Quality improvement report	No	Bacteremia in any patient with a tunneled catheter or tunnel site infection	Chlorhexidine-impregnated transparent dressing	Dry gauze dressing plus antibiotic (not specified) ointment	Bloodstream infection	0.5 95% CI, N/A <1.0/1000 catheter-d	0.04	Local allergic reaction/rash, cost

95% CI, 95% confidence interval; CDC, Centers for Disease Control and Prevention; NNIS, National Nosocomial Infections Surveillance System; N/A, not available.

Antimicrobial Dressings

In the 2017 Centers for Disease Control and Prevention’s updated guidelines, chlorhexidine-impregnated sponge dressings are recognized as an alternative to ointments for prophylactic use in short-term, nontunneled catheters (43). These recommendations were on the basis of studies performed in hospitalized patients not on hemodialysis. Chlorhexidine is a nonantibiotic antimicrobial agent; therefore, the risk of selection for resistant organisms is minimal. Chlorhexidine-impregnated dressings were associated with a 70% reduction in bloodstream infection rates (44,45). Data from studies performed in catheters used for hemodialysis are conflicting (Table 1) (46,47). A study comparing chlorhexidine-impregnated sponge dressing with a transparent dressing at the catheter exit site found no difference in bloodstream infection (46). In contrast, in a recent quality improvement project a 50% reduction in bloodstream infection was reported using chlorhexidine-impregnated transparent dressings (changed weekly) compared with the control group using dry gauze dressings with antibiotic ointment applied to the catheter exit site (changed three times weekly) (47). The weekly cost per patient of the chlorhexidine dressing regimen was twice that of the standard dressing, which may be offset by overall cost-savings from a reduction in bloodstream infections. Unfortunately, although a reduction in bloodstream infection saves overall health care costs, the savings are not realized by the dialysis provider.

Interventions Targeting the Catheter Hub and Lumen

Hub Devices

Data regarding catheter hub devices on bloodstream infection outcomes are provided in Table 2 (48–51). There are two published studies using a neutral-valve catheter hub connector, which is changed weekly and locked with either saline or heparin. The first was a small study in which bloodstream infection rates did not statistically differ between neutral-valve connector with saline lock and standard catheter caps locked with citrate (concentration 46.7%) (48). The second study was a large, retrospective trial that reported a small reduction in the use of intravenous antibiotics with the neutral-valve connector, but no difference in bloodstream infection rates (49).

Table 2.

Hub devices used for hemodialysis catheters

Study	N	Methods	Standard Definition of Bloodstream Infection Used	Definition of Bloodstream Infection	Hub Device	Control	Outcome	Rate Ratio (95% CI)	P Value	Concerns
Bonkain et al. (48)	66	Prospective, nonblinded, randomized trial	No	Same organism in two blood samples from catheter with clinical suspicion	Neutral valve connector+saline lock	Standard cap+trisodium citrate (46.7%) lock	Composite of catheter dysfunction or bloodstream infection	0.16 (0.02 to 1.38)	0.06	No difference, small underpowered study
Brunelli et al. (49)	17,145	Retrospective, observational analysis	No	(1) IV antibiotics started; (2) receipt of antibiotics course; (3) positive blood cultures	Neutral valve connector+saline lock	Standard cap+heparin lock	a) IV Antibiotic start b) Bloodstream infection	a) 0.87 (0.82 to 0.93); b) 0.95 (0.88 to 1.09)	a) <0.001; b) 0.34	Retrospective study
Hymes et al. (50)	2470	Prospective, cluster-randomized, comparative effectiveness trial	No	Positive blood cultures reported	Chlorhexidine cap+heparin lock	Standard cap+heparin lock	a) Bloodstream infection rate b) IV Antibiotic start	a) 0.44 (0.23 to 0.83); b) 0.9 (0.74 to 1.19)	a) 0.01 b) 0.6	Cost
Brunelli et al. (51)	1671	Prospective, cluster-randomized, open-label trial	Yes	NHSN (2)	Chlorhexidine cap+heparin lock	NVC+70% isopropanol alcohol cap	Bloodstream infection	0.37 (0.20 to 0.68)	0.003	Cost

95% CI, 95% confidence interval; IV, intravenous; NHSN, National Healthcare Safety Network.

Another novel hub device contains a chlorhexidine-coated rod that extends into the catheter lumen (50,51). It is changed with each hemodialysis session and used with a heparin lock. In a prospective, cluster-randomized trial, the use of the chlorhexidine cap was associated with a significant reduction in the rate of positive blood culture episodes (56%) and in hospital admissions for bloodstream infections (40%) over a 1-year follow-up period, when compared with standard caps (50). A recently published, prospective, cluster-randomized trial compared bloodstream infection rates between hemodialysis facilities using the chlorhexidine cap to a neutral-valve connector with disinfecting caps containing 70% isopropyl alcohol (51). The use of the chlorhexidine cap was associated with a 63% reduction in bloodstream infection compared with the neutral-valve connector plus 70% isopropyl alcohol cap. The chlorhexidine cap has been approved by the US Food and Drug Administration for use in hemodialysis catheters, and has recently replaced the neutral-valve connector hub device in some units operated by large dialysis organizations in the United States. The cost of chlorhexidine-coated rod hub devices varies according to the volume purchased, and is $6 per pair if purchased directly in small quantities. Hymes et al. (50) argue that the upfront cost of the chlorhexidine-coated rod hub device may be offset by reducing the total economic effect of catheter-associated bloodstream infections, and project that the estimated savings to dialysis providers and Medicare would be $2400 and $16,000 per episode, respectively.

Antimicrobial Lock Solutions

Antimicrobial lock solutions are highly concentrated antiseptic agents that are instilled in the catheter when the catheter is not in use, thereby targeting the intraluminal route of entry (52–86). An antiseptic agent is needed to prevent colonization and biofilm formation, thereby reducing catheter-associated bloodstream infections. Antiseptic catheter locking agents may consist of an antibiotic or a nonantibiotic solution, and both types of catheter locks have been shown to effectively reduce the incidence of catheter-associated bloodstream infections in clinical trials (Table 3). The routine use of an antibiotic-containing catheter lock for prevention of bloodstream infections in hemodialysis may result in the emergence of resistant organisms, and has led to the development of safer, nonantibiotic locks. The second component added to most antimicrobial locks is an anticoagulant, used to prevent catheter dysfunction. Heparin promotes biofilm formation, whereas citrate in concentrations in concentrations ≥0.2% prevents biofilm formation, making citrate advantageous (87).

Table 3.

Antimicrobial locking solutions used for hemodialysis catheters

Study	N	Methods	Standard Definition of Bloodstream Infection Used	Definition of Bloodstream Infection	Antimicrobial Locking Solution	Control	Outcome	Rate/1000 Catheter-Days	P Value	Concerns
Pervez et al. (52)	36	Prospective, randomized study	No	Fever or chills without an alternate source of infection	Gentamicin 20 mg/ml+citrate 4.67%	Heparin 1000 U/ml	Bloodstream infection	0.62 versus 2.11	N/A	Resistance not measured
Dogra et al. (53)	83	Double-blind, randomized trial	Yes	CDC/IDSA (10)	Gentamicin 27 mg/ml+citrate 1%	Heparin 5000 U/ml	Bloodstream infection	0 0.3 versus 4.2	<0.001	Vestibular toxicity, resistance not measured
McIntyre et al. (54)	50	Prospective, randomized, controlled trial	Yes	CDC/IDSA (10)	Gentamicin 5 mg/ml+heparin 5000 U/ml	Heparin 5000 U/ml	Bloodstream infection	0.3 versus 4	0.02	Resistance not measured
Nori et al. (55)	30	Prospective, open-label, randomized, controlled trial	Yes	CDC/IDSA (10)	Gentamicin 4 mg/ml+citrate 3.13%	Heparin 5000 U/ml	Bloodstream infection	0 versus 4	0.008	Resistance not measured
Venditto et al. (56)	265	Prospective, observational trial	No	Two positive blood cultures from peripheral or catheter in a symptomatic patient with fever >38°C with no other apparent source of infection	Gentamicin (N/A)+heparin	Heparin (units N/A)	Bloodstream infection	0.4 versus 2.9	0.06	Resistance not measured
Landry et al. (57)	1410	Retrospective chart review after initiation of a unit protocol	Yes	CDC/IDSA (12)	Gentamicin 4 mg/ml+heparin 5000 U/ml	Historical controls before introduction of locks	Bloodstream infection	0.83–1.2 versus 17.0	N/A	4 yr study, gentamicin resistance observed in 32%. Bloodstream infection (coagulase-negative staphylococci n=13, Enterococcus n=7, Streptococcus n=2, Staphylococcus aureus n=1)
Fernández-Gallego et al. (58)	101	Prospective, observational study	No	Clinical improvement after antibiotics in patients with a fever and with positive blood cultures taken from the circuit, excluding other possible infection sites	Gentamicin 5 mg/lumen (approximately 1.5 mg/ml)+heparin 100 U	None	Bloodstream infection and bacterial resistance to gentamicin	0.11	N/A	7 yr study, no resistance for pathogens “normally sensitive to gentamicin”, two cases of MRSA resistant to gentamicin
Moran et al. (59)	303	Randomized, prospective, single-blinded, multicenter trial	No	Presence of a positive blood culture obtained from catheter, associated with fever or chills or hypotension	Gentamicin 0.32 mg/ml+citrate 4%	Heparin 1000 U/ml	Bloodstream infection	0.28 versus 0.91	0.003	4.5 yr study, no change in gentamicin resistance
Moore et al. (60)	555	Prospective, multicenter, observational cohort study	Yes	CDC/IDSA (12)	Gentamicin 0.32 mg/ml+citrate 4%	Heparin 1000 U/ml	a) Bloodstream infection b) Mortality	a) 0.45 versus 1.68 b) 10% versus 18% HR 0.32 (0.14–0.75)	a) 0.001 b) 0.001	3 yr study, reduction in gentamicin resistance by approximately 50% (P=0.01)
Goh et al. (61)	64	Single-center, retrospective cohort study	Yes	NHSN (2)	a) Gentamicin 5 mg/lumen (approximately 1.5 mg/ml) + Heparin 1000 U/ml b) Gentamicin +Citrate	a) Heparin 1000 U/ml b) Heparin 1000 U/ml	Bloodstream infection	a) 0.66 versus 1.42 RR 0.46 (0.30 to 0.72) b) 0.16 versus 1.42 RR 0.11 (0.05 to 0.22)	0.001	Resistance not measured
Onder et al. (62)	43	Single-center, retrospective cohort study	No	Positive blood culture from the catheter with or without positive peripheral blood culture with systemic symptoms (fever, chills, vomiting, hypotension) and no other identified source of infection	Tobramycin 5 mg/dl+tissue plasminogen activator 1 mg/ml	Heparin 5000 U/ml	Bloodstream infection	6.2 versus 16.8	0.2	Resistance not measured
Bleyer et al. (63)	60	Single-center, double-blind, randomized, controlled trial	No	Catheter colonization plus a peripheral blood culture growing the same organism	Minocycline 3 mg/ml+EDTA 30 mg/ml	Heparin (dose N/A)	Bloodstream infection	0 versus 0.47	0.35	Resistance not measured
Campos et al. (64)	204	Multicenter, open-label, randomized, controlled trial	Yes	KDOQI and CDC (8,10)	Minocycline 3 mg/ml+EDTA 30 mg/ml	Heparin 5000 U/ml	Bloodstream infection	1.1 versus 4.3	0.005	Resistance not measured
Saxena and Panhotra (65)	96	Prospective, case-control study	Yes	CDC (9)	Cefotaxime 10 mg/ml+heparin 5000 U/ml	Heparin 5000 U/ml	Bloodstream infection	1.65 versus 3.13	N/A	Resistance not measured
Saxena et al. (66)	113	Single-center, double-blind, randomized, controlled trial	Yes	CDC (9)	Cefotaxime 10 mg/ml+heparin 5000 U/ml	Heparin 5000 U/ml	Bloodstream infection	1.44 versus 3.15	<0.001	Resistance not measured
Al-Hwiesh and Abdul-Rahman (67)	63	Single-center, randomized, controlled trial	Yes	CDC (9)	Vancomycin 25 mg/ml+gentamicin 40 mg/ml+heparin5000 U/ml	Heparin 5000 U/ml	Bloodstream infection	4.54 versus 13.11	0.05	Resistance not measured
Kim et al. (68)	120	Single-center, double-blind, randomized, controlled trial	Yes	CDC (10)	Cefazolin 10 mg/ml+gentamicin 5 mg/ml+heparin 1000 U/ml	Heparin 1000 U/ml	Bloodstream infection	0.44 versus 3.12	0.03	Resistance not measured
Rijnders et al. (69)	270	Multicenter, open-label, evaluator-blinded, randomized, controlled trial	No	Clinical signs or symptoms of systemic infection and one of the following: positive blood culture in combination with temperature >38°C, temperature >37.5°C (during dialysis), rigors, chills, malaise, altered mental status, or hypotension unresponsive to fluids on dialysis	Trimethoprim 5 mg/ml+ethanol 25%+EDTA 3%	Heparin 5000 U/ml	Bloodstream infection	0.09 versus 0.41	<0.03	Catheter dysfunction
Bueloni et al. (70)	145	Multicenter, open-label, nonrandomized trial	No	Presence of at least one of the signs or symptoms of infection, such as fever, tremors or hypotension without another apparent focus of infection, with a positive culture if one is performed	Cefazolin 12 mg/ml+gentamicin 7 mg/ml+heparin 500 U/ml	Taurolidine 1.35%+citrate 4%+heparin 500 U/ml	a) Bloodstream infection b) Catheter removal	a) 0.79 versus 1.1 b) 49% versus 53%	0.01 0.85	Significantly higher rate of MRSA exit-site infections with cefazolin+gentamin lock
Allon (71)	50	Prospective, case-control study	No	Positive peripheral blood cultures in a febrile patient	Taurolidine 1.35%+ citrate 4%	Heparin 5000 U/ml	Bloodstream infection	0.6 versus 5.9	<0.001	Catheter dysfunction
Betjes and van Agteren (72)	58	Single-center, randomized, controlled trial	No	Symptomatic patient with a positive bacterial blood culture drawn from the dialysis catheter with no other apparent source of infection	Taurolidine 1.35%+ citrate 4%	Heparin 5000 U/ml	Bloodstream infection	0 versus 2.1	0.05	Catheter dysfunction
Solomon et al. (73)	110	Multicenter, double-blind, randomized, controlled trial	No	A single positive blood culture	Taurolidine 1.35%+ citrate 4%	Heparin 5000 U/ml	Bloodstream infection	1.4 versus 2.4	0.1	Catheter dysfunction
Solomon et al. (74)	174	Prospective, cohort study compared with historical controls	No	A single positive blood culture	Taurolidine 1.35%+citrate 4%+heparin 5000 U/ml	a) Heparin 5000 U/ml b) Taurolidine 1.35%+ Citrate 4%	a1) Bloodstream infection a2) First use of thrombolytic b1) Blood stream infection b2) First use of thrombolytic	a1) 1.33 versus 3.25 a2) RR 1.4 (0.5 to 3.9) b1) 1.33 versus 1.22 b2) RR 0.2 (0.06 to 0.5)	a1) <0.001 a2) 0.5 b1) <0.001 b2) <0.001	Cost
Winnicki et al. (75)	106	Multicenter, randomized, controlled trial	No	A positive bacterial blood culture drawn from the dialysis catheter in a symptomatic patient with fever or chills associated with dialysis and no apparent other source of infection	Taurolidine 1.35%+citrate 4%+heparin 5000 U/ml twice a wk with taurolidine 1.35%+citrate 4%+urokinase 25,000 U once a wk	Citrate 4%	a) Bloodstream infection b) Catheter dysfunction	a) 0.67 versus 2.7 b) 18.7 versus 44.3	0.003 0.001	Cost
Al-Ali et al. (76)	164	Multicenter, single-blinded, randomized, controlled trial	No	Same organism obtained from blood aspirated through the catheter hub and from blood sample obtained from peripheral vein with no other identifiable cause of infection	Taurolidine 1.35%+citrate 4%+heparin 5000 U/ml twice a wk with taurolidine 1.35%+citrate 4%+urokinase 25,000 U once a wk	Taurolidine 1.35%+citrate 4%+heparin 5000 U/ml three times a wk	a) Catheter removal for bloodstream infection b) Catheter removal for dysfunction c) Need for tissue plasminogen activator use	a) 0 versus 3 b) 1 versus 4 c) 5 versus 12	0.08 0.17 0.61	Cost
Weijmer et al. (77)	291	Multicenter, double-blind, randomized, controlled trial	No	Fever or cold chills not during a dialysis treatment and at least one positive blood culture and no other obvious cause of infection	Trisodium citrate 30%	Heparin 5000 U/ml	Bloodstream infection	1.1 versus 4.1	<0.001	—-
Winnett et al. (78)	413	Multicenter, case-control study	No	Fever and one positive blood culture result with no other obvious source of infection	Trisodium citrate 46.7%	Heparin 5000 U/ml	Bloodstream infection	0.81 versus 2.13	<0.001	Not FDA approved, excessive overfill may result in death
Power et al. (79)	232	Single-center, randomized, controlled trial	No	Symptomatic febrile patient with positive blood cultures	Trisodium citrate 46.7%	Heparin 5000 U/ml	Bloodstream infection	0.7 versus 0.7	0.9	Lack of efficacy
Venditto et al. (56)	265	Single-center, prospective, cohort study compared with historical controls	Yes	CDC (10)	Trisodium citrate 46%	Heparin (n/a)	Bloodstream infection	3.4 versus 2.9	0.6	Lack of efficacy
Correa Barcellos et al. (80)	464	Single-center, double-blind, randomized, controlled trial	Yes	CDC (10)	Trisodium citrate 30%	Heparin 5000 U/ml	Bloodstream infection	Rate ratio 1.53 (95% CI, 0.9–2.58)	—	Lack of efficacy
Hemmelgarn et al. (81)	225	Multicenter, double-blind, randomized, controlled trial	Yes	Health Canadian 1997 Guidelines (7)	Recombinant tissue plasminogen activator once a wk +heparin 5000 U/ml twice a wk	Heparin 5000 U/ml	Bloodstream infection	0.4 versus 1.37	0.02	Cost
Maki et al. (82)	407	Multicenter, open-label, randomized, controlled trial	No	Definite catheter-related: fever with concordant positive blood cultures drawn from the catheter and a peripheral vein or a peripheral blood culture and a concordant exit-site culture; concordant catheter-related bloodstream infection: two concordant positive blood cultures but with temperature not exceeding 38.0°C; probable catheter-related bloodstream infection: fever with one positive blood culture	Citrate 7%+methylene blue+methylparabens/propylparaben	Heparin 5000 U/ml	Bloodstream infection	0.24 versus 0.82	0.005	Not FDA approved
Broom et al. (83)	49	Multicenter, open-label, randomized, controlled trial	Yes	KDOQI (8)	Ethanol 70% once a wk+heparin 1000 U/ml twice a wk	Heparin 1000 U/ml	Bloodstream infection	0.28 versus 0.85	0.12	Small proof-of-concept study. potential symptoms, hepatotoxicity and mechanical changes in catheter polymer using high-dose ethanol
Sofroniadou et al. (84)	103	Single-center, randomized, controlled trial	Yes	CDC/IDSA (11)	Ethanol 70%+heparin 2000 U/ml	Heparin 2000 U/ml	Bloodstream infection	2.53 versus 6.7	0.04	Short-term study, potential symptoms, hepatotoxicity and mechanical changes in catheter polymer using high-dose ethanol
Vercaigne et al. (85)	40	Multicenter, randomized, controlled trial	No	Two or more positive blood cultures of the same organism from any source (peripheral or intravascular device cultures) from a patient with no other source of infection	Ethanol 30%+citrate 4%	Heparin 1000 U/ml	Bloodstream infection	0 versus 0.75	0.12	Small pilot study
El-Hennawy et al. (86)	452	Single-center, open-label, randomized, controlled trial	Yes	CDC/IDSA (10)	Sodium bicarbonate (7.5% or 8.4%)	Normal saline 0.9%	a) Bloodstream infection b) Catheter loss due to thrombosis c) Catheter loss due to bloodstream infection	a) 0.17 versus 2.6 b) 0.4% versus 0.6% c) 0.4% versus 6.6%	0.01 <0.001 <0.001

N/A, not available; CDC, Centers for Disease Control and Prevention; IDSA, Infectious Disease Society of America; MRSA, methicillin resistant Staphylococcus aureus; NHSN, National Healthcare Safety Network; KDOQI, Kidney Disease Outcomes Quality Initiative; FDA, Food and Drug Administration.

Antibiotic Locks

A variety of antibiotic-containing locking agents have been studied for the prevention of catheter-associated bloodstream infection in the hemodialysis setting (Table 3) (52–70). The current Centers for Disease Control and Prevention recommendations are for the limited use of prophylactic antimicrobial lock solutions in catheter-dependent patients on hemodialysis who have a history of multiple bloodstream infections (15). The prophylactic use of combination antibiotic-anticoagulant catheter lock solutions is associated with a significant reduction in bloodstream infections (range 50%–100% reduction). The antibiotics used as a catheter locking-solution in these trials were gentamicin, tobramycin, minocycline, cefotaxime, vancomycin, cefazolin, and trimethoprim, with gentamicin being the most commonly studied. Early trials, which used a relatively high-dose gentamicin (4–27 mg/ml) lock, reported that gentamicin alone was as effective as other antibiotic combinations and had a broad spectrum of activity against both Gram-positive (including Staphylococcus aureus) and Gram-negative bacteria at drug levels achieved in the catheter lumen. The emergence of gentamicin resistant strains of Enterococcus and Staphylococcus have been associated with serious episodes of bloodstream infection, and one death has been reported using a gentamicin (4 mg/ml) lock (57). In more recent studies, using a lower concentration of gentamicin lock (range 0.32–1.7 mg/ml), no gentamicin resistance has been observed over long follow-up periods, and in one study the rate of resistance declined, although an explanation for this finding is unclear (58–60). In one study, a low-dose gentamicin (0.32 mg/ml)-citrate (4%) lock was associated with significant reduction in bloodstream infections, infection-related hospitalizations, and patient mortality (60). If a gentamicin lock is used, a low-dose formulation (0.32 mg/ml) is recommended. In a recent, retrospective, cost-effectiveness analysis from New Zealand, the routine use of a gentamicin-containing catheter lock was associated with significantly lower rates of catheter-associated bloodstream infection, which translated to significant cost-savings (61). Inpatient costs associated with the management of catheter-associated bloodstream infection were NZ$27,792 per 1000 catheter-days (heparin-only lock) versus NZ$10,608 (gentamicin-heparin) and NZ$1898 (gentamicin-citrate). Concerns about antibiotic resistance associated with prophylactic use of antibiotic catheter lock solutions have prevented their adoption by consensus guidelines.

Nonantibiotic Locks

In an attempt to avoid the development of antimicrobial resistance, more recent focus has been on the development of nonantibiotic catheter lock solutions (Table 3) (56,70–86). Taurolidine is a broad spectrum, antimicrobial agent that reduces biofilm formation and has a low-risk bacterial resistance. Although taurolidine-citrate 4% antimicrobial lock is associated with a reduction in bloodstream infections, when compared with heparin lock, the need for thrombolytic therapy is increased (71–73). Newer preparations, which have added heparin or a thrombolytic agent to taurolidine-citrate 4%, are associated with improved rates of catheter dysfunction and bloodstream infections (70,74–76) A taurolidine-based, nonantibiotic lock (taurolidine-citrate-heparin) has been shown to be as efficacious as an antibiotic-containing lock (containing cefazolin-gentamicin-heparin) for preventing catheter-associated bloodstream infections (70). In this study, a higher rate of resistant strains of methicillin-resistant organisms was reported in the antibiotic lock group. At present, taurolidine containing catheter locks are widely used in Europe, but have not yet been approved for use in the United States. A large, multicenter, randomized trial of a taurolidine-citrate-heparin catheter lock was recently completed and results are available in abstract form, however publication of the manuscript is pending (89).

Citrate 4% locks, when used alone, are not associated with reduction in bloodstream infection rates. The efficacy of catheter-associated bloodstream infection reduction using medium- and high-dose citrate (30%–47%) antimicrobial locks are conflicting: two studies reported a reduction in bloodstream infection, and three studies reported no advantage of citrate over heparin (Table 3) (56,77–80). Trisodium citrate antimicrobial locks, in concentrations of between 30% and 47%, were withdrawn by the US Food and Drug Administration after one patient death, presumably because of overfill of the catheter (88). There are additional reports of symptomatic paresthesia and cardiac and embolic complications due to precipitation of trisodium citrate (30%–47%) in the catheter, and an increase in thrombolytic requirements for catheter dysfunction (90).

Recombinant tissue plasminogen activator used weekly as a catheter locking solution, with heparin twice weekly, has been shown to be associated with a significant reduction in bloodstream infection (approximately 67%) and catheter dysfunction (approximately 50%) (Table 3) (81). Tissue plasminogen activator is not in routine use for bloodstream infection prophylaxis because of excessive immediate costs; however, when counterbalanced by the anticipated reduced cost of bloodstream infection–associated hospitalizations, a decision analysis model calculated that there would be no significant difference in the mean overall cost (91). Again, under the United States health care model, prophylactic tissue plasminogen activator would increase the cost of providing hemodialysis, but the expected savings in reducing catheter-associated bloodstream infection would not be shared with the hemodialysis provider.

A novel, chelator-based, nonantibiotic antimicrobial lock containing citrate 7%-methylene blue-methylparabens-propylparaben was associated with a significant reduction (approximately 70%) in bloodstream infections and in catheter removal for dysfunction (Table 3) (82). This chelator-based antimicrobial locking solution was not approved by the US Food and Drug Administration.

Ethanol antimicrobial lock has also been studied for use in hemodialysis catheters (Table 3) (83–85). The potential advantages of ethanol include its low cost, reduction of biofilm, lack of resistance, and its broad antimicrobial and antifungal properties. In a small study, ethanol lock 70% used once weekly (heparin 1000 U twice weekly) was associated with a reduction in bloodstream infections when compared with heparin lock alone (83). A larger, prospective, randomized study of ethanol 70% plus heparin 2000 U/ml lock instilled three times a week compared with heparin alone reported a significant reduction in catheter-associated bloodstream infections in the ethanol lock group (84). Ethanol lock in high concentrations (70%–100%) has been associated with catheter dysfunction, reports of headaches, nausea, dizziness, fatigue, hepatotoxicity, and structural changes in the catheter integrity, including elution of molecules from the catheter polymers (92). Ethanol lock in a concentration of 30% is not associated with alterations in carbothane catheter integrity (93). In a small study, a lock containing ethanol 30%-citrate 4% was associated with a reduction in the rate of catheters removed for dysfunction compared with heparin (85). The only catheter-associated bloodstream infection episode occurred in the heparin lock group. One silicone catheter was found to have a crack in the ethanol 30% group. Ethanol locks (30%) should only be used with compatible catheters, and the ethanol should be withdrawn before initiation of hemodialysis (92,93). Larger, well powered clinical trials using low concentrations of ethanol 30% locks are needed.

Finally, there was a recently published prospective trial comparing the use of a sodium bicarbonate lock solution (8.4% or 7.5%) with a control group using saline (0.9%) for use in hemodialysis catheters (Table 3) (86). Sodium bicarbonate lock was associated with a significant reduction in catheter-associated bloodstream infections, and was shown to prevent catheter loss caused by bloodstream infections and thrombosis. Bicarbonate exerts antimicrobial activity against various bacteria, including S. aureus, Pseudomonas aeruginosa, and Escherichia coli, by potentiating the antimicrobial properties of mammalian antimicrobial peptides (cathelicidins and defensins), components of innate immunity. The mechanism appears to be bicarbonate’s ability to stimulate global changes in bacterial cell walls, gene expression, and membrane permeability. Alkaline solutions interfere with Staphylococcus adherence and impede biofilm formation (94). Sodium bicarbonate has been shown to have anticoagulant properties in small in vitro studies. By chelating calcium ions, similar to citrate, sodium bicarbonate impairs conversion of fibrin to fibrin clot, and prolongs prothrombin and thrombin clotting times (95). Sodium bicarbonate lock appears to be a safe and cost-effective intervention to reduced catheter-associated bloodstream infections in this single study. The potential advantage of a sodium bicarbonate lock is its anticipated low cost and safety profile. The findings of this study are promising, and larger, well powered studies further evaluating the efficacy of sodium bicarbonate (8.4%) lock in hemodialysis catheters are needed.

In summary, nonantibiotic antimicrobial lock solutions may have an important future role as novel, low-risk approaches to catheter-associated bloodstream infection prevention. Studies combining nonantibiotic antimicrobial locks with chlorhexidine antimicrobial hub devices have not been performed, and so it is unclear if this combination offers additional efficacy.

Interventions Utilizing S. Aureus Decolonization Protocols

Intranasal Mupirocin Ointment

S. aureus nasal carriage is associated with increased risk of hemodialysis catheter–associated bloodstream infections. Decolonization protocols performed in the hemodialysis setting are described in Table 4 (96–100). Intranasal mupirocin prophylaxis was associated with a 94%–100% efficacy for nasal decolonization and a significant reduction in the incidence of S. aureus catheter-associated bloodstream infections. S. aureus decolonization protocols are efficacious and cost-effective (101). Bloom et al. (101) performed a decision analysis evaluating clinical outcomes and cost-effectiveness of three possible management strategies in patients on hemodialysis (1): screen for S. aureus nasal carriage every 3 months and treat those with a positive test result with mupirocin; (2) treat all patients weekly with mupirocin; or (3) no prevention strategy, treat infection only. Eliminating S. aureus nasal carriage with mupirocin markedly reduced the number of infections (approximately 50% reduction) and reduced health care expenditures relative to treating infections when they occur. The potential annual savings to Medicare (in 1996 US dollars) were projected to be $784,000–$1,117,000 per 1000 patients on hemodialysis, depending on which prevention protocol was used. Mupirocin has not been widely adopted in the hemodialysis setting because of concerns of emerging resistance, although reports of mupirocin resistance have largely been conducted in hospitalized patients and with long term mupirocin use. The use of S. aureus decolonization protocols in catheter-dependent patients on hemodialysis should be revisited.

Table 4.

Intranasal mupirocin Staphylococcus decolonization protocols used for the prevention of blood stream infections in the hemodialysis setting

Study	N	Methods	Standard Definition of Bloodstream Infection Used	Definition of Bloodstream Infection	Protocol	Treatment Group	Outcome	Staphylococcus aureus Infection Rate/Patient yr	P Value	Concerns
Kang et al. (95)	19	Open, prospective cohort study	No	Not reported	Mupirocin twice per d for 5 d for 6 mo	MRSA nasal carriers	Bloodstream infection	1 Bloodstream infection	N/A	Potential for resistance
Lederer et al. (96)	16	Open, prospective cohort study	No	Not reported	Mupirocin three times per d for 5 d for 12 mo	MRSA nasal carriers	Bloodstream infection	0	N/A	Potential for resistance
Kluytmans et al. (97)	226	Open, prospective cohort study compared with historical controls	No	Not reported	Mupirocin twice per d for 5 d, then weekly for 6 mo	Staphylococcus aureus nasal carriers	Bloodstream infection	0.25 versus 0.04	<0.001	No resistance observed
Boelaert et al. (98)	80	Open, prospective cohort study compared with historical controls	No	Positive blood cultures (two out of six bottles)	Mupirocin three times per d for 5 d, then three times per wk for 6 mo, then weekly for 18 mo	Staphylococcus aureus nasal carriers	Bloodstream infection	0.097 versus 0.024	0.008	One mupirocin-resistant organism identified
Boelaert et al. (99)	35	Single-center, randomized, double-blind, placebo-controlled trial	No	Not reported	Mupirocin versus placebo three times per d for 14 d, then three times per wk for 9 mo	All Staphylococcus aureus nasal carriers	Bloodstream infection	0.489 versus 0.115; 1 bloodstream infection in mupirocin-treated group	<0.05	No resistance observed

MRSA, Methicillin resistant Staphylococcus aureus; N/A, not available.

Conclusion

Catheter avoidance is the obvious best strategy for reducing bloodstream infection episodes. Interventions such as early referral of patients to nephrologists, multidisciplinary teams, and vascular access coordinators who provide education to patients; implantation of early-cannulation arteriovenous grafts for hemodialysis; and urgent-start peritoneal dialysis have been shown to be effective at decreasing catheter use. However, in those patients where the use of a catheter is unavoidable, implementation of a multitargeted approach to prevent infections is imperative. Active patient engagement and staff education about appropriate catheter care, and improved adherence to the recommended Centers for Disease Control and Prevention’s core interventions is essential. The development of electronic chair-side checklists/audit tools for catheter care may also improve staff antiseptic technique and patient education in the hemodialysis unit. Novel interventions for catheter-associated bloodstream infection prophylaxis include the following: (1) a chlorhexidine-impregnated transparent exit-site dressing; (2) a chlorhexidine-coated rod hub device; and (3) an antimicrobial lock, preferably one without an antibiotic. There may be an important, underutilized role for S. aureus nasal decolonization protocols using mupirocin for patients on hemodialysis. A combination of these interventions may be optimal. Additional factors to consider when selecting a prophylactic strategy include efficacy, cost, potential for adverse effects such as catheter dysfunction, and local availability/device approval by regulatory agencies. Finally, identifying barriers to safe practices in the hemodialysis setting, using human factors systems engineering, will no doubt be invaluable for reducing infections in the future.

Disclosures

Dr. Allon reports receiving personal fees as a consultant for CorMedix. Dr. Allon is also the Editor-in-Chief of American Society of Nephrology open-access journal Kidney360. Dr. Golestaneh reports grant support from the Cardiorenal Society of America, Horizon Pharmaceuticals, and Montefiore Care Management Organization outside of the submitted work. Dr. Mokrzycki reports stock ownership in Abbott Laboratories. Dr. Golestaneh and Dr. Mokrzycki report positions as Clinical Events Committee Members for Spyral Pivotal Hypertension On-Medications and Spyral Pivotal Hypertension Off-Medications, sponsored by Medtronic. Dr. Abreo and Dr. Fisher have nothing to disclose.