Introduction

Central-line related bloodstream infections (CLABSI) are associated with increased morbidity, mortality and costs [1].

Antibiotic failure to treat CLABSI is essentially linked to microbial biofilms, which develop when microorganisms irreversibly adhere to the catheter surface and produce extracellular polymers that facilitate adhesion and provide a structural matrix. This matrix acts as a filter and potentially reduces antibiotic capacity to inhibit microorganisms [2].

In the majority of cases, catheter removal is considered the best option to cure patients [3].

However, catheter salvage could be an option especially for situations where venous access is poor and incidence of infection is high, such as onco-hematological and hemodialysis patients, as well as for subjects receiving parenteral nutrition [4, 5].

While for Staphylococcus aureus and fungi infections catheter removal is generally considered the best strategy, conservative treatment remains more controversial in case of CLABSI due to other agents, such as coagulase-negative staphylococci (CoNS) and gram-negative bacilli (GNB) [3, 4].

Antimicrobial lock therapy (ALT) in particular appears as a promising treatment strategy for conservative management of the catheter. It consists in instilling an antibiotic solution into the hub of the central venous catheter when it is not in use, to achieve a concentration of the antibiotic that is manyfold higher than the minimal inhibitory concentration (MIC) for the bacteria involved [6, 7]. The antibiotic should remain for a specified duration, typically between 18 and 24 h and this strategy appears particularly useful for patients in whom reintroducing a catheter may be difficult [6].

Table 1 Current knowledge on available molecules for antimicrobial lock therapy
Full size table

Besides, novel molecules such as Daptomycin and Tigecycline have demonstrated potential efficacy, alone or in combination with other molecules, due to their mode of action on bacterial biofilms.

However, making evidence-based decisions regarding ALT recommendations is sometimes difficult [8–10] and protocols are usually developed on the basis of local experience.

Our aim was to review the literature concerning ALT for CLABSI, including data on novel molecules, which have added extra potential to ALT strategies.

Methods

We obtained relevant articles from the Pubmed database, using the broad search terms “central line infection” and “lock therapy”. Case–control studies evaluating two or more drugs as ALT in central venous catheter infections were included regardless of date, language and publication status. Experimental models and in vivo studies on humans or animals were considered.

The Pubmed search was performed in April 2014 and updated in December 2014. Among 221 articles selected in Pubmed, 217 contained an abstract, 204 were written in English and 103 focused more specifically on the research theme. Among these 103 papers, 54 were retained for their particular interest concerning ALT. Publication dates of selected articles ranged from 1990 to 2014. The remaining articles were excluded as their content did not provide further information with regard to the selected papers or were not considered relevant. Tables 1 and 2 resume main results of this research.

Results

Incidence of CLABSI and clinical outcome

In the United States, approximately 80,000 cases of CLABSI occur among 3 million central venous catheters (2.7 %) [5]. The same problem is encountered in Europe, where the incidence of health-care associated infection, which includes CLABSI, reaches 10.1 % [11].

A recent CDC report revealed encouraging data, with a 58 % reduction of CLABSI in ICUs from 2001 to 2009. Indeed, Fagan et al. [12] measured incidence trends in pathogen-specific CLABSI in US ICUs from 1990 to 2010, showing a decline since 2006, with the exception of infections due to S. aureus in pediatric ICUs, probably as a consequence of surveillance programs [13].

Dreesen et al. [14] evaluated differences between cancer patients and those with other conditions. Indeed, they found that in patients with a benign underlying disease, rates of CLABSI range between 0.19 and 2.441 episodes for 1,000 catheter-days, with a median of 0.82 episodes for 1,000 catheter-days. This rate increases in studies including over 50 % cancer patients, where the incidence rate of CLABSI lies between 1.9 and 6.8 episodes per 1,000 catheter-days, with a median of 2.71 [14].

According to a recent systematic review and meta-analysis, the odds ratio for in-hospital death due to CLABSI was 2.75 [CI 1.86–4.07], decreasing to 1.51 [CI 1.02–2.65] when patients were matched using an illness severity index [1].

Mechanisms for CLABSI

Endogenous microorganisms residing on the skin or in the catheter hub are the main cause of CLABSI. In this case leading pathogens are S. aureus and CoNS [3, 11, 15].

Other sources of CLABSI are organisms spread by contaminated hands or intravenous fluids, or hematogenous diffusion. In these cases, involved microorganisms are more heterogeneous, including Enterobacteriaceae, Pseudomonas and fungi [15].

Moreover, the ability of a microorganism to attach to catheters depends on the properties of the catheter surface, the microorganism itself and the host. In the first case, physical irregularities on the catheter can increase risks of adherence. Among risk factors linked to the microorganism, binding characteristics of infective agents differ, in particular their capacity to link to fibronectin and fibrinogen. Besides, bacteria can produce extracellular polymeric products, contributing to the development of the biofilm, which reduces antibiotic capacity to cure catheter infection [14].

Indeed, the biofilm is formed by bacteria in stationary phase embedded in an extracellular matrix; consequently, it protects bacteria against both the immune system and antimicrobial agents and could be responsible for recurrent bacteremia [16–20].

Moreover, Mekhni et al. [21] found that strains of Staphylococcus epidermidis isolated from CLABSI produce more biofilm than strains from non catheter-related bacteremia and those considered as commensal isolates.

Among risk factors for CLABSI linked to the host, age, co-morbid conditions, severe immunodepression and the individual capacity to produce fibronectin and fibrinogen are those most often stated.

Several studies describe potential risk factors for CLABSI. Dreesen et al. [14], focusing on patients receiving home parenteral nutrition, listed factors that were device-related, e.g., catheter caliber, education-related, e.g., lack of training with provision of detailed information to patients, patient-related, e.g., age and regular smoking and therapy-related, e.g., duration of parenteral nutrition and days of administration,

Studies of antimicrobials used for ALT

Aminoglycosides

In an animal study, Chauan et al. [22] showed that gentamicin-EDTA lock therapy is active against S. aureus and CoNS biofilms. Fortun et al. used this compound as ALT for 14 days, in association with systemic treatment, for gram-negative rod CLABSI.

Gentamicin ALT is typically administered at 2.5 or 5 mg/ml concentrations [23]. However, data about its efficacy are contradictory, as Lee et al. [24] showed that it was not effective over 5 days of ALT in in vitro biofilm models, nor were cefazolin, nafcillin or erythromycin, while Fernandez-Hidalgo et al. [25] found that gentamicin ALT was generally effective on S. aureus biofilms in rabbit catheters.

Amikacin is generally administered 12 h/day at the dose of 3 mg for a period varying from 14 to 27 days and is generally associated with systemic treatment [26].

Bestul et al. [15] reported a trial on 11 patients receiving home parenteral nutrition, where 2 ml/day of 1.5 mg/ml amikacin, locked for 12 h, resulted in 90 % successful treatment of CLABSI.

Antifungal lock therapy

Candida species are associated with the highest overall crude mortality due to bloodstream infection, comparable to that of Pseudomonas. Candida albicans is the most common fungal species responsible for CLABSI (see Table 2).

Table 2 Antimicrobial lock therapy (ALT) according to pathogen
Full size table

Although guidelines recommend catheter removal in case of CLABSI due to fungi [3, 27], there are reports of effective ALT.

If azoles generally have poor activity against Candida biofilms, lipid Amphotericin formulations, echinocandins and 25 % ethanol showed good success rates in salvage therapy of CLABSI [28]. Indeed, Ramage et al. [29] showed that liposomal Amphotericin B displays rapid dose-dependent activity against C. albicans biofilm. Raad et al. [30] compared EDTA, Amphotericin B lipid complex (ABLC) and EDTA plus ABLC against C. albicans and Candida parapsilosis, showing that EDTA plus ABLC was the most effective against biofilms.

Treatment is generally administered for 14 days, in conjunction with systemic antifungal therapy.

In case of polyurethane catheters, ethanol solutions cannot be used because of potential incompatibility [28].

Vancomycin lock therapy

Vancomycin is a glycopeptide that is active against gram-positive bacteria, including methicillin-resistant staphylocci. The antibiotic is prescribed as ALT using concentrations ranging from 1 to 5 mg/ml. Duration of locks varies from 3 to 27 days and ALT is generally associated with systemic treatment [6]. Indeed, Bookstaver et al. [23] showed that active systemic therapy, associated with vancomycin ALT, alone or together with gentamicin or ethanol, allowed good success rates while avoiding catheter removal.

Beigi et al. [31] compared a strategy with vancomycin lock and systemic ceftriaxone with a combination of systemic vancomycin and amikacin, finding that the ALT arm was superior for fever resolution and catheter retention. Vancomycin ALT was administered at the dose of 500 mg and injected in the lumen every 48 h for 7 days.

Interestingly, LaPlante et al. [19] showed that the addition of heparin to vancomycin enhanced the activity against biofilm-forming S. epidermidis, but not biofilm-forming S. aureus.

Daptomycin lock therapy

Daptomycin (DPT) is a cyclic lipopeptide antibiotic, which has a rapid, concentration-dependent bactericidal activity against most clinically important gram-positive bacteria, including methicillin-resistant Stahylococcus aureus (MRSA), CoNS, vancomycin-resistant enterocci and penicillin-resistant Streptococcus pneumoniae. Moreover, its in vitro capacity to destroy staphylococcal biofilm offers interesting potentialities for the treatment of CLABSI.

DPT is a calcium-dependent antibacterial agent, working optimally in vitro with levels of free calcium ions at physiological concentrations (50 μg/ml). Most studies are, therefore, conducted using DPT with lactated Ringer’s solutions (containing 50–80 μg/ml of calcium) [18, 32].

Estes et al. [33] compared four strategies for eradicating biofilm-producing strains of MRSA embedded on silicon disks, including DPT or minocycline, both alone or associated with 25 % ethanol. Authors found that 5 days of 4-h daily exposures to DPT (2.5 mg/ml), independently from ethanol use, and minocycline associated with ethanol were superior to minocycline alone for reducing MRSA colony counts.

Aumeran et al. [18] compared the effectiveness of vancomycin (5 mg/ml), DPT (5 mg/ml) and 40 % ethanol (with enoxaparin sodium 400 IU/ml) lock solutions in destroying a pre-established biofilm of S. epidermidis. Authors found that mean logarithmic reductions in bacterial growth were statistically greater for DPT and ethanol than for vancomycin, while bacterial reformation after 24 h of exposure to ALT was higher for vancomycin and DPT than for ethanol.

Van Praagh et al. [32] compared the efficacy of vancomycin and DPT ALT in a rat model of Staphylococcal central venous catheter biofilm. DPT and vancomycin were administered as ALT at concentrations of 5 mg/ml 18 h daily for 3 days in rats with central venous catheter infection caused by methicillin-resistant S. epidermidis (MRSE) strains susceptible to both molecules (MIC 1 µg/mL for DPT and 2 µg/mL for vancomycin). Animals were also administered the same antibiotics subcutaneously: DPT at the concentration of 40 mg/ml once daily and vancomycin 50 mg/kg twice daily. DPT and vancomycin both caused rapid elimination of MRSE from central venous catheters and although differences were not statistically significant, DPT appeared to act slightly faster. However, authors combined systemic treatment with ALT and no port-catheter was included in the experiment.

Interestingly, authors found no difference in bacterial inhibition during ALT when DPT was diluted with either Ringer’s or saline solutions, while dilution with Ringer’s was superior in sustaining bacterial clearance 7 days after the last ALT.

Dominguez-Herrera [34] also compared DPT with vancomycin against biofilm-producing MRSE in an experimental in vivo model, showing that DPT reduced bacterial concentrations in the catheter by a factor of 3 log10 CFU/mL compared to vancomycin.

Moreover, LaPlante et al. [19] compared in vitro activity of DPT and vancomycin lock solutions on staphylococcal biofilm in a central venous catheter model, showing that after 72 h of exposure (with each compound at a concentration of 5 mg/ml), they displayed equivalent capacity to eradicate S. epidermidis growth, while DPT was superior to vancomycin for inhibition of S. aureus. Authors also showed that adding heparin to DPT is not necessary to eradicate S. epidermidis or S. aureus, differing from vancomycin.

Meije et al. [35] showed that DPT achieved greater activity than vancomycin against biofilm-forming methicillin-susceptible S. aureus and MRSA, but concentrations of 50 mg/ml performed better than the usual dose of 5 mg/ml.

The combination of DPT with rifampin appears useful for treating biofilm, as a consequence of a synergistic mechanism, as showed by Cirioni et al. [36] Indeed, authors found that the combination of these molecules had higher efficacy than that of each single compound for treating a rat model of Staphylococcal infection. Authors suggest that the positive interaction between DPT and rifampin could be explained by the mechanism of action of DPT, which opens channels in the cell membrane, promoting entry of hydrophobic drugs such as rifampin.

In vivo studies on humans concerning ALT with DPT are rare. In a short series of 13 patients, Del Pozo [37] showed that DPT lock therapy is effective in CLABSI due to CoNS infection. In this study, 5-milliliter syringes were filled with DPT (5 mg/ml) plus heparin and reconstituted in Ringer’s lactate solution. The antimicrobial locks were replaced daily, for a mean duration of 14 days, associated with systemic treatment in 85 % of cases. Success was defined as fever resolution, negative blood cultures and catheter salvage within one month after the end of treatment, and was achieved for 11 out of 13 subjects. The catheter had to be removed during therapy in two patients, due to persistent fever.

In another short case series, Tatarelli et al. [38] describe eight patients with CLABSI mainly due to CoNS infection and with previous failure to vancomycin and cefazolin, where six of them were successfully treated with DPT as systemic and lock therapy. Mean duration of ALT was 13 days.

In vivo studies using ALT with DPT require the use of Ringer’s lactate solution (60–130 mg/l calcium bicarbonate) to approximate calcium concentrations used in the majority of in vitro studies [19].

Tigecycline lock therapy

Tigecycline, a derivative of tetracycline, acts against strains producing biofilms and has a wide spectrum of action, including the majority of gram positive and gram negative bacteria.

However, there are few data concerning its use in ALT.

In CoNS-infected polyurethane catheters inserted in rabbits, electron microscopy showed that Tigecycline was superior to vancomycin in reducing bacterial growth [39].

Raad et al. [17] compared the activity of various compounds against MRSA embedded in biofilm, showing that after 3 days of 4-h daily exposures, Tigecycline ALT was faster than Linezolid, Rifampin and Vancomycin in eradicating MRSA from biofilm, but slower than DPT. Furthermore, the addition of Rifampin to Tigecycline or DPT enhanced antimicrobial activity against MRSA.

Ethanol lock therapy

In vitro studies concerning ethanol, EDTA and levofloxacin lock therapy showed that ethanol, alone or combined with EDTA, was effective as lock therapy against Strenotrophomonas maltophilia CLABSI [40].

Moreover, Aumeran et al. [18] showed that ethanol and DPT were more effective than vancocymin for treating S. epidermidis biofilms.

The New York Presbyterian Hospital guidelines suggest treating catheters with 3 ml of 74 %ethanol combined with 1 ml of NaCl 0.9 %, preceded and followed by washing with NaCl, and to avoid ethanol use in case of polyurethane catheters [41]. Similarly, Dannenberg et al. [42] suggest injecting 2.3 ml 74 %ethanol and to lock for 20–24 h/day for 3 days. However, Abu-El-Haija et al. [43] found that ethanol ALT decreased the incidence of CLABSI but had a negative impact on line integrity, suggesting careful selection of patients for ethanol therapy. Similarly, a recent systemic review of ethanol ALT [44] by Mermel and Alang showed high rates of systemic side effects and device’s structural changes, especially for polyurethane catheters.

Linezolid lock therapy

Very few data are available concerning the potential role of linezolid as ALT.

Fernandez-Hidalgo et al. [25] showed that linezolid and vancomycin had equivalent activity on MSSA and MRSA strains responsible for catheter-related infections in a rabbit model.

Moreover, Leite et al. [45] reported a synergistic effect of linezolid and N-acetyl-cysteine for treating S. epidermidis biofilms.

Telavancin lock therapy

Telavancin is another antibiotic with some potential in ALT. However, to our knowledge, no consistent data have been published at present regarding its efficacy. LaPlante et al. [46] showed that Telavancin 2 5 and 5 mg/ml were physically compatible with heparin 2,500 units/ml and citrate 2.2 and 4 % over 72 h for lock solutions.

Taurolidine

Taurolidine is a drug with antimicrobial and anti-lipopolysaccharide properties, which has demonstrated its capacity to inhibit in vitro biofilm formation [47].

It is derived from the amino-acid taurine and acts via the release of three active methylol groups, which react with the bacterial cell wall, resulting in bacterial lysis and bacterial endotoxin neutralization [48].

Comparing the effect of taurolidine and heparin in biofilm formations, Handrup et al. [47] showed that taurolidine was more effective in preventing CLABSI.

These results are confirmed by Shah et al. [49], who showed that ALT with taurolidine was superior to heparin for bactericidal activity against several bacteria responsible for catheter infections.

However, comparative studies with other molecules are lacking; therefore, strategies for its use are at present based on local experience rather than on guidelines [7].

Studies evaluating ALT and systemic therapy for CLABSI

O’Horo et al. [6] analyzed studies which compared ALT to systemic therapy, finding that the combination of both was superior (OR 0.2, 95 % CI: 0.10–0.39) to systemic antibiotic alone for catheter salvage, with 10 % of locked patients requiring replacement, compared to 33 % of subjects without lock.

Bookstaver et al. [23] evaluated clinical outcomes with ALT as adjunctive treatment to systemic therapy in case of CLABSI, in a small series of patients. The main pathogens were CoNS (35 %), MRSA (23 %) and GNB (23 %), while the main drugs used for ALT were DPT, vancomycin and gentamicin. Mean duration of ALT was 8.9 days. Authors found that blood culture sterilization was obtained in 69 % of cases and catheter salvage in 42 % of cases. Moreover, longer ALT duration (>9 days) was correlated with achieving catheter sterilization with an OR of 1.367 for each day above the mean.

ALT in hemodialysis-related CLABSI

As hemodialysis catheters have certain specificities regarding type of device and patient characteristics, ALT in this context was analyzed separately.

A systematic review and meta-analyis on hemodialysis-related CLABSI recently revealed that ALT and guidewire exchange had similar rates of success and both were superior to systemic therapy alone. The highest cure proportion was reached for CoNS infection, followed by GNB and S. aureus. In case of S. aureus infection, guidewire exchange was superior to either systemic therapy or ALT [50].

Joshi et al. [51] prospectively examined the efficacy of systemic and ALT antibiotics for the treatment of tunneled hemodialysis CLABSI, showing that ALT was effective in 59 % of cases, with similar success rates in gram-positive and gram-negative infections (63 vs 62 %, respectively).

Lee et al. [24] studied different combinations of ALT in adult hemodialysis patients. Locks included vancomycin, imipenem, ciprofloxacin and ceftazidime and were instilled into the lumens of catheters on completion of each dialysis session and withdrawn immediately prior to the next session. Authors found that ALT combined with systemic treatment was superior to systemic therapy alone.

Moreover, Saxena et al. [52] showed that a cefotaxime-heparin lock was effective for preventing CLABSI in Staphylococcus aureus nasal carriers undergoing hemodialysis through tunnelled, cuffed catheters, with the exception of MRSA carriage.

Citrate has been used clinically as an anticoagulant in heparin-free catheter locks and a systematic review and meta-analysis performed by Zhao et al. [10] showed that antimicrobials associated with citrate locks in hemodialysis patients could reduce the incidence of CLABSI, whereas citrate alone was not superior to heparin. Interestingly, glyceryl trinitrate seems synergistic with citrate and this combination appears attractive as a lock solution [53].

In an observational, non-comparative, prospective cohort study on patients with GNB-CLABSI, where the majority of devices were hemodyalisis catheters, Funalleras et al. [4] showed high rates of success using ALT with ciprofloxacin or amikacin, both at 2,000 mg/L and mixed with sodium heparin (20 UI/ml). Total volume of ALT solutions depended on the type of device, usually 2 mL for tunneled catheters and 3–5 mL for implanted ports. Median days of ALT were 13, associated with systemic treatment for a median period of 15 days.

Few studies have focused on the potential side effects of aminoglycosides as ALT: Dogra et al. [54] recently showed that plasmatic concentrations of gentamicin when used as ALT for the prevention of hemodyalisis-related CLABSI could be associated with ototoxicity.

Discussion

Studies evaluating ALT for central venous catheter infections vary according to the type of antibiotic and concentrations used. Many catheter-associated infections are related to intraluminal biofilms and are often difficult to treat. The recent introduction of novel molecules has added potentialities for improving their treatment, but studies on ALT, alone or associated with systemic antimicrobials, are often limited by the small number of patients included or by the study design.

Moreover, the advantage of ALT for clinicians seems to differ according to the causative organism. Indeed, although some molecules, such as aminoglycosides, fluoroquinolones and tigecycline, have a broad spectrum of action, including gram-positive and gram-negative bacteria, their potentiality for ALT is limited by the lack of studies [55]. Some recent data, such as Funallera’s work, suggest that aminoglycosides and fluoroquinolones as ALT could allow conservative management of CLABSI due to GNB infection with no signs of severity, but further studies are needed and better knowledge about their potential systemic toxicity, especially for gentamycin, is required.

In case of fungal infections, Amphotericin B showed good activity in biofilms caused by Candida sp. infections, but guidelines suggest that catheter removal is the best option [3, 27].

Among gram-positive CLABSI, only few data exist about conservative management of S. aureus infection, such as Fernandez-Hidalgo’s work, which suggests that gentamycin could be a valid option for sterilizing catheters within 24 h of lock. However, as systemic signs and risks of complicated infections are frequent, catheter removal in case of S. aureus CLABSI is generally recommended [3]. In case of Enterococcal infections, Del Pozo et al. [56] showed the effectiveness of ALT among 3 out of 4 patients with Enterococcus faecium CLABSI, but the limited number of subjects requires prudent conclusions.

Therefore, the most promising indication for treating CLABSI with ALT appears to be CoNS-related infection, where systemic signs of infection are generally less severe and the introduction of novel molecules, such as DPT and tigecycline, improved the potential for cure without catheter removal. Although large prospective trials comparing DPT with vancomycin are lacking, the majority of comparative studies are in favor of DPT for its potent and rapid antibacterial capacity to inhibit bacterial biofilm and show a synergistic action with rifampin.

However, in vivo studies for DPT as ALT are scarce and questions concerning associated systemic treatment and duration of ALT remain unresolved.

Among other potential molecules for CoNS CLABSI, ethanol and taurolidine are two promising options, but few studies are available to confirm their place among ALT strategies and to exclude side effects.

Together with CoNS CLABSI, hemodialysis catheter infections appear as the other condition where ALT shows most chances of being successful and for which data are robust, with recent data also suggesting possible conservative management of GNB infections.

However, which kind of antibiotic, its dosage and treatment duration are points that still need to be elucidated.